Effects of Tenascin-C on Normal and Diabetic Retinal Endothelial Cells in Culture

Abstract

purpose. Tenascin-C (TN-C) is expressed in embryogenesis, tissue remodeling, and healing. It is up-regulated in retinas of patients affected by diabetic retinopathy (DR). Because TN-C may promote neovascularization, its potential angiogenic effects were examined in vitro in normal and diabetic retinal endothelial cells (RECs).

methods. Bovine and human RECs were cultured on plastic or reconstituted basement membrane (BM) matrix. Production of TN-C, capillary-like tube formation, secondary sprouting, and cell migration, survival, and proliferation were measured with or without angiogenic growth factors (GFs). Antibodies and inhibitors were used to determine the involvement of specific TN-C receptors and signaling pathways.

results. TN-C significantly delayed collapse of REC capillary-like tubes on BM matrix. It decreased tube involution associated with serum deprivation, high glucose, and exposure to TGF-β. TN-C’s enhancement of tube stability was mediated by αvβ3 integrin. TN-C increased REC viability in 0.5% serum and stimulated REC proliferation in 10% serum. It promoted REC secondary sprouting on BM matrix, which involved signaling through mitogen-activated kinase kinase (MEK) and p38 mitogen-activated protein kinase. TN-C also enhanced tube branching after treatment with VEGF and stimulated REC migration twofold. Angiogenic GF increased TN-C production by RECs in an additive manner, which may explain higher levels of TN-C deposition in DR cells.

conclusions. TN-C was overexpressed in diabetic and DR REC cultures. TN-C enhanced the sprouting, migratory, and survival effects of angiogenic GFs, and had distinct proliferative, migratory, and protective capacities. The data suggest that TN-C may act as a proangiogenic mediator in DR and other pathologic conditions involving neovascularization.

Tenascin-C (TN-C) is a large extracellular matrix (ECM) glycoprotein. It is mainly expressed in embryonic development and during adult tissue remodeling and in pathologic conditions. ¹ Increased expression of TN-C in tumors is associated with cell proliferation, neovascularization, stimulation of production of matrix metalloproteinases and decreased cell adhesion to the substratum. ^{2 3 4 5} Expression of TN-C inversely correlates with apoptosis, indicating its possible influence on cell proliferation and death. ^{4 5}  

Many factors associated with pathologic processes affect TN-C production and mRNA splicing. ⁶ Among these regulators are basic fibroblast growth factor (FGF-2), ⁷ transforming growth factor (TGF)-β1, ⁸ epidermal growth factor (EGF), and platelet-derived growth factor (PDGF)-BB, ⁹ as well as the cytokines interleukin (IL)-1, IL-4, and tumor necrosis factor (TNF)-α. ¹⁰ Alternative splicing regulates the interactions of TN-C with other proteins, altering the ultimate effect of TN-C on cellular physiology. ^{11 12} The effects of TN-C on cells appear to be mediated by its cell surface receptors, including annexin II and integrins α9β1, α8β1, α2β1, αvβ3, and αvβ6. ^{13 14 15}  

We have reported that TN-C is upregulated in retinas of patients with diabetic retinopathy (DR) and is associated with neovascularized retinal areas. ^{16 17} It is unclear why DR-affected retinas produce large amounts of TN-C and whether it exerts a beneficial or deleterious effect on diseased tissue. In this article, we present data supporting the hypothesis that TN-C is produced by retinal endothelial cells (RECs) in response to angiogenic growth factors upregulated in DR. We also demonstrate that TN-C exerts proliferative, migratory, differentiative, and protective effects by itself, and it can potentiate some of the angiogenic effects of vascular endothelial growth factor (VEGF) on RECs in vitro. 

Materials and Methods

Cell Culture

RECs were obtained from fresh bovine eyes (Sierra for Medical Science, Santa Fe Springs, CA), and from normal, diabetic, and DR-affected human donor eyes (National Disease Research Interchange, Philadelphia, PA). Cultures were used from seven nondiabetic (normal) donors (mean age, 66.1 ± 2.8 years), four diabetic donors without DR (mean age, 55.8 ± 12.7 years), and seven diabetic donors with documented DR (mean age, 63.9 ± 3.8 years; Table 1 ). Age differences between these groups were not statistically significant. The study was conducted in accordance with the provisions of the Declaration of Helsinki for research involving human tissue. 

