Biol Cell (1996) 87, 37-43 0 Elsevier, Paris

37

Original article

Internalization of oligodeoxynucleotide antisense to type-l plasminogen

activator inhibitor mRNA in endothelial cells: A three-dimensional reconstruction by confocal microscopy

Elzbieta Wyroba a, Zofia Pawlowska b, Anna Kobylanska c, Elzbieta Pluskota b, Maria Maszewska c, Wojciech J Stec c,

Czeslaw S Cierniewski b

a Nencki Institute of Experimental Biology, Polish Academy of Sciences, 3, Pasteur Str, 02-093 Warsaw; b Departmen f t o Biophysics, Medical University in Lodz;

c Center for Molecular and Macromolecular Research, Polish Academy of Sciences, L.odz, Poland (Received 26 June 1996; accepted 15 October 1996)

Summary - A three-dimensional reconstruction analysis of localization of phosphodiester and phosphorothioate oligonucleotide anti- sense to type-l plasminogen activator inhibitor (PAI-1) mlXNA within endothelial cells is described. When EA.hy 926 cells were incu- bated with fluorescently labelled phosphodiester (PO-16) or phosphorothioate (PS-16) oligonucleotides at low, not cytotoxical concen- trations, the relative brightness composition of the images of the particular samples was much higher for PS-16 than PO-16 and dependent upon the extracellular concentration and the incubation time. The 3-D reconstructions based on the series of optical sections of the samples, spaced every 1.5 ,clm, showed the punctuate accumulation of the oligonucleotides and a striking difference in a spatial distribution between PO- 16 and PS- 16 within the cytoplasm. Even after 24 h incubation of endothelial cells with 2.5 /.M of PO- 16 and PS-16 oligonucleotides, there was a predominant oligonucleotide localization within the cytoplasm and only traces of oligonucleotides could be seen in the cell nucleus and/or perinuclear organelles.

antiseme oligonucleotides / spatial distribution / eonfocal microscopy / human endothelial cell line

Introduction

Antisense oligonucleotides designed to hybridize to specific mRNA or double-stranded DNA sequences are effective tools for control of the expression of target genes [7, 191. These compounds have been used to analyse the functions of various genes, since short DNA molecules can inhibit synthe- sis of particular proteins in cells. However, before practical application of oligodeoxynucleotides, a number of problems including poor biological stability and cell delivery have to be solved. Although the fully thioated and other modified oli- godeoxynucleotides are more resistant to nucleases, they have several non-sequence-specific effects, including reduced hybridization, non-selective protein binding, altera- tion in cell morphology and proliferation [4, 10,22,23,25].

The mechanism by which antisense oligodeoxynucleo- tides enters the cells is not fully understood. Some evidence is available on the existence of receptors on the cell surface [9, 241, and consequently, the uptake of oligodeoxynucleo- tides has been shown to be time- and concentration-depen- dent [ 1, 16, 22, 241. Such a receptor with a relative molecu- lar mass of 80 kDa was isolated from the membranes of CHO fibroblasts, HL60 cells and other cell types [16, 241. In HL60 cells in addition to p80 kDa, at least one other oli- gonucleotide surface binding site was found [ll, 121. It was concluded that internalization of oligodeoxynucleotides in these cells seems to depend primarily on fluid-phase pino-

cytosis. Observations based on flow cytometry and fluores- cence microscopy show that within 0.5-2.0 h, an apparent steady-state distribution of the antisense oligodeoxyribonu- cleotide phosphorothioate (28-mer) specific for rev gene- encoded RNA of HIV-l in H9 cells can be achieved. Cellu- lar uptake as a function of the external oligonucleotide concentration was non-linear, being more efficient at con- centrations below 2 @4, and there was predominant oligo- nucleotide localization within the cell nucleus and perinu- clear organelles [ 181.

Recently, we identified antisense phosphorothioate oli- godeoxynucleotides which inhibited the synthesis of type-l plasminogen activator inhibitor in endothelial cells in a time- and concentration-dependent manner [5, 61. Due to a potent physiological role of PAI- in regulation of fibrino- lysis [ 171, the strategy to develop specific antisense therapy is of great interest.

