Oligonucleotide biological activity: Relationship to the cell cycle and nuclear transport

Biology of the Cell (1997) 89, 257-261 0 Elsevier, Paris

257

Original article

Oligonucleotide biological activity: Relationship to the cell cycle and nuclear transport Susanna Wu-Ponga, Julie Bard a, John Huffman b, Jason Jimerson b aDepartment of Pharmacy and Pharmaceutics, Box 980533; bSchool of Pharmacy, Virginia Commonwealth University, Richmond, VA 23298, USA

Previous studies suggest that oligodeoxynucleotide (ODN) cellular uptake is cell cycle-dependent which may have important implications in cancer cell targeting. To further our understanding of ODN transport and activity, this study examines the relationships between the cell cycle, ODN cellular uptake, intracellular transport, and activity. An antisense c-myc ODN 21-mer was used to study ODN cellular uptake in Rauscher erythroleukemia cells synchronized by either chemical methods or flow cytometry. ODN uptake was examined using subcellular fractionation and confocal fluorescence microscopy. Western blot analysis was used to measure ODN-mediated decreases in c-myc protein levels. Intracellular ODN distribution and extent of uptake was influenced by the phase of the cell cycle, but the mechanism of uptake was not. The relative activity of the antisense ODN was positively correlated to ODN distribution to the cytosol, but negatively correlated to total cellular uptake. Although ODN total cellular uptake is positively influenced by the cell cycle, retention of the ODN in the cytosol (presumably extra-vesicularly) appeared to be relevant in determining the activity of an antisense ODN. Novel methods to target cytosol-acting drugs to the cytoplasm may therefore be warrented.

oligonucleotide / uptake / cancer / cell cycle

INTRODUCTION Oligodeoxynucleotides (ODNs) have demonstrated remarkable potential as gene regulators in animal and in vitro studies, and are presently being tested in clinical trials. These new therapeutic agents are single-stranded DNA molecules, generally up to 25 nucleotides in length, which may be derivitized to increase stability or cellular transport. ODNs are typically designed to bind with high affinity to DNA (triplex-forming ODNs), RNA (antisense ODNs), or proteins (aptamers) resulting in specific inhibition of target genes or proteins (for review see Wu-Pong, 1994). Cellular internalization of these hydrophilic macromolecules is inefficient, yet remains poorly defined. Both endocytic and cesses have been implicated in transport, and the role of a receptor remains unclear (Loke

Oligonucleotide biological activity

non-endocytic proODN cell membrane specific membrane et al, 1989; Shoji et a2,

1991; Wu-Pong et al, 1992; Stein et aI, 1993; Zamecnik et al, 1994). Previous studies suggest that variations in uptake between cells in a single culture may

result from differences

in cell cycle phase, with

dividing cells possibly internalizing ODN with greater efficiency (Krieg et al, 1990; Noonberg et al, 1993; Wu-Pong ef al, 1994, Zamecnik et al, 1994). However, the role of the cell cycle in ODN cellular uptake is largely unknown, despite the obvious implications of ODN targeting to proliferating cancer cells. Therefore, the role of the cell cycle in ODN cellular uptake, nuclear transport and activity was examined in this study. ODN cellular uptake was most efficient in dividing, rather than resting cells, although the mechanism of uptake appeared to be independent of cell cycle phase. However, antisense ODN activity appeared to be positively correlated to cytosol uptake, not total cellular uptake, suggesting that ODNs with sites of action in the cytosol should be designed for cytosolic retention.