For all our assays, large numbers of cells were needed. To establish reproducibility, triplicate and repeated experiments had to be conducted. Therefore, bovine cells were used in some experiments, because bovine retinal tissue was available in large quantities and RECs derived from it grew better in culture than human cells derived from older patients. Human cells were mostly used in experiments comparing normal and diseased states (Fig. 1) or those involving human growth factors (Figs. 2 3) or antibodies that only recognized human proteins (Fig. 4) . However, all our results in bovine RECs were further confirmed in human RECs and demonstrate that there is no significant differences in their responses to TN-C or growth factors (data not shown). Cultures of human and bovine RECs were established and maintained as previously described. ¹⁸ For all experiments, only passages 2 to 7 were used. Cultures were often checked for purity by immunostaining for von Willebrand factor (Sigma Chemical Co., St. Louis, MO). 

Wound Migration Assay

We used a modification of a previously described method. ¹⁹ In short, bovine RECs were seeded in 24-well plates and allowed to reach confluence in growth medium. Cells were serum starved overnight in medium with 0.5% fetal calf serum (FCS). Monolayers were wounded with a sterile wooden stick of constant diameter, rinsed with low-serum medium to remove detached cells, and treated with 40 μg/mL TN-C (Invitrogen, Carlsbad, CA, and Chemicon International, Temecula, CA). On day 7, cells were rinsed three times with PBS and fixed with methanol for 15 minutes, rehydrated with dH₂O and stained with Mayer’s hematoxylin for 5 minutes to improve visualization, followed by destaining with dH₂O. All wells were photographed with a 4× or 10× objective with a digital camera (MDS-100; Eastman Kodak, Rochester, NY) attached to an inverted microscope (Leica, Deerfield, IL). The original wound area was measured at 0 hours and used as a baseline for comparison to the treated wells at the conclusion of the experiment. Counting of migrating cells was automated by computer (AAB software; Advanced American Biotechnology, Fullerton, CA). ¹⁸  

Tube Formation Assay on Reconstituted Basement Membrane Matrix ^{20 21 22}

Fifty microliters of reconstituted basement membrane BM matrix (Matrigel; BD Biosciences, Bedford, MA) were dispensed with frozen pipettes into each well of a previously frozen, sterile 96-well plate sitting on wet ice and allowed to solidify for 1 hour at 25 or 37°C. Approximately 5 × 10⁴ bovine RECs or 7.5 × 10⁴ human RECs in 100 μL were seeded into each well. To study the effects of different ECM components on tube formation, the following proteins were added to the growth medium at the time of seeding at 20 μg/mL for up to 72 hours: human TN-C, mouse laminin, human thrombospondin-1 (Invitrogen), or human cellular fibronectin (Calbiochem, San Diego, CA) in 0.5% FCS medium. In separate experiments, TN-C was combined at the time of seeding with VEGF, placenta growth factor (PlGF), or VEGF-PlGF heterodimer. All growth factors were added at 10 ng/mL. Regular microscopic monitoring and photography were used to determine the rate of tube formation and collapse. Total tube length and number of branch points were determined from digital photographs using the NIH Image 1.62 program in combination with commercial software (Photoshop 5.0; Adobe, San Diego, CA; ImageProPlus 4.0; Media Cybernetics, Silver Spring, MD, and the Image Processing Toolkit; Reindeer Games, Inc., Gainesville, FL). To determine which TN-C receptors mediated its effects, the tube formation assay was performed for up to 72 hours in the presence of 25 μg/mL exogenous TN-C and 20 μg/mL each of function-blocking antibodies to the TN-C–binding integrins αvβ3, αvβ5, αvβ6, α2 (part of α2β1) or α9β1. Control experiments included normal mouse IgG1 or antibodies to other integrins, α1 (part of α1β1) and α5β1 (all antibodies from Chemicon International). 

