AML1-like Transcription Factor Induces Serine Elastase Activity in Ovine Pulmonary Artery Smooth Muscle Cells
Abstract—In previous studies, we showed that induction of pulmonary artery (PA) smooth muscle cell (SMC) elastase activity by serum-treated elastin (STE) requires DNA transcription. We therefore used differential mRNA display to identify transcripts expressed coincident with elastase induction. Twenty-four individual transcripts were differentially expressed from a screen of ≈2000 mRNA sequences. An mRNA with sequence homology to the human transcription factor AML1 was identified and subsequently cloned from ovine PA SMCs. Since AML1 binds to a consensus sequence in the promoter of neutrophil elastase, we pursued the possibility that AML1 is a candidate transcription factor for SMC elastase. We documented by immunohistochemistry that serum stimulation induces increased expression of AML1 in the nucleus of PA SMCs. We also showed that STE induction of elastase activity is associated with early expression of AML1 mRNA and protein and that AML1 consensus sequence DNA binding activity is increased in nuclear extracts of STE-treated cells. In addition, AML1 antisense oligonucleotides reduced serum induction of elastase activity. Our study thus provides the first functional evidence of AML1 transcriptional activity related to elastase genes and offers novel insights into the broader biological significance of AML1 in nonmyeloid cells.
Ultrastructural observations in lung biopsy tissue from patients with congenital heart defects showing fragmentation of the arterial elastic lamina first suggested that heightened elastolytic activity may play a critical role in the development of pulmonary vascular disease.1 Further studies in experimental animals confirmed high elastin turnover in the pulmonary artery as a feature of monocrotaline-induced progressive pulmonary hypertension.2 Increased vascular elastolytic activity was then documented as preceding the development of sustained pulmonary hypertension and the induction of pulmonary artery hypertrophy associated with hypoxia3 or with the injection of the toxin monocrotaline.4 A cause-and-effect relationship between increased vascular elastolytic activity and the initiation of increased pulmonary pressure with its concomitant structural remodeling was suggested when elastase inhibitors were shown to prevent the development of experimental pulmonary hypertension.3 5 6 We subsequently identified, in hypertensive pulmonary arteries, a 20-kDa serine elastase that localized to SMCs and appeared to be related to the serine proteinase adipsin.7 These studies supported previous reports describing the production of a similar enzyme by aortic SMCs.8 9
Since endothelial injury is an early feature of both pulmonary10 and systemic vascular11 diseases, we further investigated whether the mechanism accounting for the induction of EVE might be related to transfer of a serum factor to the medial SMC layer as a consequence of loss of endothelial cell barrier function. We demonstrated that incubation of PA SMCs with either serum12 or STE13 resulted in induced elastolytic activity. This enzyme has the inhibitor profile of a serine elastase, and its activity appears to require the induction of a tyrosine kinase intracellular signaling pathway.12 Furthermore, similar to neutrophil elastase, this EVE releases potent mitogenic growth factors, such as basic fibroblast growth factor, from the extracellular matrix, thus establishing a mechanistic link between heightened elastase activity and the SMC proliferative response observed with pulmonary vascular hypertrophy.14 As a result of elastin degradation, the generation of elastin peptides might also stimulate SMC migration through the induction of fibronectin, linking elastase activity with neointimal formation and pulmonary vascular disease.15
Further experiments demonstrating that the induction of SMC elastase activity is inhibited by actinomycin D, as well as cycloheximide, implied transcription and translation of the elastase gene and/or expression of other gene products related to its induction or processing.13 The following investigation used a differential mRNA display to detect transcripts upregulated coincident with elastase induction. We identified one of these transcripts (227 bp) as having homology (90% over a span of 120 bp) to human AML1 and another as having a similar 120- to 126-bp region of 88% to 89% homology to human lymphotoxin–TNF-α and IL-1α.
AML1 belongs to a family of transcription factors highly conserved from Drosophila to human species.16 It was originally identified as the gene altered by the t(8;21) translocation commonly occurring in acute myelogenous leukemia, juxtaposing AML1 with the ETO gene on chromosome 8.17 18 AML1 is a putative transcription factor for the neutrophil elastase gene that contains the appropriate consensus sequence in its promoter.19 Because of the provocative potential association between the transcription of EVE and AML1, we undertook to investigate its presence and functional significance in our system. In the present study, we used 5′ and 3′ RACE to confirm the cDNA sequence of the ovine AML1. We documented that there is increased expression of AML1 in the nucleus after serum stimulation by using Northern blot and immunocytochemistry. By Western immunoblot, we showed that AML1 expression is temporally related to EVE activity. Gel shift analysis was used to document AML1 DNA binding activity in SMC nuclear extracts, and antisense oligonucleotide experiments supported a role for AML1 in inducing EVE activity. The present study provides new insight into the biological significance of AML1 and the regulation of SMC elastase activity.
Materials and Methods
PA SMC Culture
PA SMCs were isolated from the vessels of 100-day-old fetal lambs and from juvenile pigs, as previously described.12 Briefly, the central pulmonary arteries were removed and rinsed in PBS supplemented with 300 U/mL penicillin, 300 μg/mL streptomycin, and 750 ng/mL amphotericin B (Fungizone, Life Technologies). The endothelial layer of each artery was removed by scraping the luminal surface with a scalpel, and the adventitia was separated from the medial layer. The media was further dissected into small explants ≈3×3 mm in diameter, which showed outgrowth of SMCs after a few days in culture. Cells were cultured in 100-mm dishes using medium 199 (Ontario Cancer Institute) with 10% FBS (Intergen) and 100 U/mL penicillin, 100 μg/mL streptomycin, and 250 ng/mL amphotericin B in a humidified 5% CO2 environment at 37°C. Cell cultures were identified as pure SMCs by uniform immunostaining with an antibody to smooth muscle α-actin and were used for experimentation at passage 2 or 3.
