ANTISENSE RESEARCH AND DEVELOPMENT 2:199-210 Mary Ann Liebert, Inc., Publishers
(1992)
The Use of Antisense Oligonucleotides to Establish Autocrine Angiotensin Growth Effects in Human Neuroblastoma and Mesangial Cells JULIA L.
COOK, LI CHEN, SRINIVAS BHANDARU, GEORGE L. BAKRIS, and RICHARD N. RE
ABSTRACT
renin-angiotensin systems (RAS) exist in many cell types, and angiotensin II (AH) has growth regulatory effects in some tissues. We demonstrated the presence of angiotensinogen (ANG) mRNA in cultured human mesangial cells (MC) and SHSY-5Y human neuroblastoma cells using reverse transcription and the polymerase chain reaction (RT/PCR) followed by hybridization to a human ANG-specific oligonucleotide probe. We speculated, therefore, that AH might act in an autocrine or paracrine fashion to regulate the growth of mesangial cells and neuroblastoma cells. Sense and antisense oligonucleotides were next synthesized complementary to the ANG transcription start site. Antisense but not sense oligonucleotides decreased [3HJthymidine incorporation into DNA by both MC and neuroblastoma cells. Growth of antisense oligonucleotide-treated cells was restored to control levels by the addition of All but not by the addition of basic fibroblast growth factor. Neither oligonucleotide affected [3H]thymidine incorporation in mouse L929 cells. These data indicate that locally produced All can act in an autocrine or paracrine fashion to alter the growth of human mesangial and neuroblastoma cells. Therefore, they suggest a role for local RAS in the pathogenesis of growth abnormalities in the cardiovascular system as well as in some forms of malignancy. Local
INTRODUCTION has accumulated to indicate the existence of local renin-angiotensin systems (RAS) in variety of tissues and cell types, including vascular cells and rodent neuroblastoma cells (Re et al., 1982; Clemens et al., 1986; Dzau and Re, 1987; Re, 1987; Dzau, 1988; Oda et al., 1991). Although the function of these tissue systems remains, for the most part, unknown, indirect evidence suggests a role in growth regulation (Ganten et al., 1975, 1976; Khairallah et al., 1972; Campbell-Boswell and Robertson, 1981; Fernandez et al., 1985; Geisterferet al., 1988; Jackson et al., 1988; Powell et al., 1989; Baker et al., 1990; Horiba et al., 1990; Schelling et al., 1991 ). In particular, angiotensin II ( ), the effector peptide of the renin-angiotensin system, has in recent years been determined to be a growth regulatory factor in a variety of tissue types. Depending on culture conditions, All can produce hypertrophy or hyperplasia of vascular smooth
Abundant
evidence
a
Alton Ochsner Medical Foundation, New Orleans, Louisiana.
199
COOK ET AL. muscle cells and adrenal glomerulosa cells (Campbell-Boswell and Robertson, 1981; Geisterferet al., 1988; Horiba et al., 1990). AH can also stimulate cardiac myocyte hypertrophy (Baker et al., 1990; Schelling et al., 1991 ). In addition, the mas oncogene has been shown to confer enhanced sensitivity to the growth-promoting effects of All upon cells into which it is transfected (Jackson et al., 1988). The relationship of All to the mas oncogene suggests that All may be a growth regulatory factor not only in the cardiovascular and endocrine systems but also, perhaps, in certain neoplastic conditions. The growth regulatory effects of All and therefore the potential growth regulatory effects of local RAS could prove to be important determinants of the sequelae of hypertension, including vascular hypertrophy, hyperplasia, and left ventricular hypertrophy. Moreover, disorders of growth play an important role in other diseases, such as diabetic nephropathy, in which hypertrophy of the mesangial matrix and, indeed, proliferation of mesangial cells occur (Schlondoroff, 1987; Kreisberg, 1988). In addition, we recently demonstrated that All regulates the growth of cultured human neuroblastoma SHSY-5Y cells, thereby suggesting a role for All in human neoplastic disease (Chen et al., 1989, 1991). As a consequence, weelected to explore the possibility that a local RAS in either mesangial cells or neuroblastoma cells could function to regulate cell growth. In this effort, we chose to employ antisense oligonucleotides. Antisense genes, RNAs, and oligonucle¬ otides (oligomers) have been used with considerable success to inhibit both gene expression and cellular proliferation of target cells (van der Krol et al., 1988; Zon, 1988). To explore the role of local RAS in cellular growth, we employed antisense oligomers complementary to the human angiotensinogen (ANG) mRNA with the intention of depleting cells of endogenous ANG and thereby reducing the synthesis of AH by the target cells. Our previous studies demonstrated that All is a growth factor for human mesangial cells as well as for the SHSY-5Y human neuroblastoma cell line, and the present studies were designed to determine whether locally produced ANG could function in an autocrine or paracrine fashion to regulate growth (Schlondoroff, 1987; Kreisberg, 1988; Bakris et al., 1990).