In this report we attempted to characterize intracellular localization of these compounds in endothelial cells using a laser confocal microscopy. We have prepared fluorescein conjugates of both phosphodiester and phosphorothioate forms of the most inhibitory oligonucleotide antisense to PAI- mRNA and studied their uptake and distribution in endothelial cells applying a 3-D reconstruction. As far as we know such kind of spatial reconstruction has not been yet used to study cellular uptake of antisense oligonucleo- tides at low not cytotoxical concentrations.

Materials and methods

Human endothelial cell line EA.hy 926, derived by fusion 01 human umbilical vein endothelial cells with continuous human lung carcinoma cell line A549, was obtained as a gift from Pro- fessor Cora-Jean S Edge11 (Pathology Department, University of North Carolina at Chapel Hill). The cells were cultured in DMEM with high glucose, supplemented with 10% foetal calf serum, HAT (100 m hypoxanthine, 0.4 @l aminoptetin, and 16 PM thymidine), and antibiotics in a 90-95% humidified atmosphere of 5% CO, at 37°C. For microscopic examination cells were plated at a density of 5 x 104 cells/well an Thermanox Coverslips in 8-well tissue culture chamber slides (NUNC) with detachable ehambered upper strueteres. Before performance of assays the serum-containing medium was changed to a serum- free medium. The cultures were gently rinsed three times with the medium and preincubated with fluorescein-labelled antisense oligonucleotides for selected periods of time. Attached treated intact cells were maintained in CC?, incubator at 37°C. Two con- trol assays were carried out using either untreated cells or cells exposed to 0.25 PM fluorescein. After incubation cells were washed three times with PBS, fixed with a~freshly prepared 3.5% paraformaldehyde for 15 min at rooti temper+@ then washed three times with PBS and mounted in-2.5% DABCO in glycerol and processed for microscopy.

Antisense oligonucleotides

A fluorescein labelled oligonucleotide was synthesised by the phosphoramidite method on an AI31 391 synthesizer. After last condensation and final detritylation the column was dried under high vacuum. Both oligonucleotides PO-16 and PS-16 were acti- vated by 0.3 M solution of N,N’-carbonyl diimidazoIe (CDI) in dioxane. After 1 h of activation ~the support was washed with dioxane and dried under high vacuum. Then the activated oli- gomer was treated with 0.1 M solution of ethylenediamine in dioxane (2 h at room temperature). After washing with dioxane and drying, 0.0.5 M solution of fluorescein isothiocyanate in a mixture of 0.1 M TEAB and DMF (20:80) was added to the col- umn. The column was kept for 48 h at room temperature in dark- ness. Finally the support was washed with the mixture of 0.1 M TEAB and DMF (20:80) and with acetonitrile. The cleavage of corresponding oligonucleotide from the support and deprotection step were carried out according to a standard procedure described in our previous paper [5]. The sequence of the oligonucleotide was 5’-d(GAGGGCTGGAGACATC).

Confocal jl*orescence microscopy

For intracellular probe visualization the confocal laser scanning microscope CLSM Phoibos 1000 (Molecular Dynamics, USA) which is based on Nikon Optiphot microscope equipped with Nikon oil immersion objectives 60 x NA 1.4 and 100 x NA 1.4 with an argon laser with fluorescein filter (line selection 488 nm, dichroic mirror 5 10, emission LP 5 10 nm) was used. Digital pho- tographic images were printed using Kodak Dye-Sublimation Printer KLS86OOPS.

The 3-D reconstructions were prepared on Silicon Graphics Personal Iris (Silicon Graphics, Mountainview, CA) workstation. Reconstructions were based on 8-15 optical sections, spaced every 1.5 m, yielding images that were 5 12 x 5 12 pixels. For all reconstructions maximum intensity voxels placed on the view axes (rotated longitudinally 334.7”, latitudinally -52.5” and along the axis 333.9”) were shown. For 3-D reconstmction the Image Space (Molecular Dynamics) software was used.

Results

Internalization of two forms of fluorescein-labelled oligo- nucleotides antisense to PAL1 mRNA by human endothe-

i1a1 cell line % as studied under non-cytotoxical cxperimeri la1 conditions. The fluorescence signal was visual&d usi@ the laser confocal microscopy.