Measurement of Elastolytic Activity
Elastase assays were performed as previously described.12 Briefly, purified elastin from bovine nuchal ligament (Elastin Products Co) was radiolabeled using [3H]NaBH4 (Dupont/NEN) as described by Takahashi et al.20 [3H]Elastin was reconstituted at 16 mg/mL (specific activity, 2000 cpm/μg elastin) in Tris assay buffer (50 mmol/L Tris-HCl, 150 mmol/L NaCl, 10 mmol/L CaCl2, and 0.02% Brij [pH 8.0]) and stored at −20°C until assay. Before use, [3H]elastin was washed until the background (nonspecific counts released) was <100 cpm/100 μL supernatant. The [3H]elastin suspension was then diluted to 100 000 cpm/20 μL.
To assay both porcine and ovine SMC elastase activity, cells were passaged into 24-well plates (16-mm diameter) at a density of ≈105 cells per well. At confluence, cells were washed with PBS and placed in SFM 199 containing 0.1% BSA. After 24 hours of serum starvation, cells were washed with PBS and incubated with 20 μL [3H]elastin (105 cpm) in SFM or with 5% FBS for a further 24-hour period. A 600 μL aliquot of culture medium was then removed from each dish and centrifuged for 5 minutes at 10 000g to pellet insoluble elastin. The resulting supernatant (500 μL) was counted to estimate the release of soluble elastin fragments. Counts resulting from elastin degradation in cell-free wells were subtracted from each data point. All assays were performed in triplicate. Elastase assays were repeated 3 times with cells from 3 different porcine PA SMC harvests.
Generation of STE
Pretreating the elastin with FBS has been shown to induce elastin adhesion to SMC surfaces and EVE activity.13 Insoluble elastin was incubated with FBS at a concentration of 10 mg/mL, rotated overnight at 37°C, and then washed 4 times with Tris assay buffer to remove all unbound serum factors. The nonradiolabeled STE was resuspended and incubated with SMCs at a final concentration of 0.5 mg/mL, a concentration equal to that used in the elastase assay.
Differential mRNA Display
Differential mRNA display was carried out as previously described21 using a kit from Genhunter Corp. Total cellular RNA was prepared by guanidinium and phenol-chloroform extraction22 from 150-mm culture dishes of ovine PA SMCs incubated with either SFM or nonradiolabeled STE for 12 hours after an initial 24-hour period of serum starvation. RNA samples were treated with DNase for 15 minutes at 37°C to remove traces of genomic DNA. Reverse transcription of RNA was performed using poly-T12MN (T12MA, T12MC, T12MG, and T12MT). The resulting cDNA was amplified by PCR in the presence of [35S]dATP and Taq DNA polymerase (Pharmacia) on a Robocycler (Stratagene ) using arbitrary decamers as 5′ primers and the corresponding T12MN 3′ primer. Samples were denatured at 94°C for 35 seconds, annealed at 40°C for 120 seconds, and extended at 72°C for 35 seconds, for a total of 40 cycles. Radiolabeled PCR products were electrophoresed on a denaturing 6% polyacrylamide sequencing gel. Selected bands differentially upregulated in response to STE were recovered from sequencing gels, reamplified in a further 40-cycle PCR in the absence of isotope, and electrophoresed on 1.5% low-melting-point agarose gels. PCR products were then removed from the gel and purified using β-agarase 1 (New England Biolabs). Briefly, β-agarase 1 buffer was added to each gel fragment, and samples were heated to 70°C for 10 minutes until melted. After cooling to 40°C, 1.5 U β-agarase 1 per 100 mL of sample was incubated for 1 hour. Samples were then cooled on ice for 10 minutes and centrifuged at 14 000g for 5 minutes to pellet insoluble material. The remaining supernatant was precipitated overnight in 2.5 vol ethanol and 0.3 mol/L sodium acetate, dried, and resuspended in 20 μL TE buffer (10 mmol/L Tris-Cl [pH 7.6] and 1 mmol/L EDTA [pH 8.0]).
For dot-blot screening, we used a small aliquot (2 μL) of each individually amplified PCR product, which was diluted in 100 μL H2O, made up to a final concentration of 5× SSC, denatured for 10 minutes at 95°C, and then applied to nylon filters (Hybond N, Amersham) using a filtration manifold (Bio-Rad). Both the membrane and the underlying filter paper were soaked in 5× SSC for 15 minutes before assembly of the manifold. Each well was washed once with 5× SSC, and PCR fragments were applied under vacuum. After disassembly of the manifold, membranes were denatured for 10 minutes on filter paper soaked in 1.5 mol/L NaCl and 0.5 mol/L NaOH and then neutralized for 5 minutes on filter paper soaked in 1 mol/L NaCl and 0.5 mol/L Tris-HCl (pH 7.0). Membranes were dried fully and UV-irradiated to immobilize the DNA.
To check for multiple representation, membranes were screened with individually reamplified PCR fragments labeled with [32P]dCTP by random priming (Amersham). Labeling reactions contained 1 mL of the 20-mL reamplification mix. After 1 hour of prehybridization, the PCR fragments were hybridized to the membrane in a solution containing 5× SSC, 50% deionized formamide, 2% SDS, and 100 μg/mL salmon sperm DNA. Membranes were stripped for reprobing in 10 mmol/L Tris-HCl and 5 mmol/L EDTA at 65°C for a series of four 30-minute washes.
