Antisense &
`Nucleic Acid Drug Development
`(The Antisense Journal)
`
`890526854688
`
`B'iDSEEBSHEB
`
`MI I ll lllll
`
`llllilililll
`
`,
`
`Volume 7. Number 3. June 1997
`ISSN 1087-2906
`
`Editors:
`
`Arthur M. Krieg, MD.
`C. A. Stein, MD, PhD.
`
`Sarepta Exhibit 1045, Page 1 of 14
`
`

`

`GENERAL INFORMATION
`
`Antisense & Nucleic Acid Drug Development, a bimonthly journal, discusses human-made substances and their effects
`on gene expression at the RNA and DNA levels. It provides a forum for basic researchers in molecular and cell biology,
`chemical synthesis, and applied therapeutics to discuss the development of new concepts and experimental approaches to
`understand and modulate gene activity.
`
`Antisense & Nucleic Acid Drug Development (ISSN: 1087-2906) is owned and published bimonthly by Mary Ann
`Liebert, Inc., 2 Madison Avenue, Larchmont, NY 10538. Telephone: (914) 834-3100; fax: (914) 834~3582; e-mail:
`liebert@pipeline.com Copyright© 1997 by Mary Ann Liebert, Inc. Printed in the United States of America.
`
`Postmaster: Send address changes to Antisense & Nucleic Aciff Drug Development c!o Subscription Department, Mary
`Ann Liebert, Inc., 2 Madison Avenue, Larchmont, NY 10538.
`
`Subscriptions should be addressed to the Publisher and are payable in advance. Rates for subscriptions are $148 per
`volume of 6 issues in the United States and Possessions, and $196 elsewhere. Subscriptions begin with the first issue of
`the current volume. Bulk subscriptions available upon request from the Publisher.
`
`Reprints, except special orders of 100 or more, are available from the authors.
`
`Information for Manuscript Submission is given elsewhere in the publication.
`
`Business Communications should be addressed to the Publisher.
`
`Advertising inquiries from within the United States or Canada should be addressed to Mary Ann Liebert, Inc,,
`2 Madison Avenue, Larchmont, NY 10538; telephone (914) 834-3100. Advertising inquiries from Europe and else(cid:173)
`where should be addressed to Hilary Turnbull, imPRESS, 2 Penrith Avenue, Glasgow G46 6LU, UK; telephone
`+44.141.620.0106, fax: +44.141.620.0055. All advertisements are subject to approval by the Publisher.
`
`Manuscripts should be directed to Arthur M. Krieg, the Editorial Office, Department of Internal Medicine, University of
`Iowa, 540 EMRB, Iowa City, IA 52242.
`
`All authorized papers and editorial news and comments, op1mons, findings, conclusions, or recommendations in
`Antisense & Nucleic Acid Drug Development are those of the author(s), and do not necessarily reflect the views of the
`journal and its publisher, nor does their publication in Antisense & Nucleic Acid Drug Development imply any endorse(cid:173)
`ment.
`
`Antisense & Nucleic Acid Drug Development is indexed in Index Medicus, Current Contents/Life Sciences,
`EMBASE, Biotechnology Citation Index, Science Citation Index, and SciSearch.
`
`Antisense & Nucleic Acid Drug Development is a Journal Club selection.
`
`