MATERIALS AND METHODS
Polymerase
chain reaction
Taq polymerase, polymerase chain reaction (PCR) buffer, and RNasin were obtained from Promega, deoxynucleotide triphosphates from Pharmacia, and Moloney murine leukemia virus reverse transcriptase from Bethesda Research Laboratories. The PCR primers were synthesized on a Milligen 8750 DNA synthesizer using phosphoramidite chemistry by LSUMC Core Laboratories (New Orleans). The upstream and downstream oligomer primers were derived from exons 4 and 5 and have sequences 5'-TATGACCTGCAGGACCTGCTC-3' (HANG-2) and 5'-TGCTGTGCTCAGCGGGTTGGC-3' (HANG-1), respectively. Amplifications were performed under the following conditions: 94°C, 1 min, 5 s; 50°C, 1 min, 6 s; and 72°C, 2 min, 5 s. Cycle number and starting concentrations of RNA varied depending on the cell type from which the RNAs were derived and are indicated in the legend to Fig. 1. RNA
preparation and southern blot hybridization
RNAs from human liver, human primary mesangial cells (cultured by standard methods; Bakris et al., 1990), and SHSY-5 Y neuroblastoma cells were collected by the method of Chomczynski and Sacchi (1987).
Following reverse transcription and amplification, the PCR products were electrophoresed on an agarose gel and transferred to nitrocellulose. The membrane was prehybridized and hybridized overnight at 42°C by the method of Wallace et al. (1981). The hybridization probe was [y-32P]ATP end-labeled oligonucleotide (HANG-1 ). The final filter wash was performed at 61 °C in 0.90 M sodium chloride and 0.09 M sodium citrate (5°C below the calculated Tm for the duplex in 1 M Na+). 200
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH
Angiotensinogen
antisense
oligonucleotides
The antisense and corresponding control (sense) oligomers were synthesized complementary to the transcription start site by Genosys Biotechnologies, Inc. (The Woodlands, TX). The antisense and sense oligomers have sequences of 5'-CCAGAACAACGGCAGCTTCT-3' and 5'-AGAAGCTGCCGTTGTTCstarting resin and were purified by high-performance liquid TGG-3', were synthesized on a 1 µ chromatography (HPLC) to separate the final full-length product from contaminating incomplete reaction
products.
Oligonucleotide stability and uptake Oligonucleotides were tested for stability in the culture media. Primary human mesangial cells (passages 5-6), SHSY-5 Y human neuroblastoma cells, and L929 fibroblasts were maintained in RPMI, -modified Eagle's medium ( -MEM), and high-glucose Dulbecco's modified Eagle's medium (DMEM), respectively, all with 10% fetal bovine serum (FBS). Each cell type was plated at 5000 per well in microtiter wells and incubated overnight. Cells were rendered quiescent by incubation in 0.5% FBS-containing media for 24 h and 108 then treated with radiolabeled antisense and sense oligonucleotides (5 x 106 cpm/ml, specific activity cprn^g DNA) in appropriate media with 10% FBS. Fetal bovine serum for all experiments in this report was heat inactivated at 65°C for 1 h to destroy nuclease activity. Media and cells were harvested at 10 min and 4, 12, 24, and 36 h after oligonucleotide addition. Media (25 µ ) were then collected and electrophoresed on a 23% acrylamide-urea sequencing gel. Radiolabeled oligonucleotide bands were quantified (by Lousiana State University Core Laboratories) using a Molecular Dynamics Phosphorlmager and ImageQuant Version 3.15 software (Sunnyvale, CA). Cells were washed, detached using trypsin and EDTA, lysed in 1% sodium dodecyl sulfate (SDS)-containing phosphate-buffered saline (PBS), and extracted with Tris-buffered phenol, pH 8 (Wickstrom et al., 1988; Becker et al., 1989). Uptake of antisense and sense oligomers was determined —
by liquid scintillation counting of the aqueous phase of the cell pellet.
Hybrid arrest translation, preliminary studies Primary human mesangial cells (passages 5-6), SHSY-5 Y neuroblastoma cells, and L929 cells were maintained in RPMI, -MEM, and high-glucose DMEM, respectively, all with 10% FBS. Each cell type was plated at 5000 per well in nine microtiter wells, grown for 2 days, and then serum deprived for 2 days (0.5% FBS). Low-serum media were removed and replaced with appropriate media containing 10% heat-inactivated serum ± oligomer (50 or 100 µ final concentration). Antisense oligomer was added to three wells, sense oligomer was added to three wells, and vehicle (medium) was added to the final three wells. Oligomers were replaced every 12hfor48h. [3H]thymidine (New England Nuclear) was added for the final 24 h at 10µ / 1. Cells were harvested and bioassays performed using a Tomtec Harvester 96 Mach II and an LKB 1205 Betaplate scintillation counter.
Hybrid arrest translation
with
growth factor supplements
Primary human mesangial cells (passages 5-6), SHSY-5 Y neuroblastoma cells, and L929 cells were maintained in RPMI, -MEM, and high-glucose DMEM, respectively, all with 10% FBS. Each cell type was plated at 2500 per well, grown for 2 days, and then serum-deprived for 1 day (0.5% FBS). Low-serum media were removed and replaced with appropriate media containing 10% heart-inactivated serum ± oligomer (25 µ final concentration). The oligomer used for this purpose was provided by Genosys but represents synthesis and purification independent ofthat indicated earlier. Cells were treated with various combinations of antisense oligonucleotide, sense oligonucleotide, AH (Sigma), and bFGF (basic fibroblast growth factor; Collaborative Research, Inc.). All and bFGF were replaced every 2 h for 48 h, and oligonucleotides were 201
COOK ET AL.
replaced every 12hfor48h. [3H]thymidine (New England Nuclear) was added for the final 24 h at µ / . Cells were harvested and bioassays performed using a Tomtec Harvester 96 Mach II and an LKB 1205 Betaplate scintillation counter. Statistical
analyses
The Wilcoxon rank-sum test
was
used to analyze differences in
thymidine incorporation into DNA.