Wu-Pong et al

Biology of the Cell (1997) 89, 257-261

258

MATERIALS AND METHODS Materials Antisense c-myc phosphodiester oligonucleotide 21-mer (5’-GAAGTTCACGTTGAGGGGCAT) with or without the 3’ FITC label, scrambled sequence control (5’AGTGAGTCACGTAACGGTTGG) and G-rich control (5’-AGGGATTACGCTTAGGGGCGTA) were synthesized and purified by Oligos Etc (Wilsonville, USA). Azide, trypsin, ‘T4 kinase, Pactin antibody and SDS (sodiumdodecyl sulfate) were purchased from Fisher Scientific (Fairlawn, USA); dionyl phthalate was purchased from Kodak (Rochester, NY). p2P-ATP (adenosine triphosphate) was purchased from Amersham (Arlington Heights, USA); fetal calf serum was purchased from BioWhittaker (Walkersville, USA); c-myc antibody was purchased from Upstate Biotechnology (Lake Placid, USA). Lovastatin was a gift from Dr Stravitz, Massey Cancer Center, Virginia Commonwealth University. All other chemicals were purchased from Sigma (St Louis, USA).

Cell synchronization Cells were synchronized using either chemical methods or flow cytometry. In flow cytometry, cells in log phase growth were resuspended in serum-free media then incubated with a viable nuclear stain, 10 PM Hoescht 3342 for 1 h at 37°C. Cells were resuspended in dye- and serum-free media at 4-5 x 106 cells/ml, then sorted (Gl, G2/M, S phases) using the Coulter EPICS 753 dual laser flow cytometer (Coulter Corp, Miami, USA). Cells were chemically synchronized by preincubating cells in log phase growth for 16 h at 37°C with either 60 PM lovastatin (Gl), 0.65 mM hydroxyurea (GZ), or 0.06 pg/mL demecolcine (M). Flow cytometry was used for confirmation of cell cycle enrichment following chemical synchronization.

Mechanism of uptake The antisense c-myc ODN was 5’ end-labeled with 32P (32P-ODN) using standard methods (Sambrook et al, 1989). This ODN was previously shown to be stable in the Rauscher cells for at least 2 h (Wu-Pong et al, 1992). Cells were either synchronized or grown to confluency, then resuspended in serum-free media to 12 - 20 x 106 cellsfml. Cells were then pretreated with either 10 mM azide for 60 min or 0.25% trypsin for 10 min at 37°C. At time = 0, 0.1 PM 32P-ODN was added to the cells, incubated at 37”C, and triplicate 100 PL samples were removed during the 2-h incubation, since the 32PODN was demonstrated in previous studies to rapidly associate with the Rauscher cells, and reach pseudosteady state after 30 to 60 min (Wu-Pong ef ~2‘1992). Cells were centrifuged through 100 PL of 2:7 dionyl:dibutyl pthalate, clipped from the tube, and dissolved by vortexing in 0.1% SDS solution. Cell associatedradioactivity was measuredby scintillation counting.

tntracellular

location

The intracellular destination of the ODN was determined using two different methods. First, subcellular fractionation was used to measure 32P-ODN transport to the