Cloning and Sequence Analysis
PCR products identified by dot-blot screening as unique sequences were again reamplified using the original sample eluted from the differential display gel as a template, electrophoresed on 1.5% low-melting-point agarose gels, and purified with β-agarase 1 as described above. Purified PCR products were subcloned into pBSK (Stratagene), by use of TA overhang cloning as previously described,23 and positive clones were selected by blue/white screening. Insert sequences were obtained using dideoxy sequencing with fluorescein-labeled primers (T7 and T3) on a Pharmacia A.L.F. automated sequencer. Nucleotide sequences were compared with known sequences by searching GenBank and EMBL databases with BLAST software.24
Northern Blotting and Hybridization
To confirm differential expression of mRNA sequences, Northern blotting was performed using either total RNA (for all nucleotide sequences screened as described above) or purified polyA+ mRNA (for selected sequences as detailed in Results). Total RNA was prepared according to Chirgwin et al,25 and polyA+ mRNA was isolated by use of a kit from Invitrogen. Total RNA (20 μg) or polyA+ RNA (5 μg) was electrophoresed on 1% agarose gels and transferred to nylon membranes (Hybond N) by using standard techniques.26 Radiolabeling of PCR products or cloned cDNA fragments and hybridization were performed as described above for dot blots. Washes were performed at 50°C (3 times for 20 minutes each in 1× SSC/0.1% SDS followed by 3 times for 20 minutes each in 0.1× SSC/0.1% SDS), and blots were exposed to Kodak X-Omat film. RNA loading was assessed by hybridizing the membrane with [32P]dCTP-labeled GAPDH cDNA fragment. Membranes were stripped and reprobed as described above.
mRNA Preparation and Generation of Ovine PA SMC AML1 cDNA
Confluent PA SMCs in 150-mm culture dishes were serum-starved for 24 hours, followed by 12 hours of stimulation with STE. TRIzol Reagent (Life Technologies) was used to isolate RNA from the SMCs. The RNA was reverse-transcribed using an adapter-oligo dT primer, 5′-GACTCGAGTCGACATCGAT11-3′. This cDNA was used in 3′ RACE with a forward primer, 5′-GGTTTCTGTTGTGTTTAATTTC-3′, designed from the F12 sequence. The reverse primer used was a primer to the adapter oligo dT, 5′-GACTCGAGTCGACATCGA-3′. The reaction was as follows: 1 minute at 94°C followed by 39 cycles of denaturation at 94°C for 45 seconds, annealing at 55°C for 2 minutes, and extension at 72°C for 2 minutes, followed by extension at 72°C for an additional 10 minutes. The resulting PCR products were used in a nested PCR using the forward nested primer, 5′-ATTTCTCTACAGATTGTATTGT-3′, from the F12 sequence. The reverse primer used was the same as described above. The PCR produced a single product of ≈600 bp, which was then subcloned into pCR2.1 (Invitrogen) plasmid (pSY3) and sequenced on a Pharmacia A.L.F. automated DNA sequencer.
In addition, 5′ RACE was carried out. A poly dA extension of the 5′ end of the above reverse-transcribed cDNA was performed using dATP and terminal deoxynucleotidyl transferase. The PCR reaction used the forward primer 5′-GACTCGAGTCGACATCGAT11-3′ and the reverse primer 5′-CATATACATATGCTCTACTTCA-3′, which was from the F12 sequence. The resulting PCR products were used in a nested PCR using the forward primer to the adapter oligo dT, described above, and a nested reverse primer, 5′-ACAATACAATCTGTAGAGAAAT-3′, which was also from the F12 sequence. A single product of ≈200 bp was resolved on agarose gel, purified, subcloned, and sequenced as described above.
Northern Blot Using Ovine AML1 cDNA
For Northern blot analyses using the ovine AML1 cDNA, ovine pulmonary artery cells were serum-starved and stimulated with STE, and total RNA was isolated using TRIzol solution. RNA was then electrophoresed on formaldehyde agarose gels and transferred to Hybond N (Amersham) membranes. Ovine AML1 cDNA 3′ RACE product (a 577-bp sequence) was excised from pSY3 and radiolabeled as described above. Hybridization was performed using QuikHyb (Stratagene) solution for 1 hour. Membranes were washed twice for 15 minutes at room temperature with a solution containing 2× SSC and 0.1% SDS. High-stringency washes were carried out in 0.1× SSC and 0.1% SDS at 60°C.
For immunohistochemistry, porcine PA SMCs were plated at 1 to 5×105 cells/mL onto 20×20-mm glass coverslips and grown for 48 hours in medium 199 with 10% FBS. Cells were then washed twice in warm PBS, incubated in SFM with 0.1% BSA for 24 hours of serum starvation, and then either harvested or incubated in medium 199 with 10% FBS for 0, 5, 15, and 30 minutes. Cells were washed twice with warm PBS and then fixed with methanol:acetic acid (2:1) for 5 minutes at −20°C. The coverslips were dehydrated and stored at 4°C until use. Rehydration was performed with three 10-minute washes in PBS and then one 5-minute wash in PBS+1% BSA (P-BSA). Blocking was performed for 1 hour at room temperature with 10% goat serum in P-BSA, and the primary antibody, a polyclonal antibody produced using a 17-amino-acid N-terminal peptide of AML1 (N-Arg-Ile-Pro-Val-Asp-Ala-Ser-Thr-Ser-Arg-Arg-Phe-Thr-Pro-Pro-Ser-C, a gift of S. Myers and S. Heibert, St. Jude’s Hospital, Memphis, Tenn) was diluted 1:100 in P-BSA for overnight incubation at 4°C. Coverslips were then washed 3 times for 5 minutes each with P-BSA and incubated for 1 hour at room temperature with the goat anti-rabbit FITC-conjugated secondary antibody. One rinse was performed for 5 minutes in P-BSA, followed by two 10-minute washes in PBS. Coverslips were mounted in Elvanol and viewed by epifluorescence using appropriate wavelengths. IgG primary antibody was used as a negative control.
