Antisense inhibition of β-actin mRNA localization and its effect on smooth muscle cell migration

113

Biology of the Cell (1997) 89, 113-l 22 o Elsevier, Paris

Original article

Antisense inhibition of pactin mRNA localization and its effect on smooth muscle cell migration Lynette Schedlich a*, Mark Hill b and Trevor Locketta aCSIR0, Division of 6iomolecular Engineering, North Ryde, 2113; bSchool of Anatomy, University of New South Wales, Kensington, 2033, Australia

A crucial step in cell migration involves changes in the actin cytoskeleton in response to extracellular signals. We have previously shown that pactin transcripts are associated with mobile regions of mouse 3T3 fibroblasts when grown in the presence of serum. In the current study we used in situ hybridization and laser scanning confocal microscopy to show that cultured rat smooth muscle cells also localize pactin mRNA to the cell periphery and that this peripheral pool of pactin mRNA is dependent on the presence of growth factors in the culture medium. We also show that antisense phosphorothioated oligonucleotides directed against sequences in the 3’ untranslated region of rat pactin mRNA block peripheral localization of pactin mRNA while the corresponding control oligonucleotides have no effect. Time-lapse video analysis demonstrates that the antisense oligonucleotides inhibit rat smooth muscle cell migration in culture and analysis of pactin mRNA confirms this is not due to changes in pactin gene expression or instability of the message. Our results suggest that depletion of pactin transcripts from the cell periphery is associated with suppression of SMC migration. 3’UTR localization sequences / cell migration / mRNA trafficking / restenosis

INTRODUCTION Cell migration plays an important role in many biological processes including embryogenesis, wound healing and the immune response (Huttenlocher et al, 1995). Cells migrate in response to external stimuli, such as growth factors and other chemoattractant agents, through interactions with specific cellsurface receptors (Ridley and Hall, 1992; Ridley et al, 1992). This activates specific signal transduction pathways which control the cortical rearrangements required for cell mobility. The initial phase of cell migration involves protrusion at the leading edge to form lamellipodia and microspikes (Huttenlocher et al, 1995). These actin-rich protrusions * Correspondence and reprints. Present address: Kolling Institute of Medical Research, Royal North Shore Hospital, St Leonards, 2065, Australia Abbreviations: B’UTR, 3’ untranslated region; CLSM, laser scanning confocal microscopy; I’-ODN, phosphorothioate-modified oligonucleotides; SMC, smooth muscle cell.

Antisense inhibition of Dactin mRNA localization

adhere to extracellular matrix proteins on the substratum through the action of cell surface adhesion molecules, such as the integrin receptors, and generate the traction required for cell movement. Linkage between integrins and the termini of actin filament bundles occurs at focal adhesions. These structural units contain, in addition to actin, other proteins such as talin, vinculin and a-actinin (Luna and Hitt, 1992). One event that is essential to the migration of all cells is the remodelling of the actin cytoskeleton. Of the six different isoforms of actin found in cells, the cytoplasmic isoforms /?- and Factin are the major contributors to the actin cytoskeleton (Vandekerckhove and Weber, 1978). Analysis of the subcellular localization of actin mRNA using in situ hybridization has revealed that the mRNAs encoding the different isoforms of actin are sorted to different compartments within the cell. Whereas y and a-cardiac actin mRNAs are located exclusively within the perinuclear region (Hill and Gunning, 1993; KisSchedlich et al

Biology of the Cell (1997) 89, 1 id-1 22

114

lauskis et al, 1993), pactin transcripts are also found at the motile periphery in a number of cell types (Hoock et al, 1991; Hill and Gunning, 1993; Kislauskis et al, 1993; Hill et al, 1994). This strong correlation between motile regions of the cell and zones of peripheral pactin mRNA accumulation suggests that peripherally localized pactin transcripts may be required for cell migration. Intact microfilaments, which are composed of actin, are required for the transport and subsequent anchoring of actin mRNA at the cell periphery (Sundell and Singer, 1991). Sequences in the 3’ untranslated region (3’UTR)l of the pactin mRNA are essential for the localization of its transcripts to the periphery. The masking of such sequences by hybridization to antisense oligonucleotides inhibits the transport and/or anchorage of &actin transcripts to the periphery of chicken embryonic fibroblasts (Kislauskis et al, 1994), and may in turn affect cytological properties that are dependent on this localization. Recently, we have shown that an antisense oligonucleotide directed against sequences involved in the peripheral localization of rat pactin mRNA caused suppression of neurite extension in the rat neuronal cell line PC12 (M Hill, unpublished observation). Cell migration contributes to a number of disease processes, such as inflammatory disorders, the formation of atherosclerotic plaques and metastatic tumour spread (Huttenlocher cf al, 1995). In addition, cell migration may have a role in the formation of restenotic lesions following balloon angioplasty, a procedure used in the treatment of stenotic vascular lesions caused by atherosclerosis iClowes and Schwartz, 1985). During balloon angioplasty, atherosclerotic plaque material is ruptured and the vessel wall damaged, leading to the exposure of medial smooth muscle cells (SMC) to serum growth factors which can, in turn, result in phenotypic changes in these cells. In as many as 40% of cases these changes lead to reocclusion of the artery at the site of the vascular wound (Serruys cf al, 1988). Following the phenotypic change in the vascular SMC, these cells migrate to the vessel lumen where they proliferate and deposit large amounts of extracellular matrix (Ferrell rt al, 1992). To date most efforts to minimize the occurrence of restenosis have focused on antiproliferative therapies directed against SMC replication (Simons et al, 1992; Morishita et al, 1993; Bennett cf al, 1994). However, of the antiproliferative therapies trialed to date, none has been effective in preventing restenosis in humans (O’Brien and Schwartz, 1994). An alternative approach would be to inhibit SMC migration. As a number of different growth factors and redundant signalling mechanisms may be capable of stimulating the migration of SMC, a Antisenseinhibitionof pactm mRNAlocalization

