INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

AMONG THE COMPONENTS of the extracellular matrix (ECM), there are various nonstriated fibrils, known as microfibrils, which are classified by the size of their diameters (24, 27, 29, 33). Fibrillin-1 and -2 seem to be the major structural components of the beaded 10- to 14-nm microfibrils that are widely distributed in elastic and nonelastic tissues (9, 10, 18). In elastic tissues like the aorta, fibrillin-1-associated microfibrils are thought to play a role in the formation of elastic fibers, thus allowing them to generate an elastic recoil (4, 24). Conversely, in nonelastic tissues they may anchor epithelial cells to the interstitial matrix and therefore participate in various biological processes, e.g., wound healing and embryonic development (42, 43). The structure of human fibrillin-1 and -2 has been described, and mutations in the genes have been found in Marfan syndrome, which is associated with connective tissue abnormalities of the eye, skeletal system, and cardiovascular system (6, 15, 20, 23, 26, 39). The cDNA sequence of fibrillin-1 in human and mouse is highly conserved (6, 23, 42). Analysis of the latter indicates that it is a modular protein, consisting of 48 epidermal growth factor (EGF)-like domains, 7 domains homologous to transforming growth factor-1 (TGF-1) binding protein (8-cysteine motifs), a fibrillin motif, a fibrillin-like module, a 57 amino acid proline-rich domain, a single RGD coding sequence, and unique amino and carboxy termini (23, 27, 29, 34, 42). Most (43/48) of the EGF-like domains have Ca²⁺-binding (cb) sites (5, 12, 28). Such cbEGF-like domains are believed to be involved in protein:protein interactions and have been found in other ECM proteins like fibulin-1 and -2, nidogen, and versican (12, 13, 28, 30, 31, 36).

Conceivably, the structural and biochemical characteristics of fibrillin-1, i.e., RGD sequence and cbEGF-like domains, as seen in some of the ECM proteins, may be relevant to the various biological processes prevalent during embryonic development. For instance, the RGD sequence mediating fibroblast attachment to fibrillin-1 is sensitive to inhibition by antibodies to the _v3-integrin receptor (34), which has been shown to regulate metanephric development. Second, nidogen that contains cbEGF-like domains interacts with another major ECM protein, laminin, and both seem to regulate tubulogenesis during renal development (7). Third, cbEGF-like domains are present in neurogenic loci Notch and Delta, the two EGF-homologous genes in Drosophila (8, 28). Their expression in nonadhesive Drosophila Schneider's 2 (S2) cells induces aggregation, suggesting their role in protein:protein and cell:cell interactions (8), with the latter necessary for organogenesis during embryonic development. In view of the above considerations, studies were initiated to delineate the role of fibrillin-1, a protein with multiple cbEGF-like domains (27, 29), in rat development by utilizing the rat metanephric culture system.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Sprague-Dawley rats (Harlan Sprague Dawley, Indianapolis, IN) were used for paired male-female mating, and the appearance of the vaginal plug was designated as day 0 of fetal gestation. Kidneys from embryonic rats were harvested aseptically at days 15 (E15) and 18 (E18) of gestation, and in addition, kidneys were obtained from the newborn and 1-, 2-, and 3-wk-old rats.

Construction of newborn rat kidney cDNA library. Total RNA from ~100 newborn rat kidneys was isolated by the guanidinium isothiocyanate-CsCl centrifugation method (2). Poly(A)⁺ RNA was selected by oligo(dT)-cellulose chromatography. First-strand cDNA was synthesized by using Moloney murine leukemia virus reverse transcriptase (MMLV-RT, RNase H) and oligo(dT)₂₅d(G/C/A) as a primer (Clontech, Palo Alto, CA). Double-stranded cDNA was prepared by the RNase H-DNA polymerase I method for second strand synthesis (17). After ligation of EcoR I-Not I-Sal I adapters and polynucleotide kinase phosphorylation, the double-stranded cDNA was size-fractionated, and small DNA fragments, i.e., <400 bp, were removed by a Chromaspin-1000 column (Clontech). The cDNA was ligated into an EcoR I-digested and dephosphorylated -ZAP II vector and packaged using Gigapack II Gold Packaging Extract (Stratagene, La Jolla, CA). The packaged ligation product was incubated with E. coli XL1-Blue MRF' cells for plating and titration of recombinant phage plaques (17).

Screening of rat newborn rat cDNA library and isolation and nucleotide sequence analyses of fibrillin-1 cDNA clones. A mouse fibrillin-1 cDNA, previously isolated in our laboratory (GenBank accession no. U22493), was used for screening the rat newborn cDNA library. About 2 × 10⁶ recombinants were screened with [-³²P]dCTP-labeled mouse fibrillin cDNA. Nitrocellulose filter lifts (Schleicher and Schuell, Keene, NH) were made, then prehybridized and hybridized with the radiolabeled screening probe as previously described (40). Several clones were isolated, then purified by dilutional secondary and tertiary screenings. The clones containing cDNA inserts, which strongly hybridized with the screening probe, were processed for further subcloning. Eleven overlapping clones were isolated and subcloned into pBluescript II KS(+) phagemid using XL1-Blue MRF' cells (Stratagene). For rescuing of single-stranded DNA, the transfected cells were grown in the presence of VCSM13 helper phage. The supernatants of the cultures were saved, and single-stranded DNA was isolated by polyethylene glycol precipitation. After the sense or antisense orientation of various subclones was determined, nucleotide sequencing was performed by the dideoxy chain termination method (35), followed by hydropathic (19) and sequence homology analyses (22) using the Wisconsin Package Version 9.1-UNIX (Madison, WI).

Gene expression studies by Northern blot analysis and in situ hybridization. Total RNA from embryonic kidneys at various stages of gestation and from kidneys of 1- and 3-wk-old rats was extracted by the guanidinium isothiocyanate-CsCl centrifugation method (2). Equal amounts of RNA, extracted from rat kidneys at various stages of gestation and the neonatal period, were glyoxalated and subjected to 1% agarose gel electrophoresis in 10 mM sodium phosphate buffer, pH 7.0. A Northern blot was prepared by transferring the RNA to a nylon filter membrane (Amersham, Arlington Heights, IL) and hybridizing with [-³²P]dCTP-labeled rat fibrillin-1 cDNA. The filter was washed under high-stringency conditions with 0.1× SSC and 0.1% SDS at 60°C, and autoradiograms were prepared. After stripping, the same blot was also hybridized with a -actin probe (GenBank accession no. M62174; American Type Culture Collection, Rockville, MD), and the autoradiogram was prepared as described above.

