Expression of the ETS Transcription Factor ELF3 in the Retinal Pigment Epithelium

Abstract

purpose. The ETS family of transcription factors regulate several critical cellular functions. They have also been implicated in invertebrate ocular development. This work was undertaken to determine whether epithelium-specific ETS transcription factors are expressed in the retinal pigment epithelium and to investigate the possible role of these factors in retinal diseases such as age-related macular degeneration.

methods. The expression of the epithelial ETS transcription factors ELF5, ESE3, and ELF3 was assessed by RT-PCR in the human RPE cell lines D407 and hTERT-RPE1. The full-length coding sequence of rat Elf3 was isolated with 3′ rapid amplification of cDNA ends (RACE) and degenerative primers, and its expression was determined in various rat tissues, by RT-PCR and real-time PCR. A polyclonal ELF3 antibody produced from a C-terminal peptide was used to observe the distribution of the transcription factor within the retina. To assess the possible ELF3 regulation of the TIMP3 promoter, transient transfection assays were performed. Promoter activity was determined with a firefly luciferase reporter gene construct.

results. The epithelium-specific ETS transcription factor ELF3 was expressed in the D407 and hTERT-RPE1 cell lines. Neither ESE3 nor ELF5 was detected in the RPE. The cloning of rat Elf3 produced two splice variants, designated Elf3a (1786 bp) and Elf3b (1855 bp). The larger form, Elf3b, contained a 69-bp insert in the coding sequence, which showed high homology to a similar insert previously identified in murine Elf3. Both splice variants were expressed in rat lung, kidney, liver, and retina, but were absent in heart tissue. Real-time PCR analysis showed the retina to contain high levels of Elf3, which was subsequently localized to the RPE. Elf3 upregulated the TIMP3 promoter, with Elf3a and -3b inducing an approximate sixfold increase in activity.

conclusions. The ELF3 transcription factor is highly expressed in the RPE and can regulate important ocular genes, such as TIMP3, in vitro. The specific expression of ELF3 in the RPE may reflect an important role for this transcription factor in retinal function. Furthermore, its regulation of TIMP3 may have implications for degenerative retinal diseases, such as age-related macular degeneration.

The retinal pigment epithelium (RPE) is a monolayer of cells bordered basally by the choriocapillaris and apically by the distal tips of the photoreceptors. It constitutes part of the blood–retinal barrier and as such controls the access of serum-based factors to the neural retina. ¹ In addition to this role, the monolayer is involved in the oxidation of retinol, isomerization of the retinoids and phagocytosis of shed photoreceptor outer segments. ^{2 3 4} Because the pigment epithelium is critical to retinal function, factors that regulate RPE gene expression not only affect tissue homeostasis, but may also have a role in disease. 

Transcription factors are known to regulate gene expression by binding directly to promoter sites within DNA or to other transcription factors. During tissue development, these factors are the main determinants of final cellular identity. Within the eye, several transcription factors are known to be critical to proper ocular development. PAX6 has been described as the master control gene in eye formation, ⁵ whereas CHX10 has been implicated in retinal development. ⁶ Other transcription factors such as those in the Fos and Jun families, ⁷ MITF ⁸ and CRX ⁹ are also critical to eye formation. The importance of these factors is further emphasized by the fact that mutations within transcription factor genes have been associated with the development of ocular disease. ⁹  

The ETS family of transcription factors, which presently number more than 30, are important regulators of hematopoiesis, ¹⁰ angiogenesis, ¹¹ cell differentiation, ¹² and organogenesis. ¹³ Originally identified because of their homology to the v-ets oncogene, ¹⁴ all members of the family share a highly conserved DNA-binding domain known as the ETS domain and recognize the GGA(A/T) core sequence in the promoters and enhancers of various cellular genes. The importance of the ETS domain to protein function is illustrated by its high degree of conservation (49% identity) between human and Drosophila. ¹⁵ In addition to their role in development, ETS factors have been implicated in several diseases such as Down syndrome, ¹⁶ Alzheimer’s disease, ¹⁷ and tumorigenesis. ¹¹  