Confluent monolayers of porcine pulmonary arteries and lamb SMCs in 100-mm dishes were washed twice with SFM, serum-starved for 24 hours, and then exposed to STE for varying time periods (0, 1, 5, 10, and 30 minutes and 1, 3, 6, and 12 hours). Cells were washed twice with cold PBS and extracted in hot 2% SDS using a rubber policeman. Elastin fragments in the suspension were removed by brief centrifugation. Samples were normalized for total protein content (Bio-Rad), electrophoresed on either 12% or 14% polyacrylamide Tris-glycine gels, and electroblotted to Immobilon-P membranes (Millipore) at 30 V for 1.5 hours. Nonspecific binding was blocked by incubating the membrane overnight in 5% nonfat milk in PBS with 0.5% Tween 20. This was followed by incubation with the polyclonal AML1 antibody (1: 200), as described above, and then goat anti-rabbit secondary antibody (1:5000) (Bio-Rad) and detected by enhanced chemiluminescence (Amersham) when exposed to Kodak X-Omat film. The intensity of the resulting bands was quantified by scanning soft-laser densitometry. Western immunoblotting was also carried out to assess expression of AML1 after incubation of the control and serum-treated SMCs with antisense, sense, and scrambled oligonucleotides, as described below.
Nuclear Extract Preparation
Confluent serum-starved porcine SMCs were treated with either STE or SFM (control) for 30 minutes, and enriched nuclear and cytoplasmic extracts were prepared as previously described.27 Briefly, after removing medium, cells were rinsed in cold PBS and scraped into microcentrifuge tubes, and the pellets were washed in PBS. Cells were lysed in cold hypotonic buffer (10 mmol/L HEPES [pH 7.9], 1.5 mmol/L MgCl2, 19 mmol/L KCl, 0.2 mmol/L phenylmethylsulfonyl fluoride, and 0.5 mmol/L DTT) followed by 10 passes of a type B Dounce homogenizer (VWR Scientific). Nuclei were pelleted at 3300g for 15 minutes at 4°C. The pelleted nuclear proteins were further extracted with a high salt solution (20 mmol/L HEPES [pH 7.9], 25% glycerol, 1.5 mmol/L MgCl2, 0.8 mol/L NaCl, 0.2 mmol/L EDTA, 0.2 mmol/L phenylmethylsulfonyl fluoride, and 0.5 mmol/L DTT) for 30 minutes and then microcentrifuged at 16 000g for 30 minutes. The supernatants were divided into aliquots, rapidly frozen in liquid nitrogen, and stored at −70°C. The remaining pellet did not contain any AML1 immunoreactive proteins, as determined by Western immunoblotting. The nuclear extracts were then used for Western immunoblotting with the AML1 antibody or for the EMSA.
Electrophoretic Mobility Shift Assay
The protein-AML1 DNA binding site interaction was analyzed by EMSA according to the protocol described by Meyers et al.16 The wild-type AML1 binding site double-stranded DNA oligonucleotides were prepared by annealing the complementary oligomers 5′-AATTCGAGTATTGTGGTTAATACG-3′ and 5′-AATTCG-TATTAACCACAATACTCG-3′. The oligomers 5′-AATTCGAG-TATTGTTAGTAATACG-3′ and 5′-AATTCGTATTACTAA-CAATACTCG-3′ were annealed to form the AML1 mutant binding site oligonucleotides. Wild-type and mutant oligonucleotides were labeled with [α-32P]dATP by use of the Klenow fragment and DNA polymerase. Briefly, the AML1 binding-site reaction contained 10 μg total protein from nuclear extracts and 1 ng of radiolabeled oligonucleotides in binding buffer (20 mmol/L HEPES [pH 7.8], 1 mmol/L MgCl2, 0.1 mmol/L EGTA, 0.4 mmol/L DTT, 40 mmol/L KCl, 10% glycerol, and 60 μg/mL salmon sperm DNA) and proceeded for 30 minutes at room temperature. In competition studies, 50 ng of unlabeled double-stranded wild-type or mutant oligonucleotides was preincubated for 30 minutes with the nuclear extracts from STE-treated cells before the addition of the radiolabeled probe. The reactions were resolved on a 5% polyacrylamide nondenaturing gel using TBE buffer (50 mmol/L Tris-borate and 1.0 mmol/L EDTA [pH 8.0]), and the gel was dried and exposed to Kodak X-Omat film. To confirm that the protein in the nuclear extract was AML1, the nuclear extracts from STE-treated cells were preincubated with AML1 antibody (1 and 5 μL nonimmunopurified antisera) before addition of the radiolabeled oligonucleotides. The AML1 N-terminal peptide was also preincubated with the AML1 antibody to further show the specificity of the AML1:DNA gel shift complex.
AML1 Antisense Experiments
Phosphorothioate oligonucleotides designed for antisense experiments were directed to the homologous translation start site of the human and mouse AML1 genes and included sense 5′-ATGCG-TATCCCCGTAGAT-3′, antisense 5′-ATCTACGGGGATACG-CAT-3′, and scrambled 5′-CAGGTCATTGACAGCATG-3′ sequences. The scrambled oligonucleotide contained similar base composition but a sequence different from antisense. Oligonucleotides were checked in GenBank for potential interference with the translation of other mRNA species. For antisense experiments, cells were plated in 24-well 16-mm dishes and grown until confluence. Cells were then serum-starved for 24 hours in SFM. Initial dose-response experiments using 1, 3, 5, and 10 μmol/L oligonucleotide concentrations determined the 3 μmol/L dose to give an appropriate inhibition of elastolytic activity with the addition of the antisense oligonucleotide. Preincubations of oligonucleotides for 24 hours, 2 hours, and 0 hours (addition at time of serum stimulation [0 time point]) demonstrated that the latter was sufficient to show the effect of antisense oligonucleotide on elastase induction. Elastase assays were performed, as described above, after 24 hours of incubation with oligonucleotides. These experiments were carried out in triplicate and repeated with porcine cells from 3 different harvests.