Oligonucleotide biological activity

nucleus.Cellswere preincubated with 3H thymidine for 8 h, then synchronizing agents (lovastatin, demecolcine,or hydroxyurea) were added for an additional 16-h incubation. Cells were washed and resuspendedin serum-free media, then incubated with 0.1 PM 32P-ODN for 2 h at 37°C. Next, the cells were washed twice in phosphate buffered saline (pH 7.2), resuspended in RSB (10 mM Tris-HCl, pH 7.5, 10 mM KCl, 1.5 mM MgCl,, and 0.25% Triton X-100), then homogenized with a ground glass homogenizer (Salvatori et al, 1994). The nuclei were removed by centrifugation at 1000 g and washed with RSB to remove residual cytosol. RSB washes of nuclei were combinedwith cytosol fractions. To measurethe relative purity of each fraction, each fraction was separatelyassayedfor a cytoplasmic marker, lactate dehydrogenase (LD) or for a nuclear marker, 3H thymidine labeled DNA (3H-DNA). LD activity was measuredin each fraction by determining the change in absorbanceat 340 nm which resulted from conversion of 1 mg/mL NADH to NAD+ in the presenceof 22.7 mM sodium pyruvate. The relative amount of 3H-DNA in each fraction was determined by lysing labeled cells and denaturing 3H-DNA with 10% trichloroacetic acid and 1% sodium pyrophosphate, filtering the 3H-DNA through Watman GF/C glassfiber filters, washing the filters with three volumes of ethanol, and counting the radioactivity associatedwith the filters in a scintillation counter. The purity of eachfraction was estimatedbasedon LD and 3H-DNA data and used to calculate the extent of ODN transport to the cytosol or nucleus. The estimated fraction of ODN which was actually transported to the nucleus was calculated by determining the ratio of the yield of ODN in the nuclear fraction (ie nuclear 32PODN/total cellular 32P-ODN)to the yield of nuclei in the nuclear fraction (ie nuclear 3H-DNA/total cellular 3HDNA). Since almost complete recovery of cytosol was achieved in the cytosol fraction (on average > 97% of LD activity was recovered in the cytosol fraction), no further correction was necessaryto adjust for incomplete cytosol recovery or cytosol contamination of the nuclear fraction. The estimated fraction of ODN in the cytosol was calculated as the absolute difference between 1 and the fraction of ODN transported to the nucleus. The stability of the FITC-ODN was measuredby incubating cells with FITC-ODN for 2 h at 37”C, resuspending and vortexing with 0.1% SDS, then centrifuging through a 0.22 pm filter. The filtrate was then fractionated on a Sephadex G-25 column equilibrated with 10 mM tris-Cl, 1 mM EDTA, and 100mM NaCl. Fractions were assayedfor FITC using a Perkins Elmer luminescencespectrophotometer.

Antisense c-myc ODN activity ODN transport to the active site was confirmed by measuring decreasesin c-myc protein levels using Western blot analysis (Sambrook et al, 1989). Briefly, cells were treated in serum free media with either 25 @I unlabeled antisenseor scrambled sequenceODN for 4 h at 37°C. Cells were washed and lysed in low salt buffer containing Triton X-100 (Sambrooket al, 1989).Nuclei were isolated by centrifugation then solubilized to releasenuclear

Wu-Pong et al

259

Biology of the Cell (1997) 89, 257-261

Table I. ODN uptake in synchronized cells. Treatment

Cell cycle arrest

3zP ODN uptake I% G1 “1

n

FITC-ODN uptake (56 Gl a)

n

G&4 G2

299 133 f 39b 12 255 * 52 137 f 19

i 2 2

202 0.5 158 zt 23

;

Demecolcine Hydroxyurea Flow cytometry Flow cytometry

S

a Control cells svnchronized P = 0.05, ANOVA-.

in Gl using lovastatin or flow cytometry.

bstatistically

significant compared to Gl cells,

c-myc protein as previously described (Papoulas et a2, 1992). Extracts were electrophoresed in a Laemmli SDS polyacrylamide gel, then capillary-transferred to a nitrocellulose membrane. Non-specific binding sites were blocked with milk protein, probed using either the c-myc or pactin antibody, washed, then reprobed with a secondary antibody conjugated to alkaline phosphatase. A calorimetric substrate (p-nitrophenyl phosphate) was used for band visualization and densitometric analysis.

Table II. Apparent subcellular distribution of s*P-ODN as a function of cell cycle; n = 4, mean * SE.

RESULTS

aSee Materials and methods for determination of apparent subcellular distribution calculations.

The extent of ODN uptake and intracellular distribution was cell cycle dependent. For example, when measuring internalization of either FITC- or 32P-ODN, uptake was increased 2-3-fold in M phase cells and 1.5-fold in G2 phase cells relative to

cells in Gl (table I). However, the effects of trypsin and azide pretreatments on synchronized cells was similar to unsynchronized cells (data not shown). The intracellular distributions of 32P- and FITClabeled ODNs were also examined using subcellular fractionation and confocal fluorescence microscopy, respectively. Following subcellular fractionation, the yield of nuclei in the nuclear fractions was approximately 40-50%; the yield of cytosol in the cytosol fraction ranged from 90-100%. Therefore, the actual fraction of ODN in the cytosol and nuclei were estimated based on these yields (see Materials and metho&) to account for incomplete fraction recovery and fraction contamination. The ODN was distributed predominantly to the nucleus following all treatments (table II). These findings are consistent with the intracellular distribution of the FITC-ODN observed

