Modulation of Pre-mRNA Splicing and Protein Production of Fibronectin by TGF-β2 in Porcine Trabecular Cells

Abstract

purpose. To determine the effect of transforming growth factor (TGF)-β2 on the pre-mRNA splicing pattern of fibronectin, as well as on the synthesis and secretion of this glycoprotein by porcine trabecular cells.

methods. First-passage porcine trabecular cells were rendered quiescent and incubated in culture medium containing 15% newborn calf serum, in serum-free culture medium containing either activated TGF-β2 (concentration range: 0.2–2.7 ng/ml) or activated TGF-β1 (1 ng/ml), or in serum-free medium alone (untreated control samples). For investigation of alternative splicing, total RNA was extracted, and reverse transcription–polymerase chain reaction (RT-PCR) was performed with primer pairs located in exons flanking the exon (extra domain[ ED]A, or EDB) that undergoes alternative splicing. The polymerase chain reaction (PCR) products were verified by Southern hybridization and quantified by using laser densitometry. The percentage of EDA-positive (+) isoforms was compared with that of the EDB+ isoforms among the groups. To study the effect of TGF-β2 on the synthesis and secretion of fibronectin, total protein was extracted from both cultured cells and conditioned medium, Western blot analysis was performed using an anti-fibronectin antibody, and the products were quantified by laser densitometry. Immunocytochemical analysis was also performed on cultured trabecular cells to detect fibronectin.

results. Fibronectin mRNA that was detected in untreated serum-starved control cells was EDA and EDB negative. Incubation of trabecular cells in medium containing 1 ng/ml TGF-β2, 1 ng/ml TGF-β1, or 15% newborn calf serum induced the expression of EDA+ and EDB+ mRNA to varying degrees. At concentrations of 0.2, 0.5, 1.5, and 2.7 ng/ml, TGF-β2 increased the concentration of fibronectin by 2-, 3-, 3.8-, and 5-fold in the conditioned medium, and by 3-, 3.7-, 4-, and 4.3-fold in the cell extracts, respectively. The trabecular cells treated with TGF-β2 exhibited strong immunoreaction for fibronectin, whereas the cells incubated in serum-free medium showed only minimal immunoreactivity.

conclusions. Our results demonstrate that TGF-β2 and TGF-β1 modified the alternative splicing pattern of fibronectin pre-mRNA and enhanced the synthesis and secretion of this extracellular matrix molecule by trabecular cells in a dose-dependent fashion. These findings indicate a mechanism whereby TGF-β2, the concentration of which is elevated in aqueous humor of patients with primary open-angle glaucoma, contributes to the increased deposition of extracellular matrix molecules in the outflow pathway.

Preliminary studies of the human trabecular meshwork from glaucomatous eyes have shown that not only is the amount of fibronectin increased relative to the total protein, but also the sialation of the molecule, compared with age-matched normal subjects (>60 years of age) ^{1 2} These findings raise several possibilities about the nature of the defect ¹ : An upregulation of the mRNA for fibronectin in a population of senescent trabecular cells that are reduced in number leads to its increased synthesis and deposition in the extracellular matrix (ECM). ² Alternative splicing and/or alterations in the normal posttranslational modifications produce an increase in sialation of fibronectin. ³ A reduction in the activity of proteolytic enzymes and/or an increase in their inhibitors results in decreased degradation of extracellular fibronectin. 

The aqueous humor contains significant amounts of growth factors, and trabecular cells also synthesize and secrete these molecules into the microenvironment of the meshwork. ^{3 4 5 6 7} We reported previously that the total amount of transforming growth factor (TGF)-β2 and the concentration of intrinsically active TGF-β2 in samples of aqueous humor from patients with primary open-angle glaucoma (POAG) were significantly higher than those in the aqueous humor from age-matched normal eyes. ⁸ In normal human fibroblasts, the TGF-βs increase not only the synthesis of fibronectin, but also its retention at the cell surface by elevating the expression of receptors. ⁹ In addition to the upregulation of fibronectin mRNA, ^{10 11} TGF-β regulates the levels of different isoforms of the glycoprotein by increasing the amount of the alternatively spliced EDA segment in the synthesized protein. ¹² Although evidence is emerging that various growth factors have an important role in the pathophysiology of the trabecular meshwork, ^{5 13} it is not known whether the synthesis of fibronectin by trabecular cells can be altered by TGF-β, as it is in fibroblasts. 