Analysis of Data
Statistical analysis was performed on data from elastase assays using a 1-way ANOVA followed by post hoc comparison of individual groups using the Fisher protected least significant difference test. A value of P<0.05 was considered statistically significant.
Differential Display After STE Induction of Elastolytic Activity
Table 1⇓ summarizes the dot-blot analysis of the 54 differential display fragments identified as potentially upregulated genes from a screen of ≈2000 representative mRNA transcripts expressed in PA SMCs after treatment with STE. A total of 46 fragments were chosen from the identified 54 for technical reasons (eg, retrievability from the filter) and because the extent of differential expression was more marked. Of these 46 fragments, 41 were successfully reamplified. Only 26 (63.4%) of 41 represented unique sequences. The number of fragments represented in triplicate and duplicate is indicated in Table 1⇓. In Figure 1⇓, some of the sequences listed in Table 2⇓ are identified on a representative portion of the gel to show how they were chosen. Although it has been suggested that fragments <200 bp are more likely to be duplicated, we found unique sequences as small as 181 bp and a number of duplicated sequences >200 bp. Two products were lost during the second round of PCR amplification, ie, after dot blot screening and before subcloning.
The majority of subcloned PCR products were sequenced at least twice from different clones, producing in one case (F11, Table 2⇑) more than one sequence per fragment. Table 2⇑ lists the 25 different sequences obtained and the corresponding Northern blot results. As would be expected from an analysis of ovine 3′ sequences, the majority of genes isolated showed no homology to sequences in the GenBank. However, a number of fragments did show homology to sequence tags generated from both human fetal and adult brain cDNA libraries. The contaminating viral sequence reverse transcriptase from Bos taurus (F25) was presumably isolated from the FBS used to make STE. One 309-bp sequence (F24) had an interesting stretch of homology to the cytokine lymphotoxin–TNF-α (88% homology over 121 bp) and IL-1α (89% homology over 126 bp) (Figure 2⇓). A 71-bp region is 89% homologous to IL-1β. Another 227-bp sequence (F12) had a stretch of 123 bp with 90% homology to the transcription factor AML1 (Figure 3A⇓), which is thought to regulate myeloid differentiation via transcriptional control of genes such as myeloperoxidase and neutrophil elastase.19 Although there is a DNA recognition sequence for AML1 on the neutrophil elastase promoter, its functional activity vis-à-vis control of neutrophil elastase transcription has not been documented, nor has AML1 been previously described in vascular smooth muscle.
The following 3 gene sequences encoding known proteins were identified. RIP 140 (F7) is a recently identified nuclear protein involved in the modulation of transcriptional activation by the estrogen receptor,28 but it has not previously been described in vascular SMCs. S-Adenosyl-l-homocysteine hydrolase (F54) is involved in adenosine metabolism. It is upregulated with cardiac hypoxia and has been localized to cardiac myocytes.29 Spermine/spermidine N-acetyltransferase (F11a) is an enzyme important in polyamine synthesis. It is increased before the monocrotaline-induced PA SMC hypertrophic response.30 The relationship between this enzyme and the monocrotaline-induced increase in pulmonary artery elastase activity, however, is not known.4 Although some of the clones with homology to known genes did not have an identifiable polyA tail within the insert sequence, all displayed alignment to the 3′ UTR of the homologous gene or expressed sequence tag. The noted exception to this observation was the alignment of F12 to the 5′ UTR of the AML1 gene, and this will be discussed (Figure 3A⇑).
Northern Blot Analysis
The majority of the sequences that appeared to be upregulated by reverse-transcription PCR did not hybridize to any mRNA species detectable by Northern blot analysis using 20 μg total RNA from PA SMCs. With other sequence tags that did show hybridization on Northern blot, we could not confirm an appreciable change in mRNA expression after STE treatment. These transcripts were not further pursued. F24, a 309-bp fragment with an interesting region of homology to cytokines, hybridized to a 1.1-kb mRNA species and did appear to be upregulated after STE treatment by ≈67% (data not shown). Three other sequences identified as having homology to known genes of potential interest were analyzed further by Northern blots using polyA+ mRNA (5 μg). The sequence F7, 90% homologous to human RIP 140 (a 7.5-kb transcript in the human breast cancer cell line ZR75-1),29 hybridized to an ≈7-kb transcript on Northern blots of ovine PA SMC RNA but was not upregulated with STE. Two of the other sequences, F11a (with 88% homology to the human spermine N-acetyltransferase species) and F54 (with 78% homology to the human S-adenosylhomocysteine hydrolase), were confirmed on Northern blot by hybridization to mRNAs of appropriate sizes but, like RIP 140, were not upregulated with STE.
It was uncertain whether the sequence tag with homology to human AML1 actually encodes this gene product in the ovine species. There was only a 123/227-bp stretch of high homology, and the A-rich region in the 3′ end of the transcript was not observed in the corresponding human AML1 cDNA. However, the potential functional significance of AML1 in our system led us to pursue studies to document its presence in ovine and porcine SMCs and its relationship to EVE activity. AML1 has been previously identified only in mammalian myeloid cells, so studies were carried out to sequence the ovine AML1 cDNA.
With a strategy using primers derived from the F12 sequence and 5′ and 3′ RACE, partial cDNA fragments of 211 and 577 bp were obtained. Sequencing of these products resulted in a total sequence of 767 bp with 70% overall sequence homology to human AML1 (Figure 3⇑). The nonhomologous sequence in the original differential display product was not present when this strategy was used and may represent an artifact.