In order to characterise the oligonucleotlde uptake, this relative brightness composition of the cell images obtaincAt for different cell samples was compared (fig I). Endothelial cells were incubated for 4 and 24 h with PO-16 and PS-II; and an average brightness of the images, taken from the section through the middle of the cells of the particula! sample, was quantified. The average brightness of the images, reflecting the concentration of the oligonucleotides

A 50 l--_l___

4.5 I--

40 -4

F3 50 - i - - - . - - - - - - -.l~_----___l_.l

I 45 j

O-J--

Control PO-16 1.25 pM PS-16 1.25 pM

Control - FITC PO-1 6 2.5 pM 4%-16 2.5 $4

Fig 1. Comparison qf the uptake of PO-16 and PS-16 by endo. thelial cells EA. hy 926 after 4 h (A) and 24 h (B) of exposure to antisense oligonucleotides. Endothelial cells were treated ~witb 1.25 and 2.5 ,uM of oligonucleotides at 37’C as described in Muterids and methods. The image histograms of the samples taken from section through the middle of the ceils (n = 48 % 5) from different samples were compared. Plots of the relative brightness composition of the image illustrate an average bright- ness of the images of the particular sample and reflect- the ccxv centration of fhe oligonucleotides within the cells.

3-D reconstruction analysis 39

within the cells, was much higher for PS-16 than for PO-16. The intracellular accumulation of the antisense oligonucleo- tides was concentration-dependent only following a 24-h exposure to PS-16 and PO-16 (fig 1B). Comparison of the fluorescence image of cells treated with 2.5 PM of PS-16 and PO-16 for 4 h is illustrated in figures 2 and 3, respec- tively. Optical sections imaged through cells by confocal microscopy (at the increments of 1.5 ,!&n) localized oligonu- cleotide accumulations within the cytoplasm. The punctuate fluorescent pattern was found in all the tested samples. A

higher fluorescence intensity was observed in the case of PS-16 (fig 2 a-h) than with PO-16 (fig 3a-h) in accordance with the data presented in figure 1. Fluorescence was hardly visible within the cell nucleus, which is particularly clear when a 3-D reconstruction was performed. In order to fur- ther analyse the intracellular distribution of PS-16 and PO- 16 within the endothelial cells, three types of images were reconstructed (figs 4, 5): A) the maximum intensity look- through reconstruction of the sample computed from the optical slices; B) the vertical slice through each sample

Figs 2,3.2. The fluorescence image of endothelial cells treated with phosphodiester oligonucleotide (PS-16) antisense to PAI- mRNA. 2.5 FM of PS-16 labelled with fluorescein was incubated for 4 h at 37°C with endothelial cells as described in Materials and methods and analysed by confocal fluorescence microscopy. The series of the optical sections (a-h) scanned through cells at increments of 1.5 pm is displayed. Bar represents 5 pm. 3. The fluorescence image of endothelial cells treated with phosphorothioate oligonucleotide (PO-16) antisense to PAI-1 mRNA. Endothelial cells exposed to 2.5 PM of fluorescein-labelled PO-16 for 4 h at 37°C and processed as described in Materials and methods were analysed by confocal fluorescence microscopy. The series of the optical sections (a-h) scanned through cells at increments of 1.5 pm is displayed. Bar represents 5 pm.

PS- 16 1.25~M 4h

PS- 16 2.SpM 4h

PO-16 1.2SpM 4h

PO-16 2.SpM 4h

Fig 4. Three-dimensional reconstructions of the fluorescein-linked oligonucleotidts ES-16 and PO-16 distribution wjthinendott ceil s after 4 h of incubation. Three types of the images correspond to the top view of the sample (A). the vertical slices through sari rple along the line shown by arrow in the columns -A and B, and the perspective projection of the maximum intensity lobk-thr recc ,nstruction of the sample computed from the optical slices (C). Cells were treated with 1.25 or 2.5 PM of oligonucleotides at I resl ,ectiveTy as described in Matericrls md mefhodt

PS-16 1.2SyM 24h

PS- 16 2.SpM 24h

PO- 16 1.25pM 24h

PO-16 2.SpM 24h

Fig 5. Three-dimensional reconstructions of the fluorescence-linked oligonucleotides PS-16 and PO-16 distribution within endothelial cells after 24 h of incubation. Three types of the images correspond to the top view of the sample (A), the vertical slices through each sample along the line shown by arrow in the columns A and B, and the perspective projection of the maximum intensity look- through reconstruction of the sample computed from the optical slices (C). Cells were treated with 1.25 and 2.5 PM oligonucleotides at 37°C.

