nature biotechnology • VOLUME 20 • DECEMBER 2002 • www.nature.com/naturebiotechnology

RESEARCH ARTICLE

1228

Systemically delivered antisense oligomersupregulate gene expression in mouse tissues

Peter Sazani1,2, Federica Gemignani1, Shin-Hong Kang1, Martin A. Maier3, Muthiah Manoharan3, Magnus Persmark4, Donna Bortner4, and Ryszard Kole1,2*

Published online 11 November 2002; doi:10.1038/nbt759

Systemically injected 2′-O-methoxyethyl (2′-O-MOE)-phosphorothioate and PNA-4K oligomers (peptidenucleic acid with four lysines linked at the C terminus) exhibited sequence-specific antisense activity in a num-ber of mouse organs. Morpholino oligomers were less effective, whereas PNA oligomers with only one lysine(PNA-1K) were completely inactive. The latter result indicates that the four-lysine tail is essential for the anti-sense activity of PNA oligomers in vivo. These results were obtained in a transgenic mouse model designedas a positive readout test for activity, delivery, and distribution of antisense oligomers. In this model, theexpressed gene (EGFP-654) encoding enhanced green fluorescence protein (EGFP) is interrupted by anaberrantly spliced mutated intron of the human β-globin gene. Aberrant splicing of this intron preventedexpression of EGFP-654 in all tissues, whereas in tissues and organs that took up a splice site-targeted anti-sense oligomer, correct splicing was restored and EGFP-654 expression upregulated. The sequence-specificability of PNA-4K and the 2′-O-MOE oligomers to upregulate EGFP-654 provides strong evidence that sys-temically delivered, chemically modified oligonucleotides affect gene expression by sequence-specific trueantisense activity, validating their application as potential therapeutics.

Sequencing of the human genome1,2 and parallel analysis ofexpressed sequence tag (EST) libraries indicate that 35–65% of allgenes code for alternatively spliced pre-mRNAs3 that generatemultiple splice variants. Alternative splicing is thus a major con-tributor to the vast diversity of proteomes, and its modificationemerges as an important approach for controlling gene expressionwith potential therapeutic outcomes4,5. As a result of their highsequence specificity, antisense oligonucleotides are particularlywell suited for modulation of splicing. This was shown in culturedcell lines for several clinically relevant targets such as the genesencoding Bcl-x6,7, cystic fibrosis transmembrane conductance reg-ulator (CFTR)8, dystrophin9,10, tau11, the interleukin-5 receptor12,and, recently, in erythroid progenitor cells of thalassemiapatients13,14. In all cases, the oligonucleotides were targeted at oradjacent to the splice sites, and the effects relied on their efficientdelivery to the cell nuclei, where pre-mRNA splicing occurs.

Most of the antisense oligonucleotides undergoing clinical trialscontain 2′-deoxynucleotide-phosphorothioate backbones15. Thesemolecules are well characterized in terms of pharmacokinetics16 andexhibit therapeutically relevant effects in preclinical and clinicalstudies17. Nevertheless, in some cases their sequence specificity andantisense mechanisms have been questioned18. Improved chemicalmodifications have been introduced, including peptide nucleic acid(PNA), bridged or locked nucleic acids, morpholino, phospho-roamidate, methylphosphonate, base-modified, 2′-modified, and3′-conjugated derivatives19. However, knowledge of the in vivo phar-macological properties of these novel compounds is still limited.Therefore, we have tested in vivo, in a transgenic mouse modeldesigned to provide a functional assay for antisense activity, three

types of oligomers: (2′-O-MOE)-phosphorothioate, morpholinos,and PNAs. The PNAs carried one or four lysines (PNA-1K, PNA-4K) at their C terminus, which conferred a positive charge, whereasthe remaining two were negatively charged or neutral. The PNA-4Koligomer showed high activity in the kidney, liver, and small intes-tine, with lower activity in the lung, heart, and skeletal muscle. Inagreement with previous pharmacokinetic studies20, 2′-O-MOEoligonucleotides were also active, whereas morpholino oligomersshowed modest effects in most tissues. These data indicate that thesystemically delivered PNA-4K oligomer has superior properties in vivo and that the mouse model employed here provides a usefultool in development of improved antisense compounds.

