IL-1β alters the expression of the receptor tyrosine kinase gene r-EphA3 in neonatal rat cardiomyocytes

Yun You Li, Charles F. McTiernan, Arthur M. Feldman


To identify proinflammatory cytokine responsive genes in the myocardium, we used differential display to study RNA isolated from neonatal rat cardiac myocytes treated with tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β). Sequence analysis of differential display products confirmed by reverse Northern blots revealed one clone as the partial sequence of an Eph-related receptor tyrosine kinase (r-EphA3). In cardiac myocytes, 36-h exposure to TNF-α and IL-1β reduced r-EphA3 transcripts to 59.9% (P < 0.01) of control levels; this effect was largely dependent on IL-1β. Western blot analysis showed that changes in r-EphA3 protein levels reflect that seen for transcripts. Cardiac nonmyocytes expressed substantially lower levels of r-EphA3. Full-length r-EphA3 cDNA clone (3,077 base pair) yielded an amino acid sequence with 90–98% homology to the Eph receptor human EphA3, chick EphA3, and mouse EphA3. In the adult rat, r-EphA3 transcripts were most abundant in the heart, brain, and lung. These results suggest that IL-1β may exert its effect on cardiac myocytes at least in part by altering r-EphA3 expression.

  • proinflammatory cytokine
  • differential display
  • cardiac myocytes

circulating concentrations of cytokines, including tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), are significantly elevated in patients with chronic heart failure (45). TNF-α and IL-1β have profound negative inotropic effects on cardiac physiology in vitro and in vivo (14, 19, 22, 29, 35, 44, 51). Additionally, these cytokines adversely regulate the expression of proteins important in cardiac excitation-contraction coupling, including the inhibitory guanine nucleotide regulatory protein Gi, the calcium regulatory proteins, and inducible nitric oxide synthase (iNOS) (9, 10, 12, 34). Although a variety of other signaling pathways have also been implicated in mediating various cytokine effects (40), the specific effects that are regulated during cardiac myocyte exposure to cytokines are not fully defined.

Proinflammatory cytokines may exert their effects on cardiac myocytes through regulation of gene expression at the transcript level (44). Thus identification of cytokine-responsive gene transcripts may identify unique pathways by which the pleiotropic effects of cytokines on the heart are regulated. A powerful approach to identify such transcripts is mRNA differential display. This technique can identify multiple differentially expressed transcripts without a priori knowledge of the particular genes or sequences (26, 27). Differential display was first introduced in 1992 as a technique for the comparison, identification, and isolation of genes differentially expressed in cells or tissues under various conditions (27). Differential display is applicable to both simple and complex biological systems and has been successfully used to identify genes that are differentially expressed in heart disease (37, 46), cancer (24, 25), and diabetes (30).

Because the proinflammatory cytokines TNF-α and IL-1β elicit many similar biological activities and may act synergistically, we used differential display to assess gene expression in cultured rat cardiac myocytes treated with these two cytokines. In this report, we identify and characterize one such proinflammatory cytokine responsive gene as an Eph-related receptor tyrosine kinase, r-EphA3.


All chemicals were purchased from Sigma (St. Louis, MO), unless otherwise stated.

Cell culture and cytokine treatment.Cardiomyocytes were isolated from ventricles of 1-day-old Sprague-Dawley rat hearts as described (28). Twenty-four to thirty-six hours after the cells were plated, experiments were initiated by addition of fresh media containing 5 ng/ml recombinant mouse IL-1β and/or 100 U/ml recombinant rat TNF-α (both from Biosource International, Camarillo, CA) or their vehicle [phosphate-buffered saline (PBS)]. When necessary, fresh medium containing the same cytokines was added after 24 h. For comparison of different treatment durations, initiation of treatments was staggered so that collection of the cells occurred at the same time. Experiments were terminated by removing media and freezing culture plates on liquid nitrogen. Plates were stored at −80°C before isolation of RNA. Cardiac nonmyocytes were prepared from the adherent preplated cells, prepared as described above, and cultured in the same media as for the cardiac myocytes but lacking 5-bromo-2′-deoxyuridine. After two cell passages, cytokine treatments were initiated as described for the cardiac myocytes.

