TEM7 (PLXDC1) in Neovascular Endothelial Cells of Fibrovascular Membranes from Patients with Proliferative Diabetic Retinopathy

Abstract

purpose. Proliferative diabetic retinopathy (PDR) results from the formation of fibrovascular membranes (FVMs) in the posterior fundus that can lead to a severe decrease of vision. Tumor endothelial marker 7 (TEM7) is a protein that is highly expressed in the endothelial cells of tumors, but whether it plays a role in FVMs is unknown. The purpose of this study was to determine whether TEM7 is associated with the formation of FVMs.

methods. FVMs were obtained during vitrectomy from patients with PDR. RT-PCR was performed to determine the level of expression of the mRNA of TEM7. The splice variants of TEM7 were identified by direct sequencing. Immunohistochemical analyses and in situ hybridization was performed to determine the sites of TEM7 in the FVMs.

results. The level of the mRNA of TEM7 was high in 10 of 10 FVMs but was barely detectable in the five idiopathic epiretinal membranes. Direct sequencing of subcloned TEM7 PCR products revealed several splice variants (intracellular, secreted, and membrane-bound forms of TEM7) in the FVMs. Immunohistochemical analysis showed a colocalization of TEM7 and CD34, an endothelial cell marker, in most of the neovascular endothelial cells in the FVMs. Immunoelectron microscopy revealed that membrane-bound TEM7 was expressed on the luminal surfaces of the vascular endothelial cells of FVMs.

conclusions. This study indicates that TEM7 may play a significant role in the proliferation and maintenance of neovascular endothelial cells in the FVMs. If correct, TEM7 may be a molecular target for new diagnostic and therapeutic strategies for PDR.

Diabetic retinopathy is one of the leading causes of decreased vision and blindness in industrialized countries. Much of the retinal damage that characterizes advanced proliferative diabetic retinopathy (PDR) results from retinal neovascularization. ^{1 2} When the newly formed vessels are associated with fibrous proliferations that form fibrovascular membranes (FVMs), traction retinal detachments can develop, resulting in potentially severe loss of vision. ³  

Despite recent improvements in vitreous surgical techniques, panretinal photocoagulation, and antivascular endothelial growth factor drugs such as bevacizumab, the prognosis for patients with PDR is still poor, especially for those with advanced PDR at the proliferative stage. ⁴ It is therefore necessary to develop better diagnosis and treatment techniques based on the exact pathogenesis of FVMs. 

Retinal neovascularization is regulated by a balance of stimulators and inhibitors of angiogenesis that are thought to tightly control the normally quiescent capillary vasculature. When this balance is upset, as in PDR, capillary endothelial cells proliferate. Several angiogenesis-related factors, such as basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), ⁵ tumor necrosis factor (TNF)- α, ⁶ interleukin (IL)-1β, ⁷ angiogenin, ⁸ angiopoietin-2, ⁹ hepatocyte growth factor (HGF), ¹⁰ monocyte chemoattractant protein (MCP)-1, ¹¹ nuclear factor κ-B, ¹² and activator protein-1, ¹³ have been detected in FVMs and vitreous fluid collected from PDR patients. Our laboratory has demonstrated that the level of IL-8 is elevated in the vitreous fluid of patients with PDR. ¹⁴ In addition, we have found that TNF-α can up-regulate the expression of the mRNAs of IL-8, MCP-1, and bFGF in retinal glial cells and microvascular endothelial cells in an in vitro system. ^{15 16}  

Retinal neovascularization can be suppressed by inhibitory molecules, such as pigment epithelium-derived factor (PEDF) ¹⁷ and thrombospondin-1. ¹⁸ However, the exact mechanism of how retinal neovascularization develops is not fully known. 