Several of the ETS transcription factor family are specifically expressed in epithelial cells. ELF3 (ERT/ESX/ESE1) was the first described in human breast cancer cells, but has subsequently been identified in a wide range of epithelial cells. ^{18 19} Three other epithelial ETS factors, ELF5 (ESE2), ²⁰ ESE3, ²¹ and PDEF (Pse), ^{22 23} have also been identified recently; however, these factors have a more restricted pattern of expression. The presence of these ETS transcription factors may be critical for the epithelial phenotype, with alterations in expression associated with malignant transformation. ¹⁹  

There has been only limited investigation into the presence and role of ETS factors within the vertebrate eye. Recently, Yoshida et al. ²⁴ have reported the presence of Elf3 in mouse corneal epithelium. Its expression parallels that of the differentiation marker, K12 keratin and appeared critical for epithelial differentiation within the cornea. Earlier work by O’Neal et al. ²⁵ implicated another two ETS domain proteins, pointed and yan, in Drosophila photoreceptor development. Because ETS function is highly conserved between species, ²⁶ these transcription factors may have important roles in the ocular development of higher organisms. 

ETS transcription factors may also regulate genes potentially involved in retinal disease. Of particular interest is the tissue inhibitor of metalloproteinase 3 (TIMP3) gene, which has shown elevated levels in retinitis pigmentosa (RP), ²⁷ Sorby’s fundus dystrophy (SFD), ²⁷ and age-related macular degeneration (ARMD). ²⁸  

TIMP3, unlike the other members of the TIMP family, is an insoluble component of the extracellular matrix (ECM). It is synthesized in the RPE and deposited in Bruch’s membrane, where it regulates ECM turnover and limits choroidal neovascularization. Mutations within the TIMP3 gene are known to lead to the development of autosomal dominant SFD. ²⁹ Because SFD and ARMD have a similar etiology, TIMP3 may also play a role in the progression of the age-related condition. However, no mutations have been found in patients with ARMD, suggesting that other mechanisms are involved in accumulation of TIMP3 and the resultant changes in the ECM that are observed in this disease. The upregulation of TIMP3 by transcription factors could provide such a mechanism. 

The purpose of the present study was to identify whether epithelial ETS transcription factors are expressed in the RPE. The identification of retinal transcription factors may provide an insight, not only into RPE regulation, but also into the cellular disruption that occurs in retinal diseases such as ARMD. 

Materials and Methods

All experiments conformed to the ARVO statement for the Use of Animals in Ophthalmic and Vision Research. 

Cell Lines and Culture

The transformed human RPE cell line D407 was grown in Dulbecco’s modified Eagle’s medium (DMEM; Gibco-Life Technologies, Gaithersburg, MD) supplemented with 10% fetal calf serum (FCS), 100 U/mL penicillin, and 100 μg/mL streptomycin (CSL, Melbourne, Australia). The telomerase-immortalized human RPE cell line, hTERT-RPE1 (Clontech, Palo Alto, CA), was grown in DMEM-F12 (Gibco-Life Technologies) containing 10% FCS, 0.348% Na₂HCO₃, 100 U/mL penicillin, and 100 μg/mL streptomycin. Cell lines were maintained in a humidified incubator at 5% CO₂ and 37°C. Confluent cell monolayers were passaged using a 1% (wt/vol) trypsin-EDTA solution (CSL). 

