Latent Transforming Growth Factor-β Binding Protein-1, a Component of Latent Transforming Growth Factor-β Complex, Accelerates the Migration of Aortic Smooth Muscle Cells in Diabetic Rats Through Integrin-β₃

Sources of materials were as follows: streptozotocin, Sigma Chemical (St. Louis, MO); fetal bovine serum (FBS) (Lot No. US177246), Invitrogen (Carlsbad, CA); PDGF-BB, R&D Systems (Minneapolis, MN); ¹²⁵I, New England Nuclear (Boston, MA); and Bis (sulfosuccinimidyl) suberate, Pierce (Rockford, IL). LTBP-1 was purified from Escherichia coli transfected with its cDNA (LTBP-1, short form), and its purity was estimated by SDS-PAGE, silver staining, and immunoblotting using antibody to LTBP-1. The molecular weight of this recombinant LTBP-1 was ∼180 kDa.

Mouse monoclonal antibody to rat integrin-β₃ (F11) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal antibody against the free form of LTBP-1 purified from human platelets (Ab39) was a gift from Dr. Kohei Miyazono (Department of Molecular Pathology, Tokyo University, Tokyo, Japan). Ab39 cross-reacts with rat, and its specificity has been reported (27,28).

Diabetes was induced by the injection of 60 mg/kg streptozotocin into male Wistar rats. Control and diabetic rats were fed a standard diet for 4 weeks. The plasma glucose levels were as follows: control 4.51 ± 1.46 mmol/l (n = 10), diabetic 23.3 ± 3.31 mmol/l (n = 10). The body weights were 380 ± 18.7 g (n = 10) for the control and 276 ± 33.6 g (n = 10) for the diabetic. The experiments were repeated in 10 different pairs.

SMCs were explanted from the thoracic aortas of male Wistar rats essentially by the method of Fischer-Dzoga et al. (29). Primary culture and subculture were carried out in Dulbecco’s modified Eagles medium (DMEM) supplemented with 10% FBS. Cells were subcultured at a ratio of 1:2 and used at the second through fourth passages for the following experiments.

The migration of SMCs was assayed by a modification of Boyden’s chamber method using a polycarbonate filter (Neuro Probe, Gaithersburg, MD) with pores of 8.0-μm diameter (30). Cultured SMCs were trypsinized and suspended at a concentration of 5.0 × 10⁵ cells/ml in DMEM with 1 mg/ml BSA. Then the SMC suspension was placed in the upper chamber and DMEM containing 1 mg/ml BSA with LTBP-1 or PDGF-BB in the lower chamber. In some experiments, F11 was added to the upper chamber. The chamber was incubated at 37°C in 5% CO₂ and 95% air for 4 h. The filter was removed and SMCs on the upper side of the filter were scraped off. Then SMCs that had migrated to the lower side of the filter were fixed in methanol, stained with hematoxylin and eosin, and counted by light microscopy (×100) for quantitation of migration. Migration activity was expressed as the mean number of migrated cells seen in five different fields (F).

Iodination of LTBP-1 was carried out by the mild chloramine T method as described by Frolik et al. (31). The cross-linking experiment was performed as previously described (32). In brief, confluent SMCs in 75 cm² culture flasks were washed with binding buffer (PBS containing 0.9 mmol/l CaCl₂, 0.49 mmol/l MgCl₂, and 1 mg/ml BSA), preincubated on ice in the same buffer with excess unlabeled LTBP-1 or 20 μg IgG of F11 for 30 min, and then incubated on ice with 100 pmol/l of ¹²⁵I-LTBP-1 for 3 h. Cells were washed, and cross-linking was done in the binding buffer without BSA together with 0.25 mmol/l Bis for 30 min on ice. Supernatants from solubilized cells (cell lysates) containing the same amounts of protein were subjected to analysis by SDS gel electrophoresis using 4% polyacrylamide gel, followed by autoradiography.

The aortas were snap-frozen and stored at −80°C. They were sectioned serially at 6-μm thickness and fixed in acetone. For the identification of integrin-β₃ and LTBP-1, we used 100× diluted F11 and 200× diluted Ab39, respectively. The staining of sections was followed by the method of ABC peroxidase immunohistochemistry described by Waltenberger et al. (27).

The significance of differences was evaluated by two-tailed Student’s t test.

