Induction of Vascular Endothelial Growth Factor After Application of Mechanical Stress to Retinal Pigment Epithelium of the Rat In Vitro

Abstract

purpose. To investigate the response to mechanical stress of vascular endothelial growth factor (VEGF) production by cultured retinal pigment epithelial (RPE) cells.

methods. A pulsatile stretch device was used in vitro. RPE cells of the second passage were seeded onto flexible-bottomed culture plates; then, at subconfluent culture, the plates were subjected to pulsatile stretch. Culture plates prepared in the same way but not subjected to stretch were used as controls. After stretching for 1 hour or 24 hours, conditioned medium for measurement of VEGF production by RPE cells was collected using a mouse VEGF immunoassay. To study the expression of VEGF in RPE cells, passaged-cultured RPE cells were exposed to pulsatile stretch for 0, 1, 3, or 14 hours. Total cytoplasmic RNA was then prepared from the RPE cells. Northern blot analysis was performed for VEGF, with G3PDH used as an internal control.

results. The expression and secretion of VEGF in RPE cells were increased by pulsatile stretching.

conclusions. Results indicate that stretching of the RPE could result in increased production of VEGF, with associated risk for neovascularization and changes in the blood-retinal barrier.

Mechanical stress occurs in the tissue or is applied to the cells in several physiological conditions. Tension forces are present during embryologic development ¹ or during growth. Mammalian cells are known to respond to mechanical stress, which is considered to be an important factor in tissue remodeling. ² Mechanical stress also occurs in the tissues or on the cells under pathologic conditions. Lin et al. ³ demonstrated the expansion of retinal pigment epithelium (RPE) in the experimental model of myopia of the chick, suggesting that the expansion of the RPE cell is caused by passive stretching or by the growth of existing cells. In patients with proliferative diabetic retinopathy (PDR) or proliferative vitreoretinopathy (PVR), traction forces on the sensory retina are generated by contracting membranes. Vitreoretinal membranes then separate the sensory retina from the RPE cells to cause retinal detachment. Traction forces are transmitted to the RPE interface and to the RPE itself via a bridge that physically links the sensory retina to the RPE. ⁴ This bridge may prevent the detachment of the sensory retina from the RPE by the mechanical forces in vivo. And in the processes of wound healing of the RPE, it is reported that the RPE is stretched. ⁵  

Vascular endothelial growth factor (VEGF), which is also termed vascular permeability factor, is a potent angiogenic mitogen that is secreted by tumor cells and by cells exposed to hypoxia. ^{6 7} Experimental evidence has shown that VEGF causes a breakdown in the blood-retinal barrier of eyes with various ischemic or nonischemic disorders. ⁸ It was recently reported that the induction of myocardial stretching ⁹ and stretching of mesangial cells ¹⁰ increases the VEGF message. Thus, stretch-induced VEGF may play a role in various physiological and pathologic conditions in other cell types. Concerning the response of the RPE to mechanical stress, Kain and Reuter ¹¹ showed that the RPE cells and corneal fibroblasts released proteases with preserved integrity of the cell membranes during stretching, suggesting that stretching could induce active secretion of proteases in these cells. The purpose of the present study was to investigate how RPE cells respond to tension forces, in other words, how RPE cells respond and adapt to mechanical stretching by expressing or secreting humoral factors such as VEGF. In this study, we applied pulsatile mechanical stretching directly on RPE cells and investigated the secretion and expression of VEGF by RPE cells in vitro. 

Methods

Primary Culture of RPE

RPE cells were isolated by Mayerson’s method with some modification. ¹² Animals were cared for in accordance with the ARVO Statement for the Use of the Animals in Ophthalmic and Vision Research. The eyeballs of Long–Evans rats were enucleated and incubated at 37°C in 0.1% crude bacterial collagenase (type I; Sigma Chemical, St. Louis, MO) and 0.2% dispase (Godo Shusei, Tokyo, Japan) for 90 minutes, followed by incubation in 0.05% trypsin–0.02% EDTA (Gibco Lab., Grand Island, NY) for another 30 minutes. The eyeballs were immediately transferred to a growth medium that consisted of Ham’s F-12 nutrient mixture (Gibco Lab.) containing 15% fetal bovine serum (FBS; Bioserum, Victoria, Australia), 0.15% sodium bicarbonate solution (Gibco Lab.), 50 U/ml penicillin (Gibco Lab.), and 50 μg/ml streptomycin (Gibco Lab.), and then the RPE was isolated. The isolated RPE was incubated in 0.05% trypsin–0.02% EDTA at 37°C for 2 minutes. Growth medium was next added and again triturated gently. The cell suspension was centrifuged for another 10 minutes at 3000 rpm, and the pellet was resuspended in growth medium and seeded into 35-mm dishes at a density of 85,000 to100,000/2 ml. The RPE cells were incubated at 37°C in 5% CO₂ and 20% O₂. 

