Antisense Inhibition of Epidermal Growth Factor Receptor Decreases ...

Antisense Inhibition of Epidermal Growth Factor Receptor Decreases Expression of Human Surfactant Protein A Jonathan M. Klein, Troy A. McCarthy, John M. Dagle, and Jeanne M. Snyder Department of Pediatrics and Department of Anatomy and Cell Biology, University of Iowa, Iowa City, Iowa

Epidermal growth factor (EGF) stimulates surfactant protein A (SP-A) synthesis in fetal lung tissue through ligand binding to the EGF receptor. We hypothesized that inhibition of EGF receptor messenger RNA (mRNA) would block SP-A expression in human fetal lung tissue during alveolar type II cell differentiation in vitro. Midtrimester human fetal lung explants were maintained in serum-free Waymouth’s medium for 3 to 5 d in the presence or absence of an antisense 18-mer phosphorothioate oligonucleotide (ON) complementary to the initiation codon region of EGF receptor mRNA. Sense and scrambled ONs similarly modified were used as additional controls. The concentration of EGF receptor mRNA was semiquantitatively determined by reverse transcriptase/polymerase chain reaction (RT-PCR). We found a significant 3-fold decrease in EGF receptor mRNA levels in the antisense-treated groups compared with the control group with no effect in the sense condition. Immunohistochemical staining revealed a decrease in the amount of staining for EGF receptor protein in distal pulmonary epithelial cells in the antisense-treated groups compared with either control or sense conditions. Treatment with antisense EGF receptor ON decreased both SP-A mRNA and protein compared with controls with no effect in the sense condition. The ONs did not affect tissue viability as measured by the release of lactate dehydrogenase. We conclude that selective degradation of EGF receptor mRNA with antisense ON treatment results in a decrease in SP-A expression in human fetal lung. These findings support the critical importance of the EGF receptor for the regulation of SP-A gene expression during human alveolar type II cell differentiation.

Epidermal growth factor (EGF) is one of many regulatory peptide growth factors and hormones involved in fetal lung development (1–3). EGF binding results in dimerization of the EGF receptor, which in turn stimulates intrinsic tyrosine kinase activity leading to autophosphorylation of the receptor (4). The activated receptor initiates a signal transduction pathway through which mitogenesis and other cellular activities are regulated (5). Transforming growth factor-␣ (TGF-␣) (1), as well as amphiregulin (6), betacellulin (7), and heparin-binding EGF-like growth factor (8), binds to the EGF receptor, making this receptor a common pathway through which several growth factors modulate cellular activity. In fetal lambs, EGF increases alveolarization and reduces the severity of respiratory distress syndrome (9). In (Received in original form March 12, 1999 and in revised form November 23, 1999) Address correspondence to: Jonathan M. Klein, M.D., Dept. of Pediatrics, University of Iowa, Iowa City, IA 52242-1083. E-mail: jonathan-klein@ uiowa.edu Abbreviations: analysis of variance, ANOVA; complementary DNA, cDNA; epidermal growth factor, EGF; lactate dehydrogenase, LDH; messenger RNA, mRNA; oligonucleotide, ON; phosphate-buffered saline, PBS; reverse transcriptase/polymerase chain reaction, RT-PCR; standard error, SE; surfactant protein, SP; transforming growth factor, TGF. Am. J. Respir. Cell Mol. Biol. Vol. 22, pp. 676–684, 2000 Internet address: www.atsjournals.org

fetal rabbits, EGF improves pulmonary compliance (10) and stimulates surfactant phospholipid synthesis (11). In fetal rhesus monkeys, EGF stimulates surfactant protein A (SP-A) synthesis, enhances alveolar differentiation, and reduces the severity of respiratory distress syndrome (12). EGF also stimulates surfactant phospholipid synthesis (13) and alveolar type II cell differentiation (14) in fetal rat lung explants and branching morphogenesis in fetal mouse lung explants (15). In human fetal lung explants, EGF stimulates SP-A synthesis (16). Ligands for the EGF receptor have been found in human fetal lung tissue. EGF has been identified in fetal tracheobronchial and distal airway epithelium (17). After 24 wk gestation, both EGF and TGF-␣ are detectable by immunostaining in alveolar type II cells (17). TGF-␣ has also been detected in epithelial cells lining prealveolar ducts (18) and in both tracheal and distal airway epithelium throughout gestation (17). Furthermore, messenger RNA (mRNA) for both ligands is found in mesenchymal tissue from human fetal lung throughout gestation (17), observations suggestive of a paracrine effect of EGF/TGF-␣ on alveolar type II cell differentiation. EGF receptor protein and mRNA are found in alveolar epithelium from midtrimester human fetal lung explants that have undergone spontaneous differentiation in vitro (19). Thus, the presence of EGF receptor and its ligands in distal pulmonary epithelium during human fetal lung development suggests a regulatory role for EGF in type II cell differentiation. Whether EGF receptors are required for alveolar type II cell differentiation or just simply associated with alveolar epithelium during alveolar type II cell differentiation needs to be addressed. One strategy to address the issue of causality would be to block human EGF receptor activity. In other species, inhibiting the EGF receptor interferes with fetal lung development. In mutant mice in which the EGF receptor is lacking, the alveolar epithelium remains undifferentiated and the lungs are inadequately inflated (20). In a second EGF receptor gene-deletion mouse model, Miettinen and coworkers (21) showed that neonatal EGF receptor–deficient mice have impaired branching morphogenesis and alveolar type II cell immaturity. In cultured embryonic mouse lungs, exposure to a tyrosine kinase inhibitor (tyrphostin) blocks branching morphogenesis (15). In cultured rabbit type II cells, exposure to tyrosine kinase inhibitors (tyrphostin and genistein) also decreases thymidine incorporation and cell proliferation (22). Branching morphogenesis is decreased in mouse embryo lungs treated with antisense oligonucleotides (ONs) that inhibit EGF expression (23), and alveolar type II cell differentiation is inhibited in fetal mice treated with antimouse EGF antiserum (24). We have previously reported that treatment of sponta-

Klein, McCarthy, Dagle, et al.: Epidermal Growth Factor and Surfactant Protein A

neously differentiating human fetal lung explants with EGF receptor tyrosine kinase inhibitors (tyrphostin AG-1478 and genistein) resulted in a decrease in SP-A gene expression (25). This work demonstrated the importance of tyrosine kinase–dependent pathways in the regulation of SP-A during human fetal lung development. However, the possibility exists that other tyrosine kinases, independent of the EGF receptor tyrosine kinase, could also have been affected by the pharmacologic agents used, and thus the observed effect may not solely represent the function of the EGF receptor. To address the problem of specificity, we selectively inhibited EGF receptor gene expression using modified antisense ONs. We hypothesized that selective inhibition of the pulmonary EGF receptor would lead to decreased expression of human SP-A. To test our hypothesis, we measured concentrations of SP-A mRNA and protein in spontaneously differentiating human fetal lung explants cultured for 3 to 5 d in the presence of an antisense 18-mer phosphorothioate ON targeted against EGF receptor mRNA.