Isolation and Characterization of Rat Elf3 cDNA

The 3′ coding sequence of rat Elf3 was determined using the rapid amplification of cDNA ends (RACE) system (Life Technologies). The initial RACE reaction used a specific rat Elf3 primer (5′-ggagatcctggaacgggtggatgg-3′) and a universal primer (UAP; Life Technologies). The subsequent product was reamplified with a nested primer (5′-ctcgtctacaagtttggcaaaaac-3′). The specific Elf3 and nested primers were derived from a rat expressed sequence tag (EST) that showed a high homology with the human and murine ELF3 sequences. Additional 5′ coding sequence was determined with a degenerate ATG-spanning oligonucleotide (5′-atggcngcnacntgygarat-3′). All cDNA sequences were confirmed by automated sequencing of both strands. 

The full rat Elf3 coding sequence was cloned using the degenerate oligonucleotide and a primer within the 3′ untranslated region (5′-ctgacccttaattctgactctctccaacc-3′). The 1192-bp (Elf3a) and 1123-bp (Elf3b) products were subcloned into a plasmid vector (pGEM-T; Promega Corp., Madison, WI) and subsequently cloned into the NotI site of the expression vector pcDNA3.1 (Invitrogen, Carlsbad, CA), yielding the constructs rElf3a/pcDNA3.1 and rElf3b/pcDNA3.1. 

RNA Isolation and RT-PCR

Total RNA was isolated from individual rat tissues (10–30 mg) and the D407 and hTERT-RPE1 cell lines (0.5–1 × 10⁷ cells) using an extraction kit (RNeasy mini kit; Qiagen, Chatsworth, CA). The retinal samples included tissue from the posterior eye cup. Total RNA from the human breast and prostate cancer cell lines, T47D and PC3, were used as the positive control for ETS transcription factor expression. RT-PCR was performed on 0.5 μg RNA (Omniscript RT; Qiagen) using an oligo dT primer. Subsequent amplification on a thermal cycler (PCR Express; Hybaid, Ashford, UK) used HotStarTaq DNA polymerase (Qiagen) and intron-spanning primers for the ETS family members hELF3, hESE3, hELF5, and rElf3 (Table 1) . The amplification protocol consisted of 95°C (15 minutes), followed by 30 cycles of 94°C (10 seconds), 58°C (30 seconds), and 72°C (1 minute). The Elf3a and -3b splice variants were amplified from rat tissues with a 35-cycle protocol. 

Rat Tissue Elf3 Expression

The proportion of Elf3 was determined in various rat tissues using real-time PCR (LightCycler; Roche, Mannheim, Germany). RT-PCR was performed on 0.4 μg total RNA as described, and 1 μL of the resultant cDNA was added to a 19-μL PCR reaction mix (50 mM Tris [pH 8.3], 3 mM MgCl₂, 0.2 mM dNTP, 0.25mg/mL BSA, 0.8 U Taq, and a 1:3000 dilution of fluorescent green dye [SYBR Green I; Applied Biosystems, Foster City, CA] in 10 mM Tris, 0.1 mM EDTA [pH 8.0]). Rat Elf3 primers (0.5 μM; Table 1 ) were included in a 40-cycle protocol (95°C, 0 seconds; 61°C, 10 seconds; 72°C, 20 seconds). Detection of the fluorescent signal was performed at 88°C, and amplifications were performed in triplicate. The full coding sequence of rat Elf3b was amplified as described and quantified spectrophotometrically, and a standard curve (10²–10⁶ copies) was produced. Tissue Elf3 copy numbers were determined with reference to the standard curve. 

Elf3 Polyclonal Antibody Production

New Zealand White rabbits were challenged with an Elf3 peptide sequence (VDGRRLVYKFGKNSSGWKE) linked to a diphtheria toxin carrier (Chiron Technologies, Melbourne, Australia). Antibodies were affinity purified using a recombinant human ELF3 linked gel (Sepharose-4B; Amersham Pharmacia Biotech, Piscataway, NJ). The recombinant protein was generated using the His-tag expression vector pQE30 and purified on a separation column (Ni-NTA Superflow; Qiagen), with a denaturing protocol. ¹⁹  

Western Blot Analysis

Rat tissue nuclear extracts were isolated in a two-step lysis protocol. Respective tissues were washed with PBS (7.2 mM Na₂HPO₄, 2.8 mM NaH₂PO₄ [pH 7.2], and 0.15M NaCl) and homogenized in a cytoplasmic lysis buffer (10 mM Tris-HCl [pH 8], 5 mM KCl, 2 mM MgCl₂, and 0.5% Nonidet P-40). Cell nuclei were pelleted (22,000g), washed with the cytoplasmic lysis buffer, and resuspended in SDS sample buffer. 