`

`Antisense &
`Nucleic Acid Drug Development
`
`Editors
`
`Arthur M. Krieg, M.D.
`Department of Internal Medicine
`University of Iowa
`540EMRB
`Iowa City, Iowa 52242
`(319) 335-6841
`Fax: (319) 335-6887
`E-mail: arthur-krieg@uiowa.edu
`
`C.A. Stein, M.D., Ph.D.
`Department of Medicine and
`Pharmacology
`Columbia University
`College of Physicians and Surgeons
`630 West 168th Street
`New York, NY 10032
`(212) 305-3606
`Fax: (212) 305-7348
`E-mail: stein@cuccfa.ccc.columbia.edu
`Editorial Board
`Patrick Iversen
`Omaha
`
`Sudhir Agrawal
`Shrewsbury, MA
`Serge L. Beaucage
`Bethesda
`Frank Bennett
`Carlsbad, CA
`Lauren Black
`Rockville, MD
`Marvin H. Caruthers
`Boulder
`Esther H. Chang
`Stanford
`P. Dan Cook
`Carlsbad, CA
`Stanley Crooke
`Carlsbad, CA
`Fritz Eckstein
`Gottingen, Germany
`Joachim Engels
`Franlifurt
`Robert P. Erickson
`Tucson
`Sergio Ferrari
`Modena, Italy
`Michael Gait
`Cambridge, UK
`Wayne L. Gerlach
`--Canberra, Australia
`Alan Gewirtz
`Philadelphia
`Donald Grierson
`Loughborough, England
`Sergei Gryaznov
`Foster City, CA
`Claude Helene
`Paris
`William R. Hiatt
`Davis, CA
`Jeffrey T. Holt
`Nashville
`Jean-Louis Imbach
`Montpellier, France
`Masayori Inouye
`Piscataway
`
`Nava Sarver
`Bethesda
`Kevin Scanlon
`Duarte, CA
`Karl-Hermann Schlingensiepen
`Gottingen, Germany
`Georg Sczakiel
`Heidelberg, Germany
`Hartmut Seliger
`Ulm, Germany
`Zoe Shabarova
`Moscow
`Barbara Ramsey Shaw
`Durham
`Hermona Soreq
`Jerusalem
`Wojciech Stec
`Lodz, Poland
`Martin Tabler
`Heraklion/Crete, Greece
`David M. Tidd
`Liverpool, UK
`J.-J. Toulme
`Bordeaux
`Paul Ts'o
`Baltimore
`Eugen Uhlmann
`Franlifurt
`Valentin V. Vlassov
`Novosibirsk, Russia
`Gerhart Wagner
`Uppsala, Sweden
`Richard Wagner
`Foster City, CA
`Daniel Weeks
`Iowa City
`
`Eric Wickstrom
`Philadelphia
`Paul Zamecnik
`Shrewsbury, MA
`Gerald Zon
`Hayward, CA
`
`Kuan-Teh Jeang
`Bethesda
`
`Rudolph L. Juliano
`Chapel Hill
`
`Ryszard Kole
`Chapel Hill
`
`Bernard Lebleu
`Montpellier, France
`
`Lee Leserman
`Marseille
`
`Robert L. Letsinger
`Evanston, IL
`L. James Maher III
`Rochester, MN
`
`Claude Malvy
`Vi/lejuif, France
`
`Dan Mercola
`San Diego
`Paul S. Miller
`Baltimore
`Ramaswamy Narayanan
`Nutley, NJ
`L.M. Neckers
`Bethesda
`Wolfgang Nellen
`Kassel, Germany
`
`Mike Nerenberg
`San Diego
`Eiko Ohtsuka
`Sapporo, Japan
`
`Robert Rando
`The Woodlands, TX
`
`John C. Reed
`La Jolla
`John J. Rossi
`Duarte, CA
`Esther Saison-Behmoaras
`Paris
`
`Founding Editor
`James W. Hawkins
`
`

`

`Antisense &
`Nucleic Acid Drug Development
`
`VOLUME 7
`
`NUMBER 3
`
`JUNE 1997
`
`Original Articles
`Rapid Measurement of Modified Oligonucleotide Levels in Plasma Samples with a Fluorophore Specific
`for Single-Stranded DNA. G.D. GRAY and E. WICKSTROM
`
`Tissue Distribution and Metabolism of the [32P]-Labeled Oligodeoxynucleoside Methylphosphonate(cid:173)
`Neoglycopeptide Conjugate, [YEE(ah-GalNAc)J]-SMCC-AET-pUmpI7, in the Mouse. J.J. RANGELAND,
`J.E. FLESHER, S.F. DEAMOND, Y.C. LEE, P.O.P. TS'O, and J.J. FROST
`
`A Specificity Comparison of Four Antisense Types: Morpholino, 2'-O-Methyl RNA, DNA, and
`Phosphorothioate DNA. D. STEIN, E. FOSTER, S.-B. HUANG, D. WELLER, and J. SUMMERTON
`
`In Vivo Metabolic Profile of a Phosphorothioate Oligodeoxyribonucleotide. J. TEMSAMANI, A. ROSKEY,
`C. CHAIX, and S. AGRAWAL
`
`Identification of a Phosphodiester Hexanucleotide That Inhibits HIV -1 Infection In Vitro on Covalent
`Linkage of Its 5'-End with a Dimethoxytrityl Residue. H. FURUKAWA, K. MOMOTA, T. AGATSUMA,
`I. YAMAMOTO, S. KIMURA, and K. SHIMADA
`
`Delivery of Oligoribonucleotides to Human Hepatoma Cells Using Cationic Lipid Particles Conjugated to
`Ferric Protoporphyrin IX (Heme). G.B. TAKLE, A.R. THIERRY, S.M. FLYNN, B. PENG, L. WHITE,
`W. DEVONISH, R.A. GALBRAITH, A.R. GOLDBERG, and S.T. GEORGE
`
`Review Article
`Morpholino Antisense Oligomers: Design, Preparation, and Properties. J. SUMMERTON and D. WELLER
`
`Presentations from the First NIH Symposium on Therapeutic Oligonucleotides
`
`Introduction. Y.S. CHO-CHUNG
`
`Background of the Antisense Oligonucleotide Approach to Chemotherapy. P. ZAMECNIK
`
`Recruiting the 2-5A System for Antisense Therapeutics. P.F. TORRENCE, W. XIAO, G. LI, H. CRAMER,
`M.R. PLAYER, and R.H. SILVERMAN
`
`Controversies in the Cellular Pharmacology of Oligodeoxynucleotides. C.A. STEIN
`
`Recombinational Repair of Genetic Mutations. A. COL~-STRAUSS, A. NOE, and E.B. KMIEC
`
`Antisense-Protein Kinase A: A Single-Gene-Based Therapeutic Approach. Y.S. CHO-CHUNG,
`M. NESTEROVA, A. KONDRASHIN, K. NOGUCID, R. SRIVASTAVA, and S. PEPE
`
`Antisense c-myc Inhibition of Lymphoma Growth. E. WICKSTROM
`
`133
`
`141
`
`151
`
`159
`
`167
`
`177,
`
`187 .
`
`197
`
`199
`
`203
`
`207
`
`211,
`
`217
`
`225
`
`(continued)
`
`

