Dynamic PolyConjugates for targeted in vivo
delivery of siRNA to hepatocytes
David B. Rozema*†, David L. Lewis*†, Darren H. Wakefield*, So C. Wong*, Jason J. Klein*, Paula L. Roesch*,
Stephanie L. Bertin*, Tom W. Reppen*, Qili Chu*, Andrei V. Blokhin*, James E. Hagstrom*, and Jon A. Wolff‡
*Mirus Bio Corporation, 505 South Rosa Road, Madison, WI 53719; and ‡Departments of Pediatrics and Medical Genetics, Waisman Center,
University of Wisconsin, 1500 Highland Avenue, Madison, WI 53719
Edited by Inder M. Verma, The Salk Institute for Biological Studies, La Jolla, CA, and approved June 18, 2007 (received for review April 24, 2007)
Achieving efficient in vivo delivery of siRNA to the appropriate target
cell would be a major advance in the use of RNAi in gene function
studies and as a therapeutic modality. Hepatocytes, the key paren-
chymal cells of the liver, are a particularly attractive target cell type for
siRNA delivery given their central role in several infectious and
metabolic disorders. We have developed a vehicle for the delivery of
siRNA to hepatocytes both in vitro and in vivo, which we have named
siRNA Dynamic PolyConjugates. Key features of the Dynamic Poly-
Conjugate technology include a membrane-active polymer, the ability
to reversibly mask the activity of this polymer until it reaches the
acidic environment of endosomes, and the ability to target this
modified polymer and its siRNA cargo specifically to hepatocytes in
vivo after simple, low-pressure i.v. injection. Using this delivery
technology, we demonstrate effective knockdown of two endoge-
nous genes in mouse liver: apolipoprotein B (apoB) and peroxisome
proliferator-activated receptor alpha (ppara). Knockdown of apoB
resulted in clear phenotypic changes that included a significant
reduction in serum cholesterol and increased fat accumulation in the
liver, consistent with the known functions of apoB. Knockdown of
ppara also resulted in a phenotype consistent with its known func-
tion, although with less penetrance than observed in apoB knock-
down mice. Analyses of serum liver enzyme and cytokine levels in
treated mice indicated that the siRNA Dynamic PolyConjugate was
nontoxic and well tolerated.
pH labile bonds
nonviral siRNA delivery
siRNA–polymer conjugates

endosomolytic polymers
The ability of siRNA to silence specific genes has generated greatinterest in its use as a research tool and therapeutic agent for a
wide spectrum of disorders that include cancer, infectious disease,
and metabolic conditions (1–3). Effective in vivo delivery of siRNA
to the appropriate target cell is an essential component of these
siRNA-based applications. Accordingly, a variety of nonviral (4–
14) and viral (15–17) systems are being developed for delivery of
siRNA to liver, tumors, and other tissues in vivo.
In addition to their importance in many infectious andmetabolic
disorders (18), hepatocytes are a particularly attractive target cell
type for siRNA delivery given their ability to be accessed directly
by nanoparticle-sized constructs after simple intravascular injec-
tion. Initial hepatocyte delivery efforts used hydrodynamic delivery
of naked siRNA to the liver (19, 20). More recent work has used
viral vectors, such as AAV or lentivirus (16, 17), or synthetic
systems such as cholesterol–siRNAconjugates or stable nucleic acid
lipid particles (SNALPs) (21, 22). Among the nonviral approaches,
SNALP technology represents a significant advance, enabling
target mRNA knockdown in liver after i.v. injection of clinically
relevant doses of siRNA (21). More recently, another lipid-based
system termed interfering nanoparticles (iNOPs) has also demon-
strated the ability to deliver siRNA in vivo (23). A key drawback of
the SNALP and iNOP systems, however, is that the siRNA com-
plexes are only passively targeted to liver. As a result, siRNAs are
delivered to a significant number of nontarget cells in the liver,
potentially contributing to toxicity.
Hepatocyte targeting after administration into a peripheral vein
requires that the delivery vehicle avoid nonspecific interactions en
route to the target cell, which is commonly accomplished by the
attachment of polyethylene glycol (PEG) (24) or other hydrophilic,
noninteractive agents. Upon reaching the liver, the vehicle must
then exit the intravascular space to access hepatocytes. Because of
the open, fenestrated nature of the hepatic vasculature, particles

100 nm in diameter can readily exit hepatic vessels and interact
with liver parenchymal cells (25). However, avoiding uptake and
subsequent activation of Kupffer cells, the resident immune cells of
the liver, are likely essential to avoid toxicity (26). As an example,
Kupffer cell uptake of adenoviral vectors is the main cause of liver
toxicity observed when these vectors are used for delivery (27).
Galactose-derived ligands, which are recognized by the asialogly-
coprotein receptor (ASGPr), can be used to specifically target
hepatocytes (28). Certain galactose-containing ligands enable he-
patocyte uptake and avoidance of Kupffer cells if properly dis-
played on the delivery vehicle (29, 30).
Once attached to the surface of hepatocytes, siRNA-containing
complexes can enter the cells via receptor-mediated endocytosis.
The siRNAs must then escape from endosomes to elicit RNAi. To
accomplish efficient endosomal escape, we developed a strategy
that relies on the selective activation of a latent endosomolytic agent
in the acidic environment of the endosome (31). Selective activation
ensures that deleterious interactions with other membranes the
agent encounters before endocytosis are prevented. In our strategy,
amine groups on the endosomolytic agent are modified with a
maleic anhydride, creating acid-labilemaleamate bonds (32). These
bonds are cleaved within the acidic environment of the endosome,
unmasking the agent’s amines and activating its endosomolytic
capabilities (31). The endosomolytic agent used in the present study
is an amphipathic poly(vinyl ether) we previously developed termed
PBAVE, which is composed of butyl and amino vinyl ethers (33).
In this study, we use a bifunctional maleamate linkage to revers-
ibly attach the shielding agent PEG and the hepatocyte targeting
ligandN-acetylgalactosamine (NAG) toPBAVE.The siRNAcargo
itself is attached to PBAVE through a reversible disulfide linkage,
which prevents displacement of the siRNA from the polymer en
route to the target cell. We have named this delivery vehicle an
siRNA Dynamic PolyConjugate, to indicate the fact that the
siRNA, shielding agents, and targeting ligands are reversibly con-
Author contributions: D.B.R. and D.L.L. contributed equally to this work; D.B.R., D.L.L.,
D.H.W., S.C.W., J.J.K., J.E.H., and J.A.W. designed research; D.H.W., S.C.W., J.J.K., P.L.R.,
S.L.B., T.W.R., and Q.C. performed research; A.V.B. contributed new reagents/analytic tools;
D.B.R., S.C.W., and D.L.L. analyzed data; and D.B.R., D.L.L., and J.A.W. wrote the paper.
Conflict of interest statement: All of the authors except J.A.W. are employees of Mirus Bio.
This article is a PNAS Direct Submission.
Abbreviations: CDM, carboxy dimethylmaleic anhydride; iNOP, interfering nanoparticle;
NAG, N-acetylgalactosamine; SNALP, stable nucleic acid lipid particle.
†To whom correspondence may be addressed. E-mail: dave.rozema@mirusbio.com or
dave.lewis@mirusbio.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
0703778104/DC1.
© 2007 by The National Academy of Sciences of the USA
www.pnas.orgcgidoi10.1073pnas.0703778104 PNAS Early Edition
1 of 6
A
PP
LI
ED
BI
O
LO
G
IC
A
L
SC
IE
N
CE
S
A
PP
LI
ED
PH
YS
IC
A
L
SC
IE
N
CE
S