Materials and Methods Organ Culture Human fetal lung tissue was obtained under a protocol approved by the University of Iowa Human Subjects Review Committee. The explants were prepared from lung tissue obtained from midtrimester abortuses (15 to 20 wk), as previously described (26). The major airways were removed and the distal lung tissue was minced into 1-mm3 pieces with a razor blade under sterile conditions. The minced tissue was placed on a piece of lens paper that rested on a metal grid inside of a 35-mm culture dish containing 1 ml of serum-free Waymouth’s MB 752/1 medium (GIBCO BRL, Grand Island, NY) with added penicillin G (100 U/ml), streptomycin (100 ␮g/ml), and amphotericin (0.25 ␮g/ml). The explants were incubated at 37⬚C in a humidified atmosphere of 5% CO2 and 95% air for 3 to 5 d with the media changed daily. Starting tissue (human fetal lung tissue before culture) and the harvested explants were frozen in liquid nitrogen and stored at ⫺70⬚C until subsequent analysis. All experiments were conducted with explants prepared from individual fetuses.

Antisense Oligonucleotides An 18-mer antisense phosphorothioate ON of the following sequence (5⬘-GAGGGTCGCATCGCTGCT-3⬘), targeted against the human EGF receptor mRNA (27) and spanning the initiation codon, was synthesized by Oligos Etc., Wilsonville, OR. A sense ON (5⬘-AGCAGCGATGCGACCCTC-3⬘) similarly modified, a scrambled ON (5⬘-AGGGCTATCGGTGCTCGC-3⬘) continuing the same frequency of nucleotides as the antisense ON, and a carrier only condition were used as controls. Lipofectamine (GIBCO BRL) was used as a carrier of the ONs, and we modified the manufacturer’s protocol for transfection of cells as follows. We prepared the following two solutions: First, we added 45 nmol of ON to 100 ␮l of Opti-MEM I (GIBCO BRL) to create solution A. Then, we added 20 ␮l of lipofectamine to 100 ␮l of Opti-MEM I to create solution B. Solutions A and B were mixed together and then incubated at room temperature for 15 min to form DNAliposome complexes. The resulting complexes were added to 35mm culture dishes containing the explants together with serumfree Waymouth’s medium to reach a final total volume of 500 ␮l. The amount of ON used was previously determined so that the final concentration when added to the media containing the explants would be 90 ␮M. The media were changed daily at which

677

time fresh ON was added as described previously. The explants were harvested after 3 to 5 d. Cytotoxicity was assessed by measuring the release of lactate dehydrogenase (LDH) (LDH assay, LDL-20 kit; Sigma Chemical Co., St. Louis, MO) into the media.

Reverse Transcriptase/Polymerase Chain Reaction Steady-state levels of EGF receptor mRNA present were determined by reverse transcriptase/polymerase chain reaction (RTPCR). Total RNA was isolated by a single-step acid-phenol-chloroform extraction method (28). Reverse transcription to complementary DNA (cDNA) was performed using 1 ␮g of total RNA per condition, 100 ng of 18-mer oligo(dt), 500 ␮M of each deoxynucleotide triphosphate (dNTP), Moloney mouse leukemia virus (MMLV) buffer, RNase inhibitor (40 U), bovine serum albumin (BSA; 5 ␮g), MMLV-reverse transcriptase (200 U), and sterile distilled water for a total volume of 30 ␮l. The mixture was incubated for 1 h at 37⬚C, and the resulting cDNA was stored at ⫺70⬚C. All PCR procedures were carried out in a laminar flow hood using positive-pressure displacement pipettes. We used a master mix that contained PCR amplification buffer (2 mM MgCl2, 60 mM Tris-HCl [pH 8.5], 15 mM NH4SO4), dNTPs (2 mM, final concentration ⫽ 0.1 mM), together with the 5⬘ primer (5⬘-TATTGATCGGGAGAGCCG-3⬘) and the 3⬘ primer (5⬘-TGGGCAGCTCCTTCAGTC-3⬘). The primers were complementary to the sequence for the human EGF receptor cDNA (27) and designed to straddle the binding site for the antisense ON. The predicted amplification product size was 451 bp. To the master mix, 6 ␮l of reverse transcribed cDNA per condition was added together with Taq polymerase (2.5 U) to a total volume of 100 ␮l. The tubes were covered with light mineral oil and denatured at 95⬚C for 30 s, annealed at 59⬚C for 1 min, extended at 72⬚C for 1 min for a total of 35 cycles, and then extended at 72⬚C for 5 min. Preliminary experiments showed that amplification was linear with 35 cycles or less. After amplification, 10-␮l aliquots of PCR products from each condition were separated on a 1% agarose gel. ␤-actin was also amplified from each condition to control for nonspecific effects of the phosphorothioate antisense ONs. The ␤-actin PCR primers, obtained from Research Genetics, Inc. (Huntsville, AL), amplified a 289-bp region of exon 3 of human ␤-actin cDNA. The ␤-actin PCR was performed as described previously, except only 1 ␮l of reverse transcribed cDNA per condition was used. The density of the amplified PCR products was measured from photographs of DNA gels stained with ethidium using the AlphaImager IS2000 Digital Imaging System (Alpha Innotech Corp., San Leandro, CA).

Northern Blot Analysis of SP-A mRNA Total RNA was isolated by a single-step acid-phenol-chloroform extraction method (28). A total of 10 ␮g of total RNA per condition was separated by gel electrophoresis (1.2% agarose), transferred by capillary action to a nylon membrane (S&S Nytran; Schleicher & Schuell, Keene, NH), baked 30 min, UV crosslinked (UV Stratalinker 1800; Stratagene, La Jolla, CA), and prehybridized as previously described (29). SP-A and SP-B cDNA probes (kindly supplied by J. Whitsett, Department of Pediatrics, University of Cincinnati, Cincinnati, OH) were radiolabeled with [␣-32P]deoxycytidine triphosphate using a random priming kit (Amersham, Arlington Heights, IL). A ␤-actin cDNA probe (American Type Culture Collection, Rockville, MD) was used to control for loading. Hybridization was performed as described (29), and mRNA levels were quantitated by densitometry of autoradiographs (AMBIS Radioanalytic and Visual Imaging System; Ambis Inc., San Diego, CA). To control for loading artifacts, the density of the ␤-actin bands was measured, and the density of the SP-A and SP-B bands was then normalized

678

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 22 2000

to the density of their corresponding ␤-actin band. The densitometric data were then normalized to the control condition with the control condition set equal to one for each experiment.