mitogen-independent mhibition of SMC migratjor: might be achieved by targeting a common stry essential for cell locomotion. In the present study we show that, as jot othta; cell types, pactin mRNA is localized to the cell periphery in rat SMC. Furthermore, using an anti senseoligonucleotide designed to alter the distribution rather than the level of pactin mRNA in s&. we have demonstrated that interference \z~iti_ peripheral localization of pactin mRNA can Hart’ an impact on the ability of rat SMC to migrate I?Z vitro.

IUlATERWS

hETt4oDS

Phosphorothioate-modified oligonucleotides (P-i)DX) were synthesized on an Applied Biosystems391 synthc sizer by 2-cyanoethylphosphoramidite chemistry. Phob phorothioate linkageswere introduced by oxidation with TETD reagent (Perkin-Elmer, Foster City, CA). Synthes& of 5’-fluorescein labelled derivatives was performed using 6-FAM amidite (Perkin-Elmer, Foster City, CA: The antisenseP-ODNs were directed against rat j?-acti:\ mRNA sequenceslocated in the 3’UTR and the contrcii P-ODNs were scrambled sequencesof the same length and base composition. The sequence of the antisrnst: I))-ODN 1A was 5’-ACGCAGCTCAGTAACAGTCCC; CCTA-3’ (position 1203-1227) and the corresponding WI:trol P-ODN 1C was S’-ATCACTAACACGCGT CAGCCGCCAT-3’. The sequence of the antisons:;

P-ODN 2A was 5’-GTTAGGTTTTGTCAAAGAA GGGTG3’ (position 1232-1256) and the corresponding control

P-ODN

2C

CAGGTTTGTGGGT-3’.

was

~3

T’-GAAATTAGAAGI

To test for stability in st,ruE:

2 ,uM P-ODNs were 5’-end labelled with 32P and t1~ei.i incubated at 37°C in DMEM containing 10% FCS ft.: 48 h. There was no evidence of P-ODN degradation dt analyzed by 7 M urea polyacrylamide gel electrophor&: and autoradiography.

Cell culture A rat SMC line, A7r5 (ATCC CRL i444j, derived tron.: aortic tissue, was grown in a 5% CO, humidified rnvi ronment in DMEM/lO% FCS supplemented with 30 @mL penicillin and 100 PglmL streptomycin. I-W the analysis of the intracellular distribution of ,?--actin mRNA the cells were grown on collagen-coated (Caitskin collagen Type I; Sigma, St Louis, MO) glass multiwell slides (Nunc Inc, Naperville, IL). To prepare serlum starved cells, SMC were allowed to attach overnight in the presence of serum, washed with serum free medium and grown for at least 20 h in DMEM alone. For growth factor stimulation, DMEM/lO% FCS or DMEM pius 10 ng/mL PDGF-BB (Gcnzyme Carp, Cambridge, MA) was added 30 min prior to fixation. For studies to determine the effect of antisense P-ODN on pactin mRNA localization, cells were incubated in the presence of 2 PM I’-ODN for 20 h in DMEM/lO% FCS. Cells were fixed in 4% para-

Schedlich et 31

Biology of the Cell (1997) 89, 113-122

115

formaldehyde, permeabilized using cold methanol and the subcellular distribution of Pactin transcripts was analysed by in situ hybridization.