`

`Identification and Characterization of Second-Generation Antisense Oligonucleotides. N.M. DEAN and
`R.H. GRIFFEY
`
`Protein Kinase C as a Target for Cancer Therapy. R.I. GLAZER
`
`Pharmacology of Therapeutic Oligonucleotides. R.B. DIASIO and R. ZHANG
`
`In Vivo Pharmacokinetics of Phosphorothioate Oligonucleotides Containing Contiguous Guanosines.
`S. AGRAWAL, W. TAN, Q. CAI, X. XIE, and R. ZHANG
`
`Ex Vivo Bone Marrow Purging with Oligonucleotides. R.C. BERGAN
`
`Antisense Transfonning Growth Factor-.{31 in Wound Healing. H.-T. CHUNG, B.-M. CHOI, C.-D. JUN,
`S.-D. PARK, and J.-S. RIM
`
`229
`
`235
`
`239
`
`245
`
`251
`
`257
`
`Instructions for Authors
`
`

`

`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`ANTISENSE & NUCLEIC ACID DRUG DEVELOPMENT 7:187-195 (1997)
`Mary Ann Liebert, Inc.
`
`Review Article
`
`Morpholino Antisense Oligomers: Design, Preparation, and
`Properties
`
`JAMES SUMMERTON and DWIGHT WELLER
`
`ABSTRACT
`
`Antisense promised major advances in treating a broad range of intractable diseases, but in recent years
`progress has been stymied by technical problems, most notably inadequate specificity, ineffective delivery into
`the proper subcellular compartment, and unpredictable activity within cells. Herein is an overview of the de(cid:173)
`sign, preparation, and properties of Morpholino oligos, a novel antisense structural type that solves the se(cid:173)
`quence specificity problem and provides high and predictable activity in cells. Morpholino oligos also exhibit
`little or no nonantisense activity, afford good water solubility, are immu!le to nucleases, and are designed to
`have low production costs.
`
`INTRODUCTION
`
`OLlGONUCLEOTIDES, OLIGONUCLEOTIDE ANALOGS, and other
`
`sequence-specific binding polymers designed to block
`translation of selected messenger RNAs (the sense strand) are
`commonly called "antisense oligos." Development of such oli(cid:173)
`gos for therapeutic applications, which constitutes the epitome
`of rational drug design, entails selecting a target genetic se(cid:173)
`quence unique and critical to the pathogen or pathogenic state
`one wishes to treat. One then assembles an oligomer of genetic
`bases (adenine, cytosine, guanine, and thymine or uracil) com(cid:173)
`plementary to that selected sequence. When such an antisense
`oligo binds to its targeted disease-causing sequence, it can inac(cid:173)
`tivate thiiJ target and thereby; alleviate the disease.
`Antisense oligos offer the prospect of safe and effective ther(cid:173)
`apeutics for a broad rangi_ of intractable diseases. Nonetheless,
`developing therapeutics that function by a true antisense mech(cid:173)
`anism presents a number of forbidding challenges. The oligos
`should achieve adequate efficacy at a concentration attainable
`within the cells of the patient. They should inhibit their selected
`target sequences without concomitant attack on any other se-,
`quences in the patient's pool of approximately 200 million
`bases of unique-sequence RNA. They should be stable in extra(cid:173)
`cellular compartments and within cells. They must be deliver(cid:173)
`able into the cellular. compartment(s) containing their targeted
`sequences. They should be adequately soluble in aqueous solu(cid:173)
`tion. They should exhibit little or no toxicity at therapeutic con(cid:173)
`centrations. Finally, they should be affordable, reflecting the in-
`
`creasing awareness that health care, even for life-threatening
`conditions, should not expend an excessive portion of society's
`resources.
`First-generation antisense oligos comprised natural genetic
`material (Belikova et al., 1967; Zamecnik and Stephenson,
`1978; Summerton, 1979) and often contained crosslinking
`agents for binding their targets irreversibly (Summerton and
`Bartlett, 1978a,b). As the design challenges became more fully
`appreciated, a number of nonnatural antisense structural types
`(Fig. 1) were developed in an effort to improve efficacy, stabil(cid:173)
`ity, and delivery. Of particular note are the early non-ionic
`DNA analogs developed by Miller and Ts'o, including phos(cid:173)
`photriester-linked DNA (Miller, 1989a) and methylphospho(cid:173)
`nate-linked DNA (Miller, 1989b). Other nucleic acid analogs of
`note include carbamate-linked DNA (Stirchak et al., 1987),
`phosphorothioate-linked DNA (Stein and Cohen, 1989), phos(cid:173)
`phoroarnidate-linked DNA (Froehler et al., 1988), a-DNA
`(Rayner et al., 1989), and 2' -0-methyl RNA (Shibahara et al.,
`1989). Figure 1B shows several novel antisense types that no
`longer resemble nucleic acids. These oligos contain acyclic
`backbone moieties, including nylon (Weller et al. 1991; Huang
`et al., 1991), the exceptionally high-affinity peptide nucleic
`acids (PNAs) (Egholm et al., 1992), and related types (Sum(cid:173)
`merton and Weller, 1993a).
`Although each of these newer structural types provides one
`or more significant advantages over the first-generation oligos,
`none yet appear to provide the full combination of properties
`needed in antisense therapeutics for clinical applications.
`
`ANTIVIRALS Inc., Corvallis, OR 97333.
`
`187
`
`