Western Blot Immunoanalysis of SP-A Starting tissue and harvested explants were homogenized in phosphate-buffered saline (PBS) with 1 mM phenylmethylsulfonyl fluoride, leupeptin (20 ␮g/ml), soybean trypsin inhibitor (5 ␮g/ ml), and 5 mM ethylenediaminetetraacetic acid (EDTA). Samples were centrifuged (600 ⫻ g) for 5 min, and supernatant protein (75 ␮g per lane) was separated by electrophoresis on a 12.5% sodium dodecyl sulfate (SDS)-polyacrylamide gel, transferred to an Immobilon-P membrane (Millipore Corp., Bedford, MA), and blocked as described (30). The membrane was incubated for 1 h at room temperature with guinea pig polyclonal antihuman SP-A antibodies (1:1,000 dilution), rinsed with double-distilled water, then incubated with sheep antiguinea pig immunoglobulin (Ig) G conjugated to alkaline phosphatase (1:2,000 dilution; Boehringer Mannheim Corp., Indianapolis, IN) for 1 h at room temperature, and then washed as previously described (30). The immunoreactive SP-A bands were then detected by incubating the membrane at room temperature for 30 min in a solution containing 100 mM Tris (pH 9.5), 100 mM NaCl, 5 mM MgCl2, 5-bromo-4-chloro-3indoyl phosphate (165 ␮g/ml), and nitro blue tetrazolium (330 ␮g/ml). The membrane was rinsed in distilled water, dried, and photographed. The relative amount of immunoreactive SP-A present in each sample was quantitated by densitometry (AMBIS Radioanalytic and Visual Imaging System, Ambis Inc.). The densitometric data from each blot were normalized to the control condition with the control value set equal to one for each experiment.

Immunohistochemistry for EGF Receptor From three different experiments, two explants per condition after culturing for 5 d were sectioned and stained. After harvesting, the tissue was frozen immediately in liquid nitrogen and stored at ⫺70⬚C. The frozen tissue was mounted in an optimal cutting temperature compound, and 7-␮m sections were prepared with a cryostat and thaw-mounted on glass slides. Sections were fixed for 10 min at room temperature in freshly prepared 10% formalin in PBS. The sections were rinsed twice for 10 min per rinse in PBS. Endogenous peroxidase activity was quenched by incubating sections in 0.3% H2O2 in methanol for 30 min followed by rinsing twice in PBS for 10 min. The sections were stained using an avidin biotinylated complex kit (Vectastain Elite kit; Vector Labs, Burlingame, CA). Nonspecific binding sites were blocked by incubating the sections with 2% normal goat serum in PBS for 20 min at room temperature followed by a second blocking step using 2% normal goat serum and 0.25% BSA for 20 min. The sections were rinsed in PBS, then incubated for 1 h in a humidified chamber at room temperature with a mouse antihuman EGF receptor antibody (Upstate Biotechnology, Inc., Lake Placid, NY) at a dilution of 1:50 in PBS. The tissue sections were washed two times in PBS, 5 min per rinse, then incubated for 30 min in biotinylated secondary antibody, then rinsed two times in PBS, 5 min per rinse. The sections were then incubated for 45 min in avidinperoxidase reagent. After rinsing two times in PBS, 5 min per rinse, the sections were incubated in diaminobenzidine (700 ␮g/ml) for 1 to 3 min. Sections were rinsed in PBS for 5 min, rinsed quickly in distilled water, then dehydrated and mounted with glass coverslips. In some experiments, the sections were counterstained with hematoxylin for 30 s. Negative staining controls were incubated with secondary antibody alone and were performed for all experimental conditions. Sections were viewed and photographed with a Nikon FX photomicroscope (Nikon, Tokyo, Japan).

Immunoprecipitation and Immunoblotting of Phosphotyrosine Proteins The harvested explants were homogenized in PBS with 200 ␮M sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, leupeptin (20 ␮g/ml), soybean trypsin inhibitor (5 ␮g/ml), and 5 mM EDTA. The resulting homogenates (500 ␮g of total protein) were incubated with monoclonal antiphosphotyrosine agarose conjugate (20 ␮g, UBI) for 2 h at room temperature with shaking. Immune complexes were pelleted by centrifugation, and the supernatant was discarded. The pellet was resuspended by boiling, and the proteins were separated on a 7.5% SDS-polyacrylamide gel electrophoresis gel. The phosphotyrosine proteins were then transferred to an Immobilon-P membrane (Millipore Corp.). Nonspecific binding was blocked by an overnight incubation at 4⬚C with 3% nonfat dry milk in PBS. The membrane was then incubated for 3 h at room temperature with a mouse monoclonal antiphosphotyrosine antibody (1:1,000 dilution, Upstate Biotechnology, Inc., Lake Placid, NY), rinsed, and then incubated with goat antimouse IgG conjugated to alkaline phosphatase (1:2,000 dilution; Boehringer Mannheim) for 1 h at room temperature. The membrane was rinsed and the immunoreactive tyrosinephosphorylated bands were detected by incubating the membrane at 25⬚C for 30 min in 100 mM Tris (pH 9.5), 100 mM NaCl, 5 mM MgCl2, 5-bromo-4-chloro-3-indoyl phosphate (165 ␮g/ml), and nitro blue tetrazolium (330 ␮g/ml). The membrane was rinsed in distilled water, dried, and photographed. The relative amount of immunoreactive tyrosine-phosphorylated protein visualized at 170,000 D, which is the presumptive tyrosine-phosphorylated EGF receptor, was quantitated by densitometry using the AlphaImager 2000 (Alpha Innotech Corp.).