In situ hybridization A 195 bp rat fi-actin isoform-specific sequence corresponding to nucleotides 3379-3574 in the 3’UTR of rat Pactin mRNA (Genbank accession no J00691) was amplified from SMC RNA and cloned into pGEM-7Zf (Promega Corp, Madison, WI). Orientation was confirmed by sequencing. To prepare labelled pactin-specific probe, the plasmid was linearized and transcribed in vitro using a digoxygenin RNA labelling kit (Boehringer Mannheim, Sydney, Australia). The complementary RNA strand was similarly labelled and used as a control probe. Northern blot analysis confirmed the specificity of the probe and the RNA was quantified by comparison with known standards. Hybridization procedures were as previously described (Hill et al, 1994). All chemicals were of RNA grade. Solutions were diethylpyrocarbonate-treated and/or autoclaved and glassware and instruments were baked overnight at 200°C. Prehybridization was in 100 PL of hybridization solution (50% deionized formamide, 0.6 M NaCl, 10 mM Tris-HCl (pH 8.0), 2 mM EDTA, 0.5 mg/mL tRNA, 1 x Denhardts and 5% dextran sulphate) without probe at 45°C for 1 h. This was replaced with fresh hybridization solution containing probe or control RNA (0.3 ng/pL) and hybridized at 45°C overnight. Slides were washed in 2 x SSC, 1 x SSC, 1 x SSC/O.l% SDS and then 0.5 x SSC. All washes were for 10 min at 65°C. High stringency washes were in 0.1 x SSC at 65°C. Slides were then equilibrated in PBS and blocked in 0.5% blocking reagent (Boehringer Mannheim, Sydney, Australia). Fluorescently labelled anti-digoxygenin antibodies (Boehringer Mannheim, Sydney, Australia) were diluted as indicated by the manufacturer, reacted with blocked slides for 1 h at 37”C, and then washed in PBS at pH 7.5 and then pH 8 to reduce non-specific binding.

Laser scanning confocal microscopy

(CLSM)

CLSM (Leica: TCS, Heidelburg, Germany) was used to analyze label distribution within cells. Individual cells with a fully spread morphology were optically sectioned in the xy planes (parallel to the substratum) with multiple scan averaging. A single CLSM slice was collected from each cell 1 mr~ above the contact point between cell and substratum. To quantify pactin mRNA distribution all images were collected under identical, non-saturating conditions using Leica software lut-glowov where the colour shift red - yellow represents low + high intensity pixels and blue represents saturation of pixel intensity. The intemity of fluorescent labelling within cells was analysed using the program NIH Image ~1.57. Quantitation was as previously described (Hill et al, 1994). Pixel intensity, as a measure of fluorescence intensity, was measured in three separate 4 p squares within specific regions of the cell (cytoplasmic and peripheral) as well as in regions outside the cell (background). The pixel intensity from each subcellular region was averaged over many cells.

Antisense inhibitionof Dactin mRNAlocalization

Northern blot analysis Total cellular RNA was extracted from A7r5 cells (Total RNA Isolation Reagent; Advanced Biotechnologies Ltd, London, UK). RNA (5 pg/lane) was size-fractionated on a 1% agarose/2.2 M formaldehyde gel, transferred to Hybond-N (Amersham, UK) and hybridized overnight with 32Plabelled rat pactin specific sequences. @actin mRNA levels were quantified using a PhosphorImager (Molecular Dynamics), and corrected for loading differences by comparison with GAPDH mRNA. For this the

blot was stripped and re-probed with s*P-labelled GAPDH sequences.pactin mRNA levels from cells cultured under different with P-ODN were

growth conditions compared with

and cells treated the level from

untreated cellsgrown in serum (taken as100%).

Migration assays Cell migration was studied using time-lapse video analysis. A7r5 cells were seeded at 7 x 104 cells/T25 flask in

DMEM/lO% FCS.After allowing the cells to adhere and spread overnight, the P-ODN (2 PM) to be tested was added. The sealed tissue culture flask was transferred to

a heated microscope stage and video recording commenced6 h after the addition of oligonucleotide. A computer-directed microscopestage automatically collected images from 4 fields of cells. Images were recorded at half-hourly intervals for 14 h using an Image Analyser

(CambridgeInstruments: Quantimet 570). Cell migration was assessed by identifying the coordinates at the center of the nucleus for each cell in the initial field of view and comparing it with the new nuclear coordinates in each successive image of the same field. The procedure was repeated for each separate field of cells collected. The distance moved by each cell at each successive time-point was determined and expressed as micrometers moved per hour.