To determine the percentage of nonmyocyte cells in the cardiac myocyte preparations, cells were grown on coated glass coverslips and stained using a monoclonal anti-sarcomeric myosin antibody (MF20; Developmental Studies Hybridoma Bank, Johns Hopkins University School of Medicine, Baltimore, MD; University of Iowa, Iowa City, IA) in a 1:2 dilution (2). Slides were then washed with PBS and incubated with a 1:100 dilution of tetramethylrhodamine isothiocyanate-labeled goat anti-mouse antibody (Sigma). Slides were viewed with an inverted phase immunofluorescence microscope (Nikon, Melville, NY). Assessed by this method, the cardiac myocyte preparations routinely contained >95% sarcomeric-myosin-positive cells.

All experimental procedures were carried out under sterile conditions and were in accordance with the Guide for the Care and Use of Laboratory Animals (7th ed.) of the National Research Council and approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh.

Total RNA and poly A+ RNA isolation. A single-step method was used for total RNA isolation from cultured neonatal rat cardiac cells or rat tissues with guanidine isothiocyanate as described (6). Poly A+ RNA was isolated with PolyATtract mRNA isolation system (Promega, Madison, WI) with reference to the technical manual of the manufacturer.

mRNA reverse transcription and differential display. Complementary DNA (cDNA) was synthesized by reverse transcription as follows. Total RNA was treated with RQ1 deoxyribonuclease (DNase) (Promega) to remove residual genomic DNA before reverse transcription. Four reverse transcription reactions were carried out for each RNA sample with 0.25 μg total RNA in Superscript II reaction buffer (GIBCO-BRL, Frederick, MD), 10 mM dithiothreitol, 20 μM of each deoxynucleoside 5′-triphosphate (dNTP), and 1 μM of either T12VA, T12VC, T12VG, or T12VT oligo primers (where V = A, C, or G). The reaction mixture was heated to 70°C for 10 min and cooled to 42°C, and 250 U of Superscript II (GIBCO-BRL) were added. After further reaction at 42°C for 1 h, the mixture was heated to 70°C for 5 min.

mRNA differential display was performed as described (24, 26). Polymerase chain reaction (PCR) amplification of the above cDNAs was carried out with corresponding 5′-anchored primers (T12VA, T12VC, T12VG, or T12VT oligo primers) and 3′ arbitrary decamers (Operon Technologies, Alameda, CA) in the presence of α-35S-labeled dATP. For each 20-μl PCR mixture, 2 μl of the reverse transcription reaction mixture were added to 18 μl of a solution to make a final concentration of 1 μM 5′-primer, 1 μM 3′-primer, and all four dNTPs (20 μM each). Each reaction also contained 10 μCi of α-35S-dATP (1,000 Ci/mmol, NEN Life Science, Boston, MA) and 3 U ofTaq DNA polymerase (Perkin Elmer, Norwalk, CT) and was overlaid with 30 μl of mineral oil. Amplification of RNA without reverse transcription was performed to show that the PCR was reverse transcription dependent. The PCR reaction parameters were as follows: denaturation at 94°C for 55 s, annealing at 40°C for 2 min, extension at 72°C for 60 s for 40 cycles, and finally one cycle of 72°C for 8 min. DNA sequencing loading buffer was added and heated to 94°C for 4 min before loading on 7% denaturing polyacrylamide gels. Gels were run at 75 W for 140 min, dried without fixation, and exposed to BioMax film (Eastman Kodak, New Haven, CT) for 12–36 h at room temperature. Repeated reactions with RNA from five untreated and four cytokine-treated sets of cells from multiple independent preparations were performed to confirm the reproducibility of differentially displayed bands.

Cloning of the differentially displayed cDNAs. Differentially displayed bands were recovered by excising the bands of interest from the dried sequencing gels with careful alignment of the gel and film. The gels containing the band of interest were rehydrated with 10 μl TE [10 mM tris(hydroxymethyl)aminomethane (Tris) ⋅ Cl, 1 mM EDTA] and boiled for 20 min. The solution was transferred to a new tube, and cDNA was precipitated with ethanol and recovered in 10 μl TE. Four microliters of the recovered cDNA were used for reamplification with the corresponding primers and the same PCR parameters as used for differential display, except that the reaction volume was 50 μl and did not contain α-35S-dATP. Reamplified cDNA fragments were cloned into pCRII vectors (Invitrogen, San Diego, CA). At least 10 colonies from each ligation were selected for preparation of plasmid DNA and further analysis.