Tenascin-C Signaling in the Secondary Sprouting Assay on Reconstituted BM Matrix

Secondary sprouting on BM matrix (Matrigel; BD Biosciences), a novel in vitro angiogenic assay, was performed as described. ¹⁸ In this test, endothelial cell colonies spontaneously survive, proliferate, migrate, and invade the matrix after the original capillary-like tubes have collapsed. In our studies this assay appears to be a better indicator of angiogenic potential of various growth factors than the commonly used tube formation assay. ¹⁸ Briefly, bovine RECs were plated on reconstituted BM matrix in 0.5% FCS medium, with or without 25 μg/mL TN-C, as described earlier. Tubes were allowed to form and collapse, followed by the addition of 10× inhibitor stocks prepared according to the manufacturers’ instructions. Inhibitors to the following signaling molecules were used: p38 mitogen-activated protein kinase (MAPK; SB202190, 10 μM), MAPK kinase (MEK; PD98059, 10 μM), Ca²⁺ calmodulin kinase II (KN-93, 500 nM), protein kinase C (PKC; calphostin C, 2.5 μM), protein kinase A (PKA; H-89, 50 nM), casein kinase I (CKI-7; 50 μM), or phosphatidylinositol 3-kinase (PI3K; wortmannin, 100 nM). All inhibitors were from Seikagaku America (Rockville, MD), and Calbiochem. These concentrations were determined to be nontoxic for RECs (data not shown). The plates were further incubated for up to 15 days. ¹⁸ Secondary sprouting colonies were photographed (not shown), and the number of live cells was determined using an MTS assay (Promega, Madison, WI), according to the manufacturer’s instructions. 

Cell Proliferation and Survival Assay

Ninety-six–well plates were coated with various amounts of TN-C. Approximately 5 × 10³ bovine RECs were added to each well in 10% FCS medium. For the survival assay, medium with 0.5% FCS was used to induce cell death through serum starvation. In the high-glucose experiments, 30 mM glucose was used. The number of cells was determined on day 6 by MTS assay (Promega, Madison, WI) according to the manufacturer’s instructions. 

Influence of Growth Factors on TN-C Deposition

Cells were seeded into eight-well chamber slides (Nalge Nunc International, Naperville, IL) in 10% FCS medium and allowed to reach 40% confluence. Cells were serum starved overnight in 0.5% FCS medium and treated with 10 ng/mL insulin-like growth factor (IGF)-I, VEGF, FGF-2, and PlGF, alone or in combinations (R&D Systems, Minneapolis, MN). The growth factor concentrations had been optimized for maximum synergy in various angiogenic assays in vitro. ¹⁸ Medium and growth factors were replaced every 3 days. On day 9, chambers were removed, and slides were rinsed three times with PBS and fixed in 100% methanol for 5 minutes at −20°C and 100% acetone for 5 minutes at −20°C. Slides were allowed to air dry and were then rehydrated with PBS before the addition of a rabbit polyclonal affinity-purified anti-TN-C antibody (20 μg/mL; Chemicon International) for 1 hour. After extensive washing with PBS, rhodamine-labeled anti-rabbit cross-species preadsorbed IgG (Chemicon International) was added for 1 hour. After extensive washing with PBS, TN-C was visualized with a fluorescence microscope (model BH-2; Olympus, Tokyo, Japan) and photographed. 

Western Blot Analysis

The assay was performed as described, ²³ with lysates of human RECs treated with growth factors, as described earlier. Samples were normalized for total protein with a bicinchoninic acid (BCA) kit (Pierce Chemicals, Rockford, IL). Twenty-five micrograms total protein per well was run on 4% to 20% SDS-PAGE gradient gels (Bio-Rad Laboratories, Hercules, CA), transferred to nitrocellulose, blocked with 5% skim milk, incubated with anti-TN-C (clone BC-8; the kind gift of Luciano Zardi, National Institute for Research on Cancer, Genoa, Italy) and developed with alkaline phosphatase-conjugated secondary antibody (Chemicon) at 1:2000 dilution and 5-bromo-4-chloro-3-indoyl phosphate (BCIP)/NBT (nitro blue tetrazolium) substrate (Sigma). 

Statistical Analysis

All the treatment data sets were individually compared with their respective controls (unless otherwise specified) by the paired Student’s t-test by computer (Prism 3.0 program; GraphPad Software, San Diego, CA). P < 0.05 was considered significant. 