First, we examined the effects of LTBP-1 on the migration and proliferation of diabetic SMC. Migration activity of control SMCs without addition of LTBP-1 was 19.4 ± 9.66 cells/F. Migration activities of control SMCs with the addition of 100 pg/ml LTBP-1 and 10 ng/ml PDGF-BB increased significantly to 28.0 ± 12.5 and 37.7 ± 12.7 cells/F, respectively, compared with those without LTBP-1 and PDGF-BB (Fig. 1). Migration activity of diabetic SMCs without LTBP-1 was 15.4 ± 6.10 cells/F, and those with the addition of 1, 10, 100, and 500 pg/ml LTBP-1 and 10 ng/ml PDGF-BB increased significantly to 27.9 ± 11.0, 48.6 ± 16.9, 51.4 ± 15.3, 52.8 ± 15.4, and 58.7 ± 21.0 cells/F, respectively (Fig. 1). Furthermore, 10, 100, and 500 pg/ml LTBP-1 and 10 ng/ml PDGF-BB significantly stimulated migration activities of diabetic SMCs compared with control SMCs (Fig. 1). However, 1–500 pg/ml LTBP-1 had no effects on the proliferation of rat SMCs in both control and diabetes (data not shown).

It has been reported that integrin-β₃ has an important role in SMC migration (33). To clarify the participation of integrin-β₃ in the accelerated migration of diabetic SMCs, migration activities were examined using anti–integrin-β₃ antibody (F11).

Migration activity of diabetic SMCs with the addition of 100 pg/ml LTBP-1 was 57.6 ± 14.9 cells/F, whereas that with no LTBP-1 was 22.2 ± 6.47 cells/F. LTBP-1–induced migration activities decreased significantly to 29.8 ± 7.01 and 31.6 ± 8.22 cells/F with the addition of 0.2 and 1.0 μg IgG/ml F11, respectively, compared with no addition (Table 1). There were no changes in migration activities with the addition of control (nonimmune) IgG instead of F11 (Table 1). Neither F11 nor control IgG had any effect on the PDGF-BB–induced migration activities of diabetic SMCs (Table 1).

Migration activity of control SMCs with the addition of 100 pg/ml LTBP-1 was 25.5 ± 4.17 cells/F, and that without LTBP-1 was 14.8 ± 4.17 cells/F. Neither LTBP-1–nor PDGF-BB–induced migration activities of control SMCs were changed by the addition of F11 and control IgG (Table 2).

To determine the mechanism of the accelerated migration in diabetic SMC due to LTBP-1, the bindings of LTBP-1 with control and diabetic SMCs were examined by cross-linking experiments. A 350- to 400-kDa band observed in cell lysates from diabetic SMCs cross-linked with LTBP-1 was stronger than that from control SMCs (Fig. 2, lanes 2 and 5), and such a band was not observed in cell lysate from diabetic SMCs preincubated with F11 (Fig. 2, lane 6), but a week band was in cell lysate from control SMCs preincubated with F11 (Fig. 2, lane 3). These bands were not detected in cell lysates from both control and diabetic SMCs added with excess unlabeled LTBP-1 at binding (Fig. 2, lanes 1 and 4). These results suggested that bands of 350–400 kDa are specific for LTBP-1 in both control and diabetic SMCs, and is also specific for integrin-β₃ in diabetic SMC. The molecular weight of a band of SMC cross-linked with LTBP-1 is 350–400 kDa and that of the LTBP-1 used in this experiment is 180 kDa, indicating that LTBP-1 associates with a 170- to 220-kDa cell surface protein of SMC, which is probably integrin-β₃ in diabetes.

To clarify the presence of integrin-β₃ and LTBP-1 in rat aorta, their distributions were studied by immunohistochemistry. No medial layer of rat aorta was stained with nonimmune serum (Fig. 3A). The staining with integrin-β₃ of the medial layer in diabetic aorta (Fig. 3C) was much stronger than that in control aorta (Fig. 3B). However, extracellular spaces of medial layers in both control and diabetic aortas were strongly stained with antibody to LTBP-1 (Fig. 3D and E). SMC is a major cellular component in aortic medial layer, suggesting that SMCs of diabetic aorta express both LTBP-1 and integrin-β₃, but SMCs of control aorta expresses LTBP-1 and not integrin-β_3.

The present results showed that 10–500 pg/ml LTBP-1 stimulated the migration of diabetic SMCs in comparison to control SMCs in vitro, and integrin-β₃ was significantly involved in the SMC migration augmented by LTBP-1 in diabetic rats. Checker-board analysis revealed that LTBP-1–stimulated rat SMC migration was unidirectional in both control and diabetes, indicating that these SMC migrations were due to chemotaxis and not chemokinesis (data not shown).