Soluble proteins were separated by SDS-PAGE, ³⁰ and Western blot analysis was performed according to the method of Towbin and Gordan. ³¹ The Elf3 antibody was used at a concentration of 1 μg/mL, and a sheep anti-rabbit HRP conjugate (Silenus, Melbourne, Australia) was used as the secondary antibody. The membrane was developed with a luminescence detection kit (ECL Plus; Amersham, Buckinghamshire, UK). 

Immunohistochemistry

Hybrid-ready rat eye sections (Novagen, Madison, WI) were deparaffinized, rehydrated, immersed in 0.01 M sodium citrate buffer (pH 6.0), and heated for 10 minutes. ³² Tissue sections were blocked with 20% sheep serum for 1 hour, rinsed, and incubated overnight in PBS containing 1% BSA, 1% rat serum, 1% sheep serum, and 5 μg/mL Elf3 antibody. Sections were subsequently incubated with a sheep anti-rabbit AP conjugate (Silenus), followed by a rabbit anti-APAAP conjugate (Sigma Chemical Co., St. Louis, MO). The signal was detected using fast blue BB and sections were counterstained with periodic acid-Schiff stain (Sigma). For the blocking control, the Elf3 antibody was incubated for 1 hour with a 200-fold molar excess of immunizing peptide. 

Construction of Luciferase Reporter Vectors

A 3.5-kbp fragment, containing the human TIMP3 promoter was isolated from the bacterial artificial chromosome (BAC) bK766E1 after BamHI digestion. This fragment was cloned into a vector (pGem3Zf(−); Promega Corp.) and a 1210-bp fragment subsequently isolated after digestion with EcoICRI (Promega Corp.). The TIMP3 promoter fragment (−932 to +278) was finally blunt-end cloned into the HindIII site of a luciferase reporter vector (pGL3basic; Promega Corp.) yielding TIMP3-932luc. 

A fragment corresponding to the human E-cadherin promoter (−970 to +30) was isolated by PCR. Forward (5′-tcacgcctgtaatccaacac-3′) and reverse (5′-tcacaggtgctttgcagttc-3′) primers were designed with reference to the published sequence. ³³ The 1000-bp product was subcloned into a plasmid vector (pGEM-T; Promega Corp.) before being cloned into an expression vector (pGL3basic; Promega Corp.), as described earlier. This construct was designated Ecad-970luc. 

The sequence of both reporter constructs were verified by automated DNA sequencing. 

Transient Transfection Assays

Human RPE cells (hTERT-RPE1) were transfected with a lipid-based transfection reagent (Qiagen). Equal amounts (0.2 μg) of the TIMP3-932luc plasmid and rElf3a/pcDNA3.1, rElf3b/pcDNA 3.1, or the empty expression vector pcDNA3.1 was added. For control experiments, the E-cadherin promoter construct (Ecad-970luc) was used in place of TIMP3-932luc. Cells were harvested after 24 hours and lysed (25 mM Tris-phosphate [pH 7.8], and 10 mM MgCl₂, 1 mM EDTA, 1 mM dithiothreitol, 15% glycerol, and 1% Triton X-100). The cell extract was added to an assay buffer (lysis buffer without 1 mM dithiothreitol, with 1.25 mM adenosine triphosphate [ATP] and 75 μM luciferin; Promega Corp.) and luciferase activity estimated on a luminometer (Lumat LB 9507; EG&G Berthold, Bad Wildbad, Germany). Protein content was determined with a protein assay kit (Dc; Bio-Rad, Richmond, CA). The mean ± SD was obtained from triplicate samples and the data compared with ANOVA. 