In the present study, we investigated the effect of TGF-β2 on fibronectin pre-mRNA alternative splicing in trabecular cells and the biologic action of TGF-β2 on the protein synthesis of fibronectin by using reverse transcription–polymerase chain reaction (RT-PCR), Southern hybridization, and Western hybridization, respectively. Because TGF-β1 is known to modulate the alternative splicing pattern of fibronectin pre-mRNA in fibroblasts and other types of cells, ¹⁴ we included this isoform in our experiment as a control. 

Materials and Methods

Procurement of Trabecular Tissue and Initiation of Cell Culture

The trabecular meshwork was excised from fresh porcine eyes obtained from a local abattoir within 20 minutes after decapitation, and cultures of trabecular cells were initiated from the tissue explants, as described previously. ^{13 15} All experiments were performed on cells of the first passage and repeated at least three times. The surgical procedures and cell culture processes were undertaken in a sterile laminar flow hood (Nuaire, Plymouth, MN). 

Alternative Splicing of Fibronectin Pre-mRNA

Trabecular cells were rendered quiescent by serum deprivation for 48 hours, and then incubated in culture medium containing 15% newborn calf serum, or serum-free medium (SFM) containing either 1 ng/ml TGF-β2 or 1 ng/ml TGF-β1 for 48 hours. The quiescent cells in SFM alone served as untreated controls. Total RNA was extracted from cultured cells according to the method described by Chomczynski and Sacchi. ¹⁶  

First-strand cDNA was prepared from cellular RNA by using avian myeloblastosis virus (AMV) reverse transcriptase and the downstream primer specific for fibronectin, which was custom-synthesized by National Biosciences (Plymouth, MN). The reverse transcription reaction was performed in a final volume of 20 μl with 5 mM MgCl₂, 10 mM Tris-HCl, 50 mM KCl, 0.1% Triton X-100, 1 mM of each deoxyribonucleotide (dATP, dCTP, dTTP, and dGTP), 20 U rRNasin RNase inhibitor, 15 U AMV reverse transcriptase, 0.1 μM of downstream primer, and 2 μg of total RNA. The reaction was allowed to proceed at 42°C for 60 minutes, heated at 99°C for 5 minutes, and followed by incubation at 3°C for 5 minutes. 

The effect of TGF-β on fibronectin pre-mRNA splicing was examined at two regions, extra domain (ED) A and EDB, by using polymerase chain reaction (PCR) primer pairs located in exons flanking the exon (EDA or EDB) that undergoes alternative splicing. ¹⁴ Two products are expected: the longer product, which contains the alternatively spliced exon, and a shorter product, which does not have the alternatively spliced exon. The sequence of the PCR primers from 5′ to 3′ are as follows: aaacagaaatgactattgaaggcttg (EDA sense), agagcatagacactcacttcatattt (EDA antisense), attactggttatagaattaccacaacc (EDB sense), and taatatcagaaaagtcaatgccagttg (EDB antisense). These primers were selected because they had no homology to other known DNA sequences, their guanine and cytidine content was low (35%), they had a melting point of 52°C to 56°C, and they were of the desired size. ¹⁴  

The products from the RT reaction were diluted to 97.5 μl, and 20 μl was used to perform the PCR amplification in a total volume of 100 μl with 2.5 U native Taq DNA polymerase (Perkin-Elmer Cetus, Norwalk, CT), 1.5 mM MgCl₂, and 1.0 μM primers. A DNA thermal cycler (model 480; Perkin-Elmer Cetus) was used with the temperature profiles as follows: initial melting at 94°C for 4 minutes, 20 cycles of 1 minute melting at 94°C, 2 minutes of annealing at 60°C, and 3 minutes of extension at 72°C. After the last cycle, the polymerization step was extended by 10 minutes so that all strands were completed. 