For in situ hybridization studies, first a PCR product of 771 bp was prepared by using the primers, derived from rat fibrillin-1 cDNA. The respective sense and antisense primers were as follows: 5'-CATCCGCACTGGAGCTTGTC-3' and 5'-CACTCATCAACGTCGATGC-3'. This fibrillin-1 cDNA PCR product was ligated into pBluescript KS(+) and used as a template for generating sense and antisense riboprobes by using the Riboprobe In Vitro Transcription System (Promega, Madison, WI). The riboprobes were synthesized by incorporating [-³³P]UTP (Amersham), using T3 and T7 RNA polymerases and the linearized PCR product of rat fibrillin-1 cDNA. The radiolabeled riboprobes were subjected to limited alkaline hydrolysis to yield polynucleotide fragments of 100-150 bp size, which were then used for in situ hybridization with the tissue sections (17, 40).

Embryonic (E15 and E18), newborn, and 1- and 3-wk-old rat kidneys were immersed in 4% paraformaldehyde in PBS, pH 7.0, for 3 h at 4°C. The tissue specimens were then dehydrated, and embedded in paraffin. Tissue sections, 3 µm thick, were prepared and mounted on glass slides coated with Vectabond (Vector Laboratories, Burlingame, CA). The sections were deparaffinized, hydrated, treated with 0.2 N HCl, deproteinated by proteinase K treatment, and acetylated with 0.1 M triethanolamine and 0.25% acetic anhydride. After the sections were washed with 2× SSC, they were prehybridized with hybridization solution (50% formamide, 10% dextran sulfate, Denhardt's solution in 10 mM Tris · HCl, pH 8.0) at 50°C for 3 h, then hybridized with [-³³P]UTP-labeled fibrillin riboprobes at 50°C for 15 h. After hybridization, the tissue sections were washed with 50% formamide in 2× SSC, treated with RNase A, and rewashed with 0.1× SSC at 50°C. The sections were then dehydrated, air-dried, and coated with NTB2 photographic emulsion (Eastman Kodak, New Haven, CT), and tissue autoradiograms were prepared after 1-2 wk of exposure.

Generation of anti-fibrillin-1 antibody and its characterization. To raise a polyclonal antibody to fibrillin-1, a synthetic peptide was prepared with the following amino acid sequence: RPPPEYPYPSPSREPPK. This stretch of amino acids is boldly underscored in Fig. 1A. An additional lysine residue was added to the NH₂ terminus of the peptide for its conjugation with keyhole limpet hemocyanin. One milligram of conjugated peptide was mixed with complete Freund's adjuvant and used for immunizing rabbits. Booster injections of the antigen were given every 3 wk, and rabbit antisera were collected. An IgG fraction was prepared from the antisera by ammonium sulfate precipitation, as previously described (41). The purified IgG fraction was dialyzed against PBS, lyophilized, and stored at 70°C.

View larger version (88K): [in this window] [in a new window]   View larger version (91K): [in this window] [in a new window]   Fig. 1.   Nucleotide and deduced amino acid sequence of full-length rat fibrillin-1 cDNA. Shaded box (top left in A): sequence antisense oligodeoxynucleotide (ODN) used in organ culture experiment. Open box (bottom right in A): location of the RGD sequence. Bold underscore (middle left of A): amino acid sequence of the synthetic peptide used for the generation of the antibody. Underscore (right in B): sense and antisense primers used for the competitive RT-PCR analyses.

The specificity of the antibody was established by ELISA, the procedural details of which have been described previously (41). Briefly, wells of an EIA/RIA plate were coated with 5 µg of synthetic peptide, followed by washing with methanol and blocking with bovine serum albumin. After two washes with PBS, 0.5 µg of the antibody (IgG fraction) was added to the first well, and log dilutions of the antibody were made in successive wells. Incubation was carried out for 1 h, and the wells were rewashed with PBS containing 0.05% Tween-20. Horseradish peroxidase-conjugated goat anti-rabbit IgG (Cappel; Organon Teknika, Durham, NC) diluted 1:1,000 was added and incubated for 30 min. After three rewashings, a colorimetric reaction was carried out with tetramethylbenzidine solution (Bio-Rad Laboratories, Hercules, CA), and the reaction was stopped with the addition of 0.3 M H₂SO₄. Finally, readings at OD₄₉₀ were made and plotted against the log dilutions of the antibody. For a competitive inhibition ELISA assay, 250 µg of the synthetic peptide were added in the first well of the peptide-coated EIA/RIA plate along with 0.5 µg of the antibody. Log dilutions of the synthetic peptide were made in successive wells while the concentration of the primary antibody was kept constant. Conditions for incubation with secondary antibody and colorimetric reactions were the same as described above. Readings at OD₄₉₀ were made and plotted against the log dilutions of the antigen.