ResultsThe EGFP-654 transgenic mouse. We recently developed a cellularassay in which the upregulation of an enhanced green fluorescentprotein (EGFP) reporter provided direct quantitative measurementof the sequence-specific antisense activity of tested oligomers21.Here, we adapted the EGFP-654-based assay (EGFP-654, see Fig. 1)for in vivo application, generating a mouse model in which theEGFP-654 transgene, cloned under chicken β-actin promoter22, isexpressed throughout the body (Fig. 2A). As a result, the functionaleffects of the same oligomer can be simultaneously evaluated innumerous tissues. As a positive control for EGFP production, wegenerated a mouse that expressed a construct with a correctly splicedwild-type β-globin intron (EGFP-WT), as shown by RT-PCR of totalRNA from several organs (Fig. 2B, 87 base pairs). The pattern oftransgene expression was similar in the EGFP-654 mouse except thatEGFP-654 mRNA was almost exclusively spliced aberrantly (Fig. 2A,

1Lineberger Comprehensive Cancer Center and 2Department of Pharmacology, University of North Carolina, Chapel Hill, NC 27599. 3Isis Pharmaceuticals, Inc.,Carlsbad, CA 92008. 4GlaxoSmithKline Research and Development, Research Triangle Park, NC 27709. *Corresponding author (kole@med.unc.edu).

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reb

iote

chn

olo

gy

RESEARCH ARTICLE

www.nature.com/naturebiotechnology • DECEMBER 2002 • VOLUME 20 • nature biotechnology 1229

160 base pairs). In the EGFP-WT mouse, bright green fluorescencewas detected in every examined tissue (Supplementary Fig. 1online). As expected, no substantial signal was detected in similarsamples from the EGFP-654 mouse (see Fig. 4, row 1).

The results shown in Figure 2 and in Supplementary Figure 1online indicate that the transgene was expressed in all tissues inboth the EGFP-654 and EGFP-WT mice. However, because theactual target of antisense oligomers that shift splicing is pre-mRNA, its expression was examined by RT-PCR with an intron-specific primer in both mouse lines. In the EGFP-654 mouse (Fig. 2C), the pre-mRNA was readily detectable in all tissues, withsomewhat lower amounts in the bone marrow, skin, and brain. Incontrast, very little pre-mRNA from the EGFP-WT mouse wasdetected under the same RT-PCR conditions (Fig. 2D). Theseresults suggest that the wild-type intron was spliced rapidly,resulting in low steady-state levels of pre-mRNA, whereas splicingof the IVS2-654 intron was less efficient, resulting in accumulationof this species. Importantly, the results indicate that in most tis-sues of EGFP-654 mouse, the target pre-mRNA was present andshould be correctable by antisense oligomers that block aberrantsplice sites. Thus, the level of EGFP translated from the correctlyspliced mRNA generated by the antisense oligomers should beproportional to their concentration and antisense activity.

Ex-vivo treatment of cells from EGFP-654 mice. To determinewhether cells derived from the EGFP-654 mice respond to anti-sense treatment, primary fibroblasts and hepatocytes were cul-

tured and treated with an 18-mer 2′-O-methyl oligoribonucleo-side-phosphorothioate, ON-654 (oligomer 1, Table 1), delivered ina complex with cationic lipids. Mock-treated (lipid only) cellsexhibited little or no fluorescence, whereas antisense oligonu-cleotide treatment resulted in dose-dependent increases in EGFPproduction (Fig. 3A). A control oligonucleotide (oligomer 6,Table 1) targeted to a region of the intron downstream of the 654mutation was ineffective, demonstrating sequence specificity ofthe observed effects. Only cells treated with oligomer 1 showed arobust increase in the correct EGFP mRNA (Fig. 3B), with maxi-mum correction occurring at 0.1 µM oligonucleotide for fibrob-lasts and hepatocytes (Fig. 3B, lanes 4 and 10, respectively).Correction of splicing is reduced at 0.3 µM, as a result of animproper charge ratio between the oligonucleotide and the con-stant amount of lipid used in the experiment. At maximum levels,the percentage of EGFP mRNA was ∼ 40%. These data confirm that

in cells from the EGFP-654 transgenic mouse,the aberrantly spliced pre-mRNA is cor-rectable by antisense oligonucleotides, pro-viding a functional assay for in vivo studies ofthese compounds.