Reverse Northern blot analysis. All plasmid DNAs were isolated by a standard alkaline/sodium dodecyl sulfate (SDS) lysis procedure. Plasmid DNAs containing inserts were denatured with 0.25 M NaOH, 0.5 M NaCl for 15 min at room temperature, and then diluted into 0.1× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate, pH 7.0), 0.125 M NaOH. Plasmid DNAs (0.3 and 0.15 μg) were blotted on to nitrocellulose membranes (Schleicher & Schuell, Keene, NH). The membranes were neutralized in 0.5 M NaCl, 0.5 M Tris ⋅ Cl, pH 7.5, dried in air, and fixed by crosslinking with ultraviolet (UV) light irradiation.

Total cDNA probes were labeled directly by reverse transcription of total RNA from cytokine-untreated or cytokine-treated cardiac myocytes with T12VA, T12VC, T12VG, or T12VT oligo primers in the presence of [α-32P]dCTP (3,000 Ci/mmol, NEN Life Science). The probes were purified with a Sephadex G50 spin column. Prehybridization was done in 10% dextran sulfate, 1 M NaCl, and 0.1% SDS at 58°C for 10 h. The same solution was used in hybridization with the addition of 100 μg herring sperm DNA/ml and 1.2 × 106 counts per minute (cpm) probe/ml. After overnight hybridization at 60°C, the membranes were washed with 2× SSC, 0.1% SDS at 60°C, then exposed in a PhosphorImager cassette (Molecular Dynamics, Sunnyvale, CA) for 12 h. Triplicate experiments were done, and each of the signals was quantified by integrating the volume of each dot with ImageQuaNT software (Molecular Dynamics). Those that showed changes congruent with differential display results were chosen for sequence analysis.

Sequencing of the cDNA fragments of differentially displayed genes and homology search with FASTA. The differentially displayed cDNA fragment inserts were cycle sequenced with M13 universal forward primer (5′-GCC AGG GTT TTC CCA GTC ACG A-3′) and reverse primer (5′-GAG CGG ATA ACA ATT TCA CAC AGG-3′). Insert sequences [sizes ranged from 177 to 593 base pair (bp)] were used for on-line homology search with FASTA against all submitted sequences in GenBank.

RNase protection assay. RNase protection assay (RPA) was performed with a RPA II kit (Ambion, Austin, TX) following the manufacturer’s instructions. For the synthesis of run-off radiolabeled transcripts, plasmid DNA containing the r-EphA3 insert cloned by differential display was linearized by complete digestion with EcoR V. r-EphA3 cRNA antisense probe was synthesized using T7 RNA polymerase. Total RNA from control and cytokine-treated cells was hybridized overnight at 46°C with radiolabeled antisense r-EphA3 and β-actin probes. Protected fragments were resolved on a 6% polyacrylamide gel after digestion with RNase A/T1. The gels were dried and exposed in a PhosphorImager cassette for 12 h, and band intensities were quantified using ImageQuaNT software. All r-EphA3 bands were normalized to the protected fragment of β-actin in the same lane to correct for variations in loading.

Northern blot analysis. Four micrograms of poly A+ mRNA was resolved in formaldehyde-agarose gels, transferred to nitrocellulose membrane by pressure blotting, and fixed by UV crosslinking. Prehybridization was performed overnight at 42°C in 5× SSC, 50% formamide, 5× Denhardt’s reagent, 0.2% SDS, and 25 mM sodium phosphate, pH 6.5. Hybridization probes were radiolabeled to at least 1 × 108 cpm/μg using a random hexamer priming kit (Boehringer Mannheim, Indianapolis, IN). Radiolabeled cDNA probes were hybridized overnight at 42°C in a solution identical to the prehybridization solution, except for the addition of 10% dextran sulfate. Membranes were washed four times with 2× SSC, 0.1% SDS at room temperature and twice with 0.2× SSC, 0.1% SDS at 56°C. The membrane was then exposed in a PhosphorImager cassette and quantified as described above. The same blot was stripped with 2 mM Tris ⋅ Cl, pH 8.0, 0.2 mM EDTA, 0.1% SDS at 75°C for 1 h, and reprobed with a radiolabeled glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe. Membranes were then reexposed and quantified, and results of the cDNA hybridization were normalized to that of the GAPDH to correct for differences in RNA mass and efficiency of transfer.