The specificity of the antibody was also determined by immunoprecipitation methods (41). About 100 E15 metanephroi were harvested and maintained in an organ culture system, as detailed previously (21). Briefly, harvested metanephric explants were placed on a 0.8-µm pore size filter and floated onto a serum-free medium. The latter was made up of equal volumes of DME and Ham's nutrient mixture F-12, penicillin (100 µg/ml), streptomycin (100 µg/ml), and transferrin (50 µg/ml) (Sigma Chemical, St. Louis, MO). The explants were cultured in a CO₂ incubator for 4 days and then radiolabeled with [³⁵S]methionine (0.25 mCi/ml) for 16 h prior to the termination of the culture. They were lysed in an extraction buffer [6 M guanidine-HCl, 100 mM Tris · HCl, pH 7.5, 0.02% NaN₃, 10 mM -amino-n-caproic acid, 5 mM N-ethylmaleimide, and 1 mM phenylmethylsulfonyl fluoride (PMSF)] by shaking vigorously for 6 h at 4°C. The extract was centrifuged at 10,000 g for 30 min at 4°C. Ten volumes of ethanol were added to the supernatant to precipitate the extracted protein at 20°C for 15 h. The precipitate was centrifuged at 10,000 g for 30 min at 4°C, and the pellet was resuspended in the immunoprecipitation buffer and saved. The immunoprecipitation buffer consisted of 0.1% SDS, 1% Triton X-100, 50 mM Tris · HCl, pH 7.5, 50 mM NaCl, 0.02% NaN₃, 0.25 mM dithiothreitol, 10 mM benzamidine-HCl, 10 mM -amino-n-caproic acid, 5 mM N-ethylmaleimide, and 1 mM PMSF. Immunoprecipitation was performed by adding 5 µl of polyclonal anti-fibrillin-1 antibody to 0.5 ml (5 × 10⁶ dpm) of the sample. The mixture was gently swirled in an orbital shaker for 15 h at 4°C. After addition of 80 µl of protein-A-Sepharose 4B (Pharmacia LKB Biotechnology, Piscataway, NJ), the antigen-antibody mixture was further incubated for 1 h at 4°C. The antigen-antibody complex was washed three times with the buffer, then dissolved in a sample buffer (4% SDS, 150 mM Tris · HCl, pH 6.8, 20% glycerol, 0.1% bromophenol blue, and 10% -mercaptoethanol), boiled for 5 min, and subjected to 5% SDS-PAGE analysis. The gels were then fixed in 10% acetic acid and 10% methanol, treated with 1 M salicylic acid, and vacuum dried, and autoradiograms were prepared. Two control immunoprecipitation experiments were also performed. In the first, polyclonal anti-fibrillin-1 antibody, previously absorbed with the synthetic peptide, was used for immunoprecipitation; in the second, preimmune rabbit serum was used as a control. In both the controls, double the amount of immunoprecipitated radioactivity was used in SDS-PAGE analyses to ensure the specificity of the antibody.

Fibrillin-1 expression by tissue immunofluorescence. Kidneys of E15 and E18 embryos and of newborn and 1-, 2-, and 3-wk-old rats were snap frozen in chilled isopentane and embedded in OCT compound (Miles Laboratories, Elkhart, IN). Cryostat sections, 4 µm thick, were prepared and air-dried. Sections were washed with 0.01 M PBS, pH 7.4, and incubated with polyclonal anti-fibrillin-1 antibody, with 1:100 dilution, for 30 min in a humidified chamber at 37°C. After washing with PBS, sections were reincubated with goat anti-rabbit IgG antibody and conjugated with fluorescein isothiocyanate for 30 min. The sections were rewashed with PBS, covered with a drop of buffered glycerol, coverslip mounted, and examined with an ultraviolet microscope equipped with epi-illumination.

Antisense experiment. The antisense experiments were performed to determine the role of fibrillin-1 in the organogenesis of embryonic kidneys. A sense-, two nonsense-, and antisense-phosphorothioated oligodeoxynucleotides (ODN) were synthesized by an automated solid-phase synthesizer (Biotech facility, Northwestern University) and purified by high-performance liquid chromatography. The 35-mer sense/antisense ODN sequence was selected from the 5' end of the cloned rat fibrillin-1, and it is as follows: 5'-GCCAGCGCGACCTCCAGCAGCCCTCCTCGCCGCAT-3' (Fig. 1). Its specificity for the target nucleotide sequences was established by S₁ nuclease protection assays as described previously (see Fig. 5D and see RESULTS). Two nonsense 31-mer phosphorothioated ODNs were also prepared for these experiments, and their sequences were as follows: 5'-TAATGATAGTAATGATAGTAATGATAGTAAT-3' and 5'-GATCGATCGATCGATCGATCGATCGATCGAT-3'. Both the ODNs (antisense and nonsense) did not exhibit any significant homology with other mammalian nucleotide sequences available in the GenBank database.

About 600 rat embryonic kidneys at day 15 (E15) of gestation were harvested and maintained in culture for 4 days. The ODNs were added to the culture media daily at a concentration of 0.5 µM. At this concentration, the ODNs retain the translational blockade specificity with no discernible cytotoxic effects (1, 17, 40). The metanephric explants (200 kidneys per variable, i.e., sense, antisense, and nonsense) were processed for light microscopy, quantitative RT-PCR analyses, and immunofluorescence and immunoprecipitation studies. For light microscopy, the sections from the midplane of the embryonic kidneys with a maximum number of ureteric bud iterations, including both the poles and the hilus, were evaluated as described previously (17, 40).

Quantitative RT-PCR analyses of fibrillin-1 mRNA of antisense ODN-treated metanephroi. To assess the effect of antisense ODNs on mRNA expression, competitive RT-PCR analyses were carried out as described previously (17, 40). Total RNAs were isolated from 50 explants per variable by an acid guanidinium isothiocyanate-phenol-chloroform extraction method (3). Extracted RNAs were treated with RNase-free DNase (Boehringer Mannheim, Indianapolis, IN), followed by an ethanol precipitation. About 25 µg of total RNAs, from each variable, were subjected to first-strand cDNA synthesis using MMLV-RT and oligo(dT) as a primer. The cDNAs from different variables were suspended in 25 µl of deionized autoclaved water and kept at 70°C until further use.

For the analyses of fibrillin-1 mRNA, the respective sense and antisense primers were 5'-GATCATATCACTGCATCTG-3' and 5'-GAGCAACCATAACTGCAGG-3'. Their locations in the rat fibrillin-1 cDNA are indicated in Fig. 1B as underscored nucleotide sequences. For -actin, the respective sense (-SE) and antisense (-AS) primers were 5'-GACGACCATGGAGAAGATCTGG-3' and 5'-GAGGATGCGGCAGTGCGGAT-3' (38). Using these primers, we expected the PCR product sizes to be 723 bp for rat fibrillin-1 and 461 bp for -actin, and their nucleotide sequences were confirmed by the dideoxy chain termination method (35). The 723-bp PCR product was then used for the preparation of a competitive DNA template for fibrillin-1. A Hinc II site was introduced by using a nested primer with the sequence 5'-GGGTAACCACCGCTGCCAACTGG-3', and the resulting PCR product was ligated into pCR II vector (Invitrogen, San Diego, CA). The plasmid with this cDNA insert was digested with Hinc II (Boehringer Mannheim) to delete the 229-bp fibrillin-1 DNA fragment flanked by Hinc II sites. The digested plasmid was purified by agarose gel electrophoresis and subjected to self-ligation with DNA ligase (Boehringer Mannheim). The plasmid, containing the truncated 494-bp fibrillin-1 insert, was linearized with appropriate restriction enzyme digestion and used as a competitive DNA template for fibrillin-1 mRNA analyses by RT-PCR. The construction of a competitive 224-bp DNA template for -actin has been described in our previous publications (17, 40; GenBank accession no. U17140).