Recent technological advances in cellular fractionation and genomics have led to the identification of several markers preferentially expressed on vascular endothelial cells of human tumors. ¹⁹ Among these markers are the tumor endothelial markers (TEMs), which are a group of cell surface proteins strongly expressed on the endothelial cells of various cancer cells. ²⁰ Of these cell surface markers, TEM7, also known as plexin domain-containing 1 (PLXDC1), is especially interesting because it is the most abundant isoform among the TEMs. ²¹ TEM7 protein is overexpressed in the blood vessels of human solid tumors such as colon, lung, and esophageal cancers. ²² Moreover, high expression of TEM7 is associated with metastasis and poor survival of patients with osteogenic sarcoma. ²³  

The full-length form of TEM7 has sequence characteristics of cell surface proteins, including signaling peptides, plexin-semaphorin-integrin (PSI) domain, and transmembrane domain(s). ²² In addition, TEM7 is expressed as a complex pattern of transcripts derived by alternative splicing with potentially different activities and biological functions. These transcripts are predicted to be intracellular (TEM7-I), secreted (TEM7-S), or on the cell surface membrane (TEM7-M) of the endothelial cells of tumors. ²²  

Recently, we finished sequencing 7000 random sequences of the cDNA library from FVMs associated with PDR and found that TEM7 is expressed in FVMs (YY, SY, unpublished data, 2007). We hypothesized that a determination of the specific role played by TEM7 in the formation of FVMs and intensive studies of TEM7 will aid in our understanding of retinal angiogenesis and may lead to the development of new diagnoses and therapeutic agents. 

Thus, the purpose of this study was to determine whether the mRNA and protein of TEM7 is expressed in the neovascular endothelial cells of FVMs surgically removed from patients with PDR. The possible biological role of TEM7 in retinal neovascularization is discussed. 

Subjects and Methods

This study was approved by the Ethics Committee of the Kyushu University Hospital, and the FVM specimens were handled in accordance with the Declaration of Helsinki. All patients gave informed consent before inclusion in the study. Criteria for exclusion were age older than 80 years, renal or hematologic disease, uremia, previous chemotherapy, life support, and the fewest possible chronic pathologic conditions other than diabetes. ¹²  

FVMs were surgically removed from the eyes of patients with type 2 diabetes with PDR (15 eyes) undergoing pars plana vitrectomy with membrane peeling. As controls, five patients with idiopathic epiretinal membranes (ERMs) underwent ERM resectioning. Membranes were dissected from the retinal surface with horizontal scissors. Neovascular endothelial cells are among the main components of FVM in eyes with PDR, whereas an idiopathic ERM is a fibroglial membrane that lacks neovascular endothelial cells. 

The FVM specimens obtained from 10 patients with PDR (age, 62.5 ± 14.6 years; duration of diabetes, 15.2 ± 8.6 years) and five control subjects with ERM (age, 72.7 ± 14.8 years) were processed for reverse transcription–polymerase chain reaction (RT-PCR) analysis and subcloning. The remaining five FVM specimens (age, 57.2 ± 18.9 years; duration of diabetes, 18.1 ± 12.5 years) were processed for immunohistochemistry and in situ hybridization. One normal eye was also obtained by orbital surgery and fixed immediately after enucleation in 4% paraformaldehyde for immunohistochemistry. 

RNA Extraction and Semiquantitative Reverse-Transcription–Polymerase Chain Reaction

All the resected tissues for RT-PCR were snap frozen and stored at −80°C. For the preparation of total RNA, the tissue was homogenized using a kit (MagNA Lyser Green Beads kit; Roche Applied Science, Mannheim, Germany) according to the manufacturer’s instructions. Total RNA was extracted from the tissue homogenate with reagent (Trizol; Qiagen, Germantown, MD) and exposed to DNase (RNase-free DNase set; Qiagen) to eliminate potential genomic DNA contamination. 