To verify the specificity of the PCR products, Southern hybridization was performed by using enhanced chemiluminescence (ECL) 3′-oligolabeling and detection system (Amersham Life Science, Arlington Heights, IL). Briefly, horizontal 2% agarose gel electrophoresis with 8 μl of PCR products, 1 μl of 10× loading buffer, and 1 μl of 1 mg/ml ethidium bromide was run in a minisubmarine gel apparatus (model GNA-100; Pharmacia LKB Biotechnology, Piscataway, NJ) at 50 V for 2 hours. Phi X 174 DNA/HaeIII digests (Promega, Madison, WI) were used as molecular size standards. The separated DNA products were transferred to 0.2-μm (pore size) pure nitrocellulose membranes (Schleicher & Schuell, Keene, NH), with a solution of 0.4 N NaOH and 0.6 N NaCl, by means of a vacuum transfer apparatus (Hoefer Scientific Instruments, San Francisco, CA) for 1 hour and then fixed to the membrane in a UV cross-linker (Hoefer). 

Southern hybridization with 30-mer antisense oligonucleotide probes that hybridize to regions within the amplified sequences was performed. The nucleotide sequences of the probes from 5′ to 3′ are as follows: ctcgatatccagtgagctgaacattgggtg (EDA), and ctctcatgttgttcgtagacactggagaca (EDB). A tail of fluorescein-11-dUTP was introduced onto the 3′ end of the 30-mer oligonucleotide probes catalyzed by terminal deoxynucleotidyl transferase. After prehybridization at 55°C in the hybridization buffer, the blots containing transferred PCR products were incubated with the specific probe (Robbins Scientific incubator, Sunnyvale, CA) for 2 hours at 55°C. The membranes were rinsed extensively with 5× SSC and 0.1% sodium dodecyl sulfate (SDS) and then two times (30 minutes each) with 1× SSC and 0.1% SDS at 55°C. After blocking, the membranes were incubated with anti-fluorescein horseradish peroxidase conjugate, and exposed to ECL detection reagents (Amersham Life Science). The resultant light emission was detected on film (Hyperfilm-ECL; Amersham Life Science). The exposure time varied from a few minutes to 2 hours. 

The Southern blots were analyzed by laser densitometry (Ultrascan XL densitometer; Pharmacia LKB), and the density of the specific bands representing the spliced-in and spliced-out forms of fibronectin pre-mRNA was evaluated by computer software (Gelscan XL; Pharmacia LKB). The percentage of the spliced-in isoforms (EDA positive [+] or EDB+) was calculated, and the results obtained from trabecular cells receiving different treatments were compared. 

Quantitative Western Blot Analysis for Fibronectin

First passage trabecular cells were first serum starved for 48 hours in culture medium containing 0.5% newborn calf serum (NCS) followed by incubation in SFM for 24 hours. Subsequently, the cells were treated with TGF-β2 (0.2, 0.5, 1.5, and 2.7 ng/ml) in SFM, or in SFM alone for 48 hours, and the medium was collected. Then, the cells were washed, harvested by scraping with a disposable rubber policeman, and placed in ice-cold Triton X-100 and protease inhibitors. The protein from cell extracts and conditioned medium was quantified by using a protein assay kit (Bio-Rad, Richmond, CA). After heating to 100°C for 5 minutes, the samples (containing an equal amount of total protein, 40 μg) were loaded on 7.5% polyacrylamide gel, and separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) using a gel electrophoresis system (Hoefer) at a constant current of 15 mA. Rainbow molecular weight markers (Amersham) were used as reference for calculation of the size of fibronectin based on its migration distance. The fractionated polypeptides were electroblotted onto a 0.25-μm nitrocellulose membrane (Schleicher & Schuell) at 14-V constant voltage overnight at 4°C. After blocking with 3% nonfat milk in Tris-buffered saline (10 mM Tris [pH 7.4] and 150 mM NaCl) for 1 hour, the membrane was incubated with a polyclonal antibody against fibronectin (Sigma, St. Louis, MO) at a dilution of 1:1000. The primary antibody, developed in rabbits by using purified human fibronectin as the immunogen, cross reacts with porcine fibronectin, according to manufacturer’s specification. The membrane was washed, incubated with secondary antibody conjugated to peroxidase (1:200), and developed in the peroxidase substrate, diaminobenzidine (DAB). The amount of fibronectin corresponding to the intensity of the color-reaction band was quantified by laser densitometry. 