For quantitative RT-PCR analyses, a fixed amount of cDNAs (1 µl) from antisense and nonsense ODN-treated metanephroi and serial logarithmic dilutions of the competitive template DNA (500 ng/µl) of fibrillin-1 were coamplified (11). The reaction mixture included 5 µl of 10× PCR buffer, 250 µM of each dNTPs, 1 µM of sense and antisense primers, and 1 U Taq polymerase (Perkin-Elmer, Norwalk, CT) in a total volume of 50 µl. The amplification reaction was carried out for a total of 30 cycles in a DNA Thermal Cycler (Perkin-Elmer), each cycle consisting of denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and extension at 72°C for 1 min. The PCR products of wild-type and mutant fibrillin-1 (competitive truncated DNA template) were analyzed by 2% agarose gel electrophoresis and photographed using an instant positive/negative film (Polaroid, Cambridge, MA). The negatives were analyzed by a scanning densitometer (Hoefer Scientific Instruments, San Francisco, CA), and the relative area underneath the tracings was computed. Similarly, the wild-type and mutant -actin were analyzed. The ratios between the densitometric readings of wild-type and mutant PCR-DNA products were plotted using a logarithmic scale on the ordinate (y-axis) against the logarithmic dilutions of the competitive truncated template DNA on the abscissa (x-axis).

Fibrillin-1 expression in antisense ODN-treated rat metanephric explants. To assess the translational blockade of fibrillin-1, immunoprecipitation and immunofluorescence studies were performed on metanephroi treated with various ODNs, i.e., antisense, sense, and nonsense, for 48 h. For immunoprecipitation experiments, [³⁵S]methionine-labeled metanephroi were processed for SDS-PAGE analyses as described above. The controls included the untreated metanephroi and those treated with sense and nonsense ODNs. To ensure the effect of antisense ODN on the translational blockade of rat fibrillin-1, double the amount of immunoprecipitated radioactivity was used for SDS-PAGE analyses. Finally, the tissue expression of fibrillin-1 in antisense and sense or nonsense ODN was assessed by immunofluorescence microscopy as described above.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Characterization of newborn rat fibrillin-1 cDNA clones. Eleven clones were isolated from a newborn rat kidney library that had overlapping sequences and common restriction sites, indicating that they contained identical cDNA. Clone 4 had the initiation codon, and clone 7 had the termination codon. By combining the nucleotide stretches with shared sequences of various clones, an open-reading frame consisting of 8,616 nucleotides was obtained, which had a deduced translated product of 2,872 amino acids (Fig. 1). Rat fibrillin-1 had ~88% and ~96% sequence homology with human fibrillin-1 (26) at the nucleotide and amino acid levels, respectively. The sequence homology with the mouse fibrillin-1 gene was ~94% and ~98% at the nucleotide and amino acid levels, respectively (42). The amino acid sequence had similar structural domains to that of the human and mouse fibrillin-1 (26, 42). Its domain structure was characterized by a proline-rich domain stretching from amino acids 391 to 450, which contained 41.6% proline residues. This domain was flanked by 8 and 49 cysteine-rich repeats upstream and downstream, respectively. The majority of the cysteine-rich repeats had calcium binding consensus sequences, derived from cbEGF repeats of several different proteins (36). The latter, i.e., Ca²⁺-binding EGF-like domains, are thought to be involved in mediating protein:protein interactions (28). The downstream cysteine-rich repeats domain also contained six TGF binding protein repeats and a fibrillin motif. Another TGF binding protein-like repeat and a fibrillin-like module were present in the upstream cysteine-rich repeats domain.

Expression of fibrillin-1 mRNA. Northern blot analyses revealed a single mRNA transcript of ~10 kb size in kidneys harvested from E15 and E18 embryos and newborn and 1- and 3-wk-old rats (Fig. 2, top). The mRNA expression for fibrillin-1 at E15 of gestation was detectable when ~50 µg of total RNA, isolated from ~80 explants, was used. After a mild decrease at E18, the fibrillin-1 mRNA expression steadily increased during various developmental stages, and was maximal in kidneys of 3-wk-old rats. Thereafter, the mRNA expression leveled off, suggesting that fibrillin-1 is developmentally regulated. The mRNA expression of -actin in the rat kidneys was constant throughout the embryonic, neonatal, and postnatal periods (Fig. 2, bottom). Since the fibrillin-1 gene seemed to be developmentally regulated, in situ hybridization studies were performed to assess its spatial distribution in embryonic and neonatal kidneys (Fig. 3). At E15, the mRNA expression was found to be restricted to the rat metanephric loose mesenchyme, and no message was found in the ureteric bud epithelia or in the nascent nephron (Fig. 3, A and D). Interestingly, the fibrillin-1 message was observed around the ureteric bud branches and developing nephrons (Fig. 3D). Also, like some of the other ECM proteins, i.e., proteoglycans, the message was slightly concentrated on the tips of the ureteric bud branches (arrowhead in Fig. 3D). At E18, the fibrillin-1 expression was similar to that observed at E15. In newborn kidneys, the message was mainly found on the mature glomeruli, and it was somewhat concentrated in their mesangial regions (Fig. 3, B and E). Occasionally, the fibrillin-1 message was also detectable in the small intrarenal blood vessels, e.g., afferent arterioles (arrow in Fig. 3E). A mild degree of expression was observed in the interstitium in some of the tissue autoradiograms of the newborn kidneys. At 3 wk, besides glomeruli, a high degree of expression was observed in large intrarenal as well as extrarenal blood vessels (Fig. 3, C and F). The expression was notably high in hilar blood vessels lined with elastic lamina (Fig. 3F), whereas no expression was observed in tributaries of the renal vein. The expression in the renal interstitium decreased to a minimal degree in kidneys of 3-wk-old rats. In kidneys of 1- and 2-wk-old rats an increasing fibrillin-1 mRNA expression in the blood vessels was observed. The adjacent kidney sections hybridized with sense riboprobes yielded a mild background signal (Fig. 3, G-I).