Fig 6. A top view of the 3-D reconstruction of the cell (visual- ised in figure 5, column A, treated with PS-16 for 24 h) at high magnification. Note a punctate fluorescence pattern and a lack of oligonucleotide accumulations within the nucleus of this typical endothelial cell. Bar represents 50 pm.

along the line shown by arrow in the column A (the whole series of optical sections performed); and C) the perspective projection. The 3-D reconstructions (figs 4, 5, column C) showed the striking difference in the relative ‘depth’ of PO- 16 and PS-16 penetration inside the~cells. This difference is particularly well seen after 24 h. Even after such a long time the punctuate fluorescence could be seen in cytoplasm of cells treated with either PO-l 6 or PS-16. The vertical slices through the series of optical sections performed (figs 4, 5, column B) show only traces of scattered fluore- scence within and/around the cell nucleus. Lack of accumu- lation of antisense oligonucleotides within the nucleus is further confirmed in a 3-D reconstructed cell sample dis- played at high magnification (fig 6).

Discussion

The major observation of this study is that at the low con- centration gradient of oligonucIeotides antisense to PAI- 1 mRNA, both phosphorodiester- and phospliorothioate forms are taken up by endothelial cells and localized almost exclu- sively within the cytoplasm. Oligonucleotide distribution was examined by confocal microscopy which enables scarl-- ning of multiple thin (1.5 pm) parallel sections through indi- vidual cells. This facilitates optical separation .of intracyto- plasmic from nuclear fluorescence. The punctuate fluorescence distribution observed even after 24 h of expo- sure to the antisense oligonucleotides seems to suggest that

endosomal vesicles are~the primary target of the pro&% under this study:However, the relative depth of penetration inside the cell ‘is significantly different for both forms of oli- gonucleotides and primary depends upon the time of incuba- tion. Such oligonucleotide distribution within the cells uas evidenced by confocal fluorescence microscopy and particu- larly well visible on the three-dimensional recon&uctions of the cell sample images. The highest oligonucleotide con,- centration used .in this study equals 2.5 PM, ie high enough td block synthesis of PAI- in.endothelial cells ES, 6f, but significantly lower than that used in. the most studies aimed at characterization of oligonucleotides transport into the ceils [13, 181. It is noteworthy that in the experiments per- formed on the endothelial cells, phosphorothioate oligonu- cleotides may-be used only at the limited range-of concen- tration since the incubation with a dose higher than 10 PM resuIts in an increased cytotoxicity. The izytoplasmic 1OCid

ization of PO- 16 and PS- 16 appe& to be in contrast to pre- vious observations showing that antisense oligonuckotides can also-be found heavily accumulated in the nucleus. How- ever, in those studies oligonricleotides were introduced directly into the cell cytoplasm by microinjection [l;t], DOTMA treatment [2] or other cationic liposomes [t3, l.S], electroporation [3], streptolysin 0 treatment [20] or cells were treated at. high -concentrations of oligodeoxynucleo- tides, lo-fold and higher than those used in this report. Since, at the .low oligonucleotide concentration .ahsorptive endocytosis plays a major role in uptake whereas .at moder- ate and high oligonucleotide concentrations; fluid-phase cndocytosis is the predominant process [21, 241, &appears that various mechanisms of transport result in different oli-- gonucleotide distribution within the cell. Our stud& showed that due to a significant compartmentalisation oF oligonucleotides within the cells, recently published intra- cellular concentrations based ori the flow cytometric experi- ments 1181 appear to be highly underestimated. Depending on the type of hematopoietic cell lines (K562, U937. HL60. H9) it was suggested that when cells are incubated with 25 PM of oligonucleotide (27-base long) for 1 h, an ovtial! intracellular concentration varied from 0.13 to 0.3. @I [ 181. Since our observations indic&e that oligonucleotides are not equally distrib&zd, therefore the local concentration withm intracytoplasmic accumulations (vesicles) may be the some as or even higher than cell medium concentrations.