Results

We have demonstrated that TN-C mRNA is overexpressed in whole DR retinas compared with normal retinas. ¹⁷ In the present study, deposition of TN-C was increased in primary cultures of human RECs derived from patients with DR compared with normal (Figs. 1A 1B) . Usually, cells derived from diabetic donors without DR showed intermediate levels of TN-C (data not shown). By Western blot analysis, the only individual growth factor that enhanced deposition of TN-C was IGF-I. The addition of VEGF did not significantly increase this response (Fig. 1A) . By immunofluorescence (Fig. 1B) , TN-C formed a thin “carpet” on the substratum and fibrils on the cell surface in all cultures. When cells were treated with single growth factors known to be increased in eyes with DR (VEGF, PlGF, IGF-I, and FGF-2), only IGF-I consistently caused an increase in fibrillar TN-C, in keeping with Western blot data. Growth factor pairs did not significantly change deposition of TN-C in both normal and DR-derived RECs (Figs. 1A 1B) . However, treatment with all four growth factors together caused a significant increase of fibrillar TN-C above that obtained by treatment with IGF-I. This effect was much greater in DR cells compared with normal RECs (Fig. 1B) . 

To establish that deposited TN-C affects various aspects of the angiogenic process in vitro, we first studied its effects on capillary-like tube formation on BM matrix. In human REC cultures, TN-C delayed tube involution, resulting in an increase in total tube length of more than 400%, compared with 25% to 200% elicited by laminin, fibronectin, or the TN-C–like protein thrombospondin-1 (Fig. 2) . This effect was pronounced, starting at 48 hours after cell seeding, the time when tubes began to collapse spontaneously in control cultures. Bovine RECs responded less well to TN-C, with tube involution occurring earlier than in human cells, which may be due to the use of human TN-C. TN-C also increased network complexity, measured by the number of tube branch points, more than any other ECM protein tested (Fig. 3) . There were no significant differences between prior coating of the BM matrix with TN-C versus the addition of TN-C to the cells at the time of seeding. TN-C did not seem to promote tube formation per se, as determined in 12- to 24-hour experiments (see Fig. 9 , upper row). 

We also determined that TN-C enhanced the proangiogenic effects of various growth factors. As seen in Figure 3 , the addition of VEGF to human REC cultures on BM matrix increased the number of branch points by 300%, indicating that this growth factor promoted the initial sprouting of RECs while preventing them from coalescing into larger tubes (a possible indicator of cell differentiation). Concurrent addition of TN-C to VEGF-treated cultures increased this effect nearly twofold, suggesting that TN-C has the capability of cooperating with VEGF to promote angiogenesis in vitro. A similar effect was observed when TN-C was added with anti-TGF-β antibody, which greatly enhanced the initial sprouting of RECs on BM matrix (data not shown). VEGF did not increase tube length, possibly because it significantly enhanced network complexity (more branch points, but shorter tubes). 

The effect of TN-C on tube stability appeared to be primarily mediated by its receptor, αvβ3 integrin, because a function-blocking antibody to αvβ3 completely inhibited the TN-C–mediated increase in tube length (Fig. 4) . This was true in both normal and DR cells. Function-blocking antibodies to other TN-C–binding integrins—αvβ5, αvβ6, α2β1 and α9β1—as well as to the fibronectin receptor α5β1, had no effect (not shown). Our results confirmed previous reports indicating that αvβ3 integrin was involved in the endothelial cell capillary-like tube formation in fibrin gels. ²⁴ The αvβ3 integrin antibody did not block TN-C’s enhancement of secondary sprouting. In this assay, the effect was observed only with a blocking antibody to the α1 integrin subunit, part of the collagen-laminin-fibronectin–binding integrin α1β1 (not shown here). 

We next tried to determine which signaling pathways were involved in TN-C effects on cultured RECs. None of the tested inhibitors of MEK, p38 MAPK, PI3K, PKC, PKA, Ca²⁺-calmodulin kinase II, or CKI significantly counteracted TN-C effects on tube stability (data not shown). At the same time, TN-C–mediated survival of cells in the secondary sprouting assay required signaling through p38 MAPK and MEK. Specific inhibitors of MEK (PD98059) and p38 MAPK (SB202190), but not the inactive counterpart SB202474 or inhibitors to PI3K, PKC (Fig. 5) , Ca²⁺-calmodulin kinase II, CKI, or PKA (not shown) abolished the increase in cell numbers found in secondary sprouting colonies after addition of TN-C. The inhibitor of p38 MAPK decreased secondary sprouting in the presence of TN-C below the control level, and the effect was not associated with cell toxicity. This is the first demonstration of TN-C modulation of RECs behavior through specific kinases. 