Systemic delivery of antisense oligomersto the EGFP-654 mouse. To determinewhether chemically modified oligomersexhibit antisense effects in vivo, we subjectedthe EGFP-654 mice to daily 50 mg/kg i.p.oligomer injections for one or four days.Most antisense oligomers are rapidly clearedfrom circulation when administered intra-venously23–25. We therefore chose i.p. injectionassuming a broader plasma area under the

Table 1. Antisense oligomers

Oligomer Sequence 5′ → 3′ Target site Backbone

1 GCT ATT ACC TTA ACC CAG 654 2′-O-Me, P = S2 GCT ATT ACC TTA ACC CAG 654 2′-O-MOE, P = S3 GCT ATT ACC TTA ACC CAG 654 Morpholino4 H- GCT ATT ACC TTA ACC CAG –(Lys)4-NH2 654 PNA-4K5 H- GCT ATT ACC TTA ACC CAG –(Lys)1-NH2 654 PNA-1K6 CCT CTT ACC TCA GTT ACA 705 2′-O-Me, P = S7 H- GCT ACT ACA TTA AAC CAG –(Lys)4-NH2 654 3MM PNA-4K8 GCT ACT ACA TTA AAC CAG 654 3MM 2′-O-MOE, P = S9 GCT ACT ACA TTA AAC CAG 654 3MM Morpholino

Oligomer sequences are arranged according to their target sites. 654 3MM indicates three mismatches;P = S, phosphorothioate internucleotide bond.

Figure 1. The EGFP-654 splicing assay. A mutated intron 2 (thin line)of the human β-globin gene interrupts the EGFP coding sequence(green boxes). The mutation at nucleotide 654 of the intron activatesaberrant splice sites and leads to retention of the intron fragment inspliced mRNA, preventing proper translation of EGFP. An antisenseoligomer (heavy bar) blocking the aberrant splice site restores correctpre-mRNA splicing and EGFP expression. Dashed lines above andbelow intron indicate correct and aberrant splicing pathways,respectively. Lower right panel shows EGFP fluorescence in HeLa cellsstably expressing EGFP-654 construct treated with an antisenseoligomer. Left panel, mock-treated control. Figure 2. Transgene expression in EGFP-654 and EGFP-WT

transgenic mice. (A) RT-PCR of total RNA from EGFP-654 tissues.(B) RT-PCR of total RNA from EGFP-WT tissues. Equal amounts (200 ng) of total RNA from tissues indicated at the top of the panelswere subjected to RT-PCR with EGFP-specific primers. The 160 bp and87 bp bands represent aberrantly and correctly spliced EGFP mRNA,respectively. (C) RT-PCR amplification of total RNA from EGFP-654pre-RNA. (D) RT-PCR amplification of total RNA from EGFP-WT pre-mRNA. Pre-mRNA was amplified by subjecting total RNA from theindicated tissues to RT-PCR (214 bp) using an intron-specific reverseprimer. PCR of the EGFP-654 plasmid (DNA) was included as a sizemarker. Similar designations were used in legends to Figures 3B, 5A,and 5C.

A

B

C

D

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reb

iote

chn

olo

gy

RESEARCH ARTICLE

nature biotechnology • VOLUME 20 • DECEMBER 2002 • www.nature.com/naturebiotechnology1230

curve and therefore possibly increased exposure time of theoligomers to tissues. This schedule was also previously shown tobe effective with 2′-O-MOE/2′-deoxy-phosphorothioate chimerasused for downregulation of Fas in murine liver26. Here, the experi-ments included 2′-O-MOE, morpholino, PNA-1K, and PNA-4K18-mers (oligomers 2–4, Table 1) targeted to the aberrant5′ splice site of EGFP pre-mRNA. Animals were killed oneday after treatment and examined for EGFP fluorescence in10 µm frozen sections from various organs (Fig. 4).

EGFP readout. In a mock-treated mouse, all examined tis-sues showed only minimal background fluorescence (Fig. 4,row 1), as expected. This background is likely due to tissueautofluorescence or to residual correct splicing of EGFP-654pre-mRNA (e.g., in liver). In contrast, treatment with 2′-O-MOE and PNA-4K elicited high levels of EGFP in the cardiacmuscle, in the cortex of the kidney, and in the liver hepato-cytes. The strong EGFP signal was also visible in the lung ofPNA-4K and 2′-O-MOE-treated mice (Fig. 4, rows 2–4). Thatseveral organs exhibited intense production of EGFP indi-cates that the oligomers were distributed by the lymphaticsystem and bloodstream throughout the animal. The highfluorescence elicited in the small intestine by 2′-O-MOE andPNA-4K oligomers after one or four days of treatment (Fig. 4,rows 2–4) may have also been in part due to close proximityof this organ to the injection site. Nevertheless, the EGFP flu-

orescence was produced in several intestinal structures, includingthe villi, the lamina propria, and the smooth muscle lining, indicat-ing that the oligomers were taken up and distributed throughoutthis tissue. The morpholino oligomer produced similar effectsdetectable only after four days of treatment (row 6). Interestingly,although PNA-4K was the most effective of the three oligomers inall other tissues, in small intestine its effect was less than that of the2′-O-MOE oligonucleotide.