RESULTS Effect of growth factors on pactin localization

mRNA

To investigate whether SMC had pactin transcripts at the cell periphery, we examined the subcellular distribution of pactin mRNA in the rat SMC line, A7r5 using in situ hybridization. When cells were grown continuously in serum (fig la), /3-actin mRNA was observed within the perinuclear region and at the cell periphery (arrows). Cells grown under identical conditions, but incubated with control RNA had only a low level signal within the perinuclear region (fig lb). When the experiment was carried out in the absence of RNA, a distribution identical to the control probe was observed (data not shown). This indicated that non-specific binding of the antibody together with autofluorescence was responsible for the signal detected with the control probe. Cells grown in the absence of serum had no peripheral /3-actin mRNA (fig lc), Schedlich et al

116

Biology of the Cell (1997) 89, 1 i j-1 22

Fig 1. pactin mRNA localization is regulated by serum growth factors, All images represent a sirtgle coniocai sim, ‘They were collected under non-saturating conditions using Leica software lut-glowov where the colour shift red + yellow repre sents low -+ high intensity pixels. A7r5 cells grown continuously in serum (a) had strong perinuclear and peripheral signals (arrows). In the absence of serum (c) the peripheral pool was depleted (arrows). Following 30 min of PDGF (d) stimulation pactin mRNA is rapidly reestablished at the cell periphery (arrows). Cells grown in the presence af serum but incubated with control RNA showed a low level of fluorescence around the nucleus (b) Bar: 40 pm in length.

although the perinuclear pool was maintained. The stimulation of serum starved cells for 30 min with 10% serum (data not shown) or PDGF (fig Id) caused the reappearance of pactin mRNA to the cell periphery (arrows). Antisense inhibition of pactin mRNA localization

Quantitation of these findings was carried out as previously described (Hill et al, 1994). The subcellular distribution of pactin mRNA was determined from a single confocal slice which provides information from a single plane of focus. The Schedkh

et ai

Biology of the Cell (1997) 89, 113-122

result is an image where the intensity differences reflect differences in mRNA distribution rather than differences in cell thickness. To include information from all parts of the cell, the confocal slice was taken through the base of the cell. The results obtained by quantifying p-actin mRNA localization supported the qualitative findings. Cells grown in the presence of serum (fig 2a, +S) had more than a two-fold increase in the peripheral p actin signal obtained using a pactin probe when compared with a control probe (fig 2a, control RNA). Using the pactin probe, a similar distribution of pactin mRNA was observed in two other rat SMC lines: R22ClF, which is derived from cardiac tissue and a SMC line cultured from the aorta of a WKY rat (data not shown). Following 20 h of serum deprivation (fig 2a, -S), the peripheral pool of p-actin mRNA was reduced to control levels. Stimulation with PDGF (fig 2a, PDGF) or serum (fig 2a, FCS) for 30 min caused a rapid re-establishment of a pattern of peripheral fi-actin transcripts indistinguishable from that of cells grown continuously in serum. Transcription from the fi-actin gene responds rapidly to growth factor stimulation and this increase in transcript abundance might play a role in re-establishing pactin mRNA at the cell periphery. To address this issue, we used Northern blot analysis of total RNA to quantify fi-actin mRNA levels in A7r5 cells cultured under different growth conditions. pactin mRNA levels were corrected for loading differences using GAPDH mRNA as an internal standard. Following 20 h of serum starvation, the steady-state level of pactin mRNA fell to 71% of that in control cells grown in serum (fig 3a). Following 30 min of serum or PDGF stimulation, p actin mRNA levels were 108% and 67%, respectively, compared to cells grown continuously in serum. As serum stimulation increased pactin mRNA levels, an increase in gene expression appears to be playing a role in re-establishing the peripheral pool of p-actin mRNA. However, there was no significant difference in total pactin mRNA levels observed following PDGF stimulation compared with those in serum-starved cells. Unless there is a rapid turnover of pactin mRNA within the cell, this suggests that PDGF stimulation caused a rapid redistribution of pre-existing pactin mRNA to the cell periphery.

Antisense-mediated of pactin mRNA

delocalization

The results of the growth factor studies indicate that pactin mRNA can be delocalized from the cell periphery in SMC when grown in the absence of serum. To investigate whether this delocalization