Synthetic oligonucleotide primers based on the cDNA sequences of TEM7, IL-8, VEGF, VEGF receptor 2 (VEGFR2), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were prepared as follows: for TEM (expected amplicon exon 7–13), 5′- CGGATGCCTTCATGATTCTC-3′ (TEM7F) and 5′- AGCACGATGCCCACGATGGT-3′ (TEM13R); for GAPDH, 5′-GAGTCAACGGATTTGGTCGT-3′ and 5′-CTTGATTTTGGAGGGATCTCGC-3′; for IL-8, 5′-CTGCGCCAACACAGAAATTA-3′ and 5′-ATTGCATCTGGCAACCCTAC-3′; for VEGF, 5′-TGCCTTGCTGCTCTACCTCC-3′ and 5′-TCACCGCCTCGGCTTGTCAC-3′; and for VEGFR2, 5′-GATGTGGTTCTGAGTCCGTCT-3′ and 5′-CATGGCTCTGCTTCTCCTTTG-3′. 

PCR was carried out semiquantitatively essentially as described. ^{14 24} First-strand cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase (Invitrogen, San Diego, CA). RNAs (1 μg) and random hexamers (100 ng) were denatured at 65°C for 10 minutes and were added to the reverse transcription mixture according to the instructions of the manufacturer. After incubation at 37°C for 1 hour, cDNAs from the reverse transcription mixture was subjected to PCR in a 10-μL volume containing 5 pmol of the primer pair and 0.5 U Platina Taq (Applied Biosystems, Foster City, CA) using a DNA thermal cycler (Applied Biosystems). DNAs were denatured for 9 minutes at 95°C followed by 35 PCR cycles. Each cycle included a 30-second denaturation at 94°C, 30-second primer annealing at 55°C, and 45-second polymerization at 72°C. An 8-μL aliquot of each RT-PCR reaction mixture was analyzed by electrophoresis on a 2% agarose gel and stained with ethidium bromide. The density of the ethidium bromide luminescence was measured by a CCD image sensor (Densitograph AE-6920M; Atto Co., Tokyo, Japan). 

Subcloning of TEM7 PCR Products

RT-PCR was performed to amplify the region spanning exon 7 to exon 13 using TEM7F-TEM13R primer pairs to detect exons specific for TEM7-S1, -S2, -I, and -M. For amplification of the region around the 5′ transcription start sites to detect signal sequences (representing TEM7-S1, -S2, and -M), TEM1BF (5′-GCCGCAGCCTGAGCCAGGG-3′)-TEM2R (5′-ACCCTGGTCCTGTTATCTGG-3′) primer pairs were used. To detect a lack of signal sequence that reflects the transcript encoding TEM7-I, TEM1AF (5′-GGCATCTTCGTGTCCTGACC-3′)-TEM2R primer pairs were used. The amplicons were subcloned into a transmembrane protein visualization tool (TOPO-2; Invitrogen) according to the manufacturer’s protocol. ^{14 25} This was followed by nucleotide sequencing using T7 primer with the Taq Dyedeoxy Terminator Cycle Sequencing Kit (Applied Biosystems). Sequencing reactions were resolved on an automated sequencer (ABI 3130; Applied Biosystems). ^{26 27 28 29 30 31 32 33}  

Murine Model of Ischemia-Induced Retinal Neovascularization

All experimental procedures on animals were performed according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. A reproducible model of ischemia-induced retinal neovascularization has been described in detail. ^{15 16 34 35} Briefly, litters of postnatal day (P)7, C57BL/6 pups with their mothers were exposed to 75.2% oxygen for 5 days and then returned to room air at age P12, as described. ^{15 16 35} The mice were killed by an overdose of sodium pentobarbital. 

Immunohistochemistry

Immunohistochemistry was performed essentially as described. ^{36 37 38} Surgically resected membranes were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) and embedded in paraffin. Thick sections (3 μm) were cut, and after removal of the paraffin, they were rehydrated, blocked, and incubated for 1 hour at room temperature with a monoclonal antibody against CD34 (1:50 dilution; NCL Dako, Glostrup, Denmark) or a monoclonal antibody against TEM7 (1:1000 dilution; Imgenex, San Diego, CA). The bound antibody was made visible by a conventional avidin-biotin-peroxidase protocol with 3-amino-9-ethylcarbazole as the substrate. For negative controls, mouse nonimmune IgG was used as the primary antibody. 