Results

The ETS transcription factors ELF3, ESE3, ELF5, and PDEF are specifically expressed in epithelial cells. It is not known whether these factors are expressed in the RPE. To address this, specific primers were designed from the published sequences and RT-PCR performed on two human RPE cell lines: D407 and hTERT-RPE1. The expression of PDEF was not assessed in the RPE, because it is confined to the glandular epithelium. ²²  

As observed in Figure 1 , both human RPE cell lines contained specific message for ELF3, whereas neither expressed ESE3 or ELF5. The spontaneously transformed D407 cell line appeared to express significantly more ELF3 than hTERT-RPE1; however, semiquantitative PCR must be performed to estimate the difference in expression. Although ELF3 expression has been reported in tissues such as liver, lung, and kidney, ¹⁹ this is the first evidence of expression in the retina. The human breast (T47D) and prostate (PC3) cancer cell lines are known to express ELF5, ELF3, ESE3, and ELF3, ESE3, respectively, and were used as positive controls. ^{19 20}  

Although the human and murine ELF3 sequences have been determined, ^{18 19} the rat sequence is unknown. Because the rat has been a suitable model for the study of retinal anatomy and disease, elucidation of its Elf3 sequence would be invaluable. The full coding sequence of rat Elf3 was derived by using the RACE method and an ATG-spanning degenerative primer. Two splice variants of 1786 and 1855 bp were identified and designated Elf3a and Elf3b (Fig. 2A) . The two differ because of a 69-bp in-frame insertion, which probably arose from differential splicing. The sequences are predicted to encode 371 and 394 amino acid proteins, both of which contain the conserved ETS domain (Fig. 2 , shown in bold). Rat Elf3 displays a high homology to the human (90% identity) and murine (95% identity) sequences, reflecting the low interspecies variation of the ETS transcription factor family. 

The 69-bp insertion, present in rat Elf3b, was very similar to the insert found in the murine gene (Fig. 2B) . Although the murine insert did not have the initial three amino acids, those remaining displayed a 90% identity. Such high homology is rare in a differentially spliced region and may reflect a functional conservation. Although human ELF3 also shows a second splice variant, the insert is not homologous to the murine and rat Elf3b and occurs closer to the 3′ end of the transcript. ³⁴  

Elf3 expression in various rat tissues was quantified by real-time PCR, which provides an accurate measure of gene expression. ³⁵ Elf3 was detected in lung, kidney, liver, and retina, yet was absent in heart (Fig. 3A) . Previous reports have found that heart tissue does not express ELF3. ¹⁹ Although rat retina contains significant quantities of this transcription factor, its levels are approximately 50% lower than those found in lung and kidney. These tissues have previously been reported to contain high levels of ELF3. ¹⁹  

The proportions of the Elf3 splice variants in rat tissues were determined with RT-PCR (Fig. 3B) . Elf3b was the major (>95%) splice variant in rat lung, kidney, liver, and retina, with the 3a-to-3b ratios remaining constant. Neither splice variant was detected in heart tissue. Expression of Elf3a and -3b did not appear to have a developmental role in the rat retina, because proportions did not vary as a function of age (data not shown). The rElf3a and 3b expression constructs are included in Figure 3B and show sizes (269 and 338 bp, respectively) identical with the amplified product from the tissue samples. 

Because commercial Elf3 antibodies were unsuitable for use in rat tissues, a polyclonal antibody was produced after challenge from a C-terminal peptide to Elf3. The 19-amino-acid peptide (368Val–386Glu) is totally conserved among human, murine, and rat Elf3. The Western blot in Figure 4 shows the antibody to be specific for Elf3, with an approximate 43-kDa band detected for the human recombinant protein and the rat tissue samples. 