The mechanism of cell migration has not been fully understood. However, it is known that the migration process consists of four major steps (34). First, there is the forward motility of the membrane at the front of the cell, which is called “protrusion.” Then, adhesion is required for protrusion to be converted into movement along the substrate. Next occurs the process leading to forward movement of the nucleus and cell body, which is termed “traction.” Finally, the last step is “de-adhesion” and “retraction” of the rear region of the cell. This means that cell-matrix adhesion and detachment are critical processes during cell migration. In recent studies, it has been reported that SMC migration is blocked by α_vβ₃-integrin–blocking antibodies or RGD peptides in vitro and that the inhibition of α_vβ₃-integrin retards the development of arterial intimal thickening and restenosis after vascular injury (33,35–39). Our data show that the LTBP-1–induced migration of diabetic SMCs is mainly due to integrin-β_3, but that of control SMCs is not; by coincidence, integrin-β₃ increases in diabetic aorta compared with control by immunohistochemistry. These results suggest that the mechanism of migration is different between control and diabetic SMCs. Bendeck et al. (40) reported that newborn SMCs of humans express α_vβ₃-integrin but that adult SMCs do not. Liaw et al. (35) demonstrated that most human aortic SMCs expressed little or no α_vβ₃-integrin, but that those isolated from an infant did. Hoshiga et al. (41) reported that α_vβ₃-integrin was strongly expressed in SMCs of intimal thickening and plaque of atherosclerotic coronary artery from a recipient human heart, but this expression was less in the media of coronary artery. Stouffer et al. (42) showed that integrin-β₃ expression was detected in arterial SMCs 1 week after BCI of a baboon brachial artery, but not in uninjured artery. In total, these studies demonstrate that α_vβ₃-integrin accompanies changes in the SMC phenotype.

The molecular mechanism of increased expression of integrin-β₃ is not known in diabetic SMCs. But Janat et al. (43) and Basson et al. (44) reported that PDGF-BB stimulates the expression of integrin-β₃. This means that PDGF β-receptor, which binds to PDGF-BB, is an important factor to increase integrin-β₃ on SMCs. We have already reported that diabetic SMCs expressed PDGF β-receptor much more than control SMCs (2–4). This overexpression of PDGF β-receptor probably results in the increased expression of integrin-β₃ on diabetic SMCs.

Our study showed that PDGF-BB–induced SMC migration was not inhibited by anti–integrin-β₃ antibody. Kappert et al. (45) reported that PDGF-directed migration of rat and human SMCs through gelatin-coated membranes involved not only α_vβ₃-integrin but also α_vβ₅-integrin by experiments using blocking antibodies to these integrins. Moreover, Liu et al. (46) showed that integrin-α_vβ₃ is involved in the migration of rat vascular adventitial fibroblasts stimulated by basic fibroblast growth factor but not by PDGF. However, Choi et al. (47) reported that PDGF-induced human SMC migration is mediated by α_vβ₃-integrin. The reason for the divergent results regarding the role of α_vβ₃-integrin in PDGF-induced migration remains to be determined.

SMC migration from the medial to the intimal layer of an artery is an initial major process of the formation of arterial intimal thickening (48). Our data demonstrated that migration of diabetic SMCs due to LTBP-1 was accelerated compared with that of control SMCs in vitro, suggesting that LTBP-1 plays a critical role in the intimal thickening of diabetic artery through the acceleration of SMC migration. As we previously described, the latent TGF-β1 complex consists of TGF-β1, LAP, and LTBP-1. In the diabetic artery and SMCs, TGF-β1 is one of the essential factors for the development of the characteristic changes of arterial SMCs, such as an increase in fibronectin and PDGF β-receptor (5). LAP has an important role in the latency of the TGF-β1 complex (11). It is conceivable that each component of the latent TGF-β1 complex influences the initiation, progression, and regression of arterial intimal thickening in diabetic artery. In summary, LTBP-1 accelerated the migration of diabetic SMCs through integrin-β₃ in comparison to that of control SMCs, and it appears to play an essential role in the intimal thickening of diabetic artery.

This study was supported in part by the Naito Foundation, the Hamaguchi Foundation for the Advancement of Biochemistry, and the Takeda Science Foundation (to T.K.).

We thank Dr. Kohei Miyazono (Department of Molecular Pathology, Tokyo University, Tokyo, Japan) for kindly providing Ab39. We also appreciate Dr. Shiro Ueda (Chiba University, Graduate School of Pharmaceutical Sciences, Chiba City, Japan) for useful discussion and Dr. Akihiro Haruno (Taiho Pharmaceutical, Hanno City, Japan) for help with the statistical analysis.