Antisense oligonucleotides are taken up by the cells in ;# saturable, size-dependent manner compatible with receptor- mediated endocytosis. An 8O-kDa surface protein that may mediate transport was identified [l6, 241. In different hematopoietic cell lines a quasi steady-state is accorii- plished with the added oligonucleotide within l-2 h 118 1 and an efflux of oligonucleotide from the cell into the media was observed at the same time constant as that for the uptake. Previous observations showed that a major dif ference between the phosphodiester tinked and phospho- rothioate linked oligonucleotides appears to be that kinetici; of the influx and effiux of the former is approximatei> 5-fold faster than the latter. Under experimental conditions used in this study intracellular accumulation of oligonucleo- tides was time-dependent and after 24 h a fluorescencl intensity of cytoplasmic vesicles was much higher than thar after 4 h. Cytoplasmic accumulations containing phosphors diester oligonucleotides were localized much .closer :I 4 membranes than those loaded with phosphorothioate oligo- nuckotides. Based on the observations under confocar fluc~ rescence microscopy, this material which under convex+ tional ftuorescence microscopy looked like it was located within the nucleus, appeared to be associated &her with

3-D reconstruction analysis 43

nuclear membrane. Deeper optical sections through the nucleus did not show the presence of fluorescein-labelled oligonucleotides, when phosphorothioate or phosphodiester oligonucleotides were used. Such a localization suggests that translational arrest may be the dominant mode of action of antisense oligonucleotides to PAI- 1 mRNA.

Our studies showed that after 24 h, the extent of uptake of phosphorothioates by endothelial cells was much higher than that of phosphodiester-linked oligonucleotides, and this effect was already seen after 4 h. This difference may result either from higher tendency of phosphorothioates to non-specific binding to cellular components or their increased resistance to cellular nucleases. On the other hand, we do not know to what extent phosphodiester oligo- nucleotides are protected from cleavage once they are incorporated into intracellular substructures. Most experi- ments on stability of oligonucleotides which were published up to now, were performed on cell extracts in which mem- branes of intracellular compartments are disintegrated [8, 131, and thus lysosomal enzymes have free access to oligo- nucleotides. Due to compartmentalization of oligonucleo- tides within the cell, the stability of such oligonucleotides may be significantly higher.

Acknowledgments

This work was supported by the statuable grant to the Nencki Institute of Experimental Biology (EW), the project 4.P05A. 138.08 (CSC) and 4.P05A.003.10 (ZP) from the State Committee for Scientific Research.

References 17

Akhtar S, Basu S, Wickstrom E, Juliano RL (1991) Interac- tions of antisense DNA oligonucleotide analogs with phos- pholipid membranes (liposomes). Nucleic Acids Res 19, 5551-5559 Bennett CF, Chiang MY, Chan H, Shoemaker JEE, Mira- belli CK (1992) Cationic lipids enhance cellular uptake and activity of phosphorothioate antisense oligonucleotides. Mel Pharmaco141,1023-1033 Bergan R, Connell Y, Fahmy B, Neckers L (1993) Electro- poration enhances c-myc antisense oligodeoxynucleotide efficacy. Nucleic Acids Res 21,3567-3573 Cazenave C, Stein CA, Loreau N, Thoung T, Neckers LM, Subasinghe C, Helene CS, Cohen JS, Tolume JJ (1989) Comparative inhibition of rabbit globin mRNA translation by modified antisense oligodeoxynucleotides. Nucleic Acids Res 17,4255-4273 Cierniewski CS, Babinska A, Swiatkowska M, Wilczynska M, Okruszek A, Stec WJ (1995) Inhibition by modified oli- godeoxynucleotides of the expression of type-l plasminogen activator inhibitor in human endothelial cells. Eur .I Bio- them 227,494-499 Ciemiewski CS, Pluskota E, Krzeslowska J, Fijalkowska I, Kobylanska A, Koziolkiewicz M, Okruszek A, Stec WJ (1995) Inhibition by modified oligodeoxynucleotides of the expression of type- 1 plasminogen activator inhibitor in human endothelial cells. Thrombos Haemostas 73, 411 (abstract) Erickson RP, Izant JG (1992) Gene regulation: Biology of Antisense RNA and DNA. Raven Press, New York, p 384