Immunocytochemical Analysis of Fibronectin

Trabecular cells of primary cultures were trypsinized and grown on tissue culture chamber slides (Miles, Naperville, IL). After they were rinsed with phosphate-buffered saline (PBS) three times for 5 minutes each, the cells, rendered quiescent by incubation in SFM for 24 hours, were treated with TGF-β2 (1 ng/ml in SFM) or SFM alone for 48 hours. Subsequently, the slides were fixed with cold absolute methanol for 10 minutes. After three rinses with PBS, the cells were permeabilized with acetone for 7 minutes at −20°C and washed thoroughly in three changes of PBS. The cells were placed in 0.3% hydrogen peroxide-methanol solution for 30 minutes to inactivate intrinsic peroxidase and rinsed with PBS. They were then incubated sequentially with 1% normal goat blocking serum in 4% bovine serum albumin for 20 minutes, rabbit anti-human fibronectin antibody (1:500 dilution, Sigma) or normal nonimmune rabbit serum at the same dilution (negative control) for 2 hours, biotinylated goat anti-rabbit antibody (Vector, Burlingame, CA) for 30 minutes, and avidin-biotinylated horseradish peroxidase complex (Vector) for 45 minutes. After three rinses with PBS, the reaction product was developed in 0.03% hydrogen peroxide, 0.037% 3-amino-9-ethylcarbazole-dimethyl-formamide in 0.1 M acetate buffer (pH 5.2). The slides were mounted with glycerol and examined by light microscopy. 

Results

Effects of TGF-β2 on Alternative Splicing of Fibronectin Pre-mRNA in Trabecular Cells

By using the primer pairs described, RT-PCR experiments generated two products for EDA and two products for EDB from cultured porcine trabecular cells incubated in serum, TGF-β2, or TGF-β1. For EDA, the products were 604 and 334 bp, the expected sizes for the EDA+ and EDA-negative (−) isoforms, respectively. For EDB, the two PCR-amplified fragments were 775 and 502 bp, consistent with the anticipated products for the EDB+ and EDB− isoforms, respectively. Southern hybridization confirmed that these PCR products were specific for EDA and EDB (Figs. 1A 1B , respectively). 

Densitometric quantification revealed that first-passage trabecular cells cultured in SFM alone (untreated controls) produced only EDA− or EDB− fibronectin mRNA isoforms (Figs. 2 3) . Incubation of trabecular cells in medium that contained 1 ng/ml TGF-β2, 1 ng/ml TGF-β1, or 15% NCS for 48 hours induced the expression of the EDA+ and EDB+ mRNA isoforms (Figs. 2 3) . The amount of the EDA+ isoform accounted for 17%, 14%, and 11%, and that of the EDB+ isoform 21%, 15%, and 5%, of the total fibronectin mRNA after treatment of the cells with 1 ng/ml TGF-β2, 1 ng/ml TGF-β1, or 15% NCS, respectively (Table 1) . 

Quantitative Western Blot Analysis for Fibronectin

Western hybridization and densitometric analysis demonstrated that TGF-β2 enhanced the production of fibronectin by the trabecular cells and also increased the secretion of this glycoprotein into the conditioned medium. At concentrations of 0.2, 0.5, 1.5, and 2.7 ng/ml, TGF-β2 increased the amount of fibronectin in the conditioned medium by 2-, 3-, 3.8- and 5-fold, respectively, and in the cell extracts, by 3-, 3.7-, 4-, and 4.3-fold over that from trabecular cells incubated in SFM alone (untreated control samples; Figs. 4 5 ). 

Immunocytochemical Analysis of Fibronectin

The trabecular cells treated with TGF-β2 exhibited strong immunoreaction with the antibody against cellular fibronectin. Positive staining was seen predominantly in the perinuclear region of the cytoplasm (Fig. 6A ). Only a minimal immunoreaction was observed in the cells that were incubated with SFM (Fig. 6B) . 