Fluorescent EGFP signal was also detected in the cells of thepancreas and, though very light, was approximately equal for 2′-O-MOE, PNA-4K, and morpholino oligomers. Weak butnotable EGFP fluorescence was detected in the pulp of the spleenand in the cortex of the thymus, but only after four daily injectionsof 2′-O-MOE or PNA-4K oligomers. The morpholino oligomerwas relatively ineffective in both the spleen and thymus.Surprisingly, fluorescence (Fig. 4, row 7) and RT-PCR results (datanot shown) indicated that PNA-1K injected i.p. at 50 mg/kg hadno antisense effects in any tissue after four daily injections. Thesedata suggest that the four-lysine moiety contributes substantiallyto the in vivo activity of the PNA-4K oligomer.

No visible pathology was seen at the time of death as a result ofoligomer delivery or robust EGFP production. In some of thetreated tissues, namely brain, skin, and stomach, EGFP fluores-cence was at background or barely detectable. For the brain, thiswas expected, because the blood/brain barrier should preventaccess of the oligomers to the cortex. In skin, it has been shown byothers that intact phosphorothioate oligonucleotides do not accu-mulate in the dermal layers of the skin27, and that direct applica-tion is a more efficient route of oligonucleotide administration ifantisense activity in the skin is desired28.

Figure 3. Ex vivo treatment of EGFP-654 mouse primary cells with 2′-O-methyl-antisense oligonucleotides. Isolated fibroblasts (left panels) andhepatocytes (right panels) were treated with increasing concentrations ofthe oligonucleotide complexed with Lipofectamine and Effectene,respectively. (A) Fluorescence microscopy. (B) RT-PCR of total RNA withEGFP-specific primers.

Figure 4. Upregulation of EGFP expression in oligomer-treatedEGFP-654 mice. EGFP-654 mice were treated with 50 mg/kg of2′-O-MOE, PNA-4K, and morpholino oligomers once daily for oneor four days by i.p. injections. After treatment, 10 µm frozen sectionswere inspected by fluorescence microscopy. Rows: (1) Vehicle(PBS)-treated mice. (2) 2′-O-MOE, four days. (3) PNA-4K, one day.(4) PNA-4K, four days. (5) Morpholino, one day. (6) Morpholino, fourdays. (7) PNA-1K, four days.

A

B

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reb

iote

chn

olo

gy

RESEARCH ARTICLE

www.nature.com/naturebiotechnology • DECEMBER 2002 • VOLUME 20 • nature biotechnology 1231

RT-PCR readout. To confirm that the EGFP signal was indeeddue to sequence-specific shifts in splicing of EGFP-654 pre-mRNA, total RNA isolated from the tissues shown in Figure 4 wasexamined by RT-PCR (Fig. 5). As expected, PNA-4K-treated tis-sues such as heart, kidney, liver, lung, muscle, and small intestinethat showed bright EGFP fluorescence also showed robust increas-es in the levels of correctly spliced EGFP-654 mRNA. On the otherhand, tissues that showed no fluorescent response to the antisenseoligomers (e.g., brain, skin, and stomach) showed no changes inthe splicing patterns of EGFP-654. In addition, bone marrow,which was not analyzed by fluorescence microscopy, showed virtu-ally no correction of aberrant splicing in response to antisensetreatment. Because of degradation of RNA by pancreatic ribonu-clease, we were unable to subject the RNA from the pancreas toRT-PCR analysis.

Quantification of the RT-PCR results (Fig. 5B) confirmed thecorrelation between the data based on EGFP fluorescence and onRT-PCR. For example, PNA-4K was more effective than the othertwo oligomers in all tissues but small intestine, in which the highestshift in splicing was obtained after four days of treatment with the2′-O-MOE oligomer. At four days, PNA-4K elicited ∼ 40% shifts insplicing in the kidney and liver, whereas their response to 2′-O-

MOE was in the 20–30% range. Similar superiority ofthe PNA-4K oligomer was also seen in the lung andmuscle. Sequence specificity of the in vivo effects ofantisense oligomers was ascertained after i.p. injectionof mice with four daily doses of 50 mg/kg of PNA-4K,2′-O-MOE, or morpholino oligomers containingthree mismatches to the target sequence (oligomers7–9, Table 1; PNA-4K 3MM, MOE 3MM, Morpholino3MM, Fig. 5C). All of these tissues were unresponsiveto the control oligomers of any backbone.