Northern Blot Analysis of AML1 mRNA
Northern blots were then carried out with RNA extracted from ovine cells under serum-free conditions and at 30 minutes and 12 hours after STE. Two transcripts recognize the cDNA (Figure 4⇓). One (designated A on Figure 4⇓) corresponds to a transcript of ≈3.2 kb, the size of AML1 (also known as AML1A), which encodes a protein of ≈33 kDa, and the other (designated B) corresponds to a transcript of ≈7.4 kb, the size of the isoform AML1B, encoding a protein of ≈50 kDa. The steady-state mRNA levels are increased for AML1A and AML1B. Using either 18S and 28S or GAPDH to control for loading conditions, we observed a 1.5- to 2-fold increase in AML1B steady-state mRNA levels at 30 minutes and 12 hours after STE, whereas a 3-fold increase in AML1A is seen both at 30 minutes and at 12 hours after STE treatment.
Nuclear Localization of AML1 on STE Stimulation of Cells
Since there is evidence of an ovine AML1 transcript in vascular SMCs, we carried out further studies to assess the presence of the AML1 protein in the nucleus. We compared serum-starved and serum-stimulated ovine and porcine cells,12 because analysis of STE-treated cells was impaired by autofluorescence of adherent elastin. As shown in Figure 5⇓, in ovine cells, within 5 minutes of serum treatment, the cells displayed a bright nuclear staining pattern, which was persistent at 30 minutes. Similar results were obtained for porcine cells.
STE-Induced Expression of AML1 Immunoreactive Proteins
We next determined by Western immunoblot whether we could confirm an increase in AML1 protein expression after STE and the time course over which this occurred. Two major immunoreactive proteins at 33 and 50 kDa were detected in porcine PA SMC lysates (Figure 6⇓). Previous studies have attributed these molecular weights to AML1A and AML1B isoforms, respectively.18 Treatment with STE resulted in a rapid and marked elevation of the 33-kDa immunoreactive protein between 1 and 5 minutes. This induction lasted up to 1 hour, but values were similar to those at baseline by 3 to 6 hours. The 50-kDa protein also appeared to be increased transiently by STE, although the difference was less than that observed with the 33-kDa immunoreactive protein.
The relative amounts of the 2 isoforms and the increase with STE at 30 minutes corresponds to the mRNA levels. The persistent elevation in AML1 mRNA levels for both transcripts at 12 hours does not correspond to the drop in protein already evident at 6 hours, and this may reflect high turnover or degradation of the protein or reduced mRNA translational efficiency.
STE Stimulates the AML1 Consensus DNA Binding Activity
Electrophoretic mobility shift assays were therefore carried out to confirm that AML1 in the nuclear extract of PA SMCs was of functional significance in that it could form a complex with its cognate DNA consensus binding sequence. Nuclear extracts from cells treated with SFM (control) or STE for 30 minutes were incubated with a radiolabeled 24-bp double-stranded oligonucleotide containing the AML1 DNA binding site, as described in Materials and Methods, and run on a nondenaturing gel. Figure 7⇓ is an autoradiograph that demonstrates at least one slower migrating protein:DNA complex (C2) that is formed in contrast to the free DNA probe. After STE treatment, there is an increase in C2 and an additional well-resolved C1 complex formed compared with control conditions. In competition experiments, a 1-fold (not shown), as well as a 50-fold, increase in unlabeled wild-type oligonucleotide, but not the 50-fold mutant oligonucleotide, resulted in greater inhibition of complex formation using STE nuclear extracts. In addition, preincubation of increasing concentrations of AML1 antibody (1 and 5 μL of antisera) with the STE-treated nuclear extracts generated a supershifted band with a corresponding reduction in the previously observed AML1 protein:DNA complexes. The bands were not shifted with IgG (data not shown). Moreover, preincubation of the AML1 N-terminal peptide (20 μL) with the antibody (5 μL) before incubation with the nuclear extract and DNA oligonucleotide prevented the supershift. These features confirm the specificity of the interaction between DNA and AML1 extracted from PA SMC nuclear extracts.
AML1 Antisense Oligonucleotide Inhibition of Elastolytic Activity
To further determine whether AML1 functioned as a vascular elastase transcription factor, porcine PA SMCs from 3 different harvests were treated with AML1 antisense oligonucleotides. This attenuated elastase induction by 5% FBS observed in all 3 experiments implicated the specific involvement of AML1 (P<0. 05) (Figure 8A⇓), whereas a scrambled oligonucleotide of similar composition and a sense oligonucleotide did not reduce serum induction of elastase activity. Although values for elastase activity appear above basal levels with antisense oligonucleotides, there was no significant difference in elastase activity. The reduction in serum-induced elastase activity with antisense was accompanied by a comparable reduction on Western immunoblot in the 33 kDa, as well as the 50 kDa, AML1 immunoreactive proteins (Figure 8B⇓).
We applied mRNA fingerprinting methodology through differential display of PCR products in an attempt to identify genes transcribed in association with production of EVE in PA SMCs. The cells were induced to increase elastase activity with STE, with the expectation being that this was a more selective stimulus than serum that also induces elastase activity. The objective of these studies was to uncover the vascular elastase mRNA or other mRNA species in which increases in steady-state levels were necessary for the transcription, translation, processing, and secretion of the enzyme.
Our studies have identified a 227-bp PCR differential display product that showed homology to the human transcription factor AML1 over a continuous stretch of the 120-bp sequence. Since it was likely that the differential display sequence represented only part of the ovine AML1 sequence, we used 5′ and 3′ RACE to confirm that we could identify an ovine AML1 homologue in our cells. Having done so, we next showed increased expression of AML1A and AML1B mRNA transcript levels in our cells at 30 minutes and 12 hours after STE stimulation. We next addressed whether expression of AML1 protein correlated with induction of elastase after STE. We showed a transient increase in expression of AML1 protein by immunohistochemistry using serum stimulation and by Western immunoblot using STE induction. The protein levels appeared to correspond to the mRNA levels for the AML1 transcripts at 30 minutes but were reduced by 6 hours. It is possible that AML1 protein is rapidly degraded or that the mRNA is inefficiently translated after 1 hour. AML1 in nuclear extracts from PA SMCs after 30-minute stimulation with STE compared with nonstimulated cells showed increased binding complex formation, with its DNA consensus sequence identified on the neutrophil elastase promoter. Using antisense oligonucleotides, we demonstrated that inhibition of AML1 translation reduced the subsequent induction of elastase activity after serum or STE in a manner that correlated with decreased expression of the protein by Western immunoblot.