`

`188
`
`SUMMERTON AND WELLER
`
`A. DNA, RNA & ANALOGS
`
`B. ACYCLIC
`
`DNA/RNA
`
`Carbamate
`
`a-DNA
`
`H
`
`H
`
`~ B
`DNA
`H
`Phosphotriester
`H
`Methylphosphonate H
`Phosphorothioate H
`Phosphoroamidate H
`RNA
`OH
`2' 0-methyl RNA
`OCH3
`
`!
`0
`OCH3
`CH3
`s
`NHCH3
`0
`0
`
`FIG. 1. Representative antisense structural types.
`
`Herein we describe the design considerations used in develop- munity to nucleases, good aqueous solubility, and low produc(cid:173)
`tion costs.
`ing a novel Morpholino structural type (Fig. 2), which affords
`antisense oligos having very high efficacy and specificity, im-
`
`ol:1B
`o=f-N( ol:rB
`
`B = adenine, cytosine, guanine, uracil
`
`FIG. 2. Morpholino oligo structure.
`
`DESIGN
`
`Backbone structure
`
`A dominant consideration in the design of most antisense oli(cid:173)
`gos has been to devise a structure that provides resistance to nu(cid:173)
`cleases while still resembling natural nucleic acids as closely as
`possible. This conservative approach has spawned a number of
`DNA analogs (Fig. IA) that may be unduly expensive for rou(cid:173)
`tine applications requiring systemic delivery. The high cost of
`DNA and its analogs is due in part to the low abundance of
`DNA in production-scale source material and the difficulty in
`cleaving DNA to the deoxyribonucleosides required for prepar(cid:173)
`ing DNA analogs. An additional factor in their high cost is the
`complexity and expense of coupling to hydroxyls, required in
`forming the phosphoester intersubunit linkages of most DNA
`analogs.
`Rather than trying to solve inherent cost problems after a
`structural type has been developed, a better approach is to in(cid:173)
`corporate fundamental cost advantages in the initial structural
`design stage. Following this strategy, we reasoned that more af(cid:173)
`fordable antisense oligos might be possible if inexpensive ri(cid:173)
`bonucleosides could be exploited as starting material. The or(cid:173)
`der-of-magnitude cost advantage of ribonucleosides relative to
`deoxyribonucleosides (Summerton, 1992) derives from the six(cid:173)
`fold greater abundance of RNA relative to DNA in production(cid:173)
`scale source material (e.g., yeast cake) and the ease of cleaving
`
`