Statistical Analysis All data are presented as the mean ⫾ standard error of the mean. The effects of antisense and sense ONs on EGF receptor mRNA, SP-A mRNA, and SP-A protein levels were statistically evaluated by one-way analysis of variance (ANOVA). The assessment of significant differences among multiple comparisons was performed using the Dunnett’s test when we compared all the different experimental treatments to the control group alone. In the experiments in which we examined differences among all possible experimental conditions in a pairwise manner, we used the Student-Neuman-Keuls test (SigmaStat, SPSS, Inc., Chicago, IL). Significance was defined as P ⬍ 0.05. The unpaired Student’s t test was used for experiments not involving multiple comparisons.

Results Antisense Inhibition of EGF Receptor mRNA Initially, we used an antisense EGF receptor ON to selectively mediate the degradation of EGF receptor mRNA in cultured fetal lung tissue. We performed dose–response experiments in tissue cultured for 3 d under the following conditions: starting tissue (tissue before culture), control, lipofectamine (carrier), sense EGF receptor ON (15 to 90 ␮M), and antisense EGF receptor ON (15 to 90 ␮M). In dose– response experiments, we found that an antisense ON concentration of 90 ␮M would consistently eliminate EGF receptor mRNA as measured by RT-PCR (Figure 1). Treatment with antisense EGF receptor ON significantly lowered the amount of EGF receptor mRNA present in cultured human fetal lung explants in a dose-dependent manner, with message levels decreasing by over 90% at a concentration of 90 ␮M (0.08 ⫾ 0.07) compared with either control (1.0) or li-


679

Figure 2. Effect of EGF receptor antisense ON treatment (90 ␮M) on EGF receptor mRNA expression in human fetal lung tissue. Densitometric data (means ⫾ SE) of EGF receptor mRNA levels as determined by RT-PCR in fetal lung tissue cultured for 3 d in the presence or absence of 90 ␮M antisense or sense ONs. The values were normalized to the control condition for each experiment. There was a significant decrease in EGF receptor mRNA levels in the antisense treated group (n ⫽ 7, ANOVA, P ⬍ 0.001). Antisense treatment significantly decreased EGF receptor mRNA levels when compared with either the control, lipofectamine, or sense groups (Student-Newman-Keuls, *P ⬍ 0.05). Figure 1. Dose–response effect of sense and antisense EGF receptor ON treatment on EGF receptor mRNA in cultured human fetal lung tissue. Explants were cultured for 3 d in the presence of either sense or antisense EGF receptor ONs at 15, 30, 60, and 90 ␮M, and the levels of EGF receptor mRNA present were determined by RT-PCR. (A) Top panel: Representative ethidium bromide–stained agarose gel showing amplified EGF receptor (451-bp band) with size markers at 500 and 300 bp (arrows). EGF receptor mRNA clearly decreases with antisense exposure at concentrations ⭓ 30 ␮M. There is no effect with exposure to either sense EGF receptor ON or to the carrier (lipofectamine [L]). The other conditions are a PCR blank (B) without added cDNA template as well as starting tissue (St) and the untreated control cultured tissue (C). Bottom panel: Same experimental samples as in top panel, except ␤-actin RT-PCR was performed showing that ␤-actin mRNA (289-bp band) was not affected by treatment with either the sense or antisense EGF receptor ONs. (B) Densitometric data (means ⫾ SE, n ⫽ 3 experiments) of EGF receptor mRNA levels as determined by RT-PCR. Sense, solid bars; antisense, lightly shaded bars. The values were normalized to untreated controls with the control value set equal to one for each experiment. There was a significant dose-dependent decrease in EGF receptor mRNA levels with increasing concentrations of antisense EGF receptor ON (ANOVA, P ⫽ 0.001). Significant decreases in EGF receptor mRNA levels compared with either control or lipofectamine were detected at an antisense concentration ⭓ 30 ␮M (Dunnett’s test, *P ⬍ 0.05). There was no effect of sense EGF receptor ON on EGF receptor mRNA.

pofectamine (0.88 ⫾ 0.14) (Figure 1B). There was no effect on EGF receptor mRNA in explants exposed to similar concentrations of sense EGF receptor ON (ANOVA, P ⫽ 0.515, control set equal to 1.0, 90 ␮M sense equaled 0.84 ⫾ 0.10). Neither sense nor antisense EGF receptor ONs affected the level of ␤-actin mRNA (Figure 1A). Using the information derived from the dose–response experiments, in which treatment with 90 ␮M antisense EGF receptor lowered EGF receptor mRNA in cultured human fetal lung tissue, we performed seven additional experiments using both sense and antisense EGF receptor ONs at a concentration of 90 ␮M. We cultured explants for 3 d in the presence or absence of either sense or antisense EGF receptor ONs and measured the concentration of EGF receptor mRNA present by RT-PCR. Steady-state

levels of EGF receptor mRNA decreased significantly after 3 d in culture with the antisense EGF receptor ON when compared with either the control, lipofectamine, or sense ON conditions, confirming the specific effect of the antisense ON in fetal lung tissue (Figure 2). We also measured the concentration of EGF receptor mRNA present by RT-PCR in human fetal lung explants exposed to 90 ␮M antisense EGF receptor ON after 5 d in culture—the time point for all of the protein assays that were used. Again, there was complete absence of EGF receptor mRNA in the antisense group (data not shown, n ⫽ 3), consistent with the 3-d data seen in Figure 1. Effect of Inhibiting EGF Receptor mRNA on Steady-State Levels of SP-A mRNA We evaluated SP-A gene expression in explants in which EGF receptor mRNA gene expression was inhibited using antisense EGF receptor ONs. Total RNA was isolated from explants that had been cultured for 3 d with antisense EGF receptor ON (90 ␮M) in which the decrease in EGF receptor mRNA levels was confirmed by RT-PCR as previously demonstrated (Figure 1). In the cultured tissue exposed to antisense EGF receptor ON, there was a significant decrease in steady-state levels of SP-A mRNA (Figure 3). The sense EGF receptor ON did not significantly affect SP-A mRNA levels. Steady-state levels of SP-A mRNA increased with time in culture in the control, lipofectamine, and sense conditions compared with the undifferentiated starting tissue before culture (Figure 3A). Treatment with antisense EGF receptor ONs significantly decreased SP-A gene expression in cultured human fetal lung tissue compared with either control, lipofectamine, or sense conditions (Figure 3B). Steady-state levels of SP-B mRNA were increased in the control, lipofectamine, and sense cultured conditions when compared with the undifferentiated start tissue (Figure 3A). However, SP-B gene expression was not affected by exposure to the antisense EGF receptor ON (Figure 3A). In addition, antisense EGF receptor ON exposure did not alter steady-state levels of ␤-actin mRNA (Figure 3A).