DiscussionAntisense activity (i.e., upregulation of EGFP) as mea-sured with the EGFP-654 transgenic mouse is a functionof at least three parameters: the pharmacokinetic prop-erties of a given oligomer, its ability to cross through thecell membrane and into the nucleus of cells, and its abil-ity to bind to the target sequence. The activity of a givenantisense oligomer in the EGFP-654 assay thereforereflects all these parameters. Clearly, comparison ofPNA-4K and PNA-1K antisense effects (Fig. 4) demon-strates that some or all of these parameters wereimproved in PNA-4K, and that PNA-4K represents anadvance in the efficacy of PNA oligomers in vivo. ThePNA-4K oligomer also outperformed the 2′-O-MOEoligonucleotide in all tissues except small intestine,where the 2′-O-MOE was ∼ 50% more effective afterfour days of treatment. In other tissues, PNA-4K out-performed 2′-O-MOE from 1.4-fold in liver to ∼ 2-foldin the heart and kidney and 10-fold in thigh muscle(Fig. 5B). The uncharged morpholino oligomer had lowactivity in all tested tissues. This could be due to rapiddistribution to and elimination of these compounds bythe kidney, similar to the results seen with neutralmethylphosphonate oligomers29,30. The same mecha-nism may apply to the PNA oligomer23 with only onelysine (PNA-1K), which showed no detectable levels ofcorrection in any assayed tissues. That a relatively highdose (50 mg/kg) was used to achieve the above resultsindicates that the modified oligonucleotides just dis-cussed may be further improved. The EGFP-654 mouseallows for rapid assessment of such improvements.

Thorough pharmacokinetic and biodistribution studies wererecently performed for 2′-O-MOE 20-mers. Plasma clearance afteri.v. injection was similar to that of a comparable 2′-oligodeoxynu-cleotide, wherein clearance was dominated by distribution intothe liver and kidney. Overall, the researchers concluded that thebackbone (e.g., phosphodiester versus phosphorothioate) had afar greater influence on pharmacokinetics of an oligomer than its2′-O modification20,31. The pharmacokinetic properties of PNAand morpholino backbones have also been recently investigated,though not as extensively. For example, morpholino oligomerswere shown to have favorable oral bioavailability25, and at highdoses had good activity against tumor necrosis factor-α (TNF-α)in mice after intranasal insufflation32. Other studies have usedmorpholino oligomers to examine gene expression in zebrafish33.The pharmacokinetics and tissue distribution of PNA in rats havealso been recently studied23. This study showed that 90% ofuncharged PNA oligomers was secreted in urine. The remainderwas distributed at low levels primarily to the liver and kidney aswell as other tissues.

In addition to the aforementioned tissue distribution data, thefunctional results are provided by the EGFP-654 mouse model.This model allows measurement of antisense activity in most tis-

Figure 5. Visualization and quantification of splicing correction. (A) RT-PCR of totalRNA from the tissues of oligomer-treated EGFP-654 mice. Total RNA was isolated fromthe same tissues, which were analyzed in experiments illustrated in Figure 4.(B) Quantification of EGFP-654 mRNA RT-PCR products. Autoradiograms from theexperiment shown in (A) as well as from at least two additional independent oligomertreatments of EGFP-654 mice were digitized and quantified using the Scion Image beta3 software package. Percentage correction is calculated as the fraction of correctlyspliced mRNA in the total EGFP-654 mRNA. The bars represent RT-PCR of total RNAfrom mice treated with vehicle (white); 2′-O-MOE, four days (gray); PNA-4K, four days(black). Error bars represent standard deviation. (C) Sequence specificity of oligomers.EGFP-654 mice were injected daily i.p. for four days with oligomers with threemismatches (PNA-4K 3MM, MOE 3MM, or Morpholino 3MM). PNA-4K and MOE wereused as positive controls (oligomers 2, 4, 7, 8 or 9, Table 1). All oligomers were at aconcentration of 50 mg/kg.