Dual-Color Immunofluorescence Immunohistochemistry

Dual-color immunofluorescence staining was performed on paraffin sections ³⁹ by staining with anti-TEM7 clone IM193 and polyclonal anti-CD34 (1:50 dilution; Santa Cruz Biochemicals, Santa Cruz, CA). TEM7 was detected with an FITC-conjugated anti-mouse IgG (1:400 dilution; Santa Cruz Biochemicals), and CD34 was detected using a biotin anti-rabbit IgG (1:400 dilution; Santa Cruz Biochemicals) followed by rhodamine-streptavidin (Vector Laboratories, Burlingame, CA). Sections were examined with a Leica (Wetzlar, Germany) fluorescence microscope. 

Immunoelectron Microscopy

A postembedding approach was used. Fresh FVMs were fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in PBS (pH 7.2) overnight at 4°C. ³⁵ After dehydration in a graded ethanol series, the samples were infiltrated and embedded in LR white resin. Ultrathin sections were cut at 70 nm and mounted onto nickel Formvar-coated grids. The sections were blocked in 1% BSA in PBS for 30 minutes at room temperature. After incubation in IM193 monoclonal antibody overnight at 4°C and 10 nm gold-conjugated anti–mouse secondary antibody (British BioCell International, Cardiff, UK) for 1 hour at room temperature, sections were stained with uranyl acetate followed by lead citrate. All grids were examined and photographed on electron microscope (2000EX; JEOL, Peabody, MA) operating at 100 kV. 

Generation of Riboprobes

The cDNAs for TEM7 were generated by RT-PCR. The following primer pairs were designed and used for human TEM7: 5′-GCAGATCACCATAGCAACTGG-3′ and 5′-AGGTCAGGTCTGAGGACATGC-3′. The following were designed and used for mouse TEM7: 5′-GACAAGTCACAGCACAGAGAAGTG-3′ and 5′-CAATGAAGCAAAGGACCAAGG-3′. 

PCR conditions were 10 minutes at 94°C, followed by 35 cycles at 72°C for 20 seconds, 94°C for 30 seconds, and 60°C for 30 seconds, with a final extension step at 72°C for 5 minutes The resultant PCR fragments were subcloned into TOPO-2 (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Their nucleotides were sequenced, and the inserts were identified with human and mouse TEM7. Clones with both orientations were selected so that the same RNA polymerase (T7) could be used to generate antisense and sense riboprobes. The templates were linearized with EcoRV, and in vitro transcription was performed using digoxigenin (DIG)-labeled uridine triphosphate (Roche Biomedical Systems, Indianapolis, IN) according to the manufacturer’s protocol. 

In Situ Hybridization

FVMs and mouse retinas were fixed in 4% paraformaldehyde for 2 hours and embedded in paraffin. In situ hybridization was performed, as described in detail. ^{16 39 40} Briefly, samples were rehydrated and treated with proteinase K, followed by refixing in 4% paraformaldehyde in PBS for 10 minutes. Tissues were then acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine-HCl buffer (pH 8.0) for 10 minutes, dehydrated, and dried. 

Hybridization was performed with fresh hybridization buffer (600 mM NaCl, 10 mM Tris-HCl [pH 7.6], 5 mM EDTA [pH 8.0], 1× Denhardt solution, 50% formamide, 17 mg/mL yeast transfer RNA, and 10% weight/volume dextran) plus an approximately diluted sense or antisense digoxigenin-labeled RNA probe for 12 to 16 hours at 55°C. After hybridization, the samples were washed to remove nonspecific bound RNA probe, and immunologic detection was performed by anti-DIG Fab fragments conjugated to alkaline phosphatase, as described in the system protocol (Roche Biomedical Systems, Indianapolis, IN). 