RESULTS Reverse
transcription/polymerase chain
reaction
Several rodent neuroblastoma cell lines have been reported to produce ANG mRNA (Okamura et al., 1981 ; Clemens et al., 1986; Petrossian and Oliver, 1989). In addition, this laboratory has demonstrated that proliferation of the human neuroblastoma cell line SHSY-5 Y is suppressed by angiotensin converting enzyme (ACE) inhibitors and receptor antagonists, suggesting that All could be involved in autocrine-stimulated proliferation (Chen et al., 1989, 1991). To firmly establish the presence of ANG mRNA in SHSY-5Y cells, we designed PCR primers with which to amplify ANG mRNA. The primers were designed so that one hybridization site is in exon 4 and the other in exon 5. The positions of the primers in two different exons permits differentiation of RNA and DNA amplification products. Amplification of ANG cDNA following RNA reverse transcription (RT) produces a 288 bp product; amplification of genomic DNA generates an approximately 1 kb product. Contaminating DNA in RNA preparations therefore does not yield false positive signals. Figure 1 is a photograph of an autoradiogram of a Southern blot that was hybridized with radiolabeled oligomer complementary to ANG PCR products. Lanes 1 and 2 contain RT/PCR products of human liver RNA. Lanes 3, 4 and 5 contain RT/PCR products of human mesangial cells (MC), SHSY-5Y, and mouse L929 (negative control) RNAs. We used filter hybridization to magnify the signal beyond the amplification
FIG. 1. Angiotensinogen mRNA presence in human liver, mesangial cells (MC), and a neuroblastoma ( ) cell line, SHSY-5Y. Liver RNA (2 and 1 µg; LI, left to right, respectively) and 5 µg each of MC, , and L929 (LC) RNAs were reverse transcribed and amplified by PCR. The liver cDNA was amplified 30 cycles, and the MC, , and LC cDNAs were amplified 40 cycles. The products were gel electrophoresed, blotted to nitrocellulose, and hybridized to an ANG mRNA-complementary oligomer. The marker sizes (arrows) represent bands from the 1 kb ladder. The 288 bp band is the predicted size of the ANG cDNA amplification product.
202
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH obtained
by PCR alone. MC PCR products are not visible in our hands in the absence of this second amplification. Although RT/PCR is not quantitative at a fine level, it is obvious that MC possess considerably
less ANG mRNA than liver or SHSY-5 Y cells and that the neuroblastomas have less ANG mRNA than liver. Visualization of ethidium bromide-stained SHSY-5 Y RNA RT/PCR products was not consistently reproduc¬ ible in agarose gels, and MC RT/PCR products were never observed without Southern blot hybridization. ANG mRNA has, to our collective knowledge, not before been demonstrated in either cultured MC or SHSY-5 Y cells.
Oligonucleotide stability and uptake We determined the
stability and uptake of the sense and antisense ANG oligomers by administering [ -32 ] end-labeled oligomers and assaying for their presence in the media and in cell pellet extracts at several time points between 0 and 36 h. The ANG sense and antisense oligonucleotides are quite stable in media containing heat-inactivated serum (Figs. 2 and 3). Oligomers are still present at 36 h, and there are no obvious intermediary degradation products. In addition, the quantity of radiolabeled oligomer remaining in the media as a function of time is similar for antisense and sense oligomers and for all cell types. The radiolabeled oligonucleotides associated with the cell pellet extracts increases over the 36 h period and is similar for both antisense and sense oligonucleotides and in MC, neuroblastoma, and L929 cells (Fig. 4).
FIG. 2.
Autoradiogram of antisense and sense oligonucleotides present in culture media as a function of time. Media harvested from cultured cells at 10 min (lanes 1 and 2), 4 h (lanes 3 and 4), 12 h (lanes 5 and 6), 24 h (lanes 7 and 8), and 36 h (lanes 9 and 10) after oligomer addition and electrophoresed on a 23% sequencing gel. Cells were treated with antisense oligomer (odd-numbered lanes) and sense oligomers (even-numbered lanes). Oligomer present in media of neuroblastoma (A), L929 (B), and mesangial cells (C). were
203
COOK ET AL.
SASASASAS 4hr } { 12hr }{ 24hr }{ 36hr }
{10min} { 125
asasasasa
{10min} { 4hr }{ 12hr }{ 24hr }{36hr}
g S ASASASASAS
{lOminH 4hr }{ 12hr }{ 24hr }1 36hr }
Quantification of antisense and sense oligomers present in culture media as a function of time. Radioactive oligonucleotide bands presented in Fig. 2 were quantified by phosphorimage analysis. Oligomers harvested from media of SHSY-5Y neuroblastoma (A), L929 (B), and human mesangial cells (C). FIG. 3.