A potentially novel gene (F24) with homology to the 3′ UTR of cytokines that is upregulated (as confirmed by Northern blot analysis) after stimulation of elastase activity was also identified (data not shown). A second sequence tag encoded a gene product, RIP-140, involved in the modulation of transcriptional activation by the estrogen receptor. Although not upregulated with STE, it was not previously known to be expressed by SMCs, so this observation may prove to be important in the study of estrogens and vascular pathobiology. Two other sequence tags encoded interesting transcripts: S-adenosylhomocysteine hydrolase and spermine N-acetyltransferase. Both may play a role in the smooth muscle proliferation and differentiation associated with pulmonary hypertension and induction of elastase activity, but neither could be confirmed as upregulated with STE on Northern blot.
The technique of differential mRNA display has rapidly gained acceptance as a powerful tool for the identification and cloning of differentially expressed genes in a variety of biological systems. The method offers a number of technical advantages over other existing techniques for the investigation of differential gene expression. It is fast and relatively easy to perform and does not require large initial quantities on RNA. Because even semiquantification of PCR products requires stringent internal standards to control for primer attachment and amplification efficiency, the technique will lead to the sequencing of a large number of “false-positives” in that the increase in mRNA level cannot be confirmed by Northern blot. Despite this, differential mRNA display has proven useful when applied to a myriad of investigations, including those directed at the cardiovascular system. For example, the system was successfully used to identify estrogen-regulated genes in human vascular SMCs,31 glucose-induced genes in bovine aortic SMCs,32 and genes expressed after allograft cardiac transplantation in the rat33 and after experimental carotid artery balloon angioplasty.34 We now report 2 differentially displayed sequences that appeared to be upregulated by Northern blot in conjunction with increased elastolytic activity. The first of these genes, F24, contained a rather large and interesting region of homology to cytokines. It is possible that this sequence may represent a novel cytokine or, alternatively, an ovine cytokine sequence.
The second differential display fragment revealed a striking stretch of 120-bp homology with human AML1. The fragment amplified by PCR was, however, not a 3′ sequence (as predicted by the experimental design) but a 5′ UTR sequence. It was therefore essential to sequence ovine AML1 to establish its precise relationship to the transcript identified by differential display. Perhaps only fortuitous homology to AML1, a transcription factor known to bind to the neutrophil elastase promoter in immature myeloid cells, led to the investigation of AML1 in our system.
The AML1 family of transcription factors is highly homologous to the Drosophila segmentation gene runt.35 This homology suggests that AML1 is likely to be of significance in the differentiation of a variety of cell types. Two human genes have been isolated to date, including AML1A and AML1B.18 By interacting with a β component, these α proteins enhance their ability to bind DNA without changing DNA contact points.36 37 The AML1 mutation occurring in acute myelogenous leukemia, involving the t(8;21) translocation juxtaposing AML1 with the ETO gene on chromosome 8 (Reference 1818 ), generates a chimeric protein product. This product retains the known DNA binding activity of AML1 to the enhancer motif TGTGGT and the region of homology to the Drosophila segmentation gene runt but loses its potential to transactivate.
There was evidence of increased AML1 in the nucleus after serum stimulation, and we have shown that MAP kinase kinase activity manifest by phosphorylation of extracellular signal–regulated kinase-1 is necessary for both increased expression of AML1 and induction of elastase activity.38 To demonstrate the functional activity of AML1, we used oligonucleotides encoding the consensus sequence in the neutrophil elastase promoter in gel mobility shift assays. Since we have not yet cloned the gene for EVE (the PA SMC serine elastase), we can only speculate that the same promoter element is present. Although the antisense experiments showed a reduction in the level of elastase activity and thus further support AML1 as being of functional significance in its induction, it is also possible that AML1 is the transcription factor for other genes required for the processing of this enzyme. The levels of enzymatic activity observed with antisense were still higher than those under control conditions, although the difference was not statistically significant. This is reflected in the lower protein level on Western immunoblot and suggests incomplete suppression of mRNA translation. This is likely a function of incomplete oligonucleotide incorporation or stability. This represents the limitation of these experiments, since higher concentrations of oligonucleotides could be expected to result in cellular toxicity.
It would be of further interest to determine whether there are other genes upregulated by AML1 in vascular tissue either in development or in disease states. The fact that mice lacking either AML1 or its corresponding β subunit of the AML1-CBF β complex die with massive hemorrhage into the central nervous system, peritoneal cavity, and pleural space may imply a developmental vascular abnormality in addition to the hematopoietic defects noted.39 40
In summary, the present study used combined Western immunoblot and immunohistochemistry to elucidate the nuclear localization of AML1 and to establish a temporal association between induction of AML1 protein and STE stimulation of PA SMC elastase activity. We also show that antisense oligonucleotide inhibition of AML1 mRNA translation reduces the STE induction of elastase activity. The present study provides novel insight into the biological activity of a transcription factor hitherto primarily investigated in the context of the genetics of myeloid cell differentiation and leukemic reactions. It also provides new information about the regulation of the recently identified vascular elastase, important in vascular development and pathobiology.