Results

Upregulation of TEM7 in FVMs

It has been demonstrated that TEM7 is preferentially expressed in neovascular endothelial cells of tumors and in nonpathologic angiogenic tissues. ²² We detected the expression of the mRNA of TEM7 in FVMs by expressed sequence tag (EST) analysis (SY, unpublished data, 2007). The mRNA of TEM7 was detected in 10 of 10 FVMs obtained from PDR patients but was barely detected in the five idiopathic ERMs (control; Fig. 1 ). In addition, RT-PCR with the 7F and 13R primer pairs yielded multiple bands suggesting the presence of splice variants of TEM7 in the FVMs. The mRNAs of IL-8 were detected in 10 of 10 FVM specimens, VEGF in 8 of 10 FVM specimens, and VEGFR2 in 9 of 10 FVM specimens. In contrast to TEM7, these 3 angiogenic molecules were also upregulated in some of the control idiopathic ERMs (3 of 5, 3 of 5, and 2 of 5, respectively). 

Presence of Several Splice Variants of TEM7 in FVMs

It has been predicted that variants of TEM7, the intracellular (TEM7-I), secreted (TEM7-S1 and -S2), and cell surface membrane (TEM7-M) are present on the endothelial cells of tumors. ²² These variants are derived by alternative splicing. The secreted form of TEM7 uses the stop codon of either exon 10 (TEM7-S2) or exon 11A (TEM7-S1), whereas TEM7-M uses exon 11B. The presence of multiple bands in the RT-PCR using 7F and 13R primer of TEM7 (Fig. 1)led us to search for the splice variants of the TEM7 gene. Direct sequencing of the subcloned PCR products of TEM7 amplified by primer pairs 7F-13R from 10 PDR specimens revealed that TEM7-S1 (Fig. 2A) , TEM7-S2 (Fig. 2B) , and TEM7-M (Fig. 2C)were also present in all 10 FVMs. 

TEM7-M, TEM7-S1, and TEM7-S2 use exon 1B, which contains a signal sequence to initiate transcription, whereas TEM7-I uses exon 1A, which lacks a signal sequence and skips exon 1B. ²² To determine the splice variants around the 5′ transcription initiation site, we designed the forward primers for RT-PCR from sequences within exons 1A and 1B and reverse primers in exon 2 to discriminate alternative uses of these exons. Subsequent direct sequencing of subcloned PCR products from 10 PDR specimens revealed that each variant, which contained either exon 1B (Fig. 2D)or exon 1A (Fig. 2E) , did exist in all 10 FVMs. Thus, TEM7-I is also present, in addition to TEM7-M, TEM7-S1, and TEM7-S2 in the FVMs. 

Localization of mRNA and Protein of TEM7 in FVMs

To determine the location of the protein of TEM7 in FVMs, we next stained the FVM sections with an anti-TEM7 monoclonal antibody (IM193), which specifically detects TEM7-M, and an antibody to CD34, an endothelial cell marker (Fig. 3) . Consistent with previous results on the staining patterns of tumors with neovascularization, ²² the antibody specifically labeled the neovascular endothelial cells in the FVMs (Figs. 3A 3B 3D 3E)but not the normal retinal vascular endothelium (Figs. 3G 3H) . Nonimmune IgG did not label any cellular structures (Figs. 3C 3F 3I) . TEM7 protein showed similar expression in all the FVM specimens from the five PDR subjects examined, and the data from two representative patients are shown (Figs. 3A 3B 3C 3D 3E 3F) . 

After confirming the sites of TEM7 protein in the FVMs, we next performed in situ hybridization using digoxigenin-labeled antisense riboprobe specific to TEM7. Consistent with the results of immunohistochemistry, the antisense probe hybridized with the mRNA of TEM7 in the vascular endothelial cells (Fig. 3J) , whereas the sense probe did not (Fig. 3K) . We then double immunofluorescently stained the FVMs with anti-TEM7 antibody and an antibody to CD34. These results showed colocalization of CD34 and TEM7 in most of the neovascular endothelial cells in the FVMs (Fig. 4) . 