Application of Continuous Pulsatile Stretch

Second passage RPE cells were seeded in six-well BioFlex collagen-coated, silicone elastomer–bottomed culture plates (Flexcell Intl. Corp., McKeesport, PA). When the cultures were subconfluent, the culture medium was replaced with Ham’s F-12 medium containing 0.5% FBS (instead of 15% FBS). The plates were placed on a Flexercell Strain Unit (model FX-3000l) placed in the incubator. The strain unit is a modification of the unit initially described by Banes et al., which consists of a computer-controlled vacuum unit and a baseplate to hold the culture dishes. ¹³ The computer system controls the frequency of deformation and the ratio of elongation of the silicone elastomer–bottom. Vacuum is repeatedly applied to the plates, causing a pulsatile stretch in one of two patterns: 0.2 Hz for 30 seconds followed by a 30-second rest (10% elongation; pattern A) or 1.0 Hz with 15% elongation (pattern B). In the pattern B, this pulsatile stretch was continuously repeated without a rest period. Other BioFlex collagen I culture plates were prepared in parallel, but not subjected to pulsatile stretch, and served as controls. After the application of pulsatile stretch, we observed the stretched and the “unstretched” (control) RPE cells by a phase-contrast microscope before the following analyses. 

Measurement of VEGF Concentration

After the application of pulsatile stretch (pattern A for 1 hour or 24 hours and pattern B for 1 hour), the conditioned medium was pooled for VEGF assay. The concentration of VEGF was measured using the mouse VEGF immunoassay system (Quantikine; R&D Systems, Minneapolis, MN), a sandwich enzyme immunoassay technique. 

RNA Extraction and Northern Blot Analysis for VEGF

In northern blot analysis, pVEGF-1 (mouse VEGF164 cDNA, 583 bp cDNA; a generous gift from Georg Breier, Max—Planck–Institut, Bad Nauheim, Germany) ¹⁴ was used as the probe for VEGF. The expression of glyceraldehyde-3-phosphate dehydrogenase (G3PDH) was analyzed as the internal control. As the cDNA probe for G3PDH, 780-bp polymerase chain reaction (PCR) product (sense primer: 5′-ATT GTT GCC ATC AAC GAC CCC TTC-3′; antisense primer: 5′-GTT GCT GTT GAA GTC ACA GGA GAC-3′) was used. The sequence of the PCR products of G3PDH was examined and was identical to the expected sequence of nucleotide of G3PDH. Those cDNA fragments were labeled with[α -³²P]–dCTP by Klenow enzyme using Ready To Go DNA labeling kit (-dCTP; Amersham Pharmacia Biotech UK Ltd, Bucks, UK). 

The passaged-cultured RPE cells were exposed to five kinds of pulsatile stretch (pattern A for 1 hour and pattern B for 0, 1, 3, or 14 hours). Next, total cytoplasmic RNA was prepared from the RPE cells using RNA zol (CINNA/BIOTECX Laboratories International, Friendswood, TX), according to the manufacturer’s instructions. In each group, 20 μg of cytoplasmic RNA was subjected to formaldehyde–agarose gel electrophoresis and was then transferred to a nylon membrane. The integrity of the extracted RNA was confirmed by visualization of the 28S and 18S ribosomal RNA bands of ethidium bromide–staining. The nylon membrane was prehybridized in a solution containing salmon sperm DNA for 1 hour, then hybridized with the ³²P-labeled VEGF or G3PDH cDNA probe overnight at 42°C. After being washed, the nylon membrane was autoradiographed at− 80°C. For semiquantitative analysis, we measured the density of the bands using computer-based densitometric analysis. 

Statistical Analysis

Data are reported as mean ± SD. The Mann–Whitney U test was used to compare the concentration of the growth factor in the stretched and unstretched (control) cells. A level of P < 0.05 was accepted as statistically significant. 