Because the RT- and real-time PCRs showed ELF3 in the retina, the polyclonal Elf3 antibody was used to determine whether expression was ubiquitous or confined to a particular cell type. Figure 5A shows the label mostly confined to the RPE, with only diffuse staining observed in the inner nuclear layer. Preincubation of the tissue section with a 200-fold excess of neutralizing peptide abolished Elf3 immunoreactivity (Fig. 5B) . Whereas the ETS factors ESE3, ELF5, and PDEF show some homology with the immunizing peptide, PDEF is confined to the glandular epithelium ²² and ESE3 and ELF5 are not present in the RPE (Fig. 1) . 

Alterations in the RPE have been associated with several diseases such as RP, SFD, and ARMD. Because ELF3 expression was shown to be restricted to the RPE, studies were undertaken to determine whether it regulates genes involved in retinal disease. Work focused on the TIMP3 gene, because it has been implicated in RP, SFD, and ARMD. ^{27 28} Computer analysis of the human and murine TIMP3 promoters revealed several consensus binding sites for ETS transcription factors. One of these, an inverted ETS binding site (Fig. 6A) , was found at positions −177 of the human and −174 of murine TIMP3 promoters. Because inverted ETS sites are known to play a role in gene regulation, ³⁶ transient transfection assays were used to assess the regulation of the TIMP3 promoter by ELF3. 

As observed in Figure 6B , the human TIMP3 promoter (TIMP3-932luc) showed low basal level activity when the empty expression vector (pcDNA3.1) was added. However, when constructs containing either Elf3a (rElf3a/pcDNA3.1) or -3b (rElf3b/pcDNA3.1) were included, a sixfold upregulation was observed (P = 0.005, P = 0.0016, respectively). The regulatory properties of the two splice variants were similar, with the difference not statistically significant (P = 0.3). The E-cadherin promoter construct (Ecad-970luc) was included as a negative control. Transfection with either Elf3a or -3b did not affect luciferase activity (P = 0.47, P = 0.1, respectively). 

Discussion

The role of transcription factors in ocular development and disease is of great interest. Although factors such as PAX6 and CHX10 have been implicated in ocular development of the eyes, the ocular expression and role of the ETS transcription factor family has not been explored in vertebrate organisms. This study is the first to show that the RPE expresses the epithelial specific ETS transcription factor ELF3. Furthermore, transient transfection studies show this factor to up regulate the important TIMP3 gene significantly. 

ELF3 was expressed in rat retinal tissue and in the human RPE cell lines, D407 and hTERT-RPE1. The other epithelial ETS factors, ELF5 and ESE3 were not found in the RPE, reflecting their restricted expression patterns. ^{20 21 22} Histochemical staining of the retina localized the expression predominantly to the RPE, with only weak staining observed in the inner nuclear layer, specifically the horizontal cell bodies. While this staining appeared to be specific, ELF3 expression has not been reported in neuronal cells. In fact, Northern and RNase protection analysis performed by Tymms et al. ¹⁹ showed that neither fetal nor adult brain expresses ELF3. Neuronal cells, however, express other ETS family members, such as Pet-1 and Er81. ^{37 38} Although cross reactivity with these related transcription factors may have resulted in the apparent horizontal cell localization, it was not observed in the Western blot analysis. 

Alternatively spliced products have been reported for many ETS genes, including Ets1, Elk1, and Tel. ^{39 40 41} The larger splice variant (Elf3b) described in this study has an extra 69 bp adjacent to the conserved N-terminal domain (pointed-like). The apparent preference for this transcript, coupled with the high insert homology to murine Elf3b, could reflect a functional significance. A protein motif search showed the insert to contain a potential casein kinase II (CK2) or protein kinase C (PKC) phosphorylation site (SQRD). Phosphorylation of ETS transcription factors is known to regulate many of their functions, such as DNA binding, transcriptional activation, transcriptional repression, and subunit association. ⁴² Furthermore, the proximity of the insert to the pointed-like domain, may potentiate this domain’s role in dimerization and transactivation. ^{43 44} The presence of a different human ELF3 splice variant may reflect a species-specific regulatory mechanism for this transcription factor. ³⁴  