Antisense inhibition of pactin mRNA localization

117

can be induced under normal growth conditions, we used antisense P-ODN to block peripheral localization of rat pactin in A7r5 cells. Recent work using antisense I’-ODN has confirmed a role for specific 3’UTR sequences in the peripheral localization of pactin mRNA in chicken (Kislauskis et al, 1994). Antisense I’-ODNs targeted to related sequences in the rat pactin 3’UTR were used to investigate whether these sequences had a similar function. Uptake of antisense or control oligonucleotides (2 @l) by live A7r5 cells was confirmed using fluorescently labelled P-ODN and CLSM (data not shown). All cells had taken up the oligonucleotide within 4 h of oligonucleotide addition to the medium. Intense nuclear and diffuse cytoplasmic labelling could still be observed 24 h after addition of the oligonucleotide to the medium. The effect of antisense and control P-ODNs (2 PM) on pactin mRNA localization was assessed 20 h after treatment. In the presence of the antisense oligonucleotides IA and 2A (figs 4a, c, respectively) the peripheral pool of /3-actin mRNA was completely eliminated (arrows). In contrast, the subcellular distribution of pactin mRNA in cells treated with the control oligonucleotides IC and 2C (figs 4b, d, respectively) was indistinguishable from that of untreated cells (fig la). Quantification of pactin mRNA localization confirmed these observations. Fig 2b shows more than a two-fold decrease in the peripheral signal in antisense treated cells compared with untreated cells (P < 0.0001) or cells treated with control oligonucleotides (P < 0.0001). These results suggest that the 5’-most region of the rat pactin 3’UTR contains sequences responsible for directing the peripheral localization of rat pactin mRNA. The antisense oligonucleotide, while being directed against the localization pathway, could conceivably act by interfering with other, as yet undefined, sequences within the 3’UTR that regulate pactin expression, or by destabilizing pactin mRNA and causing a reduction in target message levels. To investigate this possibility, the steadystate level of pactin mRNA was examined in A7r5 cells following the addition of I’-ODNs (2 PM) for 20 h (fig 3b). After correcting for loading differences using GAPDH mRNA, there was no significant difference in pactin mRNA levels observed in cells treated with antisense or control P-ODNs. Furthermore, there was no evidence of degradation. of p-actin mRNA in antisense treated cells, indicating an absence of RNAse H-mediated activity. Taken together, these results indicate that inhibition of /I-actin mRNA peripheral localization was not the result of alteration in p-actin gene expression or changes in mRNA stability or integrity.

Schedlich et al

Biology of the Cell (1997) 89, 113-I 2:

118

a

b

tc Pi No Qtigo

IC Pi Oligo 1A

[C PI O&o 1C

SC PI OPgo2A

LC Pi @Ego 2C

RNA

Fig 2. Quantitative analysisof the intracellulardistributionof pactin mRNAin A7r5 cells. The localization of fiactrn mRNA within different regions of the cell was analysedfrom three separateexperimentsas previously described(Hill et ai, 1994). The average fluorescent intensity (*SE) was determinedfor the cytoplasmic (C) and peripheral(PI regions. a. Cells were grown in the presence f+S) (n=58) or absenceof serum Ml fn=29) and stimulatedwith PDGF(PDGF)(n=24) or serum (FCS)(n=27) and hybridized with the pactin specific probe. In addition, cells were grown in the presence of serum and hybridized to control probe (Control RNA)(n=24) b. Cellswere grown in the presence of serumwith the addition af anti, senseoligonucleotides1A In=281 or 2A (n=25) or the control oligonucleotides1C in=281 or 2C (n=24) or were untreated (n=58).

a

b

+S

-S

100

71

F c s

P D G F

N II L

A

12 C

A

2 C

p-actin

GAPDH %

108 67

100

97

97

92

92

Fig 3. Northern blot analysisof Pactin mRNA in A7r5 ceils. The level of @actin mRNAwas normalized agarnst GAPDH mRNAand the values represent the average corrected percentagesfrom two independentexperimentsrelative to untreated cells grown in the presence of serum.Total RNAwas extracted from cells grown (a) in the presence f+S) or absence of serum f-S) and following either serum(FCSI or PDGFstimulationfor 30 min or (b) in the presenceof antisense(1A or 2A) or control UC or 20 oligonucleotidesor untreatedcells (NIL). Antisense inhibitionof Dactin mRNAlocalization

Schedlich et a:

Biology of the Cell (1997) 89, 113-122

119

Fig 4. pactin mRNA localization is regulated by antisense oligonucleotides. All images represent a single confocal slice. They were collected under non-saturating conditions using Leica software lut-glowov where the colour shift red -I yellow represents low 4 high intensity pixels. When A7r5 cells were grown under normal culture conditions in the presence of the antisense oligonucleotides 1A (a) or 2A (c), the peripheral signal was completely eliminated (arrows). The corresponding control oligonucleotides 1C (b) and 2C (d) had no effect on pactin mRNA localization. Bar: 40 pm in length.