`

`MORPHOLINO ANTISENSE OLIGOS
`
`RNA to its component ribonucleosides. It is noteworthy that ri(cid:173)
`bonucleosides are now directly available from special excreting
`strains of yeast, further reducing their cost. However, the use of
`ribonucleosides for preparation of RNA and RNA analogs pre(cid:173)
`sents two serious problems. First, during oligo assembly, one
`must selectively couple either the 2' or the 3' hydroxyl. This is
`typically achieved in a relatively expensive manner by selec(cid:173)
`tively masking the 2' hydroxyl with a cleavable or noncleav(cid:173)
`able moiety. The second problem is that coupling to the 3' hy(cid:173)
`droxyl of the riboside is even more difficult and expensive than
`the corresponding coupling of deoxyribonucleosides.
`We envisioned that these problems could be circumvented by
`converting the riboside moiety to a morpholine moiety (Stir(cid:173)
`chak et al., 1989; Summerton, 1990) (Fig. 3). Although
`oligomers assembled from such Morpholino subunits differ
`substantially from DNA, RNA, and analogs thereof, our initial
`modeling studies carried out in 1985 suggested that such novel
`Morpholino-based oligomers might constitute useful and highly
`cost-effective antisense agents. The simple and inexpensive ri(cid:173)
`bose to morpholine conversion shown in Figure 3 replaces two
`poor nucleophiles (the 2' and 3' hydroxyls) with a single good
`nucleophile (the morpholine nitrogen) and allows oligo assem(cid:173)
`bly via simple and efficient coupling to the morpholine nitrogen
`without the expensive catalysts and postcoupling oxidation
`steps required in the production of most DNA-like antisense
`oligos. It is noteworthy that in spite of the relatively low nucle(cid:173)
`ophilicity of the morpholine nitrogen (pK0 = 5.75), we still typ(cid:173)
`ically achieve coupling efficiencies of 99.7% without using cat(cid:173)
`alysts.
`
`ioate-linked DNA (S-DNA), appreciably higher for DNA, and
`highest for the Morpholino oligo.
`
`189
`
`PREPARATION
`
`Oligo assembly
`
`Although phosphorodiarnidate-linked Morpholino oligos can
`be assembled by a variety of methods, one relatively simple
`method that has proved effective (Summerton and Weller,
`1993b) entails protection and activation of the Morpholino sub(cid:173)
`unit (Fig. 5A). The activated subunits can be stored at low tem(cid:173)
`peratures for many months without significant breakdown.
`Whereas they are relatively resistant to hydrolysis, they react
`rapidly (T112 of 1-2 minutes) with the morpholine nitrogen of
`growing chains on a 1 % crosslinked polystyrene synthesis sup(cid:173)
`port loaded at 500 µ,Mfg of resin, with coupling efficiencies
`typically about 99.7%. A preferred oligo assembly cycle (Sum(cid:173)
`merton and Weller, 1993b) is shown in Figure 5B. It is note(cid:173)
`worthy that in large-scale syntheses, excess activated subunit
`used in the coupling step can be recovered and reused, effecting
`a further substantial reduction in production costs.
`Because of cheaper starting materials and simpler, more effi(cid:173)
`cient oligo assembly, we estimate that in large-scale produc(cid:173)
`tion, the cost of these Morpholino antisense oligos will be at
`least an order of magnitude lower than the cost of correspond(cid:173)
`ing DNA analogs (Summerton, 1992).
`
`lntersubunit linkage
`
`PROPERTIES
`
`We have assessed a substantial number of intersubunit link(cid:173)
`age types, including the carbonyl, sulfonyl, and phosphoryl
`linkages (Fig. 4) (Summerton and Weller, 1991, 1993a,b; Stir(cid:173)
`chak et al., 1989). Although Morpholino oligos containing a
`number of such linkages provide effective binding to targeted
`genetic sequences, consideration of cost and ease of synthesis,
`chemical stability, aqueous solubility, and affinity and homo(cid:173)
`geneity of binding to RNA led us to focus on the phosphorodi(cid:173)
`amidate shown in Figure 2 as our principle linkage type for oli(cid:173)
`gos targeted against single-stranded RNA sequences. These
`non-ionic phosphorodiamidate-linked Morpholino oligos ex(cid:173)
`hibit quite good binding to complementary nucleic acids, par -
`ticularly RNA sequences. Table 1 compares the temperature of
`melting (Tm) values at physiologic salt concentration for identi(cid:173)
`cal-sequence 20-mer oligos of three different antisense struc(cid:173)
`tural types paired with their complementary RNA. As seen in
`Table 1, RNA binding affinity is lowest for the phosphoroth-
`
`Solubility
`
`For an antisense oligo to have effective access to its target se(cid:173)
`quence within the cytoplasm of a cell, the oligo should show
`reasonable water solubility. Good water solubility may also
`prove essential for systemic delivery of antisense oligos. Con(cid:173)
`ventional wisdom in the antisense field is that non-ionic anti(cid:173)
`sense oligos invariably show poor water solubility. In this re(cid:173)
`gard, it is interesting that a Morpholino dimer containing a rigid
`carbamate linkage shows little or no base stacking (Kang et al.,
`1992), and in the absence of special solubilizing groups, Mor(cid:173)
`pholino oligomers containing such carbamate linkages are quite
`insoluble in aqueous solutions (Stirchak et al., 1989). In con(cid:173)
`trast, phosphorodiarnidate-linked Morpholino oligos of the type
`shown in Figure 2 show excellent base stacking (Kang et al.,
`1992) and are several orders of magnitude more soluble in
`aqueous solutions. To illustrate the exceptional aqueous solu-
`
`B•
`
`HOH
`HO OH
`
`Nal04
`..,
`
`HOjo
`y
`~o
`
`o,,.
`
`B*
`
`0T 0
`' ( e-
`HOAN~H
`H
`e• = N4-benzoylcytosine, N6-benzoyladenine, N2-phenylacetylguanine, uracil
`
`"
`
`NaCNBH, "°lo)
`
`N
`H
`
`FIG. 3. Conversion of ribonucleoside to Morpholino subunit.
`
`