A B

C

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reb

iote

chn

olo

gy

RESEARCH ARTICLE

nature biotechnology • VOLUME 20 • DECEMBER 2002 • www.nature.com/naturebiotechnology1232

sues, rather than in one or two tissues, which express a specific tar-geted gene. With the combined data, novel antisense compoundscan be engineered to target diseases depending on their functionaldistribution. Furthermore, in the EGFP-654 assay, the testedoligomers, with different backbones or modifications, always havethe same sequence, eliminating this variable in their comparison.Once a superior chemistry is identified, the oligomers can beretargeted to any desired sequence or even modified to chimeric,RNase H-activating forms. One can expect that the latter mole-cules may retain the stability and favorable tissue distribution pro-vided by the modified monomers.

Quantitative analysis of the RT-PCR data (Fig. 5B) indicatesthat PNA-4K modified EGFP-654 pre-mRNA splicing ∼ 40% inliver and kidney and close to 20% in other tissues. 2′-O-MOEoligonucleotide had somewhat lower activity in heart, liver, andlung, but it corrected EGFP-654 splicing >50% in small intestine.Note that a partial shift in splicing had, for example, considerablebiological effects in erythroid cells from thalassemia patients14

and in cancer cell lines6,7. Thus, if these levels of modificationcould be achieved in appropriate tissues of patients, the effectcould be of clinical value in genetic diseases and cancer4. Theresults presented in this report suggest that 2′-O-MOE and PNA-4K modifications are particularly promising in this regard. TheEGFP mouse model described here should provide a convenientplatform for comparison.

Experimental protocolGlaxoSmithKline is fully accredited by the American Association forAccreditation of Laboratory Animal Care, and all procedures were conduct-ed in accordance with the Institutional Animal Care and Use Committee.

Oligomers. 2′-O-Methyl (2′-O-Me) phosphorothioate oligonucleotideswere purchased from TRI-Link, Inc. (San Diego, CA). The 2′-O-MOE andPNA oligomers were synthesized as described21. Morpholino oligomerswere obtained from Gene-Tools (Philomath, OR). The sequences of theoligomers are indicated in Table 1.

Generation of EGFP plasmids and transgenic mice. Plasmid CX-EGFP-654 was constructed from plasmid CXEGFP22 following earlier describedprocedures21. For the generation of the CX-EGFP-WT plasmid, an EcoN1-Ppum1 fragment of β-globin intron 2 from the CX-EGFP-654 plasmid wasreplaced by the same fragment from the intron 2 of the wild-type β-globingene. Transgenic mice were generated using standard techniques bymicroinjection of the 3.8 kb PstI/Sal1 fragments of CX-EGFP-654 and CX-EGFP-WT into fertilized FVB/N mouse embryos.

Hepatocyte and fibroblast cultures. For hepatocyte collection, livers ofEGFP-654 mice were perfused with RPMI medium containing 0.53 mg/ml of

collagenase (Worthington Type 1, code CLS). After perfusion, the cell suspen-sion was collected and seeded in a stop solution of RPMI with 10% (vol/vol)FBS and 0.5% penicillin–streptomycin plus 1 nM insulin and 13 nM dexam-ethasone. Approximately 3 × 105 cells were seeded on a six-well collagen-coated plate. The seeding medium was replaced 1 h later with maintenancemedium consisting of seeding medium without the 10% (vol/vol) FBS. Forfibroblasts, tail clippings were digested in PBS containing 0.125% (wt/vol)trypsin and 1.2 U/ml dispase. The cells were placed in DMEM–F-12 mediumcontaining 20% (vol/vol) FBS and seeded in 24-well plates at 1 × 105

cells/well. For both, varying amounts of oligonucleotide–lipid complexeswere applied 24 h later. Cells were subjected 24 h after transfection to lysiswith TRI-Reagent (MRC, Cincinnati, OH).

Treatment of EGFP transgenic animals. Transgenic mice were injected i.p.with a 200 µl solution of the oligomers in PBS at indicated concentrations.Injections were given daily, at the same time each day, for the indicatednumber of days. The mice were killed 24 h after the last injection, and tis-sues were collected. A portion of each tissue was sliced into small (2-mm-thick) pieces and fixed in 2 ml of 4% (wt/vol) paraformaldehyde in PBS.Frozen 10 µm tissue sections were prepared in O.C.T. mounting medium(Miles Scientific, Naperville, IL), thaw-mounted onto glass slides, and keptat –20°C. Images of each slide were taken by fluorescence microscopy anddigitized with Scion Image software. The remainder was snap-frozen inliquid nitrogen.