Discussion

The morphologic organization of the trabecular system in porcine and primate eyes exhibits several similarities. ^{17 18} Also, the cellular and ECM components of the porcine meshwork react with a battery of antibodies to human epitopes. ^{15 19 20 21} Comparison of confluent primary cultures initiated from fresh eyes of pigs with those from human eyes obtained after death reveals significant similarities in cell morphology, immunoreactivity, biochemistry, and pharmacology. ^{15 19 22} However, the availability of human trabecular cells was limited because of the inadequate supply of fresh donor eyes. In addition, the postmortem changes that occur not only in the aqueous humor, but also in the tissues that border the anterior chamber liberate proteases, ribonucleases, and other endogenous enzymes that degrade both polypeptides and mRNAs within the cells. To overcome these constraints, we used our established tissue culture model of trabecular cells from porcine eyes. ¹⁵  

Treatment of trabecular cells with TGF-β2 at concentrations that are detected in the aqueous humor of normal and glaucomatous human eyes increased their production and secretion of fibronectin in a dose-dependent manner. This finding illustrates the importance of increased levels of both total and intrinsically active forms of TGF-β2 in patients with POAG and could explain the excessive deposition of fibronectin in the aqueous outflow pathway of these patients. ^{1 24 25} The enhanced production of fibronectin as demonstrated by Western hybridization may be the result of an increased synthesis, decreased degradation, or both. The strong reaction product of the antibody against cellular fibronectin as observed in the perinuclear region of trabecular cells after incubation with TGF-β2 suggests that the increased accumulation of fibronectin is due at least in part to augmented protein synthesis. 

Laser densitometry has been used routinely to quantify autoradiograms, gels, and blots. ^{25 26 27 28 29} In our present study, the percentage of fibronectin mRNA that contained an alternatively spliced exon were determined by laser densitometric measurements of the Southern blot analysis. The results were calculated based on the ratios of the density of the bands obtained from the same PCR reaction with a single pair of primers, and each data point is, therefore, internally controlled. Our findings demonstrated for the first time that TGF-β2 is able to regulate the alternative splicing pattern of fibronectin pre-mRNA in trabecular cells and significantly increases the proportions of EDA+ and EDB+ isoforms. 

Rat liver and skeletal muscle consistently express EDA− and EDB− isoforms of fibronectin, and rat spleen, liver, skeletal muscle, and lung express EDB− isoforms only. ¹⁴ In our current experiments, serum-starved porcine trabecular cells expressed only the EDA− and EDB− isoforms of fibronectin. This reflects either the actual synthetic activity of trabecular cells in vivo, or it is the result of serum starvation. Future studies using trabecular tissue ex vivo will help resolve this question. We chose to use serum-starved cells and SFM to study the effect of TGF-β2 because serum is known to increase or even induce EDA+ and EDB+ mRNA. ¹⁴ Our finding of 11% EDA+ and 5% EDB+ in cells incubated in 15% NCS probably reflects the effect of a variety of growth factors and cytokines present in serum, including TGF-β1 (but not TGF-β2 which is not detectable in serum). The effect of TGF-β1 alone may have been diminished by other unknown serum components. 

Alternative splicing of the fibronectin gene during its translation is a well-known phenomenon and accounts at least in part for the known different forms of this protein. ³⁰ The functional significance of the alternatively spliced regions of fibronectin is not fully understood, but the presence of the spliced-in fragments in the protein could influence the extent of its posttranslational modifications such as sialation of the molecule. An aberration in either or both of these processes may result in an increase in sialation of the molecule. Additional terminal sialic acid residues would render fibronectin more resistant to the action of proteolytic enzymes, so that its normal degradation and turnover would be disrupted and manifest as increased deposition in the ECM of the trabecular meshwork–Schlemm’s canal system. 

An accumulation of fibronectin could also occur because of an imbalance in the activity of the degradative enzymes and their inhibitors. Proteases, such as tissue plasminogen activator (tPA), ^{15 31 32} urokinase, and plasmin, and metalloproteases, especially stromelysin, ³³ that are present in the aqueous humor and/or are produced locally by trabecular cells, are thought to contribute to the removal of fibronectin from the aqueous outflow pathway. Limited proteolysis of fibronectin by these various enzymes would result in the formation of fibronectin fragments that express functions not ordinarily associated with the intact protein and acquire functional activities de novo. ^{34 35 36 37 38}  