Results

Phase-Contrast Microscopic Observation

The pulsatile stretching of the silicone elastomer–bottom of the culture plates resulted in the stretching of the RPE cells, with some RPE cells becoming spindle-shaped and separated from each other (Fig. 1) . 

Measurement of VEGF Concentration in Conditioned Medium

The concentration of VEGF in the conditioned media of cells stretched 10% under pattern A conditions (0.2 Hz for 30 seconds followed by a 30-second rest period with 10% elongation) for 1 hour was 72.73 ± 7.73 pg/ml, whereas that from unstretched control dishes was 62.93 ± 4.85 pg/ml. Although the concentration of VEGF appeared to be increased in the stretched samples compared with the control samples, the differences were not statistically significant (P > 0.05, n = 4). The concentration of VEGF in the conditioned media of cells stretched 10% under pattern A conditions for 24 hours was 321.02 ± 15.92 pg/ml and that from the control dishes was 137.52 ± 10.41 pg/ml. The concentration of VEGF was significantly increased in the stretched samples compared with the control samples (P < 0.01, n = 4; Fig. 2 ). The concentration of VEGF in conditioned media of cells stretched 15% under pattern B conditions (1.0 Hz with 15% elongation) for 1 hour was 43.29 ± 4.14 pg/ml, whereas that from the control dishes was 38.74 ± 2.32 pg/ml. The concentration of VEGF was significantly increased in the stretched samples compared with the control samples (P < 0.05; Fig. 3 ). 

Northern Blot Analysis for VEGF

In Figure 4 , we present the most typical pattern of northern blot analysis. We confirmed this pattern by at least 3 independent experiments. By densitometric analysis, the ratio of VEGF/G3PDH for lanes 1, 2, 3, 4, and 5 in Figure 4 was 1, 40, 37, 0.057, and 28, respectively. Under pattern B conditions, the expression of VEGF was increased by short periods of stretching for 1 and 3 hours and was not increased by a prolonged period of stretching for 14 hours (Fig. 4 , lanes 1 through 4). Under pattern A conditions, the expression of VEGF was also increased by a short period of stretching (1 hour; Fig. 4 , lane 5). When the lanes 2 and 5 are compared, the density is higher in lane 2, indicating that the expression of VEGF was more strongly induced by pattern B than by pattern A condition within 1 hour. 

Discussion

The present study demonstrated that pulsatile stretching induced the production and secretion of VEGF by the RPE cells in vitro. We performed this study to know whether VEGF levels in the eye change when RPE cells are exposed to mechanical stress in such a condition as ocular enlargement or tractional retinal detachment. When the physiological condition in the eye is considered, the RPE is in a mechanically active condition because of ocular pulse and accommodation. Therefore, in the present study, we applied pulsatile stretching directly on cells by a Flexercell Strain Unit, which is appears to be the best way at the moment. And it is difficult to produce physiological conditions by this method, we tried two patterns of stretching, pattern A and pattern B, in this study. Our results indicate that VEGF is possibly secreted by RPE cells in vivo in the conditions in which mechanical stress is loaded on RPE cells. 

The RPE is thought to be stretched in processes of normal or pathologic ocular enlargement, tractional retinal detachment, and wound healing of itself. ⁵ In normal eye development or pathologic ocular enlargement of high myopia, traction forces loaded on the RPE may increase the secretion of VEGF by RPE cells, suggesting that stretch-induced VEGF secretion may be involved in angiogenesis in normal eye development as well as in angiogenesis (Fuch’s spot) ¹⁵ or in the increased permeability observed in patients with high myopia. ¹⁶ In the tractional retinal detachment observed in PDR and PVR, the traction forces generated by contracting membranes are transmitted to the RPE interface and RPE via a physical bridge between the sensory retina and RPE. ⁴ It has been reported that the concentration of VEGF is increased in the intraocular fluid of patients with PDR. ¹⁷ Recently, Armstrong et al. ¹⁸ reported that VEGF levels of PVR membranes are moderately elevated. In PDR, the hypoxic retina leads to the production of angiogenic factors including VEGF, resulting in angiogenesis, and the vitreous hemorrhage along with the glial cell proliferation leads to the formation of epiretinal membranes. On the other hand, PVR results from migration and proliferation of dispersed RPE cells. Although the primary mechanism of membrane formation is different between PDR and PVR, membrane contraction is a common result, so the following events, such as stretch of RPE cells, could occur commonly. Therefore, stretch-induced VEGF secretion can, at least in part, contribute to the increased levels of VEGF, which in turn plays a role in the angiogenesis, in the breakdown of the blood–retinal barrier, or in the proliferation of RPE in PDR or PVR. ¹⁹ However, the role of stretch-induced VEGF could be smaller compared with the role of VEGF primarily produced after hypoxia. In tractional retinal detachment, the sensory retina, as well as the RPE, would be involved in traction forces. We think that the study about the sensory retina in response to mechanical stress is meaningful and significant, as well as the RPE. The Müller cells, in particular, which are thought to be involved in formation of epiretinal membrane, are our next target. 