As in previous studies, ELF3 was detected in the liver, lung, and kidney, but was absent from the heart. ¹⁹ Quantification of the relative copy numbers showed lung and kidney to express the highest levels of ELF3, followed by retina and liver. However, because ELF3 is expressed only in epithelial cells, the proportion of epithelia in the various tissues has a large impact on the apparent levels. Whereas the lung and kidney contain high proportions of specialized epithelial cells, the retina has only a single, noninvaginated layer of epithelial cells. It is therefore likely that retinal expression of ELF3 is underestimated. The existence of high expression levels in the RPE could reflect an important role for this factor in retinal gene regulation. 

Although the part played by ELF3 in tissue development is yet to be determined, its expression has been associated with mammary gland development and the occurrence of the epithelial phenotype. ^{45 46} ELF3 is known to regulate numerous genes such, as keratin 4, c-met and Erb-B2. ^{12 18 47} Recently, it was also shown to induce expression of the TGF-β receptor type II. ⁴⁸ The TGF-β family, including their receptors, are known to play important roles in the development and functioning of the retina. 

The in vitro transfections showed ELF3 to upregulate the TIMP3 gene, presumably through one of the identified ETS binding sites. Although previous reports have identified an ETS binding site in the response element for TIMP1, ⁴⁹ this is the first evidence of ELF3 regulation of the TIMP3 gene. TIMP3 is localized to the extracellular matrix and is able to inhibit all the major classes of matrix metalloproteinases (MMPs), ⁵⁰ making it a very powerful modulator of extracellular matrix (ECM) turnover. The ELF3 regulation of TIMP3 would have implications for development, morphogenesis, and tissue remodelling in epithelial cells. Furthermore, because alterations in ECM have been observed in several diseases such as cancer, arthritis, and cardiovascular disease, ⁵¹ this regulatory mechanism could be of great interest. 

In the retina, the accumulation of TIMP3 is thought to play a significant role in the progression of the phenotypically similar SFD and ARMD. ^{27 28} Although the development of SFD has been attributed to several point mutations within the TIMP3 gene, the mechanisms underlying ARMD have yet to be elucidated. ELF3 regulation of TIMP3 may provide such a mechanism. If the regulation observed in vitro is representative of in vivo regulatory properties, then increases in ELF3 expression within the retina would result in an accumulation of TIMP3. Such an accumulation may play a role in the thickening of Bruch’s membrane and the progression of ARMD. A similar ELF3-related pathway could also explain the increased TIMP3 levels observed in patients with RP. ²⁷ Although this study shows TIMP3 to be under the control of ELF3 in vitro, further work is needed to determine whether such regulation occurs in vivo. Furthermore, the existence of ELF3 within the retina raises the possibility that other genes involved in retinal disease may be under its control. 

This study has shown the RPE to express significant quantities of the ETS transcription factor ELF3. Although the role of this factor in retinal development is unclear, there are a number of critical genes that it may regulate. The upregulation of TIMP3 by ELF3 may have implications for the ECM changes observed in diseases such as RP and ARMD. 

 Supported by grants from The National Health and Medical Research Council of Australia and from the Lions Vision Research Fellowship Fund.

 Submitted for publication November 27, 2001; revised February 11, 2002; accepted February 28, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Specific PCR Primers Used for the Amplification of the ETS Factors ELF3, ESE3 and ELF5

Target Gene	Primers			Product Size (bp)
		Upstream	Downstream
hELF3	gatggggccaccctctgcaattgtg	ccctcagttccgactctggagaacctc	827
hESE3	tttcccacccagaatctttag	ccaaagtattggcagcttcag	953
hELF5	gaaagcctcctctttggacc	gcaatagacattcgaaaggctt	890
rElf3	gttgaccctgaacaaccaac	cttcgggacctcacctcca	272, 341

 Primers for the human genes were designed from previously published sequences, and those for rat were determined as outlined in Materials and Methods. The sizes of the two splice variants amplified from rat Elf3 are indicated.