`

`190
`
`SUMMERTON AND WELLER
`
`carbonyl
`
`sulfonyl
`
`phosphoryl
`
`B = purine or pyrimidine
`X=O,S
`Y= 0 , N-CH3
`Z = alkyl
`O-alkyl
`S-alkyl
`NH2
`NH(alkyl)
`NH(O-alkyl)
`N(alkylh
`N(alkyl)(O-alkyl)
`(alkyl includes substituted alkyl]
`
`Ol01 B
`x=c
`o'(r B
`
`N
`I
`
`I
`
`olo1 B
`Ol01 B
`I x=P-Z
`o=s=o
`B Vt) B
`y'(r
`
`N
`I
`I
`
`N
`
`I
`
`FIG. 4.
`
`Intersubunit linkage types for Morpholino oligos.
`
`bility of Morpholino oligos of this type, we have dissolved 263
`mg of a heteromeric 22-mer of the sequence 5'-GCUCGCA(cid:173)
`GACUUGUUCCAUCAU in 1 ml of water (36 rnillimolal) at
`20°c without reaching saturation.
`We suggest that the poor water solubility of the carbamate(cid:173)
`linked Morpholino oligos results at least in part from the diffi(cid:173)
`culty of inserting the hydrophobic faces of the unstacked bases
`into an aqueous environment. In contrast, it seems likely that
`the excellent water solubility of the phosphorodiarnidate-linked
`Morpholino oligos is a consequence of effective shielding of
`these hydrophobic faces from the polar solvent because of good
`stacking of the bases.
`
`Biologic stability
`
`To achieve reasonable efficacy, an antisen~e oligo should not
`be degraded rapidly either extracellularly or within cells. In this
`regard, it has been demonstrated that DNA and 2' -O-methyl
`RNA are rapidly degraded and phosphorothioate DNA is
`slowly degraded by nucleases in blood and within cells (Hoke
`et al., 1991; Morvan et al., 1993). Although resistance to nucle(cid:173)
`olytic degradation can be improved by adding special groups to
`the termini (Cazenave et al., 1987) or by incorporating a few
`nuclease-resistant intersubunit linkages near each end (Larrouy
`et al., 1992), we believe a better solution, on the basis of both
`function and cost, is to use a backbone structure that is inher(cid:173)
`ently immune to a broad range of degradative enzymes present
`in the blood and within cells. A further advantage of using a
`backbone structure that is not degraded in the body is that it
`avoids concerns that modified nucleosides or nucleotides re(cid:173)
`sulting from degradation of an antisense oligo might be toxic or
`might be incorporated into cellular genetic material and thereby
`lead to mutations or other undesired biologic effects.
`In experiments detailed elsewhere (Hudziak et al., 1996), it is
`demonstrated that Morpholino phosphorodiamidate oligos of
`
`TABLE 1. MELTING TEMPERATURES OF RNA/Ouoo DUPLEXES
`
`RNA/S-DNA
`RNA/DNA
`RNA/Morpholino
`
`68.5°C
`77.3°C
`g1.3oc
`
`the type shown in Figure 2 are immune to a wide range of nu(cid:173)
`cleases, including DNase I (an endonuclease that cleaves both
`single-stranded and double-stranded DNA), DNase II (cleaves
`between the 5' oxygen and the phosphorus of DNA linkages),
`RNase A (cleaves on the 3' side of pyrimidines), RNase Tl
`(cleaves on the 3' side of guanines), nuclease Pl (cleaves sin(cid:173)
`gle-stranded RNA and DNA), phosphodiesterase (3' exonucle(cid:173)
`ase for both RNA and DNA), Mung bean nuclease (cleaves sin(cid:173)
`gle-stranded RNA and DNA), and benzonase (cleaves both
`single-stranded and double-stranded RNA and DNA, including
`linear, circular, and supercoiled). These Morpholino oligos
`have also been found to be immune to pronase E, proteinase K,
`and pig liver esterase, as well as degradative enzymes in serum
`and a liver homogenate.
`
`Antisense efficacy
`
`Because of the excellent RNA binding affinity of oligos of
`this phosphorodiamidate-linked Morpholino structural type, it
`seemed likely Morpholino oligos would be effective in block(cid:173)
`ing translation of their targeted mRNAs, and this has been
`found to be the case. In cell-free translation experiments using a
`sensitive luciferase reporter, we have demonstrated that a Mor(cid:173)
`pholino oligo 25 subunits in length, in both the presence and ab(cid:173)
`sence ofRNase H, inhibits its targeted mRNA somewhat better
`than the corresponding S-DNA oligo in the presence of added
`RNase H, with both showing good efficacy at concentrations of
`10 nM and above. Representative translational inhibition re(cid:173)
`sults are shown in Figure 6 (Summerton et al., 1997). A similar
`comparison of Morpholino and S-DNA antisense oligos tar(cid:173)
`geted against murine tumor necrosis factor-a (TNF-a) mRNA
`in a cell-free translation system also showed greater activity for
`the Morpholino oligos (Taylor et al., 1996).
`
`Specificity
`
`In the early days of anti sense research, one of the most com(cid:173)
`pelling arguments for antisense therapeutics was their promise
`of exquisite specificity for their targeted genetic sequences.
`However, as the most synthetically accessible antisense struc(cid:173)
`tural types (DNA and S-DNA) have come into broad use, it has
`become clear that these two structural types provide reasonable
`
`