To determine a more exact site of the TEM7 protein within the endothelial cells, we performed immunoelectron microscopy using monoclonal anti-TEM7 antibody (IM193). Electron microscopy revealed that TEM7-M was expressed at the tight junctions and at the luminal surfaces of the vascular endothelial cells (Fig. 5) . 

Low Levels of Expression of mTEM7 in Neovascular Endothelial Cells in a Rodent Model of Neovascularization

It has been reported that human TEM7 is abundantly expressed in human tumor endothelial cells, whereas its mouse counterpart, mTEM7, cannot be detected in neovascular endothelial cells of mouse. ^{19 40} To determine whether mTEM7 is expressed in the endothelial cells of mouse during neovascularization, neovascularization was produced by oxygen-induced retinopathy and used for in situ hybridization. We were able to detect low levels of mTEM7 in the vessels (Fig. 6) , despite the fact that its human counterpart TEM7 was abundantly expressed in FVMs when assessed by the same in situ hybridization methods (Fig. 3G) . On the other hand, higher signals were seen in retinal ganglion cells and cells in the inner nuclear layer in the mouse model of oxygen-induced retinopathy and control P17 mouse (Figs. 6A 6C) . As a positive control, we stained the mouse cerebral cortex, and positive staining was detected (Fig. 6D) . Therefore, the low level of mTEM7 signal in retinal neovascular vessels was unlikely caused by technical problems. 

Discussion

Our results demonstrated that the mRNA of TEM7 is expressed in the FVMs collected from human eyes with PDR. In addition, in situ hybridization and immunohistochemical analysis of the FVMs showed positive staining in the neovascular endothelial cells of the FVMs. In contrast to the extensively studied angiogenic factors, such as VEGF and their receptors whose expressions are also elevated in some nonangiogenic ERMs, TEM7 expression was specifically enhanced in neovascular endothelial cells in the FVMs associated with PDR (Fig. 1) . These observations indicate that TEM7 may play a critical role in the proliferation and maintenance of neovascular endothelial cells in the FVMs. 

The presence of TEM7 in neovascular endothelial cells of FVMs further supports the idea that they are markers of neoangiogenesis, including the endothelial cells found during wound healing and in the corpus luteum. ²⁰ Therefore, our results indicate that human retinal neovascularization exploits the same genes used by the developing normal endothelial cells and tumor angiogenesis. It would thus be interesting to determine whether other neovascular endothelial cells express TEM7 in other diseased eyes, such as corneal and choroidal neovascularization. 

The function of TEM7 has not been completely determined. The sequencing of subcloned PCR products from the cDNAs of FVMs confirmed the predicted structure of TEM7, which contains a type 1 transmembrane protein with a large extracellular region, a hydrophobic transmembrane domain, and a cytoplasmic tail (data not shown). ⁴⁰ In support of this, electron immunohistochemistry showed that the membrane-bound TEM7 was on the luminal surfaces of the tumor endothelial cells (Fig. 5) . These findings suggest that the extracellular domain of TEM7 containing a PSI domain plays a role in ligand binding or binding to an extracellular matrix protein, and the intracellular tail may initiate downstream signaling. ²¹ Additional studies on the role of TEM7 and the identification of the ligand of TEM7 in the development of retinal neovascularization remain to be performed. 

Alternative RNA splicing plays a major role in modulating gene function by expanding the diversity of expressed mRNA transcripts, and it may determine cell fate in numerous contexts. ^{41 42} The recent genomewide transcriptome analysis revealed that the differential use of endoplasmic reticulum signaling peptides and transmembrane domains are common occurrences in the different protein output of transcripts. ⁴³ Consistent with this, sequencing of TEM7 PCR products revealed that, in addition to TEM7-M, other forms of TEM7, including TEM7-I and TEM7-S, were also present in the fibrovascular endothelial cells (Fig. 2) . These different transcripts of TEM7 may be used to vary the membrane organization of TEM7 proteins. ⁴³ This can then lead to the deregulation of crucial cellular processes such as adhesion, proliferation, differentiation, death motility, and invasion. ⁴⁴ How the alternative splicing of TEM7 is related to retinal neovascularization awaits further studies. 