Antisense inhibition of SMC migration Migration of all cells is dependent upon the reorganization of the actin cytoskeleton (Ridley and Hall, 1992; Ridley et al, 1992) and pactin mRNA at the motile periphery of cells may be associated with this reorganization (Hill et al, 1994). As antisense P-ODNs were capable of blocking peripheral localAntisense inhibition of pactin mRNA localization

ization of jl-actin mRNA under standard culture conditions, we investigated whether SMC migration could be suppressed using the same antisense I?-ODNs Using time-lapse video analysis, images were collected for analysis from 6-20 h after the addition of antisense or control P-ODNs. The results in figure 5 represent the mean distance moved by cells following the different treatments. Schedlich et al

Ecology of the Cell (1997) 89, 113-l Li

120

In the presence of the antisense P-ODN 1A (2 @I), migration of A7r5 cells was reduced by 50% of the control P-ODN 1C (P < 0.0001). The degree of SMC migration in the presence of control I’-ODN (2 PM) was not significantly different from untreated cells (P = 0.3). Migration was reduced by 40% when SMC were incubated with the antisense P-ODN 2A (2 ,IJM) compared to control I’-ODN 2C (P < 0.001). Therefore, the antisense P-ODNs are exerting an effect on cell migration in a sequence-specific manner.

DiSCUSS#~ A significant proportion of cellular actin is kept in a monomeric form through its interaction with actin binding proteins. These actin monomers continually undergo polymerization and depolymerization to generate protrusions such as the lamellipodia required for cell migration. In quiescent 3T3 fibroblasts, actin polymerization can be stimulated by growth factors resulting in the rapid formation of stress fibers and membrane ruffling (Ridley and Hall, 1992; Ridley et al, 1992). These effects have been shown to be mediated by the small GTPases Rho and Rat, respectively. Recently, we (Hill et al, 1994) and others (Latham et aI, 1994) have shown that the distribution of pactin mRNA in mouse 3T3 fibroblasts and chicken embryonic fibroblasts is also under growth factor control. We have now shown that growth factor regulation of intracellular /l-actin mRNA trafficking occurs in rat vascular SMC. Cells grown in the absence of serum were depleted of peripheral p-actin mRNA. Within 30 min of stimulation with PDGF, these quiescent cells had re-established their peripheral pool of P actin mRNA. As PDGF stimulation was not associated with an increase in transcript abundance during the time studied, the rapid appearance of pactin mRNA at the cell periphery may represent a redistribution of pactin transcripts already present in the serum-starved ceils. The observation that both translocation and anchorage of actin mRNA at the cell periphery in fibroblasts is independent of protein synthesis (Sundell and Singer, 1990) together with the rapidity with which redistribution of pactin mRNA occurs, suggested that newly synthesized proteins are not involved in this process. The response in SMC to PDGF is particularly interesting since following balloon injury to the rat carotid artery, SMC infiltration of the intima appears to be dependent on PDGF (Ferns et ai, 1991). Considering our results in this light, we predict that PDGF’s ability to stimulate the migratory response in SMC in uizlo may, in part, be a consequence of this factor’s ability to increase the peripheral pactin mRNA pool in these cells. A corollary to this would be that if peripheral localization of Antisense inhibitionof Bactin mRNAlocalization

25

-r-i

20

15

10

Nil

1A

1C

2A

2C

Fig 5. Effect of antisenseoligonucleotidetreatment on the migrationof A7r5 cells. Cell migrationwas anajysedin the presence of antisenseoligonucleotides1A (n=651 or 2A (n=49) and control oligonucleotides 1C (n=851 or 2C (n=45) and in untreated cells (n=98). The results are pooled from at least three separate experiments and repre. sent the average distance moved (Z&E) by cells following the different treatments.

pactin transcripts could be inhibited, SMC migration might also be inhibited. Sequences in the chicken pactin mRNA involved in the subcellular compartmentalization of these transcripts have recently been mapped to two regions in the 3’UTR of the mRNA (Kislauskis et uli 1994). These regions comprise a 54-nt ‘zipcode’ segment, which has been shown to be necessary and sufficient to localize heterologous transcripts to the periphery of chicken embryonic fibroblasts, and r: less active 43-nt segment. Furthermore, the anti-sense P-ODNs targeted to these regions were ahlc to inhibit peripheral localization of pactin mRNA in chicken fibroblasts. We have now shown tha: antisense oligonucleotides directed against a region in a rat p-actin S’UTR can block the peripheral localization of ,+actin mRNA in rat SMC. Tht region targeted by the antisense P-ODNs i.ics within the rat equivalent of the 54-nt RNA zipcode of chicken pactin 3’UTR. These sequences share 64% homology with the corresponding region ir the chick pactin mRNA and contain the GGAC I motif shown to be important for localization of thi:. Schedlich erai