`

`MORPHOLINO ANTISENSE OLIGOS
`
`191
`
`A.
`
`Cl
`
`I O=P-N/
`I
`'
`
`Cl
`
`B* = N4-benzoylcytosine, N6-benzoyladenlne, N2-phenylacetylguanine, uracil
`
`8.
`
`Cl
`
`I
`o=P-N(
`I
`
`o-"(j""
`
`s-
`
`N
`
`"- I
`
`I.
`
`FIG. S. Protection, activation, and coupling of Morpholino subunits.
`
`100
`
`80
`
`C
`
`0 .. 60
`
`..... Morpholino
`-0- S-DNA
`
`sequence specificity within only a very narrow concentration
`range (ANTIVIRALS Inc., 1993; Stein and Cheng, 1993).
`We believe a key factor responsible for the low specificity of
`DNA and S-DNA oligos is their RNase H competency; that is,
`DNA and S-DNA form duplexes with complementary RNA
`that are readily cleaved by RNase H, an enzyme widely distrib(cid:173)
`uted in living organisms. The specificity problem arises be(cid:173)
`cause DNA/RNA and S-DNA/RNA duplexes as short as 5 base
`pairs in length are cleaved by RNase H (Crouch and Dirksen,
`1982). Presuming about 6% of the genome is transcribed in
`higher animals, the patient's RNA pool will comprise about 200
`million bases of unique-sequence RNA. With this level of se(cid:173)
`quence complexity, it is inevitable that antisense oligos will
`form many short transient duplexes with partially complemen(cid:173)
`tary nontarget sequences of inherent cellular RNAs. Cleavage
`of the RNA strand of such nontarget duplexes by endogenous
`RNase H (Larrouy et al., 1992; Cazenave et al., 1989) is ex(cid:173)
`pected to cause significant disruption of normal cellular transla(cid:173)
`tion. As this cleavage process releases the DNA or S-DNA in
`0
`100
`200
`300
`its original form, such oligos can continue the cycle of tran(cid:173)
`siently pairing with additional nontarget cellular RNA se(cid:173)
`Oligo Concentration (nM)
`quences, cleavage of the RNA strand, and release of the anti(cid:173)
`sense oligo. As a consequence, essentially every RNase
`FIG. 6. Cell-free efficacy of Morpholino and S-DNA antisense oli- H-competent oligo is expected to cleave hundreds to thousands
`gos.
`of species of inherent cellular RN As.
`
`·-..Q
`
`.c
`C
`
`40
`~ 0
`20
`
`0
`
`