with

confocal

fluorescence

microscopy

(Wu-Pong et al, 1994) and with the electron microscopy autoradiographic data in cultured cells and animal tissues (Zamecnik et al, 1994; Plenat et al, 1995), although others have observed ODN confinement in

vesicles only (Loke et ~2, 1989; Tarrason et al, 1995). Some degradation of the FITC-ODN was observed in the cells following a 2-h incubation as demonstrated by a broad peak which spanned fractions 3-12, although free FlTC was found primarily in the supernatent resulting in a second peak in fraction 17 Oligonucleotide biological activity

Treatment

Cell cycle Cytosolic uptake Nuclear uptake arrest f%P f%P

Unsynchronized Lovastatin Hydroxyurea Demecolcine

N/A :: M

47~ 16 14 f 14 42* 17 25* 9

53 i 16 86 f 14 58* 17 75* 9

(fig 1). FITC-ODN also showed similar ability to inhibit c-myc expression compared to unlabeled ODN (65% + 8 of control, mean + SE, n= 3) suggesting the FITC-ODN is transported to the active site within the cell. Any improvement in stability afforded by the 3’ FITC modification did not result in a measurable increase in antisense activity. The relationship between ODN uptake and activity as a function of cell cycle was examined by measuring the inhibition of c-myc protein synthesis following incubation with 25 @I antisense c-myc ODN in synchronized cells as described previously. Cell cycle-related differences in ODN activity appeared to exist since the ODN was ineffective in reducing c-myc protein levels in Gl and M phases, despite enhanced total cellular uptake. Overall, both total cellular (Rx = 0.62) and nuclear (R2 = 0.66) uptake in synchronized cells were negatively correlated to c-myc protein reduction (fig 2, insert). In contrast, cytosol uptake exhibited a positive correlation with activity (fig 2 insert, R2 = 0.66). Fractionation of cells following cellular pretreatments with either azide, trypsin, or MgCl, resulted in comparable results and an improved correlation between cytosol uptake and ODN activity (fig 2, R2 = 0.83).

DISCUSSION Reports in the literature suggest that dividing cells may preferentially internalize ODNs from the extracellular media (Krieg et ~2, 1990; Noonberg et Wu-Pong et al

Biology of the Cell (1997) 89, 257-261

260

al, 1993; Wu-Pong et al, 1994; Zamecnik et al, 1994). Such a finding could have significant implications in selective ODN targeting to rapidly dividing can-

cer cells. Therefore, the role of the cell cycle was examined to determine whether ODNs may be differentially targeted to either dividing or quiescent cells. Cells were synchronized in Gl, G2, M, or S phases either using chemical cell cycle inhibitors or sorting by flow cytometry. ODN uptake and activity was measured and compared in the resulting synchronized cell populations. ODN cellular uptake was increased up to 2-3-fold by cells in G2 or M phases (table I), but the mechanism of uptake was similar between each cell cycle phase (data not shown) compared to unsynchronized cells (WuPong et at, 1992). The ODN was ineffective in further reducing c-myc expression in synchronized cells (data not shown), possibly caused by a significant reduction in c-myc concentrations which resulted from cell cycle arrest. In addition, cells synchronized in Gl and M phases also had relatively less cytoplasmic retention of the ODN compared to unsynchronized cells (table II), which may also partially account for the lack of effect of the ODN in synchronized cells. Next, intracellular distribution was examined during Gl, G2, and M phases of the cell cycle. The radiolabeled ODN was rapidly transported to the nucleus, accounting for more than 50% of total uptake (table II). Since ODNs are rapidly transported to the nucleus by a passive mechanism following microinjection into the cytosol (Chin et al, 1990; Fisher et ~2, 1993), it is likely that significant fractions of 32P-ODN become extra-endosomal in the cytosol prior to nuclear transport. The possibility that ODN redistribution to the nucleus has occurred during fractionation cannot be completely ruled out, but this possibility is unlikely given the lack of cytosol contamination of the nuclear fraction. Lysis of endocytic vesicles during fractionation may also result in some contamination of the nuclear fractions, but is not expected to account for a large fraction of nuclear uptake because of the rapid