In addition to the biological interest of the TEM7 gene, the presence of a secreted form of TEM7 in FVMs raises a clinically interesting possibility for the development of novel serum markers that can be used to predict retinal angiogenic activity, which may reflect disease severity in patients with PDR. 

Considerable effort has been invested recently to develop agents that block the formation of new blood vessels. For example, bevacizumab, a selective VEGF inhibitor, was recently found to be effective in the regression of retinal and iris neovascularization secondary to PDR, but because of its cytostatic property, its effect may be limited to established vasculature. ⁴⁵ Therefore, it has become apparent that targeted destruction of the established vasculature is another avenue for therapeutic opportunities. ⁴⁶ In this regard, the presence of membrane-bound TEM7 on the luminal surfaces of neovascular endothelial cell is of interest (Fig. 5) . Recently, cortactin ²² and nidogen ⁴⁷ were identified as proteins capable of binding to the extracellular region of TEM7. Because the retinal vessels are the only vessels that can be observed in situ, it may be possible to use a more selective ligand-based neovascular endothelial cell targeting strategy by delivering the bioactive molecules to the blood-retinal neovascular endothelial interface. For example, photodynamic therapy using such TEM7-binding partners labeled with photosensitive biomolecules may open new possibilities for a lower invasive therapy of retinal neovascularization. This possibility is of interest because, unlike choroidal neovascularization as a target of available photodynamic therapies, retinal neovascularization associated with PDR generally occurs in the midperipheral retina, avoiding possible severe side effects of damage to the central visual function. 

The low level of expression of mTEM7 in mouse retinal neovascular endothelial cells (Figs. 6A 6C)is also notable. This is consistent with previous findings that the expression of the mRNA of mTEM7 in the tumor endothelium of mice seems to be variable and highlights potentially important differences between mouse and human angiogenesis. ⁴⁰ This indicates that TEM7 may be involved in a species-specific cascade associated with retinal neovascularization. This could partly explain the pronounced effects of angiogenesis inhibitors observed in rodent models compared with humans. ⁴⁰ Moreover, rodent TEM7 is expressed in some neuronal cells, such as retinal ganglion cells (Figs. 6A 6C) , suggesting a specific role in several types of neuronal cells of the nervous system. Our observations agree with the previous suggestions that TEM7 plays a role in some neuronal populations of the vertebrate central nervous system. ⁴⁸ Differential expression patterns in humans and rodent models indicate the importance of considering models other than rodent models ⁴⁹ to examine the possibility of TEM7 as a therapeutic and diagnostic molecular target to combat retinal neovascularizations associated with PDR. 

 Supported in part by grants from the Ministry of Education, Science, Sports and Culture, Japan (TI, SY); the Japan Diabetes Foundation (SY); and the Kanae Foundation for Life and Socio-medical Science (SY).

 Submitted for publication September 25, 2007; revised February 3, 2008; accepted May 1, 2008.

 Disclosure: Y. Yamaji, None; S. Yoshida, None; K. Ishikawa, None; A. Sengoku, None; K. Sato, None; A. Yoshida, None; R. Kuwahara, None; K. Ohuchida, None; E. Oki, None; H. Enaida, None; K. Fujisawa, None; T. Kono, None; T. Ishibashi, None

 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.

 

Figure 1.

 

RT-PCR analysis of TEM7, IL-8, VEGF, VEGFR2, and GAPDH in fibrovascular membranes derived from patients with proliferative diabetic retinopathy (PDR; patients 1–10) and in epiretinal membranes from eyes with idiopathic epiretinal membranes (iERMs; patients 11–15). After 35 cycles, 8 μL each sample was electrophoresed through a 2% Tris-acetate-EDTA agarose gel, and the fractionated products were stained with ethidium bromide. Note the distinct high expression of the mRNA of TEM7 in the fibrovascular membranes derived from patients with PDR compared with control iERMs.