Isolation of total RNA and RT-PCR. Tissues (25 mg) or cultured cells werehomogenized in 1 ml of TRI-Reagent. The samples were then centrifuged for 2 min to remove any cellular debris, and the supernatant was transferred intoa new tube. RNA isolation, RT-PCR, and RNA visualization was carried out asdescribed34, except that EGFP-specific forward (5′-CGTAAACGGCCA-CAAGTTCAGCG-3′) and reverse (5′-GTGGTGCAGATGAACTTCAGGGTC-3′) primers were used. To detect pre-mRNA only, a reverse primer with thesequence 5′-GATCCTGAGACTTCCACACTGATG-3′ was used.

Note: Supplementary information is available on the Nature Biotechnologywebsite.

AcknowledgmentsWe are grateful to Dr. Masaru Okabe, Osaka University, Japan, for a giftCXEGFP plasmid. We thank J. Summerton and P. Morcos for morpholinooligomers. We would like to thank Y. McNeill, K. Powell, C. Wilson, B. Spencer,D. Banner, M. Cameron, and E. Smith for their invaluable technical contribu-tions, and F. Bennett, R. Griffey, and S. Crooke for critical review of this manu-script. This work was supported by the National Institutes of Health (PO1-GM59299) and GlaxoSmithKline grants to R. K.

Competing interests statementThe authors declare competing financial interests: see the Nature Biotechnologywebsite (http://biotech.nature.com) for details.

Received 24 July 2002; accepted 4 October 2002

1. Lander, E. S. et al. Initial sequencing and analysis of the human genome.Nature 409, 860–921 (2001).

2. Venter, J. C. et al. The sequence of the human genome. Science 291,1304–1351 (2001).

3. Graveley, B. R. Alternative splicing: increasing diversity in the proteomic world.Trends Genet. 17, 100–107 (2001).

4. Mercatante, D. & Kole, R. Modification of alternative splicing pathways as apotential approach to chemotherapy. Pharmacol. Ther. 85, 237–243 (2000).

5. Kole, R. & Sazani, P. Antisense effects in the cell nucleus: modification of splic-ing. Curr. Opin. Mol. Ther. 3, 229–234 (2001).

6. Mercatante, D. R., Bortner, C. D., Cidlowski, J. A. & Kole, R. Modification ofalternative splicing of Bcl-x pre-mRNA in prostate and breast cancer cells.Analysis of apoptosis and cell death. J. Biol. Chem. 276, 16411–16417 (2001).

7. Taylor, J. K., Zhang, Q. Q., Wyatt, J. R. & Dean, N. M. Induction of endogenousBcl-xS through the control of Bcl-x pre-mRNA splicing by antisense oligonu-cleotides. Nat. Biotechnol. 17, 1097–1100 (1999).

8. Friedman, K. J. et al. Correction of aberrant splicing of the cystic fibrosis trans-membrane conductance regulator (CFTR) gene by antisense oligonucleotides.J. Biol. Chem. 274, 36193–36199 (1999).

9. Dunckley, M. G., Manoharan, M., Villiet, P., Eperon, I. C. & Dickson, G.Modification of splicing in the dystrophin gene in cultured Mdx muscle cells byantisense oligoribonucleotides. Hum. Mol. Genet. 7, 1083–1090 (1998).

10. Wilton, S. D. et al. Specific removal of the nonsense mutation from the mdx dys-

trophin mRNA using antisense oligonucleotides. Neuromuscul. Disord. 9,330–338 (1999).

11. Kalbfuss, B., Mabon, S. A. & Misteli, T. Correction of alternative splicing of tauin frontotemporal dementia and parkinsonism linked to chromosome 17. J.Biol. Chem. 276, 42986–42993 (2001).

12. Karras, J. G., McKay, R. A., Dean, N. M. & Monia, B. P. Deletion of individualexons and induction of soluble murine interleukin-5 receptor-α chain expres-sion through antisense oligonucleotide-mediated redirection of pre-mRNAsplicing. Mol. Pharmacol. 58, 380–387 (2000).

13. Suwanmanee, T. et al. Restoration of human β-globin gene expression inmurine and human IVS2-654 thalassemic erythroid cells by free uptake of anti-sense oligonucleotides. Mol. Pharmacol. 62, 545–553 (2002).

14. Lacerra, G. et al. Restoration of hemoglobin A synthesis in erythroid cells fromperipheral blood of thalassemic patients. Proc. Natl. Acad. Sci. USA 97,9591–9596 (2000).

15. Tamm, I., Dorken, B. & Hartmann, G. Antisense therapy in oncology: new hopefor an old idea? Lancet 358, 489–497 (2001).