`

`192
`
`SUMMERTON AND WELLER
`
`A second factor expected to contribute to superior specificity
`of Morpholino oligos relative to RNase H-competent types is
`that RNase H-independent oligos have far fewer potential tar(cid:173)
`gets in the inherent pool of cellular RNA. This is because most
`antisense structural types that do not support RNase H cleavage
`of their RNA targets have been found to be effective in block(cid:173)
`ing translation of their targeted mRNAs only when said oligos
`are complementary to sequences in the 5' leader region of that
`mRNA or when they are targeted against other special sites,
`such as splice junctions and transport signals [e.g., methylphos(cid:173)
`phonate DNA (Walder and Walder, 1988), a-DNA (Rayner et
`al., 1989), 2' -O-methyl RNA (Shibahara et al., 1989), and Mor(cid:173)
`pholino (Summerton et al., 1997)]. We estimate that such spe(cid:173)
`cial targetable regions constitute on the order of 2%-5% of the
`sequeces in the cellular RNA pool. Presumably, this targeting
`limitation reflects the ability of ribosomes to displace essen(cid:173)
`tially all antisense oligos during translocation down the coding
`region of mRNAs.
`Because an antisense oligo that does not support RNase H
`cleavage cannot effectively block functioning of an RNA when
`said oligo is bound to sequences outside of special targetable
`regions, such an oligo only needs to distinguish its target se(cid:173)
`quence from those 2%-5% of the cellular RNA sequences com(cid:173)
`prising special targetable regions. In contrast, antisense oligos
`that form RNase H-cleavable duplexes with RNA can be effec(cid:173)
`tive when targeted essentially anywhere along an RNA tran(cid:173)
`script (Walder and Walder, 1988), presumably because RNase
`H cleavage at the target site of the antisense oligomer destroys
`the RNA, rendering moot the oligo displacement capability of
`translocating ribosomes. Accordingly, RNase H-competent oli(cid:173)
`gos (DNA and S-DNA) face the much greater specificity chal(cid:173)
`lenge of distinguishing selected target sequences from essen(cid:173)
`tially the entire pool of cellular RNA sequences. As a
`consequence, RNase H-independent oligos, such as Morpholi(cid:173)
`nos, should enjoy a 20-fold to SO-fold advantage in sequence
`
`specificity because of this more than order-of-magnitude reduc(cid:173)
`tion in the number of inherent nontarget cellular sequences of
`any given length that they can inhibit.
`A third factor compromising the specificity of S-DNA oligos
`is their promiscuous binding to proteins (Krieg and Stein,
`1995), including components of the cell's replication, transcrip(cid:173)
`tion, and translation machinery.
`Given these factors expected to limit the sequence specificity
`ofRNase H-competent antisense structural types, particularly S(cid:173)
`DNA, we set out to compare sequence specificities of S-DNA
`anft Morpholino antisense oligos. To this end, we carried _out
`stringent specificity assays in a cell-free translation system using
`two oligos of each structural type (Summerton et al., 1997). In
`these experiments, one oligo was perfectly complementary to its
`target mRNA to provide a measure of the total inhibition af(cid:173)
`forded by that oligo type. The other oligo incorporated 4 mis(cid:173)
`pairs to that same mRNA target sequence, with the longest run
`of perfect pairing comprising 10 contiguous base pairs, to prer
`vide an estimate of the low-specificity component of the inhibi(cid:173)
`tion. The difference between these two inhibition values at each
`concentration than provided a measure of the high-specificity
`component, which we denote as "sequence-specific inhibition."
`Figure 7 (experimental as in Summerton et al., 1997) shows
`that the S-DNA oligo achieved reasonable efficacy at concen(cid:173)
`trations above about IO nM, but the sequence-specific compo(cid:173)
`nent of its inhibition dropped below 50% at concentrations of
`only 100 nM and higher. The corresponding Morpholino oligo
`achieved even better efficacy at IO nM while maintaining good
`sequence specificity through 10,000 nM, the highest concentra(cid:173)
`tion tested. Thus, in this stringent test of specificity, the Mor(cid:173)
`pholino oligo achieved highly effective and specific antisense
`activity over a concentration range more than two orders of
`magnitude greater than the concentration range wherein the
`corresponding S-DNA achieved reasonable efficacy and speci(cid:173)
`ficity.
`
`(.)
`
`5
`(.)
`(1>
`C. C:
`cp 0
`(1> · -
`( . )~
`C: .c
`(1> · -
`::::, .c
`C" .5
`(1>
`U)
`
`~ 0
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`♦ Morpholino
`
`-0- 5-DNA
`
`0
`
`2
`
`4
`
`6
`
`8
`
`10
`
`Oligo Concentration ( µ M )
`
`FIG. 7. Sequence specificity of Morpholino and S-DNA oligos.
`
`

`

`MORPHOLINO ANTISENSE OLIGOS
`
`193
`
`Taylor et al., (1996) have reported that S-DNAs targeted
`against TNF-a mRNA showed very poor sequence specificity
`in a cell-free translation system, whereas the corresponding
`Morpholino oligos afforded good specificity over the full range
`tested.
`
`Activity in cells
`
`For effective biologic activity, an antisense oligo must gain
`entry into the cellular compartments where the target genetic
`sequence is