ABSTRACT
To understand the parameters required for designing potent and specific
antisense C-5 propynylpyrimidine-2'
-deoxyphosphorothioate-modified oligonucleotides (C-5 propyne ONs), we have utilized a HeLa line that stably
expresses luciferase under tight control of a tetracycline-responsive promoter. Using this sensitive and regulatable cell-based system we have identified five distinct antisense ONs
targeting luciferase and have investigated the role that ON length, target
mismatches, compound stability and intracellular RNA levels play in affecting
antisense potency. We demonstrate that C-5 propyne ONs as short as 11 bases retained 66% of the potency
demonstrated by the parent 15 base compound, that a one base internal mismatch
between the antisense ON and the luciferase target reduced the potency of the
antisense ON by 43% and two or more mismatches completely inactivated the
antisense ON and that C-5 propyne ONs have a biologically active half-life in tissue culture of 35 h. In addition, by regulating the
intracellular levels of the luciferase mRNA over 20-fold, we show that the potency of C-5 propyne ONs is unaffected by changes in the expression level of
the target RNA. These data suggest that low and high copy messages can be
targeted with equivalent potency using C-5 propyne ONs.
Traditional pharmaceutical drug candidates are identified by screening large
chemical libraries. A promising compound is characterized by its ability to
bind to a target protein and alter its normal biological function. In contrast,
antisense-based therapeutic compounds, oligonucleotides (ONs), can be rationally
designed to inhibit protein expression from any gene for which a partial
sequence is known.
Recent technological advances, (i) the incorporation of nuclease-resistant backbone modifications into ONs, (ii) the discovery that C-5 propynylpyrimidine-modified phosphorothioate oligonucleotides have enhanced
affinity for their RNA target and (iii) the development of cationic lipids to
efficiently deliver ONs to cells, have overcome many of the barriers limiting
the use of antisense in biological systems (
1
-
8
). It now appears that antisense-based drugs will be viable human therapeutics. In fact, clinical trials
using antisense ONs containing phosphorothioate backbones have been initiated
for several human diseases (
4
,
9
).
Despite the rapid progress in developing antisense ONs as human therapeutics,
there is concern that many of the biological and clinical effects observed are
not due to an antisense mechanism alone (
10
-
13
). Recently, Kreig
et al
. showed that oligonucleotides containing CpG (cytosine-phosphate-guanine) can mimic bacterial DNA and trigger a potent immune
response (
14
). Such non-antisense effects, while clinically benefical, undermine the promise of
antisense agents as sequence-specific therapeutic agents. Clearly, the ultimate success of antisense
therapeutics as rationally designed drugs relies on demonstrating that the
clinical benefits can be attributed to an antisense mechanism of action.
In an effort to understand the parameters that influence the potency and
specificity of C-5 propyne antisense ONs, we have used the HeLa X1/5 cell line, which
stably expresses luciferase under control of a tetracycline-responsive promoter, to study the effects that ON length, non-complementary RNA/ON binding, compound stability and intracellular
RNA levels have on C-5 propyne-modified antisense-mediated inhibition of luciferase expression.
The HeLa X1/5 cell line stably expresses luciferase and has been described in
detail (
15
). Briefly, HeLa X1/5 cells contain a tetracycline-controlled transactivator that was produced by fusing the
tet
repressor with the activating domain of virion protein 16 (VP16) of herpes
simplex virus. The tet-VP16 transactivator is constitutively expressed in the cells. This
transactivator stimulates transcription of a stably integrated luciferase gene
by binding to tetracycline operator sequences found upstream of the luciferase
gene. In the presence of tetracycline (1 [mu]g/ml), the
tet
repressor binds tetracycline and blocks binding to the
tet
operator sequences. Thus, the presence of tetracycline in the HeLa X1/5 growth
medium inhibits luciferase gene expression. In the absence of tetracycline,
HeLa X1/5 cells maximally express luciferase. Cells were grown in Dulbecco's
modified Eagle's medium (DMEM) + 10% fetal bovine serum (FBS).
ONs were synthesized using the H-phosphonate method using standard procedures (
16
-
18
). The location of the ONs is based on the sequence of
Photinus pyralis
luciferase (accession no. M15077;
19
).
To prepare the cytofectin/ON complexes for transfection in one well of a 6-well tissue culture plate, the ON was first diluted into 100 [mu]l pre-warmed Opti-MEM (Gibco BRL) in polystyrene plastic. Similarly, the
cytofectin was diluted in another container (
5
). The pre-diluted ON and cytofectin mixtures were next combined (again in
polystyrene plastic) and within 15 min 800 [mu]l of the appropriate pre-warmed (37oC) medium (with serum) was added. Medium was next removed from
the cells and replaced with the medium containing the ON/cytofectin complex.