680


Figure 3. Effect of antisense EGF receptor treatment (90 ␮M) on surfactant protein gene expression in human fetal lung explants after 3 d in culture. (A) Top panel: Representative autoradiogram of Northern blot of total RNA (10 ␮g/lane) hybridized with [32P]-labeled human SP-A cDNA. Lane 1, starting tissue before culture; lane 2, untreated control explants; lane 3, lipofectamine treated; lane 4, sense EGF receptor ON treated; lane 5, antisense EGF receptor ON treated. There is a significant decrease in the level of SP-A mRNA in cultured tissue exposed to antisense EGF receptor ON compared with the control, lipofectamine, and sense ON conditions. Middle panel: The same blot stripped and reprobed with [32P]-labeled human SP-B cDNA. SP-B mRNA levels were not affected by exposure to the antisense EGF receptor ON. Bottom panel: The same blot stripped again and reprobed with [32P]-labeled human ␤-actin cDNA. Levels of ␤-actin mRNA also were unaffected by exposure to antisense EGF receptor ONs. (B) Densitometric data (n ⫽ 5, means ⫾ SE) of steady-state levels of SP-A mRNA normalized to ␤-actin in each lane to control for loading and then normalized to the control condition for each experiment, with the control value set equal to one. Treatment with antisense EGF receptor ON significantly decreased the level of SP-A mRNA compared with either the control, lipofectamine, or sense conditions (one-way ANOVA, P ⫽ 0.004; *Student-Newman-Keuls, P ⬍ 0.05).

Effect of Inhibiting EGF Receptor mRNA on SP-A Content Immunoreactive SP-A protein was measured in human fetal lung explants cultured for 5 d with antisense EGF receptor ON. There was a significant 50% reduction in SP-A protein in tissue exposed to antisense EGF receptor ON compared with control (Figure 4). This decrease was consistent with the SP-A mRNA data (Figure 3). SP-A content was not significantly decreased with 5 d of exposure to the sense EGF receptor ON or to lipofectamine (Figure 4B). There was an apparent 21% reduction in SP-A content in the vehicle (lipofectamine) condition; however, this decrease was not statistically significant. Because the antisense plus vehicle (lipofectamine) condition showed a significant decrease, but the vehicle (lipofectamine) condition alone did not, the effect observed was most likely due to the presence of the antisense EGF receptor ON and not due to the vehicle. However, if there was a small additive, nonspecific effect of lipofectamine on SP-A synthesis in addition to the effect on SP-A synthesis from antisense in-

Figure 4. Effect of antisense EGF receptor ON on SP-A content (35-kD band) in cultured human fetal lung tissue. (A) Representative SP-A immunoblot of the effect of antisense EGF receptor ON on levels of SP-A in human fetal lung explants cultured for 5 d. In the tissue exposed to antisense EGF receptor ON, there is a clear decrease in the amount of SP-A protein (35-kD band) compared with the sense condition. (B) Densitometric data of SP-A protein levels after 5 d of exposure to antisense EGF receptor ON, normalized to control with the control value set equal to one for each experiment (means ⫾ SE, n ⫽ 3). There is a significant decrease in SP-A protein in tissue exposed to antisense EGF receptor ON when compared with the control condition (ANOVA, P ⫽ 0.017; Dunnett’s test, *P ⬍ 0.05). SP-A protein content is not significantly decreased with exposure to either the carrier (lipofectamine) or the sense EGF receptor ON when compared with control.

hibition of EGF receptor gene expression, our study would not have had the power to prove this effect significant. As an additional control, the tissue was cultured for 4 d with a scrambled EGF receptor ON that did not decrease SP-A protein as compared with the antisense condition (data not shown). Effect of Inhibiting EGF Receptor mRNA on Immunolabeled EGF Receptor Immunohistochemical staining for EGF receptor was performed on frozen sections of cultured fetal lung tissue in order to observe the post-translational effects of using antisense EGF receptor ONs to selectively inhibit EGF receptor mRNA. EGF receptor immunostaining was detected in epithelial cells lining the prealveolar ducts of the undifferentiated start tissue before culture (data not shown). Immunostaining for the EGF receptor was clearly observed in distal pulmonary epithelium lining the ducts of control fetal lung explants that had been cultured for 5 d (Figure 5A). However, in the explants cultured for 5 d with antisense EGF receptor ONs, there was a minimal amount of EGF receptor immunostaining detected in the distal pulmonary epithelial cells that lined the ducts (Figure 5C). This attenuation of immunostaining for the EGF receptor was not seen in explants cultured with either the sense EGF receptor ON (Figure 5D) or lipofectamine alone


681

Figure 5. Representative photomicrographs showing immunohistochemical staining for EGF receptor protein in cultured human fetal lung tissue. (A) Immunostaining for the EGF receptor was observed in the distal pulmonary epithelial cells (arrows) lining the lumen (asterisks) of prealveolar ducts in control explants cultured for 5 d. CT, connective tissue. (B) Immunostaining for EGF receptor was similarly detected in epithelial cells (arrows) lining the lumen (asterisks) of ducts in explants cultured for 5 d in the presence of lipofectamine (the carrier). (C) In the tissue cultured for 5 d with antisense EGF receptor ON, EGF receptor immunostaining was decreased to absent in the majority of epithelial cells (arrows) that lined the prealveolar ducts. There was also decreased cellularity of the CT compared with the control or lipofectamine conditions. (D) In tissue cultured for 5 d with sense EGF receptor ON, there was no decrease in EGF receptor immunostaining in the epithelial cells (arrows); however, the CT had decreased cellularity. Original magnification of all panels: ⫻40.