121

Biology of the Cell (1997) 89, 113-l 22

transcript. Our results have identified a region within a mammalian pactin 3’UTR necessary for peripheral localization of its transcripts, consistent with previous work on chicken pactin mRNA (Kislauskis et al, 1994). There are a number of possible mechanisms that could account for the loss of peripheral p-actin mRNA following treatment with antisense P-ODNs. The oligonucl.eotides may bind specifically to the complementary sequences in the pactin 3’UTR to prevent its interaction with the translocation or attachment machinery. This would either halt the movement of /3-actin mRNA at the perinuclear boundary or prevent its anchorage at the cell periphery. An alternative mechanism involves the activation of RNAse H activity by the DNA/RNA duplex causing the degradation of pactin mRNA. Our results suggest that the effect of antisense I’-ODNs on pactin mRNA localization involves the interruption of transport of pactin mRNA to the cell periphery or its anchorage at the cell periphery. We base our conclusion on two observations. Firstly, the inhibition of localization was dependent on oligonucleotide sequence (control P-ODNs of identical base composition and length as the antisense P-ODNs did not block peripheral localization of p-actin mRNA) and secondly, the analysis of total cellular RNA from SMC treated with the antisense oligonucleotides showed no significant change in the level of pactin mRNA compared with untreated cells or cells treated with control P-ODNs, nor was there any evidence of degraded products. Evidence for the importance of actin at the cell periphery in cell migration originally came from studies in living cells where fluorescently labelled actin molecules were microinjected into migrating fibroblasts (Wang, 1984). The labelled actin was found to be incorporated into the filaments at the leading edge of the cell. More recently, the use of experimental models which examine wound healing have shown epithelial cells in viva localize actin at the cell periphery, generating actin cables at the wound edge as part of the process of wound closure (Martin and Lewis, 1992). Similarly, vascular endothelial cells in vitro localize pactin mRNA and protein peripherally (Hoock et al, 1991). Following injury to the endothelial cell monolayers, cells at the wound edge respond by moving rapidly to heal the wound. pactin mRNA has been localized to the motile regions of the cytoplasm in cells abutting the wound. In addition, pactin mRNA has been localized to the motile regions of myoblasts, fibroblasts and now in vascular SMC. The function of the association of pactin mRNA with motile regions of these cells is unknown. However, one plausible explanation is that it can provide new free pactin protein monomers locally at sufficiently high concentration Antisense inhibition of pactin mRNA localization

to support the rapid polymerization of actin filaments required for the cytoskeletal changes that occur in these motile regions (Hill et al, 1994). Antisense oligonucleotides which blocked peripheral localization of pactin mRNA in chicken embryonic fibroblasts also depleted the peripheral actin pool, This was associated with an altered lamellipodia structure, changes in actin stress fibre organization and a non-polarized phenotype (Kislauskis et al, 1994). This work suggests that loss of pactin mRNA from the cell periphery may be associated with the reorganization of the actin cytoskeleton. Our studies have demonstrated that loss of pactin mRNA from the cell periphery is associated with suppression of SMC migration. However, there was a difference in the degree to which the antisense oligonucleotides exerted its effect. Whereas the oligonucleotides were able to completely eliminate peripheral p-actin mRNA, it suppressed SMC migration by only 40-50%. This may suggest that the peripheral pool of pactin mRNA augments already existing pactin monomers at the cell periphery and that the loss of pactin mRNA from this region causes only a partial inhibition of SMC migration. In a number of disease states, inappropriate cell migration is involved in lesion formation. The use of antisense P-ODN to block the peripheral localization of pactin mRNA provides the opportunity to suppress cell migration and, together with an appropriate delivery technique, may be useful in a therapeutic context. The sorting of mRNAs to different regions of the cell is emerging as an important mechanism for targeting proteins to their functional domains. This localization of transcripts has been shown to be particularly important in early development where maternal mRNAs are directed to different regions in the embryo for targeted protein synthesis (Kim-Ha et aI, 1993; Macdonald e: al, 1993). In addition, we have now shown the importance of transcript localization in somatic cells where the loss of peripheral pactin mRNA is associated with suppression of cell migration. However, differences in transcript localization may not always be associated with a particular functional role. Antisense oligonucleotides targeted to localization sequences may be useful in determining whether transcript localization is of functional significance or simply an example of gratuitous mRNA localization.

ACKNOWLEDGMENTS The authors gratefully acknowledge Dr Mark Dziegielewski for assistance with time-lapse video analysis and Drs Phil Hendry and Minoo Moghaddam for advice and synthesis of phosphorothioated oligonucleotides. We also thank Professor Peter Rowe, Children’s Medical Research Schedlich et al

hology of the Ceil (1997) 89, %13-i 2/’

122

Institute, available

RE

Sydney, for making to us.

the facilities

of the Institute

ES

Bennett MR, Anglin S, M&wan JR, Jagoe R, Newby AC and Evans GI (1994) Inhibition of vascular smooth muscle cell proliferation in vitro and in viva by c-myc antisense oligodeoxynucleotides. JCfin Invest 93,820-828 Clowes AW and Schwartz SM (1985) Significance of quiescent smooth muscle migration in the injured rat carotid artery.