Selected Abbreviations and Acronyms
|EMSA||=||electrophoretic mobility shift assay|
|EVE||=||endogenous vascular elastase|
|PA SMC||=||pulmonary artery SMC|
|PCR||=||polymerase chain reaction|
|RACE||=||rapid amplification of cDNA ends|
|SMC||=||smooth muscle cell|
|TNF||=||tumor necrosis factor|
This study was supported by grant T724 from the Heart and Stroke Foundation of Ontario. Dr Thompson is a Fellow and Dr Rabinovitch is a Career Investigator of the Heart and Stroke Foundation of Ontario.
Reprint requests to Marlene Rabinovitch, MD, Division of Cardiovascular Research, The Hospital for Sick Children, 555 University Ave, Toronto, Ontario, Canada M5G 1X8.
Presented in part at the American Society for Cell Biology, San Francisco, Calif, December 7–11, 1996.
- Received November 24, 1997.
- Accepted April 15, 1998.
- © 1998 American Heart Association, Inc.
Rabinovitch M, Bothwell T, Hayakawa BN, Williams WG, Trusler GA, Rowe RD, Olley PM, Cutz E. Pulmonary artery endothelial abnormalities in patients with congenital heart defects and pulmonary hypertension: a correlation of light with scanning electron microscopy and transmission electron microscopy. Lab Invest. 1986;55:632–653.
Maruyama K, Ye C, Woo M, Venkatacharya H, Lines LD, Silver MM, Rabinovitch M. Chronic hypoxic pulmonary hypertension in rats and increased elastolytic activity. Am J Physiol. 1991;261:H1716–H1726.
Todorovich-Hunter L, Dodo H, Ye C, McCready L, Keeley FW, Rabinovitch M. Increased pulmonary artery elastolytic activity in adult rats with monocrotaline-induced progressive hypertensive pulmonary vascular disease compared with infant rats with nonprogressive disease. Am Rev Respir Dis. 1992;146:213–223.
Ilkiw R, Todorovitch-Hunter L, Maruyama K, Shin J, Rabinovitch M. SC-39026, a serine elastase inhibitor, prevents muscularization of peripheral arteries, suggesting a mechanism of monocrotaline-induced pulmonary hypertension in rats. Circ Res. 1989;64:814–825.
Ye C, Rabinovitch M. Inhibition of elastolysis by SC-37698 reduces development and progression of monocrotaline pulmonary hypertension. Am J Physiol. 1991;261:H1255–H1267.
Zhu L, Wigle D, Hinek A, Kobayashi J, Ye C, Zuker M, Dodo H, Keeley FW, Rabinovitch M. The EVE that governs development and progression of monocrotaline-induced pulmonary hypertension in rats is a novel enzyme related to the serine proteinase adipsin. J Clin Invest. 1994;94:1163–1171.
Hornebeck W, Brechemier D, Soleilhac JM, Bourdillon MC, Robert L. Studies on rat aorta smooth muscle cells’ elastase activity. In: Reddi AH, ed. Extracellular Matrix: Structure and Function. New York, NY: Alan R Liss; 1985;25:269–282.
Rosenberg HC, Rabinovitch M. Endothelial injury and vascular reactivity in monocrotaline pulmonary hypertension. Am J Physiol. 1988;255:H1484–H1491.
Thompson K, Kobayashi J, Childs T, Wigle D, Rabinovitch M. Endothelial and serum factors which include apolipoprotein A1 tether elastin to smooth muscle cells inducing serine elastase activity via tyrosine kinase-mediated transcription and translation. J Cell Physiol. 1998;174:78–89.
Meyers S, Downing JR, Hiebert SW. Identification of AML-1 and the (8;21) translocating protein (AML-1ETO) as sequence-specific DNA-binding proteins: the runt homology domain is required for DNA binding and protein-protein interactions. Mol Cell Biol. 1993;13:6336–6345.
Meyers S, Lenny N, Hiebert SW. The t(8;21) fusion protein interferes with AML-1B-dependent transcriptional activation. Mol Cell Biol. 1995;15:1974–1982.
Nuchprayoon I, Meyers S, Scott LM, Suzow J, Hiebert SW, Friedman AD. AML1/CBF, altered by chromosomal breaks in myeloid leukemia cells, regulates the murine myeloperoxidase and neutrophil elastase genes in immature myeloid cells. Mol Cell Biol. 1994;14:5558–5568.
Liang P, Pardee AB. Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science. 1992;257:967–970.
Marchuk D, Drumm M, Saulino A, Collins A. Construction of T-vectors, a rapid and general system for direct cloning of unmodified PCR products. Nucleic Acids Res. 1991;19:1154–1154.
Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1989.
Dignam JD, Lebovits RM, Roeder RG. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 1983;11:1475–1487.
Lloyd HG, Schrader J. Adenosine metabolism in the guinea pig heart: the role of cytosolic S-adenosyl-l-homocysteine hydrolase, 5′-nucleotidase and adenosine kinase. Eur Heart J. 1993;14(suppl I):27–33.
Nishio Y, Aiello LP, King GL. Glucose induced genes in bovine aortic smooth muscle cells identified by mRNA differential display. FASEB J. 1994;8:103–106.
Utans U, Liang P, Wyner LR, Karnovsky MJ, Russell ME. Chronic cardiac rejection: identification of five upregulated genes in transplanted hearts by differential display. Proc Natl Acad Sci U S A. 1994;91:6463–6467.
Wang S, Wang Q, Crute BE, Melnikova IN, Keller SR, Speck NA. Cloning and characterization of subunits of the T-cell receptor and murine leukemia virus enhancer core-binding factor. EMBO J. 1993;13:3324–3339.
Thompson K, Rabinovitch M. Induction of vascular elastase is mediated by integrins and is associated with phosphorylation of focal adhesion kinase and activation of the transcription factor AML-1 [abstract]. Mol Biol Cell. 1996;7(suppl):588A.