The migratory capacity of RECs in a monolayer wound assay was increased more than twofold after addition of purified TN-C (Fig. 6) , confirming previous reports indicating that TN-C promotes migration of cells derived from a variety of tissues. ¹⁰ The αvβ3 integrin was involved in the promigratory effects of TN-C on RECs, as determined in the experiments with function-blocking antibodies (not shown). Attachment and migration of RECs on TN-C resulted in a spindle-shaped cell appearance (data not shown), similar to the sprouting phenotype described earlier. ²⁵ It has been reported that this effect of TN-C is mediated by α2β1 and αvβ3 integrins. ^{14 15} An elongated, sprouting phenotype is also produced when endothelial cells are exposed to FGF-2; this effect has been attributed to the fibrinogen globe of TN-C. ²⁶  

Experiments described in Figure 2 demonstrated that the presence of TN-C delayed tube involution (by promoting tube stability), possibly by preventing collapse-related cell death. Cells cultured in high-glucose (30 mM) or serum-free medium on BM matrix showed accelerated tube collapse and cell death. ²⁷ Addition of TN-C abrogated the effect of both proapoptotic culture conditions (Fig. 7) . Similarly, addition of TN-C to RECs grown on plastic in low-serum medium enhanced their survival in a dose-dependent manner (Fig. 8) . Moreover, TN-C was found to exert a mild proliferative effect on bovine RECs grown under standard 10% serum conditions, indicating that TN-C was also capable of promoting endothelial cell division in a less stressful environment (Fig. 8) . Because TGF-β1 and -β2 also exerted a collapsing (and proapoptotic) effect on REC networks on BM matrix, ²⁸ the addition of TN-C was expected to attenuate this effect. Figure 9 demonstrates that this was indeed the case, supporting the hypothesis that TN-C exerts a protective effect on RECs exposed to a variety of death-inducing conditions. 

Discussion

Preretinal neovascularization is a hallmark of proliferative DR. ^{29 30} It is apparently promoted by several angiogenic factors that are elevated in diabetes, such as VEGF (and its family members including PlGF), FGF-2, IGF-I, hepatocyte growth factor (HGF), and PDGF. ^{27 29 31 32 33 34 35 36 37} VEGF has been considered the main mediator of neovascularization in proliferative DR. ³⁸ However, these growth factors may well act in concert to fully promote the angiogenic phenotype of cultured RECs, as suggested by our recent data. ¹⁸  

Changes of ECM and BM components, proteinases, and integrins have also been implicated in angiogenesis and pathologic neovascularization. ^{17 39 40 41 42 43 44 45} These changes may be induced by abnormal expression of growth factors. Our previous data indicate that TN-C is increased in DR-affected retinas, especially in neovascular areas. ^{16 17} Because TN-C is upregulated by FGF-2, TGF-β1, EGF, PDGF-BB, IL-1, IL-4, and TNF-α, ^{7 8 9 10} some of which are elevated in DR, we proposed that TN-C may enhance the angiogenic phenotype. We tested this hypothesis with in vitro angiogenic assays in the presence of TN-C and certain angiogenic growth factors. We also used a variety of chemical inhibitors and antibodies to identify the molecular mechanisms of TN-C’s effect on endothelial cells. 

First, we determined whether RECs could have been responsible for the increase in TN-C seen in the neovascular areas of retinas affected by proliferative DR. Consistently, RECs derived from DR-affected retinas demonstrated increased immunostaining for TN-C when compared with RECs from normal retinas, whereas diabetic RECs showed intermediate staining. Furthermore, IGF-I, and to a lesser extent VEGF, increased deposition of TN-C in these cultures. These results were corroborated by Western blot analysis. In accordance with our most recent data, ¹⁸ a combination of four growth factors was much more effective than single ones in promoting deposition of TN-C. This effect was most pronounced in RECs in DR. These data suggest that TN-C may contribute to the angiogenic phenotype promoted by growth factors that are overproduced in DR. 

Alternatively, TN-C may also be induced by matrix metalloproteinases (MMPs), ⁴⁶ which are upregulated in eyes with DR. ^{47 48} These MMPs were able to liberate ECM-bound growth factors that could synergize with TN-C to increase cell proliferation. ⁴⁶ Other evidence, however, suggests the importance of the plasminogen activator system but not of MMPs in the in vitro angiogenesis. ⁴⁹ In support of these data, a broad spectrum MMP inhibitor, galardin/GM6001, did not alter TN-C and growth factor effects on RECs in our experiments (not shown). The role of proteinases in modulating expression of TN-C on endothelial cells warrants further studies. 