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Northern blot analyses of rat fibrillin-1 mRNA expressed in various developmental stages of the rat kidney. A major transcript of ~10 kb for rat fibrillin-1 is observed at day 15 (E15) of the gestation. After a mild decrease at day 18 (E18), its expression increases progressively during the gestational and postnatal periods in developing kidneys. Expression of -actin remains constant. Lanes 15d and 18d: renal mRNA from 15- and 18-day-old fetuses. Lanes NB, 1W, and 3W: newborn and 1- and 3-wk-old rat renal mRNAs.

View larger version (126K): [in this window] [in a new window]   Fig. 3.   Dark-field light microscopic autoradiogram of renal tissue sections in situ hybridized with specific fibrillin-1 riboprobe. Kidney sections from E15 (A, D, and G), newborn (B, E, and H), and 3-wk-old rats (C, F, and I) were hybridized with [-35S]dUTP antisense (A-F) or sense (G-I) riboprobes. At E15 (A and D), fibrillin-1 mRNA is expressed in the metanephric mesenchyme (m) surrounding the nascent nephron (n) and ureteric bud branches (u). At places the message was seen concentrated at the tips of ureteric bud branches (arrowhead in D). In newborn rat kidney (B and E), fibrillin-1 mRNA is expressed in glomeruli (g) and occasionally in afferent arterioles (arrow in E). In the 3-wk-old rat kidney, mRNA is heavily expressed in the intrarenal (C) and extrarenal (F) arterial blood vessels (v), in addition to the glomeruli. A background signal is observed in kidney tissues hybridized with sense probe (G-I). Arrows in G mark the boundary of the embryonic metanephros, and dots in I outline a venous channel.

Characterization of anti-fibrillin-1 antibody. A polyclonal anti-fibrillin-1 antibody was raised using a synthetic peptide, the sequence of which is boldly underscored in Fig. 1A. This sequence is identical in both the mouse and rat species. The specificity of the antibody was characterized by ELISA and immunoprecipitation methods. In the ELISA assay, a fixed amount of antigen, i.e., synthetic peptide, and serial log dilutions of the antibody were used. With increasing dilutions of the antibody, a proportional decrease in OD₄₉₀ readings was observed (solid line in Fig. 4A). To confirm the antibody specificity, an inhibition ELISA assay was performed in which synthetic peptide was used as a competitive antigen. With increasing dilutions of the competitive antigen, a proportional increase in OD₄₉₀ readings was observed (broken line in Fig. 4A), indicating the specificity of the anti-fibrillin antibody. No change in the optical density readings was observed when synthetic peptide with sequences derived from fibrillin-2 was used. For immunoprecipitation, the embryonic explants were radiolabeled with [³⁵S]methionine, and the tissue extract was prepared. The extract was immunoprecipitated with anti-fibrillin-1 antibody prepared in two different rabbits. The immunoprecipitates were then subjected to 5% SDS-PAGE analyses. A single ~300-kDa band, similar to the size of human fibrillin-1, was observed for antibodies prepared in both the rabbits (Fig. 4B, lanes 2 and 3). This indicated that the antibody bound specifically to the putative fibrillin-1 polypeptide. No autoradiographic band was observed when the immunoprecipitation was performed with preimmune serum (Fig. 4B, lane 1), or when the antibody was preabsorbed with the synthetic peptide (Fig. 4B, lane 4). In these latter two controls, double the amount of radioactivity was loaded in lanes 1 and 4 to ensure that there was no detectable nonspecific binding to the fibrillin-1 polypeptide.

View larger version (95K): [in this window] [in a new window]   Fig. 4.   A: ELISA and competitive-inhibition ELISA profiles of fibrillin-1 antibody and the synthetic peptide, RPPPEYPYPSPSREPPK. A decrease in OD490 optical density readings is observed with increasing log dilutions of the anti-fibrillin-1 antibody (solid line), whereas an increase in OD490 readings is observed with increasing log dilutions of the competitive antigen, i.e., synthetic peptide (dotted line). No change in optical density readings was observed when synthetic peptide with sequences derived from fibrillin-2 was used. B: SDS-PAGE autoradiogram of de novo synthesized and immunoprecipitated fibrillin-1. Metanephroi were radiolabeled with [35S]methionine, and extracts were immunoprecipitated with anti-fibrillin-1 antibody. De novo synthesized fibrillin-1 is visualized as a band of ~300 kDa Mr as indicated by the large arrow for lanes 2 and 3. No band is seen either in preimmune serum (lane 1) or in antiserum preabsorbed with the synthetic peptide (lane 4). Small arrow indicates the point of application. C-H: immunofluorescence micrographs representing kidney tissues harvested from E15 (C) and E18 fetuses (D), newborn (E), and 1- (F), 2- (G), and 3-wk-old (H) rats. Immunofluorescence findings confirm the results obtained in gene expression studies. At E15 of gestation (C), immunoreactivity is confined to mesenchyme (*) and tissues surrounding the ureteric bud branches (u) and nascent nephrons (n). At E18 (D), reactivity is reduced, and it is confined to residual mesenchyme or interstitium (*) and around the blood vessels (v) and nephrons (arrows). In newborn kidney (E), immunoreactivity is mainly confined to glomerular mesangium (g) and blood vessels (v), while mild reactivity is seen in interstitium surrounding the tubules (arrows). Immunoreactivity in glomeruli (g) and blood vessels (v) is progressively accentuated in kidneys of 1- (F), 2- (G), and 3-wk-old (H) rats.