Hybrid arrest translation We next synthesized an antisense oligomer complementary to the ANG mRNA to test the mitogenic effects of lowering steady-state ANG translation product levels in these two cell types. The oligomer was synthesized complementary to the transcription start site. One study has shown that ß-globin antisense oligomers are effective in blocking translation only if they are made complementary to the cap site (van der Krol et al., 1988; Walder and Walder, 1988). Those made complementary to the translation start site or the coding sequence are effective only in that the duplexes are substrates for RNAse-H, an enzyme that digests RNA present in 204
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH
a,
o
0
5 10 15 20 25 30 35 40 Hours
0
5
10 15 20 25 30 35 40 Hours
0
5
10 15 20 25 30 35 40 Hours
FIG. 4. Quantification of radiolabeled antisense or sense oligomers present in extracts of cell neuroblastoma (A), L929 (B), and human mesangial cells (C).
pellets.
SHSY-5Y
DNA/RNA duplexes. We therefore designed the ANG antisense oligomer complementary to the cap site to maximize efficiency. A sense oligomer, the inverse complement of the antisense oligomer, was synthesized and used as a control. Neuroblastoma and MC were treated with the antisense and sense oligomers described in Materials and Methods. Antisense oligomers were effective in reducing [3H]thymidine uptake into both cell types, but sense oligomers demonstrated no significant effects. [3H]thymidine incorporation into DNA of SHSY-5 Y cells was decreased 30% when the cells were treated for 48 h with a 50 µ concentration of antisense oligomer. Sense oligomer treatment for the same interval and at the same concentration had no effect (Fig. 5A). Because in this and subsequent experiments the sense oligomers produce no change in [3H]thymidine incorporation, for the purpose of statistical analysis these data were combined with values obtained in the absence of all oligonucleotides. The antisense strand was found to significantly reduce [3H]thymidine incorporation (p < 0.05). MC proliferation was decreased by 28% when the cells were treated for 48 h with a 50 µ concentration of oligomer and by 35% when treated for 48 h with a 100 µ concentration (Fig. 5B and C, < 0.05 for each concentration). Again, sense oligomer treatment was without effect. We next tested the possibility that the antimitogenic effects of the antisense oligomer could be the result of a contaminating organic or other toxin in the antisense DNA preparation that was not present in the sense material. Although the oligomers were commercially synthesized, HPLC eluted, and guaranteed to be of high purity, we elected to test the preparations in a heterologous system in which we expected no effects. Mouse L929 cells were incubated with a 50 µ concentration of either sense or antisense oligomer for 48 h, and proliferation was measured by [3H]thymidine bioassay. Sense and antisense treatments were not significantly different (Fig. 5D).
Hybrid arrest translation with growth factor supplements We next performed experiments designed to show the specificity of the antisense oligomer. If the antisense oligomer mediates its antimitogenic action by by specifically blocking ANG translation, then addition of All should restore proliferation. Addition of other growth factors might not restore proliferation, depending on the importance of All and the ability of other growth factors to compensate for an All deficiency.
because sense oligomers produced no change in [3H]thymidine incorporation into DNA, for the statistical analyses, data obtained from sense oligonucleotide treatments were combined with of purpose values obtained for vehicle treatment. All and bFGF treatment of SHSY-5Y cells increased [3H]thymidine incorporation into DNA by 35 and 26%, respectively (p < 0.05, data from bFGF and sense + bFGF
Again,
205
COOK ET AL.
FIG. 5. [3H]thymidine incorporation into DNA of various cell types treated with "sense" and "antisense" oligonucle¬ otides. Bars represent the mean and standard deviation of triplicate readings. (A) SHSY-5Y neuroblastoma cells treated with oligonucleotides at a 50 µ concentration. (B) Human primary mesangial cells treated with oligonucleotides at a 50 µ concentration. (C) Human mesangial cells treated with oligonucleotides at a 100 µ concentration. (D) L929 cells treated with oligonucleotides at a 50 µ concentration.
treatments were combined; Fig. 6A). Antisense oligonucleotide decreased proliferation to 33% of control levels (p < 0.05). The growth inhibitory effect of All antisense oligonucleotide was completely eliminated by
the concurrent addition of All, but bFGF only partially compensated for the effects of antisense oligonucle¬ otide. Human mesangial cells responded in a similar fashion to oligomer and growth factor supplements (Fig. 6B). All stimulated MC proliferation by 30% (p < 0.05), as did bFGF. Antisense oligomer decreased proliferation to 39% of control levels (p < 0.05), and All restored proliferation to normal levels. bFGF was
incapable of restoring proliferation significantly. 206
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH
25000 8000·
-
-,-
12000
_.
10000
20000 6000 ·
4000
2000-
15000-
10000-
5000-
nj
D
«
FIG. 6. [3H]thymidine incorporation into DNA of three cell types. A2 (angiotensin II) and bFGF were administered at 6 10 and 2 x -9 M final concentrations, respectively. They were each freshly applied every 2 h for 48 h. Sense and antisense oligonucleotides, at a 25 µ final concentration, were reapplied every 12 h for 48 h. (A) SHSY-5Y neuroblastoma cells. (B) Human primary mesangial cells. (C) L929 fibroblasts. Bars represent the mean and standard deviation of five separate treatments.
respond significantly to All sense 0.01) stimulatory effect to bFGF (Figure 6C).