Because TN-C has multiple, often opposing, effects on cell behavior, ^{50 51} it was important to identify whether it would influence various aspects of the angiogenic process in vitro. We first studied its effects on capillary-like tube formation on reconstituted BM matrix. This is a well established and widely used assay for the screening of potential angiogenic agents. ^{20 21 22} Compared with the commonly used collagen gels, ^{52 53} this reconstituted BM matrix contains natural substrata that endothelial cells encounter under physiological conditions (i.e., BM components). In this assay, TN-C was a better promoter of tube branching and stability than other ECM proteins tested (Figs. 2 3) . This effect was apparently mediated by the TN-C receptors αvβ3 integrin, as determined with blocking antibodies. In an unexpected finding, TN-C’s enhancement of secondary sprouting seemed to be mediated by collagen-laminin-fibronectin receptor, α1β1 integrin. Possibly, TN-C can modify the α1β1 integrin’s binding to laminin and/or type IV collagen, ⁵⁴ which is abundant in BM matrix, resulting in secondary sprouting inhibition. Alternatively, α1β1 integrin may represent an as yet unrecognized TN-C receptor in RECs. 

Although able to promote capillary-like tube stability by itself, TN-C also enhanced the effect of VEGF and PlGF on tube branching. Growth factors and TN-C could synergize to promote the angiogenic phenotype by the concurrent activation of common cell signaling cascades. Ligation and clustering of integrins (including αvβ3) activates the MAPK cascade, which is also used by mitogenic growth factors. ^{45 55} Therefore, integrins and growth factors are likely to exert additive effects on activation of downstream genes. Our data show that TN-C effects on secondary sprouting were blocked by inhibitors of both p38 MAPK and MEK. These kinases are key elements in two important signaling pathways linked to cell survival, proliferation, and apoptosis and used by various angiogenic growth factors. ^{56 57} This may be the first demonstration of TN-C’s action through integrin-dependent signaling by the p38 MAPK pathway. A puzzling aspect is that p38 MAPK is usually linked to induction of stress responses and apoptosis, ⁵⁶ but activation of the MEK pathway is considered promitogenic. ⁵⁷ Yet, both these pathways similarly abrogated TN-C effects on REC secondary sprouting. It can be suggested that there is considerable cross talk between these pathways, as was recently emphasized. ⁵⁷ Also, if TN-C binding to integrins can induce a stress response through the p38 MAPK pathway, the ultimate effect of this cellular reaction would be to promote cellular regeneration and, hence, mitogenesis. However, further experimentation is needed to dissect the role of individual signaling pathways in various effects of TN-C on the angiogenic behavior of RECs. 

It is possible that some angiogenic effects of growth factors and/or TN-C may be related to the enhancement of cell survival. In the present study, we demonstrated a direct effect of TN-C on endothelial cell survival in two- and three-dimensional cultures. Addition of TN-C to RECs on BM matrix significantly increased the number of surviving cells after tube involution, a process that involves apoptosis and TGF-β action. ^{28 58} These effects (Figs. 7 8 9) were also seen under proapoptotic culture conditions (high glucose, serum starvation, addition of TGF-β). We also demonstrate that TN-C increases the number of live cells in secondary sprouting colonies (Fig. 5) . These data are supported by reports indicating that the presence of TN-C plays a role in cell survival in other systems. ^{59 60 61 62 63} This effect seemed to extend beyond the attachment-promoting activity of TN-C, because it also decreased cell apoptosis in suspension cultures. ⁶⁰  

Confirming previous reports ^{12 50} we also demonstrated that TN-C increases cell migration (Fig. 6) , again suggesting that its influence on angiogenesis occurs at various steps. The TN-C–mediated increase in cell migration may be related to its antiadhesive properties. ^{50 51} Such properties may be exerted through integrins, cell surface disialogangliosides, ⁶⁴ and the suppression of the activity of RhoA, a small guanosine triphosphatase (GTPase) that induces formation of actin stress fibers. ⁶⁵  

Besides its direct proangiogenic effects and cooperation with growth factors, TN-C may relate to the diabetic state and DR in other ways. A hyperglycemic environment leads to nonenzymatic glycosylation (glycation) of various molecules producing advanced glycation end products (AGEs). High levels of AGEs in the ECM proteins of diabetic patients have been correlated with the severity of diabetic eye disease. ⁶⁶ Induction of experimental diabetes in rats enhanced the accumulation of AGE-vitronectin in the areas surrounding retinal blood vessels. This modified vitronectin exhibited alterations in its capacity to bind TN-C. ⁶⁷ Pentosidine, an AGE, induced the expression of PDGF-B in retinal pigment epithelial cells. ⁶⁸ Because PDGF-B can upregulate TN-C, the angiogenic potential of RECs may be augmented through a positive feedback loop. This mechanism may be responsible for the increased production of TN-C by RECs derived from DR-affected retinas (Fig. 1) . 