The data of the present study and other studies ^{9 10} strongly suggest that in general mechanical stretch induces secretion and expression of VEGF by VEGF-expressing cells such as cardiac myocytes, mesangial cells, and RPE cells. There is increasing evidence concerning the intracellular response to mechanical stress or the regulation of VEGF secretion or production in response to mechanical stress. Li et al. ⁹ reported that myocardial stretch increases mRNA expression of VEGF and secretion of transforming growth factor-β (TGF-β) in the heart and that inhibition of TGF-β activity by a neutralizing antibody eliminates the stretch-induced increase in VEGF expression. This indicated that the stretch-induced increase in VEGF expression was mainly mediated by secretion of TGF-β and that the VEGF was actively secreted via the mechanism mainly dependent on TGF-β, but not by the passive release caused by the loss of cell membrane integrity under mechanical stress. Furthermore, it was found that the inhibition of both protein kinase C and protein tyrosine kinase abolishes the VEGF response to mechanical stretch in human mesangial cells. ¹⁰ Therefore, it is likely that a similar mechanism is involved in the stretch-induced secretion of VEGF by RPE cells. Further studies are needed to clarify the mechanism in RPE cells. 

From the results of northern blot analysis in the present study, the expression of VEGF was increased by short periods of stretching for 1 hour and 3 hours, but not by prolonged stretching (for 14 hours). We speculate that during the prolonged period of stretching, RPE cells adapted to mechanical stretching by altering cell morphology or cell adhesion–to-extracellular matrix in this in vitro model. However, the real mechanism of this downregulation by prolonged stretching is unknown and remains to be clarified. In conclusion, we have demonstrated for the first time that the mRNA expression and secretion of VEGF by RPE cells, which are known to be induced by hypoxia, can also be induced by mechanical stretch. 

 Supported by a Grant-in-Aid for Scientific Research (No. 10770922) from the Ministry of Education, Science, Sports and Culture of Japan.

 Submitted for publication February 8, 1999; revised June 10, 1999; accepted July 23, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

RPE before (A) and after (B) the application of pulsatile stretch (pattern A, for 24 hours). RPE cells are shown before (A) and after (B) the application of pulsatile stretch in the same culture dish. The pulsatile stretching of the silicone elastomer–bottom of the culture plates resulted in the stretching of the RPE cells, with some RPE cells becoming spindle-shaped and separated from each other. (Magnification, ×86.)

Figure 2.

 

Concentration of VEGF in the conditioned medium after the application of pulsatile stretch (Pattern A). Data represent mean ± SD. After the Pattern A stretching for 1 hour (A), the concentration of VEGF was increased, but not significantly so, in the stretched samples compared with the control samples (P > 0.05). After the pattern A stretching for 24 hours (B), the concentration of VEGF was significantly increased in the stretched samples compared with the control samples (P < 0.01).

Figure 3.

 

Concentration of VEGF in the conditioned medium after the application of pulsatile stretch (pattern B). Data represent mean ± SD. After the pattern B stretching for 1 hour, the concentration of VEGF was significantly increased in the stretched samples compared with the control samples (P < 0.05).

Figure 4.

 

Northern blot analysis for VEGF and G3PDH. Under pattern B conditions, compared with the control (lane 1), the expression of VEGF was increased by short periods of stretching for 1 hour (lane 2) and 3 hours (lane 3) and was not increased by a prolonged stretching for 14 hours (lane 4). Under pattern A conditions, the expression of VEGF was also increased by a short period of stretching for 1 hour (lane 5). When lanes 2 and 5 are compared, the density is higher in that shown in lane 2, indicating that the expression of VEGF was more strongly induced by pattern B than pattern A condition within 1 hour.