Western blot analysis. Control and cytokine-treated cells were lysed into RIPA buffer [1× PBS, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, and protease inhibitor cocktail (Sigma cat. P2714)]. Protein concentrations were measured using the Bradford method with Bio-Rad protein assay reagent (Bio-Rad, Hercules, CA). Equal amounts (120 μg) of cell lysates were separated by SDS-polyacrylamide gel (6%) electrophoresis and electroblotted onto nitrocellulose membranes. r-EphA3 protein was probed with rabbit polyclonal immunoglobulin G (IgG; Santa Cruz Biotechnology, Santa Cruz, CA) raised against the synthetic peptide IISSIKALETQSKNGPVPV corresponding to the deduced h-EphA3 amino acid 965–983 (the peptide is identical to r-EphA3 corresponding region 966–984). Goat anti-rabbit IgG conjugated with horseradish peroxidase was used as the secondary antibody. The antibody reaction was developed using an enhanced chemiluminescence reagent (NEN Life Science).

cDNA cloning and sequence analysis of r-EphA3. The full-length cDNA of the rat Eph-related receptor tyrosine kinase gene r-EphA3 was cloned with a long PCR approach (21, 23). cDNA was reverse transcribed from total RNA of neonatal rat cardiac myocytes with Superscript II and was used as template for long PCR amplification. Two overlapping fragments of the entire coding sequence were amplified by using high-fidelity rTth DNA polymerase (Perkin Elmer, Norwalk, CT) with the following conditions: denaturation at 94°C for 1 min and annealing and extension at 67°C for 3 min for 30 cycles after an initial “hot start” at 78°C for 3 min. The upstream primer for the 2.2-kb 5′-fragment and the downstream primer for the 1.1-kb 3′-fragment were designed based on the conserved regions within the 5′- and 3′-untranslated sequences of receptor tyrosine kinases, whereas the other primers for each of the fragments were chosen from the sequence of the r-EphA3 fragment cloned by differential display. The two overlapping fragments were fused together to generate full-length cDNA by repeating the above PCR reactions for 30 cycles without primers and templates but mixing equimolar amounts of the two fragments into the reaction. The full-length cDNA was cloned into a pCRII vector, and the complete sequence was analyzed with a 377 automatic DNA sequencer (Perkin Elmer, Applied Biosystems, Foster City, CA). The resultant sequence has been deposited in GenBank (accession number U69278). The prediction of hydrophobic domains of r-EphA3 was performed with on-line TMpred program (16). The multiple sequence alignment was made with the on-line program Blitz (7, 41).

Data processing and statistics. The RPA and Northern blot results were quantified with ImageQuaNT software and normalized to that for β-actin or GAPDH signals obtained from the same lane. The Western blot films were digitized with an HP ScanJet 4C scanner (Hewlett-Packard, Englewood, CO) and then quantified with ImageQuaNT software. The quantitative results were presented as percent change compared with untreated cells. The results are given as means ± SD. One-way analysis of variance was applied to compare these changes in different experimental groups. When a significantF value was obtained, comparison among the means was made by use of the post hoc Student-Newman-Keuls analysis test. Statistical significance was considered atP < 0.05.


Identification of novel proinflammatory cytokine-responsive genes by mRNA differential display. Differential display with 20 arbitrary 10-nucleotide primers paired to all four T12VN primers revealed 14 upregulated and 6 downregulated PCR products when RNA from cardiac myocytes treated with proinflammatory cytokines was compared with that from untreated samples (Fig. 1). Repetitive experiments using the same primers and different batches of total RNA produced almost equivalent banding patterns. Fifteen bands excised from the differential display gel were successfully eluted and reamplified with PCR; 12 of these bands were cloned after reamplification. For each successfully reamplified and cloned product, plasmid DNA was isolated from 10 individual bacterial colonies and the expression patterns were analyzed with reverse Northern hybridization (Fig. 2). Each of the above differentially displayed bands contained a number of distinct cDNA fragments derived from different transcripts. Not all of the cDNA fragments represented transcripts that showed changes resembling the differential display pattern. However, from 8 of the 12 reamplified bands, at least one cloned product was in accordance with the differential display results. Some bands contained multiple gene products, each of which demonstrated differential regulation. Therefore, from eight of the reamplified bands that agree with differential display results, a total of 12 genes was identified that appeared to be true positive (Table1).