(Figure 5B). The decreased degree of detectable EGF receptor immunostaining in explants exposed to antisense EGF receptor was replicated in two additional experiments using different starting tissue. An additional control, consisting of a scrambled antisense EGF receptor ON, also did not affect EGF receptor immunostaining, results similar to the sense ON condition (data not shown). In the explants treated with EGF receptor ON, there was decreased cellularity of the connective tissue compared with the control and lipofectamine conditions, with the prealveolar ducts remaining intact (Figure 5). Negative staining controls using secondary antibody alone were performed for all experimental conditions and resulted in an absence of immunostaining (data not shown). Effect of Inhibiting EGF Receptor mRNA on Basal EGF Receptor Tyrosine Phosphorylation To determine the post-translational effects of using antisense EGF receptor ONs to selectively inhibit EGF receptor mRNA, we measured basal levels of tyrosine-phosphorylated EGF receptor in human fetal lung explants cultured for 5 d in the presence or absence of antisense EGF receptor ON (90 ␮M). Densitometric analysis of ty-

rosine-phosphorylated EGF receptor was performed, and the densities of the 170-kD bands from each blot were normalized to the lipofectamine (vehicle-control) condition, which was set at a value of one, and compared between blots. The lipofectamine control condition was chosen as the condition for normalization of the data between blots in order to minimize any nonspecific effects of lipofectamine. There was a significant reduction in tyrosine-phosphorylated EGF receptor in tissue exposed to antisense EGF receptor ON compared with lipofectamine (vehicle) (Figure 6). These data were consistent with the decrease in EGF receptor mRNA seen after 3 d of antisense EGF receptor ON treatment (Figure 1). The level of tyrosinephosphorylated EGF receptor was not significantly decreased after 5 d of exposure to the sense EGF receptor ON compared with either lipofectamine or the control condition (Figure 6). Additonally, there was no difference between the control and lipofectamine conditions in the amount of tyrosine-phosphorylated EGF receptor present. Cytotoxicity The release of LDH into the media from fetal lung tissue cultured for 3 d in the presence of 90 ␮M antisense EGF

682


Figure 6. Tyrosine phosphorylation of EGF receptors from fetal lung explants cultured for 5 d in the presence of either lipofectamine, sense EGF receptor ON, or antisense EGF receptor ON. EGF receptor tyrosine phosphorylation was quantitated by densitometry after explants were homogenized, and then immunoprecipitated and immunoblotted with antibodies to phosphotyrosine. There was a significant decrease in tyrosine-phosphorylated EGF receptor after exposure to the antisense EGF receptor ON (ANOVA, P ⫽ 0.026; *Student-Newman-Keuls, P ⬍ 0.05 antisense group compared with vehicle control [lipofectamine]). There were no significant differences between either control, lipofectamine (vehicle control), and sense ON conditions. Data are the mean ⫾ SE; n ⫽ 4 separate experiments.

receptor phosphorothioate ON was measured to assess toxicity. There were no significant differences in LDH levels (IU/liter) between control (10.5 ⫾ 4.1), antisense EGF receptor ON (11.2 ⫾ 3.7), sense EGF receptor ON (10.2 ⫾ 3.1), and lipofectamine (8.2 ⫾ 2.4) conditions (means ⫾ standard error [SE], n ⫽ 3 experiments).

Discussion SP-A is the most abundant of the surfactant proteins and serves many functions, including helping to form tubular myelin, facilitating the surface tension reducing properties of surfactant phospholipids, regulating the recycling and secretion of surfactant phospholipids, and contributing to local host-defense mechanisms in the lung (31). In human fetal lung explants, EGF has been shown to significantly increase SP-A expression and synthesis (16). EGF mediates its effects through a cell-surface receptor dependent on intrinsic tyrosine kinase activity for initiation of its signal transduction pathways (32, 33). We have previously shown that inhibition of EGF receptor tyrosine kinase with tyrphostin blocks the expression of SP-A during spontaneous differentiation of cultured human fetal lung explants (25). However, the possibility exists that other receptor tyrosine kinases (e.g., platelet-derived growth factor [PDGF] receptor, insulin-like growth factor 1 receptor, or fibroblast growth factor receptor [4]), independent of the EGF receptor tyrosine kinase, could also have been affected by tyrphostin, and thus the observed effects may not purely represent the function of the EGF receptor alone. Because pharmacologic agents have limitations regarding their degree of specificity, we decided to inhibit expression of the EGF receptor gene by using antisense ONs to selectively mediate the degradation of EGF receptor mRNA. This, in turn, would result in a selective decrease in EGF receptor content, and thus the observed effects of exposure to antisense EGF receptor ONs on SP-A

should be specific to the EGF receptor and not to other receptor or nonreceptor tyrosine kinases. Antisense ON strategies to inhibit growth factors or their receptors have been performed successfully in embryonic tissue cultures. Potts and colleagues (34) used phosphoramidate-modified ONs at a concentration of 1 ␮M to inhibit transforming growth factor ␤3 in embryonic chick heart explants. Seth and coworkers (23) used unmodified ONs at a concentration of 30 ␮M to inhibit EGF expression in cultured embryonic mouse lungs. Souza and associates (35) used phosphorothioate-modified ONs at a concentration of 10 ␮M to inhibit PDGF receptor expression in cultured embryonic rat lungs. We found that we needed a slightly higher concentration of the phosphorothioate-modified ONs than others have used owing to the thickness of the human fetal lung explants compared with embryonic chick, mouse, and rat explants. Still, the concentration of ON that we used was at a similar order of magnitude to that used in previous studies in embryonic lung tissue. The 1-mm cubic thickness of these explants was also the most likely cause for some variability in the degree to which we were able to eliminate EGF receptor from within each epithelial cell. The farther the epithelial cells were from the periphery of the explant, the more difficult it would have been for the antisense ON to diffuse into the cell. Thus, this would result in a few of the epithelial cells still having detectable immunoreactive EGF receptor as seen in the immunohistochemistry (Figure 5C). We chose to evaluate the role of the EGF receptor on the developmental regulation of SP-A in human tissue using cultured fetal lung explants as an in vitro model of human lung development. Previously, it has been shown that epithelial cells within undifferentiated midtrimester human fetal lung tissue spontaneously differentiate into alveolar type II cells with the ability to synthesize SP-A after 3 to 4 d in explant culture (25, 30). Additionally, these distal lung epithelial cells express EGF receptors on their cytoplasmic membranes (19), which supports a role for the EGF receptor in the developmental regulation of SP-A synthesis. Thus, we evaluated the effects of selectively inhibiting EGF receptor gene expression on SP-A content and gene expression in cultured human fetal lung tissue. We found that treatment with antisense EGF receptor ONs significantly decreased SP-A mRNA and protein in cultured human fetal lung explants compared with the control, sense, and scrambled ON conditions. This is an important finding because SP-A content would be predicted to greatly increase in the control explants with time in culture (25), whereas treatment with antisense EGF receptor ON limited the increase in SP-A content by over 50% compared with control explants. Treatment with antisense EGF receptor ON also resulted in decreased expression of SP-A mRNA, consistent with an inhibitory effect at the pretranslational level. Our observation that SP-A gene expression was decreased by antisense inhibition of EGF receptor agrees with previous work by Raaberg and coworkers (36), who described decreased immunostaining for SP-A in EGF-deficient newborn rats, thus supporting the importance of the EGF receptor pathway for SP-A synthesis. The effect of antisense EGF receptor ON on SP-A