The HeLa X1/5 cells were seeded onto 6-well tissue culture plates at a density of 5 * 10
5
cells/well in DMEM + 10% FBS. The following day the cells were transfected for 4
h using either antisense ONs or mismatch sequence ONs (see Table
1
, underlining indicates mismatch positions). At defined time points, cell
extracts were made using Reporter Lysis buffer (Promega Corp., Madison, WI).
Luciferase enzyme activity was quantified using a single photon liquid
scintillation counter (LS6500, Beckman, Palo Alto, CA) as described (
20
). Percent luciferase activity is expressed as the ratio of the relative light
units detected from the luciferase enzyme assays for the ON-treated samples relative to the control cells treated with GS2888
cytofectin alone.
RNA was prepared from the treated cells using Trizol reagent (Gibco BRL)
according to the manufacturer's instructions. Total RNA (10 [mu]g/lane) was electrophoresed through a 1.2% agarose-6% formaldehyde gel, transferred to Hybond-N (Amersham) and UV-crosslinked to the membrane using a UV-Stratalinker (Stratagene). The blot was hybridized
as described (
21
) and probed overnight at 42oC with a random primer (Promega)-generated luciferase DNA probe. A glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probe (Ambion Inc.) was used as
an internal control. Following hybridization, the blot was washed three times
with 2* SSC, 0.1% SDS at 65oC and exposed to X-ray film. Luciferase mRNA levels were determined using a
radioanalytic imager and normalized to GAPDH RNA levels (Ambis, San Diego, CA).
Table 1
.
Oligonucleotides used in this study
We screened five different 15 base C-5 propyne antisense ONs directed against luciferase mRNA and found that
all the ONs demonstrated sequence-specific antisense inhibition. The RNA location, size, C-5 propyne content and sequence of the ONs are shown in Table
1
. The inhibitory concentration (IC
50
), where 50% of the luciferase enzyme activity is inhibited following antisense
ON treatment, was determined. The IC
50
for the antisense ONs Lucif-1, Lucif-2, Lucif-3 and Lucif-4 was between 1 and 2 nM when delivered to the HeLa X1/5
cells using GS2888 cytofectin. In the absence of GS2888 cytofectin, no
antisense activity was observed for any of the ONs. The IC
50
for antisense ON Lucif-5 was 5 nM.
Dose-response curves for two of the antisense ONs, Lucif-2 and Lucif-5, and their corresponding mismatch ONs are shown in Figure
1
. The range of the doses was from 1 to a maximum of 27 nM of antisense or
mismatch ONs. A dramatic decrease in luciferase activity was observed at all
concentrations of the antisense ONs. As indicated above, the IC
50
of Lucif-2 is between 1 and 2 nM and that of Lucif-5 is 5 nM. Essentially complete inhibition of luciferase activity
was demonstrated at 27 nM Lucif-2. The IC
50
of Lucif-2.MM, an 8 base mismatch of Lucif-2, and Lucif-5.MM, a 4 base mismatch, was estimated to be 180 nM
(extrapolated IC
50
values; Fig.
1
A;
5
). Only a modest decrease in luciferase activity was observed with between 1 and
9 nM Lucif-2.MM. Both mismatch ONs are 20- to 30-fold less active at inhibiting luciferase enzyme activity than
the antisense ONs, consistent with previous microinjection data using C-5 propyne-modified ONs (
1
,
3
). For the remainder of the results, data will be shown for Lucif-2, although similar results were obtained for the other luciferase
antisense ONs.
To investigate the relationship between ON length and C-5 propyne antisense activity, we tested two 13mer, 11mer and 9mer
antisense ONs made by shortening the Lucif-2 15mer by 2 base intervals from either the 5'- or 3'-end of the ON (see Table
1
for the sequences). Previously, we have shown that shorter ONs persist in the
nucleus of the cell for less time than longer (15mer) ONs (
22
), therefore, cellular extracts were prepared 6 h rather than 24 h following ON
delivery, so that we could observe the full antisense effect of the shorter
ONs.
As seen in Figure
2
, Lucif-2 used at 25 nM demonstrated a significant antisense-specific decrease (59%) in luciferase activity after only 6 h.
Shortening the Lucif-2 antisense ON to a 13mer by removing 2 bases from either the 5'- or 3'-end had only a modest effect on the potency of
the ONs. Lucif-2.5'.13mer and Lucif-2.3'.13mer inhibited luciferase enzyme activity by 54 and
48% respectively (Fig.
2
). Further shortening of Lucif-2 to an 11mer by removing 4 bases at either the 5'- or 3'-end still resulted in antisense-specific inhibition of luciferase
activity, with Lucif-2.5'.11mer and Lucif-2.3'.11mer inhibiting luciferase activity by 39 and 33%
respectively. Remarkably, when Lucif-2 was trimmed to 9 bases, Lucif-2.5'.9mer and Lucif-2.3'.9mer still inhibited luciferase activity by 10
and 17% respectively. Lucif-2.MM, the mismatch ON, had no effect on luciferase activity under these
conditions (Fig.