Fig. 1.

mRNA differential display of cardiac myocytes treated with tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β). Repeated reactions with RNA from 5 untreated and 4 cytokine-treated sets of cells from independent preparations of cardiac myocytes were performed to confirm the reproducibility of differentially displayed bands. A, C, G, T: differential display with 5′-anchored primers of either T12VA, T12VC, T12VG, or T12VT. First 3 lanes (left toright) in each set show differential display of mRNA from untreated control cardiac myocytes, and next 2 lanes show results from cardiac myocytes treated with TNF-α and IL-1β for 36 h. Arrowhead points to the band that was identified to be r-EphA3.

Fig. 2.

Reverse Northern hybridization. Equal amounts of plasmid DNA (300 ng) were applied to duplicate membranes. One membrane (left) was hybridized with a reverse transcription labeled cDNA probe derived from RNA of control cardiac myocytes, whereas the second membrane (right) was hybridized with a cDNA probe derived from RNA of cardiac myocytes treated with TNF-α and IL-1β. The 2 membranes were depicted such that dots at corresponding positions represented the same clone. Results shown are representative of 3 independent experiments.

View this table:
Table 1.

TNF-α- and IL-1β-responsive genes in cardiac myocytes identified by differential display

DNA sequence analyses of the clones confirmed as having differential expression revealed that all insertions had one or both of the flanking sequences complementary to the appropriate primer pairs. The sizes of cloned inserts ranged from 177 to 593 bp. By comparison with sequences in the GenBank database, one clone (17-4) contained a 374-bp sequence that was 93% identical to bases 1945 to 2318 of the mouse Eph-related receptor tyrosine kinase m-EphA3 (38). Deduced amino acid sequence of this clone had 95% identity with the corresponding region of m-EphA3. Hereafter, we identify this receptor tyrosine kinase as r-EphA3 following the nomenclature of the Eph Nomenclature Committee (11).

Differential regulation of r-EphA3 gene by TNF-α and IL-1β. In addition to differential display and reverse Northern blot hybridization, RPA assay and Northern blot analyses were performed to further confirm the expression pattern of r-EphA3 in neonatal rat cardiac myocytes exposed to proinflammatory cytokines. RPA assay with the cloned 374-bp r-EphA3 probe showed downregulation of r-EphA3 transcripts by combined TNF-α and IL-1β treatment of cardiac myocytes (64.1 ± 18.9% that of untreated cells,n = 13,P < 0.01). IL-1β treatment of the cells had a similar effect (67.7 ± 11.9% that of untreated cells,n = 4,P < 0.01) (Fig.3). Northern blot hybridization detected two bands of 6.5 (major) and 1.4 kb (minor) (Fig.4 A). A time course study showed a significant decrease in the 6.5-kb r-EphA3 transcript after 36 h of treatment with TNF-α and IL-1β (59.9 ± 16.3% that of untreated cells, n = 13, P < 0.01), whereas expression of the 1.4-kb transcript appeared unchanged. Separate treatment with TNF-α or IL-1β showed that the reduction of the 6.5-kb r-EphA3 transcripts occurred only in the IL-1β-treated cells (60.3 ± 17.4% that of untreated cells, n = 10, P < 0.01, Fig.4 B).

Fig. 3.

RNase protection assay. [α-32P]UTP labeled r-EphA3 antisense probe was transcribed from linearized plasmid DNA containing the r-EphA3 inserts using T7 RNA polymerase. Total RNA (27 μg) from control and cytokine-treated cells was hybridized with antisense r-EphA3 and β-actin probes. Protected fragments were resolved on a 6% polyacrylamide gel after digestion with RNase A/T1. The band intensities were quantified and normalized to the protected fragment of β-actin to correct variations in loading.A: representative RNase protection assay. Marker sizes at left are in base pairs.B: summary of quantitative data (n = 12). * P < 0.05 compared with control.

Fig. 4.