Immunofluorescence studies. Tissue expression was assessed by using the polyclonal anti-fibrillin-1 antibody in embryonic and neonatal rat kidneys. At E15 of gestation, the immunoreactivity of fibrillin-1 was observed in the loose metanephric mesenchyme. The fibrillin-1 was expressed as wrinkled filamentous structures, which seemed to surround the developing nephrons and the ureteric bud iterations (Fig. 4C). This expression was comparable to the spatiotemporal mRNA expression of fibrillin-1 (Fig. 3, A and D). No immunoreactivity was observed in the epithelial components of the embryonic metanephros. At E18, immunoreactivity was notably reduced, and it was confined to the interstitium and loose mesenchyme that had not yet been populated with developing nephrons (Fig. 4D). Interestingly, mild immunoreactivity with newly formed small blood vessels was also noted. At day 22 (newborn), immunoreactivity was seen in mature glomeruli and blood vessels (Fig. 4E). The vessels with elastic lamina exhibited intense reactivity. Mild immunoreactivity of the antibody could be observed in the interstitium as well. At 1, 2, and 3 wk, the immunoreactivity in the blood vessels and in the glomeruli was further accentuated, whereas that in the interstitium was considerably diminished (Fig. 4, F-H). The glomerular mesangium also exhibited a steady increase in the expression of fibrillin-1. The findings of the protein expression studies are comparable to the results obtained by in situ experiments. Both these studies indicate that fibrillin-1 has a spatiotemporal expression in the mammalian embryonic metanephros and thus may play a role in metanephric development.

Role of fibrillin-1 in rat renal development (antisense experiments). To study the role of fibrillin-1 in rat metanephrogenesis, gene disruption experiments, employing an antisense ODN strategy, were performed. Fibrillin-1-specific antisense ODN was used for these experiments. Its nucleotide sequence was derived from the 5' end of the fibrillin-1 gene (shaded box in Fig. 1). The specificity of fibrillin-1 antisense ODN was established by S₁ nuclease protection assay, as detailed in previous publications (17, 40). A single band of radioactivity, corresponding to the size of the ODN, i.e., 35-mer, was observed at 35°C, 40°C, 45°C, and 50°C hybridization temperatures (Fig. 5D; lanes 1, 3, 5, and 7, respectively). No band of radioactivity was seen in the control (nonsense ODN) samples (Fig. 5D; lanes 2, 4, 6, and 8). The morphological changes induced by the antisense ODN exposure to the embryonic metanephroi are depicted in Fig. 5C. An overall moderate reduction in the size of the metanephric explant was observed compared with the untreated control (Fig. 5A) or samples treated with sense/nonsense ODN (Fig. 5B). A notable decrease in iteration of the ureteric bud branches was seen. Also, the ureteric bud branches were disorganized, and there was a loss of acuteness of their tips (arrow heads in Fig. 5C). The metanephric mesenchyme was loosely organized and expanded, and mesenchymal cells appeared to be shrunken or atrophic. With the disorganization of the ureteric bud branches and atrophy of the mesenchymal cells, the population of glomerular and tubular elements remarkably decreased. A mild reduction in the size of the metanephric explant treated with nonsense ODN was observed; however, no abnormalities in the ureteric bud iterations or decrease in the population of nascent nephrons was noted. No discernible cytotoxic effects, in the form of necrosis, were apparent in explants treated either with antisense or sense/nonsense ODNs.

View larger version (128K): [in this window] [in a new window]   Fig. 5.   A-C: light photomicrographs of untreated (A), sense/nonsense ODN-treated (B), and antisense ODN-treated (C) rat metanephric explants. Antisense ODN-treated explant shows dysmorphogenesis of the ureteric bud branches (u) and a reduced population of glomerular and tubular elements in the loose uninduced metanephric mesenchyme. In addition, there is a loss of acuteness of the tips of ureteric bud branches (arrowheads in C), the site where epithelial:mesenchymal interactions take place. D: autoradiogram of S1 nuclease digests of antisense ODN:RNA hybrids. Lanes 1, 3, 5, and 7 represent 5 µg of metanephric kidney mRNA hybridized with antisense ODN at 35°C, 40°C, 45°C, and 50°C, respectively, followed by S1 nuclease digestion. A single band of radioactivity, corresponding to the size of antisense ODN (~35 mer), is observed in all the 4 lanes. No band of radioactivity is seen in the control mRNA samples (lanes 2, 4, 6, and 8) that were incubated with nonsense ODN, indicating the susceptibility of the unprotected radiolabeled ODN to S1 nuclease digestion.

To assess the transcriptional and translational fibrillin-specific blocking activities of phosphorothioated antisense ODN, competitive PCR, immunoprecipitation, and immunofluorescence studies were performed. The competitive RT-PCR method was chosen instead of Northern blot analysis since only a minute amount total RNA can be extracted from the E15 rat metanephric explant, i.e., ~0.6 µg per explant. In both the nonsense (control) and antisense ODN-treated groups, a linearity in the ratio of wild to mutant fibrillin-1-DNA could be maintained when plotted against 10² to 10⁷ serial logarithmic dilutions of the competitive template DNA (Fig. 6A). Within this range of dilution, the bands of wild-type and mutant DNA were discernible for densitometric analyses to calculate a ratio. A ratio of 1 was obtained at a dilution of 10⁴ of the competitive DNA in explants of the nonsense ODN-treated group (control). For the cDNA from explants treated with antisense ODN, a ratio of 1 was obtained at a dilution of 10⁶ with competitive DNA. However, for the -actin, no significant differences in the linearity relationship, within the range of 10¹ to 10⁶ dilutions of competitive DNA, or in the ratio of wild-type to mutant DNA were observed between the two groups (control and antisense) (Fig. 6B). These data suggest that the steady-state mRNA levels of fibrillin-1 were selectively reduced with the exposure of metanephric explants to antisense ODN, whereas that of the -actin were unaffected.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Competitive RT-PCR of fibrillin-1 (A) and -actin (B) cDNAs, prepared from sense/nonsense (control) and fibrillin-1 antisense ODN-treated kidney explants. Serial logarithmic dilutions of mutant competitive DNA template of fibrillin-1 and -actin (see MATERIALS AND METHODS) were coamplified with a fixed amount (1 µl) of first-strand cDNA prepared from kidney explants. Amplified DNAs were subjected to 2% agarose gel electrophoresis, and densitometric tracings were prepared. Then, ratios between the densitometric readings of wild-type and mutant PCR products were plotted on logarithmic scales on the ordinate (y-axis) against the logarithmic dilutions of competitive DNA on the abscissa (x-axis). A ratio of 1 is obtained at a dilution of 106 of the competitive DNA in the antisense ODN explant compared with the dilution of 104 in the control explants. This indicates a reduction in the amplification of wild-type fibrillin-1 DNA in the antisense-treated kidney explants compared with the control (A). No differences in the amplification of -actin between the two groups are observed (B).