2
).
The potency of an antisense ON is dependent on precise binding to its target.
Mismatches between the antisense ON and its target reduce the effectiveness of
the antisense ON to inhibit gene expression (
1
). To determine how mismatches affect the potency of C-5 propyne-modified antisense ONs we synthesized 1, 2 and 8 base mismatches of
Lucif-2 and tested their ability to inhibit luciferase activity.
Two different 1 base mismatch ONs were synthesized and tested for antisense
activity at 10 nM ON concentration. Lucif-2.MM1A substituted the guanosine at position 7 with a C-5 propyne uridine. Lucif-2.MM1B substituted the guanosine at position 7 with a C-5 propyne cytosine (see Table
1
). The mismatch ONs retained ~60% of the activity of Lucif-2, despite the 1 base mismatch (Fig.
3
). Mismatch ONs containing 2 base changes at positions 7 and 8 failed to
demonstrate any antisense activity (Fig.
3
). The 8 base mismatch (Lucif-2.MM) also failed to inhibit luciferase activity.
To determine the potency of C-5 propyne-modified antisense ONs observed over time in tissue culture, we
delivered Lucif-2 (30 nM) or Lucif-2.MM, the mismatch ON, into HeLa X1/5 cells using GS2888 cytofectin
and monitored luciferase enzyme activity at various times following
transfection (Fig.
4
). Antisense-specific inhibition of luciferase activity was seen at 6 h following
transfection, the earliest point tested, and persisted for at least 48 h.
Maximum inhibition of luciferase activity was observed between 18 and 24 h
following transfection and Lucif-2 inhibited luciferase activity by at least 50% for 35 h after
transfection (Fig.
4
).
HeLa X1/5 cells stably express luciferase under tight control of a tetracycline-responsive promoter. By varying the tetracycline concentration in the
tissue culture medium (0-1 [mu]g/ml), the luciferase mRNA levels can be regulated over a wide range
(
15
). HeLa X1/5 cells were treated with varying concentrations of tetracycline (0-1 [mu]g/ml) for 24 h and RNA was isolated for Northern blot analysis (Fig.
5
A). Luciferase RNA was maximally expressed in the absence of tetracycline and at
a tetracycline concentration of 0.0001 [mu]g/ml (Fig.
5
A, lanes 1 and 2). A 10-fold increase in tetracycline (0.001 [mu]g/ml) resulted in a 2-fold decrease in luciferase mRNA levels (compare lanes 1 and 2
with lane 3, Fig.
5
). Increasing tetracycline concentrations to 0.01 [mu]g ml reduced luciferase mRNA levels 20-fold (Fig.
5
A, lane 4). At 0.1-1.0 [mu]g/ml tetracycline, luciferase mRNA expression was completely
inhibited (Fig.
5
A, lanes 5 and 6).
To examine the relationship between intracellular mRNA levels and antisense
potency, HeLa X1/5 cells that had been treated with either 0.001 or 0.01 [mu]g/ml tetracycline for 24 h were then transfected with Lucif-2 or Lucif-2.MM (1-27 nM ON) using GS2888 cytofectin. In HeLa X1/5 cells (no
tetracycline treatment) maximally expressing luciferase mRNA the IC
50
of Lucif-2 was previously demonstrated to be between 1 and 2 nM (Fig.
1
). Surprisingly, the IC
50
of Lucif-2 antisense ON in HeLa X1/5 cells treated with 0.001 [mu]g/ml tetracycline, which expressed 2-fold less luciferase RNA, was also between 1 and 2 nM (Fig.
5
B). Moreover, HeLa X1/5 cells treated with 0.01 [mu]g/ml tetracycline, which reduced luciferase RNA levels 20-fold as compared with untreated HeLa X1/5 cells (Fig.
5
A), demonstrated an IC
50
of between 1 and 2 nM Lucif-2 (Fig.
5
B). These data indicate that the potency of the Lucif-2 antisense ON is independent of luciferase mRNA levels over a 20-fold range of RNA concentration. Futhermore, these data suggest that
poorly expressed as well as highly expressed mRNA transcripts can be potently
targeted by C-5 propyne-modified antisense ONs.
Using microinjection, we have previously shown that the biological potency of C-5 propyne antisense ONs is dependent on two critical parameters: ON length
and precise binding of the antisense ON to its target (
1
,
3
). In the present study, we have extended these results and demonstrated (i)
that antisense-specific inhibition is dependent on ON length, but that C-5 propyne-modified ONs as short as 11 bases demonstrate significant
antisense-specific activity, (ii) that antisense inhibition can be demonstrated 6 h
after delivering antisense ONs to cells, (iii) that antisense ONs retain at
least 50% of their biological activity for 35 h in tissue culture and (iv) that
the potency of C-5 propyne-modified antisense ONs is independent of the intracellular level of
the target RNA.
REFERENCES
Return