Time-dependent and differential regulation of r-EphA3 expression in cardiac myocytes by TNF-α and IL-1β. Northern blot analysis of r-EphA3 gene expression was carried out with 4 μg poly A+ RNA from rat neonatal cardiac myocytes. Blots were sequentially hybridized with r-EphA3 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probes. Differences in RNA mass were corrected by normalizing the signal for r-EphA3 to that for GAPDH. A: time-dependent downregulation of r-EphA3. Left, representative Northern blot; right, summary of the Northern blot quantitative data (n = 6–13).B: differential effect of TNF-α and IL-1β on the expression of r-EphA3.Left, representative Northern blot;right, summary of quantitative data (n = 7–13). * P < 0.05 compared with control.

r-EphA3 protein expression in cardiac myocytes. Western blot with polyclonal anti-EphA3 antibodies detected a robust 130-kDa signal in whole cell lysates of cardiac myocytes, whereas this 130-kDa signal was not measurable in nonmyocytes. This band was completely abolished by preincubating the r-EphA3 antibody with the r-EphA3 antigenic peptide (blocking peptide). The time-dependent regulation of r-EphA3 protein reflected the changes at the mRNA level as measured by RPA and Northern blot analyses (Figs.3, 4, and5 A). The downregulation of r-EphA3 was observed only when IL-1β was present in the treatment (Fig. 5 B).

Fig. 5.

TNF-α and IL-1β regulation of r-EphA3 protein in cardiac myocytes. Myocytes were stimulated with TNF-α (100 U/ml) and/or IL-1β (5 ng/ml). Whole cell lysates (120 μg) were separated by sodium dodecyl sulfate-polyacrylamide gel (6%) electrophoresis and electroblotted onto nitrocellulose membrane. r-EphA3 protein was probed with rabbit polyclonal immunoglobulin G (IgG) raised against the synthetic peptide IISSIKALETQSKNGPVPV (corresponding to the deduced r-EphA3 amino acids 966–984). Goat anti-rabbit IgG conjugated with horseradish peroxidase was used as a secondary antibody. Bands were detected using a chemiluminescence reagent.A: time-dependent effect of cytokines on the expression of r-EphA3. Note band was completely blocked by preincubating the antibody with r-EphA3 peptide.B: differential effects of TNF-α or IL-1β. Quantitative data presented in bar graph were obtained from 4 independent experiments. * P < 0.01 compared with control.

r-EphA3 is predominantly expressed in cardiac myocytes. Northern blot analyses were also performed to assess the expression of r-EphA3 transcripts in cardiac nonmyocytes. Hybridization to equal amounts of RNA repetitively showed that the r-EphA3 transcripts were expressed at much lower levels in nonmyocytes relative to cardiac myocytes (Fig.6 A). Although expression at such low levels made assessment of alterations in transcript levels difficult, we could not detect any alteration in r-EphA3 transcript levels in cardiac nonmyocytes exposed to proinflammatory cytokines. At the protein level, r-EphA3 was detected in whole cell lysates of cardiac myocytes, but barely detectable in nonmyocyte lysates (Fig. 6 B).

Fig. 6.

r-EphA3 is predominantly expressed in cardiac myocytes.A: Northern blot analysis of equal amounts (4 μg) of poly A+ RNA from cardiac myocytes and cardiac nonmyocytes either untreated or treated with TNF-α and IL-1β detected the full-length r-EphA3 transcripts in cardiac myocytes but not in nonmyocytes.B: Western blot of 120 μg whole cell lysates with polyclonal IgG against EphA3 detected a 130-kDa band in cardiac myocytes, but not in nonmyocytes.

Cloning and characterization of r-EphA3 cDNA. To further characterize r-EphA3, we applied a PCR-based approach to clone the full-length r-EphA3 cDNA. Using primers based on the 17–4 clone sequence and previously reported sequences for 5′- and 3′-untranslated regions of the receptor tyrosine kinases in the Eph subfamily, two overlapping fragments of 2,241 and 1,156 bp were amplified from cDNA reverse transcribed from RNA of neonatal rat cardiac myocytes. These two fragments were then fused by coamplification with PCR to yield a 3,077-bp product. All PCR reactions for this cloning used rTth DNA polymerase to reduce amplification errors. In addition, two partial clones containing additional sequences extending from the coding region through the putative 3′-untranslated region were isolated by screening a rat heart cDNA library with the 17–4 cloned fragment. Of the 200 sequenced nucleotides that overlapped with the clones generated through rTth-based PCR, no variations were observed, suggesting minimal PCR-generated alterations in nucleotide sequence. After cloning, the nucleotide sequence of the full-length coding region together with partial 5′- and 3′-untranslated sequence was determined.