Figure 2.

 

Nucleotide sequence showing several splice variants of TEM7 expressed in FVMs. Vertical line represents exon boundaries. (A) Subcloned sequence of the PCR amplicon spanning exon 7 to exon 13 of the TEM7 cDNA from FVM contains exon 9 just before exon 11A demonstrating the existence of TEM7-S1 in the FVM. (B) Another subcloned sequence of the same amplicon contains exon 10 after exon 9, which is specific to TEM7-S2. (C) Another subcloned sequence of the same amplicon shows transcripts consisting of exon 9 and exon 11A representing TEM7-M. (D) Subcloned sequence of PCR amplicon spanning exon 1B and exon 2 of the TEM7 cDNA from the FVM confirmed exon 1B, which includes signal sequence. Sequencing from antisense strand is shown. (E) Another subcloned sequence of PCR amplicon spanning exon 1A and exon 2 of the TEM7 cDNA demonstrating a skipping of exon 1B, resulting in transcripts that lack signal sequence and that reflect transcript encoding TEM7-I. Sequencing from antisense strand is shown.

Figure 3.

 

TEM7 expression in fibrovascular membranes (FVMs) from a 42-year-old patient with a 7-year history of diabetes (A–C) and a 56-year-old patient with 15-year history of diabetes (D–F, J, K) and in the normal retina from a 46-year-old control subject (G–I). (A, D, G) Immunohistochemical staining with IM193 antibody to detect TEM7. (A, D) Positive staining is seen in neovascular endothelial cells. (G) No signals are seen in normal retinal vasculature. (B, E, H) Anti-CD34 antibody was used as a control for vessel staining. (C, F, I) Signals are not seen with negative control of FVM stained with mouse nonimmune IgG. (J, K) In situ hybridization was performed with antisense (J) or sense (K) probes specific for human TEM7. (J) Positive staining is seen in vascular endothelial cells. (K) No signals are seen with TEM7 sense probe. Scale bars, 50 μm.

Figure 4.

 

Double staining for TEM7 and CD34 in the FVM. (A) Neovascular endothelial cells are visible after specific staining with CD34 in the FVM. (B) Specific staining for TEM7 in the same section shows an identical staining pattern. (C) Double staining for TEM7 and vascular endothelial cells in the same sample shows positive cells for both antibodies. The yellow staining is caused by the overlapping of the red and the green colors, showing colocalization of TEM7-M with the pan-endothelial marker CD34. Sale bars, 50 μm.

Figure 5.

 

TEM7 staining for transmission electron microscopy with IM193 antibody to detect TEM7-M. Staining is observed on the tight junctions (arrows) and on the luminal surfaces of endothelial cells (ECs; arrowheads). Scale bar, 200 nm.

Figure 6.

 

Retinal sagittal section of mice with oxygen-induced retinopathy. Animals were subjected to 7 days of normoxia, followed by 5 days of hyperoxia, and then by 5 days of normoxia. In situ hybridization was performed with antisense (A, C, D) or sense (B) probes specific for mouse (m)TEM7. (A) Retina 5 days after ischemia (P17). Staining is observed predominantly in the ganglion cell layer and cells in the INL, and a lower level of expression on the neovascular endothelium can be seen (arrowheads). (B) Signals are not seen with mTEM7 sense probe. (C) Retina of normal control animal exposed only to room air (P17). Staining is observed in the ganglion cell layer and cells in the inner nuclear layer. (D) Positive signals are seen in the cells of the mouse brain (cerebral cortex). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. Scale bar, 50 μm.

The authors thank Akifumi Ueno, Yasuaki Hata, Yusuke Murakami, Hiroshi Yoshikawa, and Yasutaka Mochizuki (Kyushu University, Fukuoka, Japan) for collecting tissue samples; Brad St. Croix (National Cancer Institute, Frederick, Maryland) for several helpful comments on TEM7; and Mari Imamura and Naomi Higuchi (Kyushu University) for their excellent technical assistance.