Al 120, *r" =

100:

aa

.-;

80-

60 -

al

E s ii 9

ODN-FITC/wlls

4020-

E:

0’ 0

, . , . , 10 20 30 Fraction number

0

10

.

, 40

B)

20

Fraction

30

40

number

Fig 1. Stability of FITC-ODNin Rauschercells. A. FITC-

ODNwas incubated with cells for 2 h, lysed, filtered, then fractionated on a Sephadex G-25 column. B. Stock FITCODNand FITCfractionated on SephadexG-25.

300

3oo-

I

nuclear

4 0 -20

-10

0

10

relative

Oligonucleotidebiological activity

20

activity

30

40

50

uptake

Fig 2. Relationshipbetweenuptake and activity of the antisense c-myc ODN in synchronized and non-synchronized cells. Relativeactivity is defined here as (100 - c-myc levels). Insert. Relationshipbetween uptake and activity in synchronizedcells only (table I).

Wu-Ponget al

261

Biology of the Cell (1997) 89, 257-261

dilution of any released ODN during cell fractionation. In addition, Chin et a2 (1990) report that 40% of microinjected ODN accumulate in the nucleus after 5 min. Therefore, because of both the rapid dilution of the cytoplasm (resulting in a large diffusional pathway compared to a cell) during fractionation and the brief time period (approximately l-2 min) during which fractionation occurs, endosome lysis alone does not appear to completely account for the observed nuclear uptake. Moreover, nuclear transport cannot be explained by ODN degradation, since the 32Plabeled ODN used in this study was previously found to be undegraded in the Rauscher cells after a 2-h incubation, with degradation products (primarily free phosphate) appearing primarily in the supernatent (Wu-Pong et al, 1992). ODN nuclear transport is consistent with previous reports of nuclear transport of either fluorescent- or radiolabeled-phosphodiester ODNs (Wu-Pong et al, 1994; Zamecnik et al, 1994). These findings are also consistent with the numerous reports in the literature citing specific biological activity of antisense and triplexforming ODNs which have sites of action in the cytosol or nucleus (for review see Wu-Pang, 1994), and confocal fluorescence microscopic data using various derivitizations of the antisense c-myc ODN in the Rauscher cells (Wu-Pong et al, 1994). In contrast, a number of investigators report that fluorescentlylabeled ODNs were confined primarily to endocytic vesicles after extracellular administration (Lake ef al, 1988; Shoji et al, 1991), although these ODNs were primarily phosphorothioate derivatives and had unknown biological activity. Such ODN or cell differences were likely to contribute to the observed discrepancies between these reports. This study demonstrated that total cellular and nuclear ODN uptake was dependent on the cell cycle (tables I, II). However, ODN activity was negatively correlated to nuclear and total cellular uptake; rather, activity appeared to be positively related to cytosolic uptake (fig 2). These findings support the concept of a cytoplasmic, rather than nuclear, site of action for this antisense ODN designed to inhibit c-myc translation in the cytosol. Examination of the relationship between activity and uptake of an ODN which acts in the nucleus (such as a triplex-forming ODN) is the logical successionto this study. While the current study does not demonstrate the anticipated potential of a ubiquitous targeting system to cancer cells, additional insight into the transport variables which influence ODN biological activity has been gained. Clearly, endosome release and nuclear transport must be balanced to optimize antisense ODN delivery to the cytosol. Designing new methods to retain antisense ODNs within the cytosol compartment without reducing activity or membrane transport will be a new delivery challenge.