The predicted translation product of this sequence shows an open reading frame of 984 amino acids (nucleotides 35–2989), yielding a protein typical of Eph-related receptor tyrosine kinase (Fig.7). Two hydrophobic regions are consistent with a signal peptide from amino acid 1 to 18 and a transmembrane domain from amino acid 542 to 564. The extracellular domain of r-EphA3, which exists between the signal peptide and the transmembrane region, is believed to constitute the ligand binding domain and contains characteristic structural motifs: one immunoglobulin-like domain, two fibronectin III-like repeats, and an enrichment of cysteine residues. Similar features are found in other receptor tyrosine kinases (50). In addition, the presumed cytoplasmic domain (amino acid residues 565–984) contains a characteristic protein tyrosine kinase catalytic domain, including an Mg2+-ATP binding site, a catalytic loop, and a putative tyrosine autophosphorylation site at position 780, also consistent with the structure of receptor tyrosine kinases. The overall nucleotide sequence homology between r-EphA3 and m-EphA3 is 94%, and the deduced amino acid sequence homology is 97.5% to m-EphA3, 96% identity to h-EphA3 (48), and 90% identity to c-EphA3 (38).

Fig. 7.

Multiple alignments and comparison of amino acid sequence of r-EphA3 with its homologs. Deduced r-EphA3 amino acid sequence was aligned with sequences of mouse (m)-EphA3, human (h)-EphA3, and chick (c)-EphA3. Potential signal peptide and functional domains are underlined.

r-EphA3 mRNA expression in rat tissues. To further understand the relative importance of r-EphA3 in the heart, the expression of r-EphA3 in various neonatal and adult rat tissues was studied by Northern blot analysis. The 6.5-kb full-length transcript was most prominent in adult brain, heart, and lung (Fig. 8). It was expressed in both neonatal and adult rat hearts. Other tissues, such as lung and kidney, showed developmental alterations in the expression of the r-EphA3 gene. The 1.4-kb transcript was most prominent in neonatal skeletal muscle tissue. Subsequent attempts to find an r-EphA3 transcript corresponding to the 1.4-kb band by rapid amplification of cDNA ends, screening of a rat heart cDNA library, and slot blot analyses with selected regions of the r-EphA3 cDNA suggested that the 1.4-kb transcript did not arise from an alternatively spliced product of the r-EphA3 gene (data not shown).

Fig. 8.

Tissue distribution of r-EphA3 mRNA. A multiple tissue Northern blot containing ∼4 μg of poly A+ RNA from rat neonatal and adult tissues indicates greatest r-EphA3 expression is in the heart, brain, and lung. Blot was probed with GAPDH to ensure that the RNA in each lane was intact. A: neonatal tissues. B: adult tissues.


Use of differential display to identify novel cytokine-responsive genes in cardiac myocytes. Differential display is a method with great potential for discovering alteration in gene expression and novel gene transcripts. The high sensitivity of this technique makes it prone to false positives; therefore, it is necessary to repeat the same experiment with multiple samples and to verify differentially displayed genes by complementary methods. For this purpose we repeated differential display with RNA from different preparations of cytokine-treated cell cultures and identified similar changes in the pattern of amplified products. The reverse Northern blot procedure was applied to screen and analyze all the candidate differentially displayed cDNA clones. The verified clones congruent with the results of differential display were used in RPA and Northern blot analyses to quantitate mRNA expression. Our results suggest that a variety of genes are differentially regulated when cultured cardiac myocytes are treated with proinflammatory cytokines. Indeed, in addition to a series of novel genes, we have identified an Eph-related receptor tyrosine kinase, r-EphA3, not previously investigated in the myocardium.

r-EphA3 transcript and protein are regulated by IL-1β in neonatal rat cardiac myocytes. Differential display identified r-EphA3 as being downregulated by cotreatment with TNF-α and IL-1β. RPA and Northern blot analyses demonstrated that these cytokines produced ∼40% reduction in r-EphA3 transcript levels. This reduction was apparently mediated by IL-1β, as TNF-α alone did not change the expression of r-EphA3, whereas IL-1β alone mimicked the effect of cotreatment of the cells with TNF-α and IL-1β at both the RNA and protein levels. This effect is not immediate and appears to take several hours to develop. The r-EphA3 protein changes are temporally related to changes in transcript levels, suggesting the r-EphA3 is transcriptionally regulated. However, the mechanism by which IL-1β downregulates r-EphA3 transcript levels remains unknown.