L929 cells did not
(p
(1992)
The Use of Antisense Oligonucleotides to Establish Autocrine Angiotensin Growth Effects in Human Neuroblastoma and Mesangial Cells JULIA L.
COOK, LI CHEN, SRINIVAS BHANDARU, GEORGE L. BAKRIS, and RICHARD N. RE
ABSTRACT
renin-angiotensin systems (RAS) exist in many cell types, and angiotensin II (AH) has growth regulatory effects in some tissues. We demonstrated the presence of angiotensinogen (ANG) mRNA in cultured human mesangial cells (MC) and SHSY-5Y human neuroblastoma cells using reverse transcription and the polymerase chain reaction (RT/PCR) followed by hybridization to a human ANG-specific oligonucleotide probe. We speculated, therefore, that AH might act in an autocrine or paracrine fashion to regulate the growth of mesangial cells and neuroblastoma cells. Sense and antisense oligonucleotides were next synthesized complementary to the ANG transcription start site. Antisense but not sense oligonucleotides decreased [3HJthymidine incorporation into DNA by both MC and neuroblastoma cells. Growth of antisense oligonucleotide-treated cells was restored to control levels by the addition of All but not by the addition of basic fibroblast growth factor. Neither oligonucleotide affected [3H]thymidine incorporation in mouse L929 cells. These data indicate that locally produced All can act in an autocrine or paracrine fashion to alter the growth of human mesangial and neuroblastoma cells. Therefore, they suggest a role for local RAS in the pathogenesis of growth abnormalities in the cardiovascular system as well as in some forms of malignancy. Local
INTRODUCTION has accumulated to indicate the existence of local renin-angiotensin systems (RAS) in variety of tissues and cell types, including vascular cells and rodent neuroblastoma cells (Re et al., 1982; Clemens et al., 1986; Dzau and Re, 1987; Re, 1987; Dzau, 1988; Oda et al., 1991). Although the function of these tissue systems remains, for the most part, unknown, indirect evidence suggests a role in growth regulation (Ganten et al., 1975, 1976; Khairallah et al., 1972; Campbell-Boswell and Robertson, 1981; Fernandez et al., 1985; Geisterferet al., 1988; Jackson et al., 1988; Powell et al., 1989; Baker et al., 1990; Horiba et al., 1990; Schelling et al., 1991 ). In particular, angiotensin II ( ), the effector peptide of the renin-angiotensin system, has in recent years been determined to be a growth regulatory factor in a variety of tissue types. Depending on culture conditions, All can produce hypertrophy or hyperplasia of vascular smooth
Abundant
evidence
a
Alton Ochsner Medical Foundation, New Orleans, Louisiana.
199
COOK ET AL. muscle cells and adrenal glomerulosa cells (Campbell-Boswell and Robertson, 1981; Geisterferet al., 1988; Horiba et al., 1990). AH can also stimulate cardiac myocyte hypertrophy (Baker et al., 1990; Schelling et al., 1991 ). In addition, the mas oncogene has been shown to confer enhanced sensitivity to the growth-promoting effects of All upon cells into which it is transfected (Jackson et al., 1988). The relationship of All to the mas oncogene suggests that All may be a growth regulatory factor not only in the cardiovascular and endocrine systems but also, perhaps, in certain neoplastic conditions. The growth regulatory effects of All and therefore the potential growth regulatory effects of local RAS could prove to be important determinants of the sequelae of hypertension, including vascular hypertrophy, hyperplasia, and left ventricular hypertrophy. Moreover, disorders of growth play an important role in other diseases, such as diabetic nephropathy, in which hypertrophy of the mesangial matrix and, indeed, proliferation of mesangial cells occur (Schlondoroff, 1987; Kreisberg, 1988). In addition, we recently demonstrated that All regulates the growth of cultured human neuroblastoma SHSY-5Y cells, thereby suggesting a role for All in human neoplastic disease (Chen et al., 1989, 1991). As a consequence, weelected to explore the possibility that a local RAS in either mesangial cells or neuroblastoma cells could function to regulate cell growth. In this effort, we chose to employ antisense oligonucleotides. Antisense genes, RNAs, and oligonucle¬ otides (oligomers) have been used with considerable success to inhibit both gene expression and cellular proliferation of target cells (van der Krol et al., 1988; Zon, 1988). To explore the role of local RAS in cellular growth, we employed antisense oligomers complementary to the human angiotensinogen (ANG) mRNA with the intention of depleting cells of endogenous ANG and thereby reducing the synthesis of AH by the target cells. Our previous studies demonstrated that All is a growth factor for human mesangial cells as well as for the SHSY-5Y human neuroblastoma cell line, and the present studies were designed to determine whether locally produced ANG could function in an autocrine or paracrine fashion to regulate growth (Schlondoroff, 1987; Kreisberg, 1988; Bakris et al., 1990).