gene expression in fetal lung explants is due to its direct and specific effect on the EGF receptor itself and not to a global effect of the ONs on overall gene expression. We showed that treatment for 3 d with antisense EGF receptor ON significantly decreased EGF receptor gene expression without affecting the expression of ␤-actin. Additionally, in the explants treated for 5 d with antisense EGF receptor ON, there was an obvious qualitative decrease in the level of immunoreactive EGF receptor, implying a decrease in EGF receptor protein is consistent with a drop in EGF receptor mRNA. Immunostaining for the EGF receptor was only seen in the epithelial cells lining the ducts in the fetal lung explants, and thus the effects of inhibiting EGF receptor gene expression was specific to these distal pulmonary epithelial cells. Additionally, we found that exposure for 5 d to antisense EGF receptor ON significantly reduced basal levels of tyrosine-phosphorylated EGF receptor by 43% (Figure 6). These data confirm that a quantitative reduction in EGF receptor protein is found 2 d after the 3-d time period it takes for the antisense ON to significantly decrease EGF receptor mRNA levels. This time frame is consistent with the half-life of the EGF receptor, which is estimated to be 10 h in human fibroblasts (32), and thus there would be adequate time for the protein to begin to degrade from baseline. In explants exposed to phosphorothioate ONs, there was a decrease in the cellularity of the connective tissue, whereas the epithelial cells lining the prealveolar ducts remained intact. This was not a result of cytotoxicity because there was not an increase in the release of LDH into the media. Furthermore, the ONs did not inhibit the expression of either SP-B or ␤-actin mRNA. Thus, the decreased connective tissue cellularity was most likely a nonspecific (sequence independent) effect of the phosphorothioate ONs. These compounds have been previously shown to interact nonspecifically with a number of cellular proteins (37). The mechanism of interaction relates, in part, to the polyanionic nature of the phosphorothioates, which mimics the charge density of heparin (38, 39). Sequence independent binding of phosphorothioate ONs to both laminin and fibronectin, extracellular matrix proteins possessing heparin-binding domains, has been previously reported (40). Decreasing the availability of heparin-binding sites on these important extracellular proteins in the lung appears in our explants to inhibit the normal migration and attachment of cells within the connective tissue. We have observed decreased cellularity of connective tissue between alveolar ducts after culturing with five different phosphorothioate ONs (experimental and control sequences). These findings are consistent with the hypothesis that phosphorothioate ONs decrease connective tissue cellularity by binding to and altering the extracellular matrix. In conclusion, we have demonstrated a direct cell-specific link between EGF receptor and the regulation of SP-A gene expression during human type II cell differentiation in vitro. Thus, the presence of intact EGF receptors is critical for the induction of SP-A in the fetal lung. This effect appears to be specific to SP-A because we saw no change in either SP-B expression or in the progression of the pulmonary epithelium from columnar to cuboidal cells in the

683

antisense ON condition. How a decrease in EGF receptors leads to decreased SP-A mRNA levels (i.e., decreased transcription or decreased message stability) will have to be addressed in future experiments. Acknowledgments: This work was supported in part by National Institutes of Health grants HL-52055 (J.M.K.) and HL-50050 (J.M.S.). The authors thank Kelli Goss and Pepper Kerr for their technical assistance.

References 1. King, R. J., M. B. Jones, and P. Minoo. 1989. Regulation of lung cell proliferation by polypeptide growth factors. Am. J. Physiol. 257:L23–L38. 2. Gross, I. 1990. Regulation of fetal lung maturation. Am. J. Physiol. 259: L337–L344. 3. Warburton, D., M. Lee, M. A. Berberich, and M. Bernfield. 1993. Molecular embryology and the study of lung development. Am. J. Respir. Cell Mol. Biol. 9:5–9. 4. Cadena, D. L., and G. N. Gill. 1992. Receptor tyrosine kinases. FASEB J. 6:2332–2337. 5. Todderud, G., and G. Carpenter. 1989. Epidermal growth factor: the receptor and its function. Biofactors 2:11–15. 6. Shoyab, M., G. D. Plowman, V. L. McDonald, J. G. Bradley, and G. J. Todaro. 1989. Structure and function of human amphiregulin: a member of the epidermal growth factor family. Science 243:1074–1076. 7. Shing, Y., G. Christofori, D. Hanahan, Y. Ono, R. Sasada, K. Igarashi, and J. Folkman. 1993. Betacellulin: a mitogen from pancreatic ␤ cell tumors. Science 259:1604–1607. 8. Higashiyama S., J. A. Abraham, J. Miller, J. C. Fiddes, and M. Klagsbrun. 1991. A heparin-binding growth factor secreted by macrophage-like cells that is related to EGF. Science 251:936–939. 9. Sundell, H. W., M. E. Gray, F. S. Serenius, M. B. Escobedo, and M. T. Stahlman. 1980. Effects of epidermal growth factor on lung maturation in fetal lambs. Am. J. Pathol. 100:707–726. 10. Catterton, W. Z., M. B. Escobedo, W. R. Sexson, M. E. Gray, H. W. Sundell, and M. T. Stahlman. 1979. Effect of epidermal growth factor on lung maturation in fetal rabbits. Pediatr. Res. 13:104–108. 11. Haigh, R. M., M. Hollingsworth, L. A. Micklewright, R. D. H. Boyd, and S. W. D’Souza. 1988. The effect of human urogastrone on lung phospholipids in fetal rabbits. J. Dev. Physiol. 10:433–443. 12. Goetzman, B. W., L. C. Read, C. G. Plopper, A. F. Tarantal, C. GeorgeNascimento, T. A. Merritt, J. A. Whitsett, and D. Styne. 1994. Prenatal exposure to epidermal growth factor attenuates respiratory distress syndrome in rhesus infants. Pediatr. Res. 35:30–36. 13. Gross, I., D. W. Dynia, S. A. Rooney, D. A. Smart, J. B. Warshaw, J. F. Sissom, and S. B. Hoath. 1986. Influence of epidermal growth factor on fetal rat lung development in vitro. Pediatr. Res. 20:473–477. 14. Fraslon, C., and J. R. Bourbon. 1992. Comparison of effects of epidermal and insulin-like growth factors, gastrin releasing peptide and retinoic acid on fetal lung cell growth and maturation in vitro. Biochim. Biophys. Acta 1123:65–75. 15. Warburton, D., R. Seth, L. Shum, P. G. Horcher, F. L. Hall, Z. Werb, and H. C. Slavkin. 1992. Epigenetic role of epidermal growth factor expression and signalling in embryonic mouse lung morphogenesis. Dev. Biol. 149: 123–133. 16. Whitsett, J. A., T. E. Weaver, M. A. Lieberman, J. C. Clark, and C. Daugherty. 1987. Differential effects of epidermal growth factor and transforming growth factor-beta on synthesis of Mr ⫽ 35,000 surfactant-associated protein in fetal lung. J. Biol. Chem. 262:7908–7913. 17. Ruocco, S., A. Lallemand, J. M. Tournier, and D. Gaillard. 1996. Expression and localization of epidermal growth factor, transforming growth factor-␣, and localization of their common receptor in fetal human lung development. Pediatr. Res. 39:448–455. 18. Strandjord, T. P., J. G. Clark, W. A. Hodson, R. A. Schmidt, and D. K. Madtes. 1993. Expression of transforming growth factor-␣ in mid-gestation human fetal lung. Am. J. Respir. Cell Mol. Biol. 8:266–272. 19. Klein, J. M., B. L. Fritz, T. A. McCarthy, C. L. Wohlford-Lenane, and J. M. Snyder. 1995. Localization of epidermal growth factor receptor in alveolar epithelium during human fetal lung development in vitro. Exp. Lung Res. 21:917–939. 20. Sibilia, M., and E. F. Wagner. 1995. Strain-dependent epithelial defects in mice lacking the EGF receptor. Science 269:234–238. 21. Miettinen, P. J., D. Warburton, D. Bu, J. S. Zhao, J. E. Berger, P. Minoo, T. Koivisto, L. Allen, L. Dobbs, Z. Werb, and R. Derynck. 1997. Impaired lung branching morphogenesis in the absence of functional EGF receptor. Dev. Biol. 186:224–236. 22. Chess, P. R., R. M. Ryan, and J. N. Finkelstein. 1994. Tyrosine kinase activity is necessary for growth factor-stimulated rabbit type II pneumocyte proliferation. Pediatr. Res. 36:481–486. 23. Seth, R., L. Shum, F. Wu, C. Wuenschell, F. L. Hall, H. C. Slavkin, and D. Warburton. 1993. Role of epidermal growth factor in branching morpho-