Role of Eph-related receptor tyrosine kinases in response to proinflammatory cytokines. The Eph family of receptors constitutes the largest subgroup of the receptor tyrosine kinases family (47). The functional significance of the Eph subfamily of receptor tyrosine kinases is not well defined. Some (including m-EphA3, the murine homologue of r-EphA3) may play a role in signal transduction during differentiation, development, carcinogenesis, neuron axon bundle formation, and control of cytoskeletal architecture (15, 17, 49). Amino acid sequence comparisons demonstrated very high homology among r-EphA3 and mouse (m-EphA3), chick (c-EphA3), and human (h-EphA3) receptor tyrosine kinases, implying a similar function among these genes. The identification of r-EphA3 as a cytokine-responsive gene in cardiac myocytes suggests the presence of additional roles for Eph receptors in both inflammation and cardiac diseases.

It is of interest to note that B61, the prototypic member of the family of ligands for Eph receptors, was first identified as a protein whose expression was upregulated in response to TNF-α (3). Indeed, B61 has been proposed to participate in the angiogenesis associated with inflammation (36). In addition to these studies with B61, several additional observations suggest that receptor tyrosine kinases including r-EphA3 might participate in cytokine signal processing: 1) IL-1β-inducible expression of gro-β and iNOS are dependent on tyrosine kinase signaling pathways (8, 20, 39, 43),2) cytokines can induce rapid tyrosine phosphorylation of several signaling molecules through tyrosine kinases (42), 3) the effect of IL-1β on the hypertrophic growth of neonatal rat cardiac myocytes can be blocked by the tyrosine kinase inhibitor genistein (44), and4) this report observed that IL-1β exposure produces significant decrease in r-EphA3 expression in cardiac myocytes at both the transcript and protein levels, although the pathway by which this occurs is unclear. Because the signal transduction pathways activated through EphA3 are unknown, it is unclear how or even whether a 50% reduction in the level of r-EphA3 protein may alter cell and organ physiology. It is reported that IL-1β induces hydrolysis of GTP (32, 33), and the activation of the related Eph receptor EphB2 leads to multiple and sequential kinase reactions and modulates hydrolysis of GTP by regulating the activities of Ras-GTPase activating protein (17, 18). Thus it is possible that IL-1β-induced changes in the level of r-EphA3 could significantly modify downstream signal transduction cascades and cellular physiology through a similar pathway. We note that, by analogy, a 50% reduction in the level of β1-adrenergic receptor is found in the failing human myocardium and is associated with profound functional alterations (4, 5).

Tissue distribution of r-EphA3 transcripts. Having demonstrated that r-EphA3 was an IL-1β-responsive gene in neonatal rat cardiac myocytes, we isolated full-length cDNA clones to more definitively identify the potential protein that this transcript encodes and examined the tissue-specific expression of r-EphA3 transcripts. We found that r-EphA3 is expressed in both neonatal and adult rat hearts. Additionally, some organs (such as lung and kidney) show marked differences in the expression of r-EphA3 between the neonatal and adult organism (Fig. 8). However, the function of this receptor tyrosine kinase in the heart remains undefined. In other organs, receptor tyrosine kinases play a role in cellular differentiation and proliferation (1, 13, 31). As cardiac myocytes are terminally differentiated nonproliferative cells, the signaling events mediated by r-EphA3 may stimulate responses other than mitogenesis or differentiation.

In conclusion, differential display identified the r-EphA3 receptor tyrosine kinase as an IL-1β-responsive gene in neonatal cardiac myocytes. r-EphA3 expression is downregulated in cardiac myocytes at both the mRNA and protein levels after 36-h treatment of the cells with IL-1β. This receptor and other members of the Eph-receptor family have not been previously studied in the myocardium. Our results suggest that IL-1β may exert its effects on cardiac myocytes at least in part by modulating r-EphA3 expression. However, the physiological relevance of the ∼50% reduction in r-EphA3 expression and the pathway by which IL-1β alters r-EphA3 expression remain to be determined.


We thank Bonnie Lemster for help with the cell preparations.


  • Address for reprint requests: C. F. McTiernan, 1744.1 Biomedical Science Tower, Division of Cardiology, Univ. of Pittsburgh Medical Center, 200 Lothrop St., Pittsburgh, PA 15213.


View Abstract