MATERIALS AND METHODS
Polymerase
chain reaction
Taq polymerase, polymerase chain reaction (PCR) buffer, and RNasin were obtained from Promega, deoxynucleotide triphosphates from Pharmacia, and Moloney murine leukemia virus reverse transcriptase from Bethesda Research Laboratories. The PCR primers were synthesized on a Milligen 8750 DNA synthesizer using phosphoramidite chemistry by LSUMC Core Laboratories (New Orleans). The upstream and downstream oligomer primers were derived from exons 4 and 5 and have sequences 5'-TATGACCTGCAGGACCTGCTC-3' (HANG-2) and 5'-TGCTGTGCTCAGCGGGTTGGC-3' (HANG-1), respectively. Amplifications were performed under the following conditions: 94°C, 1 min, 5 s; 50°C, 1 min, 6 s; and 72°C, 2 min, 5 s. Cycle number and starting concentrations of RNA varied depending on the cell type from which the RNAs were derived and are indicated in the legend to Fig. 1. RNA
preparation and southern blot hybridization
RNAs from human liver, human primary mesangial cells (cultured by standard methods; Bakris et al., 1990), and SHSY-5 Y neuroblastoma cells were collected by the method of Chomczynski and Sacchi (1987).
Following reverse transcription and amplification, the PCR products were electrophoresed on an agarose gel and transferred to nitrocellulose. The membrane was prehybridized and hybridized overnight at 42°C by the method of Wallace et al. (1981). The hybridization probe was [y-32P]ATP end-labeled oligonucleotide (HANG-1 ). The final filter wash was performed at 61 °C in 0.90 M sodium chloride and 0.09 M sodium citrate (5°C below the calculated Tm for the duplex in 1 M Na+). 200
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH
Angiotensinogen
antisense
oligonucleotides
The antisense and corresponding control (sense) oligomers were synthesized complementary to the transcription start site by Genosys Biotechnologies, Inc. (The Woodlands, TX). The antisense and sense oligomers have sequences of 5'-CCAGAACAACGGCAGCTTCT-3' and 5'-AGAAGCTGCCGTTGTTCstarting resin and were purified by high-performance liquid TGG-3', were synthesized on a 1 µ chromatography (HPLC) to separate the final full-length product from contaminating incomplete reaction
products.
Oligonucleotide stability and uptake Oligonucleotides were tested for stability in the culture media. Primary human mesangial cells (passages 5-6), SHSY-5 Y human neuroblastoma cells, and L929 fibroblasts were maintained in RPMI, -modified Eagle's medium ( -MEM), and high-glucose Dulbecco's modified Eagle's medium (DMEM), respectively, all with 10% fetal bovine serum (FBS). Each cell type was plated at 5000 per well in microtiter wells and incubated overnight. Cells were rendered quiescent by incubation in 0.5% FBS-containing media for 24 h and 108 then treated with radiolabeled antisense and sense oligonucleotides (5 x 106 cpm/ml, specific activity cprn^g DNA) in appropriate media with 10% FBS. Fetal bovine serum for all experiments in this report was heat inactivated at 65°C for 1 h to destroy nuclease activity. Media and cells were harvested at 10 min and 4, 12, 24, and 36 h after oligonucleotide addition. Media (25 µ ) were then collected and electrophoresed on a 23% acrylamide-urea sequencing gel. Radiolabeled oligonucleotide bands were quantified (by Lousiana State University Core Laboratories) using a Molecular Dynamics Phosphorlmager and ImageQuant Version 3.15 software (Sunnyvale, CA). Cells were washed, detached using trypsin and EDTA, lysed in 1% sodium dodecyl sulfate (SDS)-containing phosphate-buffered saline (PBS), and extracted with Tris-buffered phenol, pH 8 (Wickstrom et al., 1988; Becker et al., 1989). Uptake of antisense and sense oligomers was determined —
by liquid scintillation counting of the aqueous phase of the cell pellet.
Hybrid arrest translation, preliminary studies Primary human mesangial cells (passages 5-6), SHSY-5 Y neuroblastoma cells, and L929 cells were maintained in RPMI, -MEM, and high-glucose DMEM, respectively, all with 10% FBS. Each cell type was plated at 5000 per well in nine microtiter wells, grown for 2 days, and then serum deprived for 2 days (0.5% FBS). Low-serum media were removed and replaced with appropriate media containing 10% heat-inactivated serum ± oligomer (50 or 100 µ final concentration). Antisense oligomer was added to three wells, sense oligomer was added to three wells, and vehicle (medium) was added to the final three wells. Oligomers were replaced every 12hfor48h. [3H]thymidine (New England Nuclear) was added for the final 24 h at 10µ / 1. Cells were harvested and bioassays performed using a Tomtec Harvester 96 Mach II and an LKB 1205 Betaplate scintillation counter.
Hybrid arrest translation
with
growth factor supplements
Primary human mesangial cells (passages 5-6), SHSY-5 Y neuroblastoma cells, and L929 cells were maintained in RPMI, -MEM, and high-glucose DMEM, respectively, all with 10% FBS. Each cell type was plated at 2500 per well, grown for 2 days, and then serum-deprived for 1 day (0.5% FBS). Low-serum media were removed and replaced with appropriate media containing 10% heart-inactivated serum ± oligomer (25 µ final concentration). The oligomer used for this purpose was provided by Genosys but represents synthesis and purification independent ofthat indicated earlier. Cells were treated with various combinations of antisense oligonucleotide, sense oligonucleotide, AH (Sigma), and bFGF (basic fibroblast growth factor; Collaborative Research, Inc.). All and bFGF were replaced every 2 h for 48 h, and oligonucleotides were 201
COOK ET AL.
replaced every 12hfor48h. [3H]thymidine (New England Nuclear) was added for the final 24 h at µ / . Cells were harvested and bioassays performed using a Tomtec Harvester 96 Mach II and an LKB 1205 Betaplate scintillation counter. Statistical
analyses
The Wilcoxon rank-sum test
was
used to analyze differences in
thymidine incorporation into DNA.