Figure 1.

 

Expression of ETS transcription factors in human RPE cell lines. Expression of ELF3, ESE3, and ELF5 was assessed by RT-PCR in the RPE cell lines D407 and hTERT-RPE1. The cancer cell lines T47D and PC3 were used as positive controls for the ETS transcription factors. A 1-kb (250–10,000 bp) DNA ladder is included between hTERT and t47D.

Figure 2.

 

cDNA coding and amino acid sequence for rat Elf3. (A) The complete nucleotide and corresponding amino acid (single-letter code) sequences are shown. Shaded area: inserted bases identified in the longer Elf3 transcript (Elf3b). The amino acids comprising the ETS domain are shown in bold, and a putative polyadenylation sequence (AATAAA) is also identified. (B) Comparison of the murine and rat Elf3b inserts. Single-letter codes are shown for the extra amino acid sequences found in murine (m) and rat (r) Elf3b. Boxed regions: amino acid identity; dashes: amino acids not present in the mouse insert.

Figure 3.

 

Elf3 expression in rat tissues. (A) Real-time PCR analysis of Elf3 expression in rat tissues. Rat heart (lane 1), liver (lane 2), lung (lane 3), kidney (lane 4), and retina (lane 5) were assessed for Elf3 expression. Copy numbers were estimated after comparison to an external standard (cloned rElf3b). (B) RT-PCR of Elf3a and -3b in various rat tissues. The expression of the two Elf3 splice variants was assessed in heart (lane 1), liver (lane 2), lung (lane 3), kidney (lane 4), and retina (lane 5). Also included are the rElf3b/pcDNA3.1 and rElf3a/pcDNA3.1 expression constructs (lanes 6 and 7, respectively).

Figure 4.

 

Western blot analysis of rat Elf3. Recombinant human Elf3 (pQE30; lane 1), rat retina (lane 2), and kidney (lane 3) extracts were separated by 10% SDS-PAGE and probed with the polyclonal antibody derived from the C-terminal peptide. Molecular weight markers are indicated at left.

Figure 5.

 

Elf3 distribution in the rat retina. The polyclonal Elf3 antibody raised against the C-terminal peptide sequence was used to stain rat retinal sections. (A) Specific labeling of the RPE layer with the Elf3 antibody. (B) The control incubation with the immunizing peptide. INL, inner nuclear layer; IPL, inner plexiform layer; ONL, outer nuclear layer; OPL, outer plexiform layer.

Figure 6.

 

Regulation of the TIMP3 promoter by Elf3. (A) Identification of a putative Elf3 consensus binding sequence within the human and murine TIMP3 promoters. The region corresponds to −177 to −147 (human promoter) and −174 to −144 (murine promoter). (B) Regulation of the human TIMP3 promoter by rat Elf3a and Elf3b. Human RPE cells (hTERT-RPE1) were cotransfected with TIMP3-932luc and rElf3a/pcDNA 3.1 ( ), rElf3b/pcDNA 3.1 ( ), or pcDNA3.1 (□). For control experiments TIMP3-932luc was replaced with Ecad970-luc. Triplicate samples were used, and luciferase activity was corrected for protein content. Multiples of upregulation are relative to the empty expression vector (pcDNA3.1).

The authors thank David Nikolic-Paterson, Department of Nephrology, Monash Medical Centre, Clayton, Australia, for assistance with the histochemical staining of the rat retinal sections and Paul Hertzog, Center for Functional Genomics and Human Disease, Monash Institute of Reproduction and Development, Clayton, Australia, for donation of the T47D and PC3 total RNA samples.