TGF-βs have a strong regulatory function on the accumulation of ECM macromolecules. They activate gene transcription, increase the synthesis and secretion of matrix proteins and protease inhibitors, and decrease the synthesis of degradative proteolytic enzymes. ^{39 40} In addition, TGF-βs have a potent inhibitory effect on the rate of proliferation and motility of trabecular cells in vitro. ¹³ Thus, the elevated amounts of both intrinsically active and total TGF-β2 present in the aqueous humor of patients with POAG, could have a significant role in the pathogenesis of this disease that is characterized by an excessive build-up of ECM proteins and decreased cellularity in the trabecular meshwork. ^{41 42 43 44}  

Senescent trabecular cells in vitro showed enhanced production of fibronectin, type VI collagen, and thrombospondin, ²⁶ supporting the contention that POAG is an exaggerated aging process. It would be interesting to determine whether TGF-β2 in aqueous humor is elevated in elderly eyes compared with that in youth. TGF-β positively regulates its own gene expression, ²⁵ which may lead to an autocatalytic and autoinductive cascade of TGF-β amplification. Thus, this cytokine may be the common pathway for the changes in the aqueous outflow pathway in both aging and glaucomatous eyes. 

Future studies are needed to determine whether the increased concentration of TGF-β2 in the aqueous humor of patients with glaucoma has a clinically significant impact on the structure and function of trabecular cells so that aqueous outflow resistance and intraocular pressure are also influenced. Such experiments will enhance our understanding of the role of TGF-β2 in the disease process of POAG and can open up the prospect for use of its antagonists in the treatment and prevention of increased resistance to aqueous outflow. 

 Supported by National Institutes of Health Grants EY-08707 and EY-03747, Research to Prevent Blindness, and Vision Research Foundation.

 Submitted for publication September 1, 1999; revised February 29 and June 6, 2000; accepted June 9, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Splicing of fibronectin pre-mRNA at the (A) EDA and (B) EDB regions analyzed by Southern hybridization of the RT-PCR products. Lanes 1 through 4: Blots from cells in SFM, treated with TGF-β1 (1 ng/ml), TGF-β2 (1 ng/ml), or 15% NCS, respectively. Length in base pairs corresponding to the DNA standard size markers, left. (A) The 334-bp and the 604-bp bands represent EDA− and EDA+ isoforms, and (B) the 502-bp and the 775-bp bands represent EDB− and EDB+ isoforms.

Figure 2.

 

Relative densitometric quantification of Southern blot analysis at the EDA region in trabecular cells incubated in SFM, TGF-β1 (1 ng/ml), TGF-β2 (1 ng/ml), or 15% NCS. Values represent the EDA+ isoform as a percentage of the total fibronectin mRNA (mean ± SE) from triplicate experiments.

Figure 3.

 

Relative densitometric quantification of Southern blot analysis at the EDB region in trabecular cells incubated in SFM, TGF-β1 (1 ng/ml), TGF-β2 (1 ng/ml), or 15% NCS. Values represent the EDB+ isoform as a percentage of the total fibronectin mRNA (mean ± SE) from triplicate experiments.

Table 1.

 

Effects of TGF-β and NCS on Fibronectin Pre-mRNA Alternative Splicing in Porcine Trabecular Cells

	EDA+ Isoform (%)	EDB+ Isoform (%)
SFM	0 (EDA− only)	0 (EDB− only)
TGF-β2	17	21
TGF-β1	14	15
NCS	11	5

Figure 4.

 

Western blot analysis of fibronectin in trabecular cells treated with TGF-β2. (A) Cell extract; (B) conditioned medium. Lane 1: cells in SFM; Lanes 2 through 5: cells incubated in 0.2, 0.5, 1.5, and 2.7 ng/ml TGF-β2, respectively. Arrow: 220-kDa fibronectin band.

Figure 5.

 

Relative densitometric quantification of intracellular and secreted fibronectin, expressed as the increase over that in SFM, after incubation of trabecular cells in TGF-β2. Values represent the mean ± SE from triplicate experiments.

Figure 6.

 

Light micrographs show immunocytochemical staining for fibronectin in first-passage trabecular cells. (A) Cells treated with TGF-β2 (1 ng/ml) for 24 hours show an intense positive immunoreaction, especially in the perinuclear region. (B) Cells incubated in SFM show minimal immunoreactivity (negative control). Magnification, ×300.