RESULTS Reverse
transcription/polymerase chain
reaction
Several rodent neuroblastoma cell lines have been reported to produce ANG mRNA (Okamura et al., 1981 ; Clemens et al., 1986; Petrossian and Oliver, 1989). In addition, this laboratory has demonstrated that proliferation of the human neuroblastoma cell line SHSY-5 Y is suppressed by angiotensin converting enzyme (ACE) inhibitors and receptor antagonists, suggesting that All could be involved in autocrine-stimulated proliferation (Chen et al., 1989, 1991). To firmly establish the presence of ANG mRNA in SHSY-5Y cells, we designed PCR primers with which to amplify ANG mRNA. The primers were designed so that one hybridization site is in exon 4 and the other in exon 5. The positions of the primers in two different exons permits differentiation of RNA and DNA amplification products. Amplification of ANG cDNA following RNA reverse transcription (RT) produces a 288 bp product; amplification of genomic DNA generates an approximately 1 kb product. Contaminating DNA in RNA preparations therefore does not yield false positive signals. Figure 1 is a photograph of an autoradiogram of a Southern blot that was hybridized with radiolabeled oligomer complementary to ANG PCR products. Lanes 1 and 2 contain RT/PCR products of human liver RNA. Lanes 3, 4 and 5 contain RT/PCR products of human mesangial cells (MC), SHSY-5Y, and mouse L929 (negative control) RNAs. We used filter hybridization to magnify the signal beyond the amplification
FIG. 1. Angiotensinogen mRNA presence in human liver, mesangial cells (MC), and a neuroblastoma ( ) cell line, SHSY-5Y. Liver RNA (2 and 1 µg; LI, left to right, respectively) and 5 µg each of MC, , and L929 (LC) RNAs were reverse transcribed and amplified by PCR. The liver cDNA was amplified 30 cycles, and the MC, , and LC cDNAs were amplified 40 cycles. The products were gel electrophoresed, blotted to nitrocellulose, and hybridized to an ANG mRNA-complementary oligomer. The marker sizes (arrows) represent bands from the 1 kb ladder. The 288 bp band is the predicted size of the ANG cDNA amplification product.
202
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH obtained
by PCR alone. MC PCR products are not visible in our hands in the absence of this second amplification. Although RT/PCR is not quantitative at a fine level, it is obvious that MC possess considerably
less ANG mRNA than liver or SHSY-5 Y cells and that the neuroblastomas have less ANG mRNA than liver. Visualization of ethidium bromide-stained SHSY-5 Y RNA RT/PCR products was not consistently reproduc¬ ible in agarose gels, and MC RT/PCR products were never observed without Southern blot hybridization. ANG mRNA has, to our collective knowledge, not before been demonstrated in either cultured MC or SHSY-5 Y cells.
Oligonucleotide stability and uptake We determined the
stability and uptake of the sense and antisense ANG oligomers by administering [ -32 ] end-labeled oligomers and assaying for their presence in the media and in cell pellet extracts at several time points between 0 and 36 h. The ANG sense and antisense oligonucleotides are quite stable in media containing heat-inactivated serum (Figs. 2 and 3). Oligomers are still present at 36 h, and there are no obvious intermediary degradation products. In addition, the quantity of radiolabeled oligomer remaining in the media as a function of time is similar for antisense and sense oligomers and for all cell types. The radiolabeled oligonucleotides associated with the cell pellet extracts increases over the 36 h period and is similar for both antisense and sense oligonucleotides and in MC, neuroblastoma, and L929 cells (Fig. 4).
FIG. 2.
Autoradiogram of antisense and sense oligonucleotides present in culture media as a function of time. Media harvested from cultured cells at 10 min (lanes 1 and 2), 4 h (lanes 3 and 4), 12 h (lanes 5 and 6), 24 h (lanes 7 and 8), and 36 h (lanes 9 and 10) after oligomer addition and electrophoresed on a 23% sequencing gel. Cells were treated with antisense oligomer (odd-numbered lanes) and sense oligomers (even-numbered lanes). Oligomer present in media of neuroblastoma (A), L929 (B), and mesangial cells (C). were
203
COOK ET AL.
SASASASAS 4hr } { 12hr }{ 24hr }{ 36hr }
{10min} { 125
asasasasa
{10min} { 4hr }{ 12hr }{ 24hr }{36hr}
g S ASASASASAS
{lOminH 4hr }{ 12hr }{ 24hr }1 36hr }
Quantification of antisense and sense oligomers present in culture media as a function of time. Radioactive oligonucleotide bands presented in Fig. 2 were quantified by phosphorimage analysis. Oligomers harvested from media of SHSY-5Y neuroblastoma (A), L929 (B), and human mesangial cells (C). FIG. 3.