For translational blockade studies, the antisense and sense/nonsense-treated metanephric explants were radiolabeled with [³⁵S]methionine, and extracts were immunoprecipitated with anti-fibrillin-1 antibody and subjected to 5% SDS-PAGE. Under reducing conditions, extracts from the control untreated explants yielded a single band of radioactivity with ~300-kDa size (Fig. 7D, lane 1). No decrease in the intensity of the band was observed in the extracts of the explants treated with sense/nonsense ODN (Fig. 7D, lane 2). However, a marked decrease in the intensity of the band was observed in metanephric explants treated with the antisense ODN, even when double the amount of radioactivity, compared with the controls, was loaded for gel electrophoresis (Fig. 7D, lane 3). The tissue immunofluorescence studies confirmed the fibrillin-1 translational blockade observed in the immunoprecipitation experiments. The control explants showed a normal population of the nascent nephrons and iterations of the ureteric bud branches (Fig. 7A). The immunoreactivities of the mesenchyme surrounding the developing nephrons could be well elucidated. No significant decrease in the population of nephrons and the degree of anti-fibrillin-1 immunoreactivity was observed in the explants treated with nonsense ODN (Fig. 7B). However, a remarkable decrease in the immunoreactivity was observed in the explants treated with antisense ODN, and also, the population of the developing nephrons was also reduced (Fig. 7C). In aggregate, both the immunofluorescence and the immunoprecipitation studies indicated that the antisense ODN interferes with the translation of the fibrillin-1 gene.

View larger version (119K): [in this window] [in a new window]   Fig. 7.   A-C: immunofluorescence photomicrographs of the control untreated (A) and sense/nonsense- (B) and antisense-treated (C) explants, which were subsequently stained with anti-fibrillin-1 antibody. Untreated (A) and nonsense ODN-treated (B) explants reveal normal immunoreactivity of the mesenchyme surrounding the nascent nephrons (n) and ureteric bud branches (u). Antisense ODN-treated explant reveals a generalized reduced immunoreactivity with anti-fibrillin-1 antibody. D: SDS-PAGE autoradiogram of the extracts of the untreated (lane 1), nonsense ODN-treated (lane 2), and antisense ODN-treated explants (lane 3), which were immunoprecipitated with anti-fibrillin-1 antibody. A marked decrease in the autoradiographic intensity of the band in lane 3 is observed (large arrow), indicating that the translational blockade in the de novo synthesis of fibrillin-1 occurred as a result of antisense ODN treatment to the rat metanephric explants. Small arrow indicates the point of application.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The results of this investigation indicate that the cDNA sequence of fibrillin-1 is highly conserved across human, mouse, and rat species. In the human sequence, a multitude of mutations throughout the entire open-reading frame of the fibrillin-1 cDNA is associated with Marfan syndrome (27). These mutations may either involve EGF- or cbEGF-like domains or latent TGF--binding protein motifs. However, most severe malformations are seen in mutations involving the central portion of the molecule, i.e., a stretch between amino acid residue ~1050 to ~1350. Whether such mutations can lead to defective fibril formation or perturb elastogenesis in mice and rats needs to be determined by transgenic or knockout experiments. Nevertheless, an internal duplication in the mouse fibrillin-1 gene has been found to be associated with "tight skin" (Tsk) phenotype (37). The heterozygous Tsk mice exhibit cutaneous and skeletal abnormalities, whereas homozygous Tsk embryos die in utero due to failure in early organogenesis. In rat, so far no genetic mutation studies have been done. However, one may assume that differences observed at the amino acid residues 416, 437, and 438 of fibrillin-1 between rat, mouse, and human are probably not associated with any disease process, because the rat cDNA exhibiting deletion or addition at these sites has no abnormal phenotype in the rat. In addition, it is interesting to note that, although substitution of tyrosine at amino acid residue 2113 is seen in Marfan syndrome, in the rat this substitution (tyrosine  phenylalanine) has no detectable phenotypic change.

One would expect the role of fibrillin-1 in various biological processes to be common to all the mammalian species, because the putative protein's structural characteristics, i.e., conserved RGD sequence and cbEGF-like domains (Fig. 1), are similar. Fibrillin may play a role in fibrillogenesis, elastogenesis, and vasculogenesis, processes in which it can confer biomechanical properties on a variety of connective tissues (24, 27, 29). In addition, because of the presence of cbEGF-like domains that mediate protein:protein interactions, it may be involved in aggregation of macromolecules or cells during early embryogenesis (8). Moreover, the RGD sequences that mediate anchorage dependence may further facilitate cell:matrix interactions prevalent during organogenesis in embryonic life (16, 32, 34). Since fibrillins are also developmentally regulated, this leads one to propose an emerging theme from the above studies that they may play a vital role in embryogenesis. The finding that there is an increasing expression of fibrillin-1 mRNA in rat fetal and neonatal renal tissues (Fig. 2) would support the notion that it, like other ECM proteins, plays a role in rat metanephric development by acting as a morphogen.