Oligonucleotide biological activity

ACKNOWLEDGMENTS This work was supported by the Burroughs Wellcome Fund, the American Foundation for Pharmaceutical Education, the AACP Grant Program for New Investigators, and the National CancerInstitute no 2R25CA22032.

REFERENCES Chin DJ, Green GA, Zon G, Szoka FC and Straubinger RM (1990) Rapid nuclear accumulation of injected oligodeoxyribonucleotides. Nau BioZ2,1091-1100 Fisher TL, Terhorst T, Cao X and Wagner RW (1993) Intracellular disposition and metabolism of fluorescently-labeled unmodified and modified oligonucleotides microinjected into mammalian cells. Nucleic Acids Res 21,3857-3865 Krieg AM, Gmelig-Meyling F, Gourley MF, Kisch WJ, Chrisey LA, and Steinberg AD (1990) Uptake of oligodeoxyribonucleotides by lymphoid cells is heterogeneous and inducible. Antisense Res Dev 1,161-171 Loke SL, Stein C, Zhang X, Avigan M, Cohen J and Neckers LM (1988) Delivery of c-myc antisense phosphorothioate 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 Topics Microbial hnnutw1141,282-289 Loke SL, Stein CA, Zhang XH, Mori K, Nakanishi M, Subasinghe C, Cohen JS and Neckers LM (1989) Characterization of oligonucleotide transport into living cells. Proc Nat1 Acad Sci USA 86,3474-3478 Noonberg SB, Garovoy MR and Hunt CA (1993) Characteristics of oligonucleotide uptake in human keratinocyte cultures. J Invest Derm 101,727-731 Papoulas 0, Williams NG and Kingston RE (1992) DNA binding activities of c-myc purified from eukaryotic cells. J Biol Chem 267,10470-10480 Plenat F, Klein-Monhoven N, Marie B, Vignaud JM and Duprez A (1995) Cell and tissue distribution of synthetic oligonucleotides in healthy and tumor-bearing nude mice. Am J Puthol147,12&135 Salvatori R, Primorac D and Lichtler AC (1994) An efficient procedure for separate extraction of nmuclear and cytoplasmic RNA from cell culture. Biotechniques 16,374375 Sambrook J, Fritsch EF and Maniatis T (1989) Molecular Cloning, A Laboratory Manual. 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York Shoji Y, Akhtar S, Periasamy A, Herman B and Juliano RL (1991) Mechanism of cellular uptake of modified oligodeoxynucleotides containing methylphosphonate linkages.

Nucleic Acids Res 19, 5543-5550 Stein CA, Tonkinson JL, Zhang LM, Yakubov Taub R and Rotenberg SA (1993) Dynamics ization of phosphodiester oligodeoxynucleotides

L, Gervasoni J, of the internalin HL60

cells. Biochemistry 32,4855-4861 Tarrason G, Bellido D, Eritja R, Vilaro S and Piulats J (1995) Digoxigenin-labeled phosphorothioate oligonucleotides: a new tool for the study of cellular uptake. Antisense Res Dev 5,193-201 Wu-Pong S (1994) Oligonucleotides: opportunities for drug therapy and research. Pharm Tech 18,102-112 Wu-Pong S, Weiss TL and Hunt CA (1992) Antisense c-myc oligodeoxyribonucleotide cellular uptake. Pharm Res 9, 1010-1017 Wu-Pang S, Weiss TL and Hunt CA (1994) Cellular uptake and activity of a c-myc antisense oligodeoxyribonucleotide.

Antisense Res Dev 4,155-163 Zamecnik P, Aghajanian J, Zamecnik M, Goodchild J, Witrnan G (1994) Electron micrograph studies of transport of oligodeoxyribonucleotide across eukaryotic cell membranes. Proc Nat1 Acad Sci USA 91,3156-3160 Received

1 March

1997; accepted

15 September

1997

WuPong et al