Hybrid arrest translation We next synthesized an antisense oligomer complementary to the ANG mRNA to test the mitogenic effects of lowering steady-state ANG translation product levels in these two cell types. The oligomer was synthesized complementary to the transcription start site. One study has shown that ß-globin antisense oligomers are effective in blocking translation only if they are made complementary to the cap site (van der Krol et al., 1988; Walder and Walder, 1988). Those made complementary to the translation start site or the coding sequence are effective only in that the duplexes are substrates for RNAse-H, an enzyme that digests RNA present in 204
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH
a,
o
0
5 10 15 20 25 30 35 40 Hours
0
5
10 15 20 25 30 35 40 Hours
0
5
10 15 20 25 30 35 40 Hours
FIG. 4. Quantification of radiolabeled antisense or sense oligomers present in extracts of cell neuroblastoma (A), L929 (B), and human mesangial cells (C).
pellets.
SHSY-5Y
DNA/RNA duplexes. We therefore designed the ANG antisense oligomer complementary to the cap site to maximize efficiency. A sense oligomer, the inverse complement of the antisense oligomer, was synthesized and used as a control. Neuroblastoma and MC were treated with the antisense and sense oligomers described in Materials and Methods. Antisense oligomers were effective in reducing [3H]thymidine uptake into both cell types, but sense oligomers demonstrated no significant effects. [3H]thymidine incorporation into DNA of SHSY-5 Y cells was decreased 30% when the cells were treated for 48 h with a 50 µ concentration of antisense oligomer. Sense oligomer treatment for the same interval and at the same concentration had no effect (Fig. 5A). Because in this and subsequent experiments the sense oligomers produce no change in [3H]thymidine incorporation, for the purpose of statistical analysis these data were combined with values obtained in the absence of all oligonucleotides. The antisense strand was found to significantly reduce [3H]thymidine incorporation (p < 0.05). MC proliferation was decreased by 28% when the cells were treated for 48 h with a 50 µ concentration of oligomer and by 35% when treated for 48 h with a 100 µ concentration (Fig. 5B and C, < 0.05 for each concentration). Again, sense oligomer treatment was without effect. We next tested the possibility that the antimitogenic effects of the antisense oligomer could be the result of a contaminating organic or other toxin in the antisense DNA preparation that was not present in the sense material. Although the oligomers were commercially synthesized, HPLC eluted, and guaranteed to be of high purity, we elected to test the preparations in a heterologous system in which we expected no effects. Mouse L929 cells were incubated with a 50 µ concentration of either sense or antisense oligomer for 48 h, and proliferation was measured by [3H]thymidine bioassay. Sense and antisense treatments were not significantly different (Fig. 5D).
Hybrid arrest translation with growth factor supplements We next performed experiments designed to show the specificity of the antisense oligomer. If the antisense oligomer mediates its antimitogenic action by by specifically blocking ANG translation, then addition of All should restore proliferation. Addition of other growth factors might not restore proliferation, depending on the importance of All and the ability of other growth factors to compensate for an All deficiency.
because sense oligomers produced no change in [3H]thymidine incorporation into DNA, for the statistical analyses, data obtained from sense oligonucleotide treatments were combined with of purpose values obtained for vehicle treatment. All and bFGF treatment of SHSY-5Y cells increased [3H]thymidine incorporation into DNA by 35 and 26%, respectively (p < 0.05, data from bFGF and sense + bFGF
Again,
205
COOK ET AL.
FIG. 5. [3H]thymidine incorporation into DNA of various cell types treated with "sense" and "antisense" oligonucle¬ otides. Bars represent the mean and standard deviation of triplicate readings. (A) SHSY-5Y neuroblastoma cells treated with oligonucleotides at a 50 µ concentration. (B) Human primary mesangial cells treated with oligonucleotides at a 50 µ concentration. (C) Human mesangial cells treated with oligonucleotides at a 100 µ concentration. (D) L929 cells treated with oligonucleotides at a 50 µ concentration.
treatments were combined; Fig. 6A). Antisense oligonucleotide decreased proliferation to 33% of control levels (p < 0.05). The growth inhibitory effect of All antisense oligonucleotide was completely eliminated by
the concurrent addition of All, but bFGF only partially compensated for the effects of antisense oligonucle¬ otide. Human mesangial cells responded in a similar fashion to oligomer and growth factor supplements (Fig. 6B). All stimulated MC proliferation by 30% (p < 0.05), as did bFGF. Antisense oligomer decreased proliferation to 39% of control levels (p < 0.05), and All restored proliferation to normal levels. bFGF was
incapable of restoring proliferation significantly. 206
ANTISENSE OLIGONUCLEOTIDES AND CELL GROWTH
25000 8000·
-
-,-
12000
_.
10000
20000 6000 ·
4000
2000-
15000-
10000-
5000-
nj
D
«
FIG. 6. [3H]thymidine incorporation into DNA of three cell types. A2 (angiotensin II) and bFGF were administered at 6 10 and 2 x -9 M final concentrations, respectively. They were each freshly applied every 2 h for 48 h. Sense and antisense oligonucleotides, at a 25 µ final concentration, were reapplied every 12 h for 48 h. (A) SHSY-5Y neuroblastoma cells. (B) Human primary mesangial cells. (C) L929 fibroblasts. Bars represent the mean and standard deviation of five separate treatments.
respond significantly to All sense 0.01) stimulatory effect to bFGF (Figure 6C).
L929 cells did not
(p