8

9

10

11

12

13

14

15

16

18

19

20

21

22

23

24

25

Gao WY, Jaroszewski JW, Cohen JS, Cheng YC (1990) Inhibition of herpes simplex virus type-2 growth by phos- phorotioate oligodeoxynucleotides. J Biol Chem 265, 20172-20178 Geselowitz D, Neckers L (1992) Analysis of otigonucleotide binding, internalization and intracellular trafficking utilizing a novel radiolabelled crosslinker. Antisense Res Dev 2, 17-25 Ghosh M, Ghosh K, Cohen JS (1992) Translation inhibition by phosphorothioate oligodeoxynucleotides in cell-free systems. Antisense Res Dev 2, 11 l-l 18 Goodarzi G, Watabe M, Watabe K (1991) Binding of oligo- nucleotides to cell membranes at acidic pH. Biochem Biophys Res Commun 181,1343-1351 Kitajima I, Shinohara T, Minor T, Bibbs L, Bilakovic J, Nerenberg M (1992) Human T-cell leukemia virus type-l tax transformation is associated with increased uptake of oli- godeoxynucleotides in vitro and in vivo. J Biol Chem 267, 25881-25888 Lappalainen K, Urtti A, Soderling E, JPaskellinen I, Syrj&nen K, Syrjanen S (1994) Cationic liposomes improve stability and intracellular delivery of antisense oligonucleo tides into CaSki cells. Biochem Biophys Acta 1196,201-208 Leonetti JP, Mechti N, Degols G, Gagnor C, Lebleu B (1991) Intracellular distribution of microinjected antisense oligonucleotides. Proc Nat1 Acad Sci USA 88,2702-2706 Loke SL, Stein CA, Zhang X, Avignan M, Cohen IS, Neck- ers LM (1988) Delivery of c-myc antisense phosphorothi- oate oligodeoxynucleotides to hematopoietic cells in culture by liposome fusion: specific reduction in c-myc protein expression correlates with inhibition of cell growth and DNA synthesis. Curr Top Microbial Immunol141,282-289 Loke SL, Stein CA, Zhang XI-I, Mori K, Nakanishi M, Sub- asinghe C, Cohen JS, Neckers LM (1989) Characterization of oligonucleotide transport into living cells. Proc Nat1 Acad Sci USA 86,3474-3478 Loskutoff DJ, Sawdey M, Keeton M, Schneiderman J (1993) Regulation of PAI- gene expression in vivo. Thromb Hae- mostas 70, 135-l 37 Marti G, Egan W, Noguchi P, Zon G, Matsukura M, Broder S (1992) Oligodeoxynucleotide phosphorothioate fluxes and localization in hematopoietic cells. Antisense Res Dev 2, 27-39 Murray JA, (1992) Antisense RNA and DNA. Wiley-Liss, New York Spiller DG, Tidd DM (1995) Nuclear delivery of antisense oligodeoxynucleotides through reversible permeabilization of human leukemia cells with streptolysin 0. Antisense Res Dev 5, 13-21 Stein CA, Mori K, Loke SI, Subasinghe C, Shinozoka K, Cohen JS, Neckers, LM (1988) Phosphorothioate and nor- mal oligodeoxyribonucleotides with 5’-linked acridine: char- acterization and preliminary kinetics of cellular uptake. Gene 72,333-341 Stein CA, Subasinghe C, Shinozuka K, Cohen JS (1988) Physicochemical properties of phosphorothioate oligonu- cleotides. Nucleic Acid Res 16, 3209-3221 Watson PH. Pon RT, Shiu RP (1992) Inhibition of cell adhe- sion to plastic substratum by phosphorothioate oligonucleo- tides. Exp Cell Res 202,391-397 Yakubov LA, Deeva EA, Zarytova VF, Ivanovna EM, Ryte AS, Yurchenko LV, Vlassov VV (1989) Mechanism of oli- gonucleotide uptake by cells: involvement of specific recep- tors? Proc Nat1 Acad Sci USA 86,6454--6458 Yaswen P, Stampfer MR, Ghosh K, Cohen JS (1993) Effects of sequence of thioated oligonucleotides on cultured human mammary epithelial cells. Antisense Res Dev 3, 67-77