Circ Res 56,139-145 Ferns GAA, Raines EW, Sprugel KH, Motani AS, Reidy MA and Ross R (1991) Inhibition of neointimal smooth muscle accumulation after angioplasty by an antibody to PDGF. Science 253,1129-1132 Ferrell M, Fuster V, Gold HK and Chesebro JH (1992) A dilemma for the 1990s. Choosing appropriate experimental animal model for the prevention of restenosis. Circulu-

tion 85,1630-1631 Hill

MA and Gunning P (1993) Beta and gamma actin mRNAs are differentially located within myoblasts. I Ceil Bioi 122, 825-832 Hill MA, Schedlich L and Gunning I-’ (1994) Serum-induced signal transduction determines the peripheral location of pactin mRNA within the cell. J Ceil Biol126,1221-1230 Hoock TC, Newcomb PM and Herman IM (1991) /l Actin and its mRNA are localized at the plasma membrane and the regions of moving cytoplasm during the cellular response to injury. J Cell Biol 112653-664 Huttenlocber A, Sandborg RR and Horwitz AF (1995) Adhesion in cell mieration. Curr Chin Cell Biol 7,697-706 Kim-Ha J, Webs‘;er PJ, Smith JL and Macdonald PM (1993) Multiple RNA regulatory elements mediate distinct steps in localization of oskar mRNA. Development (Camb) 119, 169-178 Kislauskis EH, Li Z, Singer RH and ‘Taneja KL (1993) Isoformspecific 3’ untranslated sequences sort cc-cardiac and Pcytoplasmic actin messenger RNAs to different cytoplasmic compartments. J Cell Biol 123,165-172 Kislauskis EH, Zhu X and Singer RH (1994) Sequences responsible for intracellular localization of pactin messenger RNA also affect cell phenotype. 1 Cell Biol 127,441-451 Latham VM, Kislauskis EH, Singer RH and Ross AF (19943

Antisense Inhibitionof Pactin mRNA localization

pactin mRNA localization is regulated by signai transouc tion mechanisms. 1 Cell Biol126,127 l-121Y Luna EJ and Hitt AL (1992) Cytoskeleton-plasma mernbran:. interactions. Science 258,955-964 Macdonald PM, Kerr K, Smith JL and Leask A (1993 j i',P;.A reg ulatory element BLEI directs the early steps of bicciti mRNA localization. Dezlelopment (Camb) 118.1233-I 243 Martin P and Lewis J (1992) Actin cables and epidermal moor ment in embryonic wound healing. Nafttrc 360,179-183 Morishita R, Gibbons GH, Ellison KE, Nakajima M. Zhang L Kaneda Y, Ogihara T and Dzau VJ (1993) Single intraluminal delivery of antisense cdc2 kinase and proliferating-cell nuclear antigen oligonucleotides results in chronic inhibi. tion of neointimal hyperplasia. Proc Natl Acad Sci LISA 90 8474-8478 O’Brien ER and Schwartz SM (1994) Update on the biology. and clinical study of restenosis. TCM 4,169--178 Ridley AJ and Hall A (1992) The small GTP-binding protei:i rho regulates the assembly of focal adhesions and a&t stress fibers in response to growth factors. Cell 70, 389-399 Ridley AJ, Paterson HF, Johnston CL, Diekmann D and Hall :‘i (1992) The small GTP-binding protein rat regulates growth factor-induced membrane ruffling. Ceil 70; 401410 Serruys PW, Luijten HE, Beatt KJ, Geuskens R, de Feyter Pi. van den Brand M, Reiber JH, ten Katen HJ, van Es GA am 1 Hugenholtz PG (1988). Incidence of restenosis after successful coronary angioplasty: a time-related phenomenon. A quantitative angiographic study in 342 consecutive patients at 1,2,3, and 4 months. Circulation 77, 361-,371 Simons M, Eddman ER, DeKeyser JL, Langer R and Rosenberg RD (1992) Antisense c-myb oligonucleotides inhibit intimai arterial smooth muscle cell accumulation irr zG7w liafulzF 359,67-71, Sundell CL and Singer RH (1990) Actin mRNA IocaliLes m the absence of protein synthesis. J Cell Biol 111,2397-2403 Sundell CL and Singer RH (1991) Requirement of microtiiaments in sorting of actin messenger RNA. Sr:ence 2% 1275-1277 Vandekerckhove J and Weber K (1978) Mammalian cytopldsmic actins are the products of at least two genes and differ in primary structure in at least 25 identified positions from skeletal muscle actins. Proc Nut! Acad Sci USA 75,1106-! 7! 0 Wang YL (1984) Reorganization of actin filament bundles :n living fibroblasts 1 Cell Biol99, 1478-1485 Received

7 April

1997; accepted

30 May

1997

Schedltch et ai