In summary, we present evidence that TN-C is overexpressed in RECs cultures from diabetic and DR donors. This upregulation is mediated by angiogenic growth factors known to be overexpressed in DR. These growth factors mediate proangiogenic events in RECs such as capillary tube formation, cell migration, and secondary sprouting, ¹⁸ which are further enhanced by TN-C. The survival-enhancing capacity of TN-C under various proapoptotic conditions is also demonstrated, indicating that TN-C may cooperate with growth factors in the prevention of cell death. Different integrins and signal transduction pathways mediate TN-C effects on various steps of the angiogenic process. Therefore, a distinct role for TN-C in angiogenesis and its cooperation with growth factors was demonstrated. TN-C may thus be actively involved in promoting diabetic retinal neovascularization. 

 Presented at the annual meetings of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 1999, 2000, and May 2002.

 Supported by National Eye Institute Grants EY12605 (AVL) and EY07739 (MBG).

 Submitted for publication January 25, 2002; revised April 2, 2002; accepted April 17, 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.

 

Clinical Characteristics of the Human Donor Eyes

Type	Duration of Diabetes	Age (y)	Sex	Race	Cause of Death
Normal	N/A	67	F	W	To be determined
Normal	N/A	70	F	W	Cardiac arrest
Normal	N/A	50	M	W	Aortic aneurysm
Normal	N/A	71	M	W	Cardiac arrest
Normal	N/A	67	F	W	Pneumonia
Normal	N/A	70	F	W	Cardiac arrest
Normal	N/A	68	M	B	Cardiac arrest
IDDM	20 Years	65	F	W	Respiratory failure
IDDM	7 Years	18	M	W	Unknown
NIDDM	5 Years	74	M	W	Hypoxia, untreated lung cancer
NIDDM	Unknown	66	M	W	Cardiac arrest
IDDM-DR	Unknown	72	M	W	Unknown
IDDM-DR	27 Years	76	F	W	Congestive heart failure
NIDDM-DR	Unknown	64	M	W	Respiratory failure, pneumonia
IDDM-DR	5 Years	51	M	W	Dilated cardiomyopathy
IDDM-DR	Unknown	61	M	W	Cardiac arrest
IDDM-DR	5 Years	51	F	W	Cardiac arrest
IDDM-DR	6 Months	72	M	W	Cardiopulmonary arrest

 W, White; B, Black; N/A, not applicable; IDDM, insulin-dependent diabetes mellitus; NIDDM, non–insulin-dependent diabetes mellitus; DR, diabetic retinopathy.

Figure 1.

 

Angiogenic growth factors stimulated deposition of TN-C by endothelial cells. Human normal (N) and DR-derived REC cultures were treated with angiogenic growth factors in 0.5% FCS medium at 10 ng/mL for 9 days. (A) TN-C was visualized by Western blot analysis. Lane 1: untreated culture; lane 2: FGF-2; lane 3: VEGF; lane 4: IGF-I; lane 5: IGF-I+VEGF; and lane 6: FGF-2+PlGF. TN-C migrated as a series of bands in the range of 220 to 260 kDa; arrow: major isoform. DR cells expressed more TN-C than normal RECs. Deposition of TN-C was significantly increased only by IGF-I. (B) TN-C visualized by immunofluorescence. TN-C had higher expression in DR-derived cells than in normal cells. Again, IGF-I increased fibrillar deposition of TN-C. The highest increase was caused by a combination of IGF-I+VEGF+FGF-2+PlGF (Four GFs) and was more pronounced in DR-derived cells.

Figure 2.