16. Juliano, R. L., Alahari, S., Yoo, H., Kole, R. & Cho, M. Antisense pharmacody-namics: critical issues in the transport and delivery of antisense oligonu-cleotides. Pharm. Res. 16, 494–502 (1999).

17. Marcusson, E. G., Yacyshyn, B. R., Shanahan, W. R., Jr. & Dean, N. M.Preclinical and clinical pharmacology of antisense oligonucleotides. Mol.Biotechnol. 12, 1–11 (1999).

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reb

iote

chn

olo

gy

RESEARCH ARTICLE

www.nature.com/naturebiotechnology • DECEMBER 2002 • VOLUME 20 • nature biotechnology 1233

18. Lebedeva, I. & Stein, C. A. Antisense oligonucleotides: promise and reality.Annu. Rev. Pharmacol. Toxicol. 41, 403–419 (2001).

19. Mercatante, D., Sazani, P. & Kole, R. Modification of alternative splicing path-ways as a potential chemotherapy for cancer and other diseases. Curr. CancerDrug Targets 1, 211–230 (2001).

20. Geary, R. S. et al. Pharmacokinetic properties of 2’-O-(2-methoxyethyl)-modifiedoligonucleotide analogs in rats. J. Pharmacol. Exp. Ther. 296, 890–897 (2001).

21. Sazani, P. et al. Nuclear antisense effects of neutral, anionic and cationicoligonucleotide analogs. Nucleic Acids Res. 29, 3965–3974 (2001).

22. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T. & Nishimune, Y. ‘Green mice’as a source of ubiquitous green cells. FEBS Lett. 407, 313–319 (1997).

23. McMahon, B. M. et al. Pharmacokinetics and tissue distribution of a peptidenucleic acid after intravenous administration. Antisense Nucleic Acid Drug Dev.12, 65–70 (2002).

24. Agrawal, S., Temsamani, J. & Tang, J. Y. Pharmacokinetics, biodistribution, andstability of oligodeoxynucleotide phosphorothioates in mice. Proc. Natl. Acad.Sci. USA 88, 7595–7599 (1991).

25. Arora, V., Knapp, D. C., Reddy, M. T., Weller, D. D. & Iversen, P. L. Bioavailability andefficacy of antisense morpholino oligomers targeted to c-myc and cytochrome P-450 3A2 following oral administration in rats. J. Pharm. Sci. 91, 1009–1018 (2002).

26. Zhang, H. et al. Reduction of liver Fas expression by an antisense oligonucleotideprotects mice from fulminant hepatitis. Nat. Biotechnol. 18, 862–867 (2000).

27. Mehta, R. C. et al. Intercellular adhesion molecule-1 suppression in skin by topi-cal delivery of anti-sense oligonucleotides. J. Invest. Dermatol. 115, 805–812(2000).

28. Brand, R. M. & Iversen, P. L. Transdermal delivery of antisense compounds. Adv.Drug. Deliv. Rev. 44, 51–57 (2000).

29. DeLong, R. K. et al. Comparative pharmacokinetics, tissue distribution, andtumor accumulation of phosphorothioate, phosphorodithioate, and methylphos-phonate oligonucleotides in nude mice. Antisense Nucleic Acid Drug Dev. 7,71–77 (1997).

30. Agrawal, S., Temsamani, J., Galbraith, W. & Tang, J. Pharmacokinetics of anti-sense oligonucleotides. Clin. Pharmacokinet. 28, 7–16 (1995).

31. Geary, R. S. et al. Absolute bioavailability of 2’-O-(2-methoxyethyl)-modified anti-sense oligonucleotides following intraduodenal instillation in rats. J. Pharmacol.Exp. Ther. 296, 898–904 (2001).

32. Qin, G., Taylor, M., Ning, Y.Y., Iversen, P. & Kobzik, L. In vivo evaluation of a mor-pholino antisense oligomer directed against tumor necrosis factor-α. AntisenseNucleic Acid Drug Dev. 10, 11–16 (2000).

33. Corey, D. R. & Abrams, J. M. Morpholino antisense oligonucleotides: tools forinvestigating vertebrate development. Genome Biol. 2, REVIEWS1015 (2001).

34. Sierakowska, H., Sambade, M. J., Agrawal, S. & Kole, R. Repair of thalassemichuman β-globin mRNA in mammalian cells by antisense oligonucleotides. Proc.Natl. Acad. Sci. USA 93, 12840–12844 (1996).

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://w

ww

.nat

ure

.co

m/n

atu

reb

iote

chn

olo

gy