A morphogen is defined as a molecule that expresses its concentration gradient in strategic locations of a given tissue and alters the fate of target cells in a dose-dependent manner (16). Given this definition, molecules other than ECM proteins, e.g., protooncogenes, growth factors and their receptors, and ECM receptors, i.e., integrins, can be classified as morphogens (16, 25). Some of them are expressed in the mesenchyme, others in the epithelia, and still others at the epithelial:mesenchymal interface. All these diverse groups of molecules can conceivably participate in epithelial:mesenchymal interactions that are essential to the morphogenesis of various mammalian organs (16, 25). The epithelial:mesenchymal interactions can be viewed as paracrine or juxtacrine interactions, where the ligand may be expressed in epithelium or mesenchyme, and the opposite would be the anticipated expression of the receptor. For instance the _v3-integrin receptor is exclusively expressed in the ureteric bud epithelia and induced mesenchyme, whereas its ligands like fibronectin, vitronectin, type I and IV collagens, and laminin are present in the mesenchyme or at the epithelial:mesenchymal interface (16, 40). Thus the fact that fibrillin-1 mediates attachment of cells by the involvement of _v3-integrin receptors (34) would suggest that, like _v3, fibrillin-1 may be relevant in the organogenesis of the kidney. This notion is supported by the fact that gene expression of fibrillin-1 is developmentally regulated in strategic locations, i.e., the metanephric mesenchyme, as indicated by our in situ data (Fig. 3). Also, the fact that protein expression of fibrillin-1 is exclusively observed in the mesenchyme during the early phase of metanephrogenesis and is developmentally regulated, as well (Fig. 4), further strengthens this notion. Here, one may suggest that fibrillin-1 could serve as a ligand for the _v3-integrin receptor to mediate paracrine epithelial:mesenchymal interactions during metanephrogenesis. Indeed, such a suggestion has been made recently in studies in which it was evaluated whether fibrillin specifically interacts with the receptors or binding proteins on cells in tissues with a high expression of fibrillin (34).

In a manner similar to ligand:receptor (_v3:fibrillin-1) interaction in the early phases of mammalian metanephric development when epithelial:mesenchymal interactions are prevalent, fibrillin may also modulate the later events like vascularization of the kidney. The fact that a rising gene as well as protein expression of fibrillin, similar to that of _v3 (40), was observed in the vascular elements of the developing metanephros, i.e., arteries and glomeruli (Figs. 3 and 4), suggests that it is likely that it plays a role in the vascularization phase of the kidney. In such a role, one may envisage the involvement of "anchoring filaments" that have been identified in the subendothelial matrix of blood vessels, where they connect the endothelia to the surrounding elastic fibers (34). These microfibrillar filaments contain fibrillin, and they are anchored at the cell surface in the plasmalemmal domains occupied by intracellular face membrane-associated dense plaques, sometimes referred to as focal contact points (34). Since _v3 is known to associate into focal contacts to organize the cytoskeletal assembly, there is a good possibility that this integrin receptor via its interactions with fibrillin modulates the later phases of metanephric development that are related to vascularization of the fetal kidney. The vascularization of the mammalian kidney involves two interlinked processes, i.e., vasculogenesis (in situ blood vessel formation) and angiogenesis (sprouting of preexisting capillaries) (14). Both the processes are highly complex in nature and difficult to study in an organ culture system, and thus this study was restricted to elucidate the role of fibrillin-1 in the early phases of metanephric development only.

The role of fibrillin-1 in metanephric development was investigated by employing antisense ODN. The specificity of the ODN was confirmed by subjecting RNA:DNA hybrids to S₁ nuclease digestion (Fig. 5), as described previously (17, 40). Also, the specificity was maintained by using them at a relatively low concentration (0.5 µM) in in vitro conditions. Usually, the antisense or sense ODNs have nonspecific translational inhibitory effects when used above a concentration of 1 µM in the medium (1). Moreover, at a higher concentration (>2.5 µM) they may be cytotoxic (1). However, at a relatively low concentration (<0.25 µM) the effects in the target tissue may not be discernible, since they are readily susceptible to nuclease degradation. The latter difficulty can be overcome by employing phosphorothioated ODNs, which are quite resistant to nuclease degradation (17, 40). The inclusion of phosphorothioated antisense ODN in the medium induced notable changes in metanephroi. They included a reduced population of nascent nephrons, disorganization of the ureteric bud iterations, and loss of the acuteness of their tips (Fig. 5). Such a loss of the acuteness of the tips of the ureteric bud branches has been reported to perturb epithelial:mesenchymal interactions with ultimate arrest in the formation of nascent nephrons (16). The mesenchyme is expanded, but the mesenchymal cells, toward which the fibrillin-1 antisense ODN is directed, seem to be quite atrophic. These observations suggest that fibrillin-1 induced perturbation in the biology of mesenchymal cells that led to an interference in the epithelial:mesenchymal interaction and ultimately dysmorphogenesis of the metanephros. It is interesting to mention here that similar delayed morphogenesis of the rat metanephros has been reported in experiments in which _v3 antisense ODN was used (40), thus suggesting an interactive role of fibrillin-1 as a ligand and _v3 as a receptor in metanephric development. The notion that fibrillin-1 acts as a ligand for _v3 has been well-documented in cell culture studies (34). The specificity of the effect of fibrillin-antisense ODN was also supported by the gene expression studies, in which RT-PCR analyses were carried out (Fig. 6). A reduced mRNA expression of fibrillin-1 was observed, whereas no change in the expression of -actin was seen in the competitive RT-PCR analyses with the antisense ODN treatment. Here, the question of whether antisense ODN caused any translational blockade in the de novo synthesis of fibrillin, which is ultimately responsible for dysmorphogenesis of the metanephric kidney, needs to be addressed. The fact that, along with the reduced mRNA levels, there was a concomitant reduction in the immunoreactivity of anti-fibrillin-1 antibody suggests that a decreased de novo synthesis of fibrillin-1 has indeed occurred. Finally, the immunoprecipitation studies confirmed the translational blockade of fibrillin-1, in which a notable decrease in the intensity of a ~300-kDa band was observed in the SDS-PAGE autoradiogram. Thus it is reasonable to propose that the dysmorphogenesis of the metanephros is associated with gene-disruption of the ECM macromolecule, i.e., fibrillin-1.

In summary, this study reemphasizes the relevance of epithelial:mesenchymal/paracrine or juxtacrine interactions in rat metanephric development. Also, this investigation adds another ECM molecule to our realm of knowledge, i.e., fibrillin-1, that seems to play a role in the organogenesis of the kidney, perhaps by acting as ligand for _v3 integrin, the latter being also a known receptor for a well-established renal tubular morphogen, i.e., laminin (7).