 

Exogenous TN-C enhanced capillary-like tube stability. Normal human RECs were cultured on BM matrix coated with various proteins (40 μg/mL) in 1% FCS medium for 48 hours. In some cases (cells+TN-C), TN-C was added to the medium at the time of seeding. Total tube length was measured from photographs taken from nine fields per treatment. TN-C was more potent than other ECM proteins in increasing tube length. Data are the mean ± SD of three separate experiments. LN, laminin; FN, fibronectin; TSP-1, thrombospondin-1.

Figure 3.

 

TN-C and angiogenic growth factors increased complexity in the tube network. Normal human RECs were cultured on BM matrix coated with 40 μg/mL of TN-C or laminin (LN). VEGF, PlGF, or VEGF/PlGF heterodimer (10 ng/mL) was added to the 1% FCS medium at the time of cell seeding. The number of branch points was counted after 48 hours from photographs taken from nine fields per treatment. Data are the mean ± SD of three experiments.

Figure 4.

 

TN-C modulated stability of capillary-like tubes through αvβ3 integrin. Normal or DR-derived human RECs were cultured for 48 hours on BM matrix coated with 40 μg/mL TN-C ± 20 μg/mL anti-αvβ3 integrin antibody or normal mouse IgG1, as a negative control. Data are expressed as the mean ± SEM of five experiments performed in duplicate. Every counted field covered an entire well of a 96-well plate. The total tube length on TN-C–coated BM matrix was significantly greater than in untreated cultures. Anti-αvβ3 antibody blocked this effect. Normal and DR-affected cells behaved very similarly in these experiments, and the data from the two cell types were therefore combined.

Figure 5.

 

TN-C modulated secondary sprouting through p38 MAPK and MEK. Bovine RECs were seeded on BM matrix coated with 25 μg/mL TN-C. Tubes were allowed to form and collapse, followed by the addition of the p38 MAPK inhibitor SB202190, the MEK inhibitor PD98059, the PKC inhibitor calphostin-C, or the PI3K inhibitor wortmannin. The inactive compound SB202474 served as a negative control for MEK and p38 MAPK inhibitors. The number of cells in sprouting colonies was measured 14 days after tube collapse. Data are the mean ± SEM in two experiments performed in duplicate. Only inhibitors of p38 MAPK and MEK blocked the sprouting effect of TN-C.

Figure 6.

 

TN-C significantly enhanced REC migration. Confluent bovine REC monolayers were wounded with a wooden stick, and medium was supplemented with 40 μg/mL TN-C or nonimmune mouse IgG (negative control) for 7 days. Fixed and hematoxylin-stained cells migrating into the wound were quantified on computer. Data are the mean ± SEM of six experiments performed in duplicate.

Figure 7.

 

TN-C counteracted the tube-collapsing effect of serum-free and high-glucose media. Bovine RECs were seeded on BM matrix, with or without TN-C coating (at 40 μg/mL), and total tube length was measured. Top: tube length in 10% FCS, serum-free medium, or high-glucose (30 mM) cultures without TN-C. Middle: tube length in serum-free medium with and without (control) TN-C. Bottom: tube length in high-glucose medium with and without TN-C. Data are the mean ± SEM of seven experiments. At 48 hours, tube formation was dramatically decreased under proapoptotic culture conditions (serum-free or high-glucose medium). This effect was reversed by TN-C.

Figure 8.

 

TN-C increased REC survival and proliferation. Bovine RECs were seeded in 96-well plastic plates precoated with various amounts of TN-C. Medium with 10% FCS was used in proliferation studies, and medium with 0.5% FCS was used in survival experiments. The number of living cells was determined on day 6. Data are mean ± SEM of two experiments in triplicate. TN-C increased survival and proliferation in a dose-dependent manner, with maximum activity around 25 μg/mL.

Figure 9.

 

TN-C decreased the tube collapse induced by TGF-β isoforms. Bovine RECs were seeded on BM matrix, with or without TN-C coating (at 40 μg/mL). Tubes formed by bovine RECs were photographed by 24 hours before the start of spontaneous tube collapse. Cultures were kept in medium with 0.5% FCS ± TN-C ± 10 ng/mL TGF-β1 or -β2. TN-C counteracted the tube-collapsing effect of both TGF-β isoforms. C, control.

While this paper was in press, a report was published showing that TN-C promoted migration of endothelial cells and that they migrated better towards TN-C than towards fibronectin. Moreover, TN-C enhanced focal adhesion kinase phosphorylation in these cells. ⁶⁹