684

24. 25. 26. 27.

28. 29. 30. 31. 32. 33.


genesis of early mouse embryo lung in culture: antisense oligodeoxynucleotide inhibitory strategy. Dev. Biol. 158:555–559. Yasui, S., A. Nagai, A. Oohira, M. Iwashita, and K. Konno. 1993. Effects of anti-mouse EGF antiserum on prenatal lung development in fetal mice. Pediatr. Pulmonol. 15:251–256. Klein J. M., L. J. DeWild, and T. A. McCarthy. 1998. Effect of tyrosine kinase inhibition on surfactant protein A gene expression during human lung development. Am. J. Physiol. 274: L542–L551. Snyder, J. M., J. M. Johnston, and C. R. Mendelson. 1981. Differentiation of type II cells of human fetal lung in vitro. Cell Tissue Res. 220:17–25. Xu, Y. H., I. Shunsuke, A. J. L. Clark, M. Sullivan, R. K. Wilson, D. P. Ma, B. A. Roe, G. T. Merlino, and I. Pastan. 1984. Human epidermal growth factor receptor cDNA is homologous to a variety of RNAs overproduced in A431 carcinoma cells. Nature 309:806–810. Chomczynski, P., and N. Sacchi. 1987. Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156–159. Dekowski, S. A., and J. M. Snyder. 1992. Insulin regulation of messenger ribonucleic acid for the surfactant-associated proteins in human fetal lung in vitro. Endocrinology 131:669–676. Snyder, J. M., and C. R. Mendelson. 1987. Insulin inhibits the accumulation of the major lung surfactant apoprotein in human fetal lung explants maintained in vitro. Endocrinology. 120:1250–1257. Weaver, T. E., and J. A. Whitsett. 1991. Function and regulation of expression of pulmonary surfactant-associated proteins. Biochem. J. 273:249–264. Carpenter, G. 1984. Properties of the receptor for epidermal growth factor. Cell 37:357–358. Chen, W. S., C. S. Lazar, M. Poenie, R. Y. Tsien, G. N. Gill, and M. G.

34.

35. 36. 37. 38. 39.

40.

Rosenfeld. 1987. Requirement for intrinsic protein tyrosine kinase in the immediate and late actions of the EGF receptor. Nature 328:820–823. Potts, J. D., J. M. Dagle, J. A. Walder, D. L. Weeks, and R. B. Runyan. 1991. Epithelial-mesenchymal transformation of embryonic cardiac endothelial cells is inhibited by a modified antisense oligodeoxynucleotide to transforming growth factor ␤3. Proc. Natl. Acad. Sci. USA 88:1516–1520. Souza, P., A. K. Tanswell, and M. Post. 1996. Different roles for PDGF-␣ and -␤ receptors in embryonic lung development. Am. J. Respir. Cell Mol. Biol. 15:551–562. Raaberg, L., E. Nexø, P. E. Jørgensen, S. S. Poulsen, and M. Jakab. 1995. Fetal effects of epidermal growth factor deficiency induced in rats by autoantibodies against epidermal growth factor. Pediatr. Res. 37:175–181. Stein, C. A., and Y. C. Cheng. 1993. Antisense oligonucleotides as therapeutic agents—is the bullet really magical? Science 261:1004–1012. Yakubov, L., Z. Khaled, L. M. Zhang, A. Truneh, V. Vlassov, and C. A. Stein. 1993. Oligodeoxynucleotides interact with recombinant CD4 at multiple sites. J. Biol. Chem. 268:18818–18823. Guvakova, M. A., L. A. Yakubov, I. Vlodavsky, J. L. Tonkinson, and C. A. Stein. 1995. Phosphorothioate oligodeoxynucleotides bind to basic fibroblast growth factor, inhibit its binding to cell surface receptors, and remove it from low affinity binding sites on extracellular matrix. J. Biol. Chem. 270:2620–2627. Benimetskaya, L., J. L. Tonkinson, M. Koziolkiewicz, B. Karwowski, P. Guga, R. Zeltser, W. Stec, and C. A. Stein. 1995. Binding of phosphorothioate oligodeoxynucleotides to basic fibroblast growth factor, recombinant soluble CD4, laminin and fibronectin is P-chirality independent. Nucleic Acids Res. 23:4239–4245.