Presence of Laminin α5 Chain and Lack of Laminin α1 Chain during Human Muscle Development and in Muscular Dystrophies
1997; Elsevier BV; Volume: 272; Issue: 45 Linguagem: Inglês
10.1074/jbc.272.45.28590
ISSN1083-351X
AutoresCarl‐Fredrik Tiger, Marie-France Champliaud, Fátima Pedrosa Domellöf, Lars‐Eric Thornell, Peter Ekblom, Donald Gullberg,
Tópico(s)Gear and Bearing Dynamics Analysis
ResumoThere is currently a great interest in identifying laminin isoforms expressed in developing and regenerating skeletal muscle. Laminin α1 has been reported to localize to human fetal muscle and to be induced in muscular dystrophies based on immunohistochemistry with the monoclonal antibody 4C7, suggested to recognize the human laminin α1 chain. Nevertheless, there seems to be no expression of laminin α1 protein or mRNA in developing or dystrophic mouse skeletal muscle fibers. To address the discrepancy between the results obtained in developing and dystrophic human and mouse muscle we expressed the E3 domain of human laminin α1 chain as a recombinant protein and made antibodies specific for human laminin α1 chain (anti-hLN-α1G4/G5). We also made antibodies to the human laminin α5 chain purified from placenta. In the present report we show that hLN-α1G4/G5 antibodies react with a 400-kDa laminin α1 chain and that 4C7 reacts with a 380-kDa laminin α5 chain. Immunohistochemistry with the hLN-α1G4/G5 antibody and 4C7 revealed that the two antibodies stained human kidney, developing and dystrophic muscle in distinct patterns. Our data indicate that the previously reported expression patterns in developing, adult, and dystrophic human muscle tissues with 4C7 should be re-interpreted as an expression of laminin α5 chain. Our data are also consistent with earlier work in mouse, indicating that laminin α1 is largely an epithelial laminin chain not present in developing or dystrophic muscle fibers. There is currently a great interest in identifying laminin isoforms expressed in developing and regenerating skeletal muscle. Laminin α1 has been reported to localize to human fetal muscle and to be induced in muscular dystrophies based on immunohistochemistry with the monoclonal antibody 4C7, suggested to recognize the human laminin α1 chain. Nevertheless, there seems to be no expression of laminin α1 protein or mRNA in developing or dystrophic mouse skeletal muscle fibers. To address the discrepancy between the results obtained in developing and dystrophic human and mouse muscle we expressed the E3 domain of human laminin α1 chain as a recombinant protein and made antibodies specific for human laminin α1 chain (anti-hLN-α1G4/G5). We also made antibodies to the human laminin α5 chain purified from placenta. In the present report we show that hLN-α1G4/G5 antibodies react with a 400-kDa laminin α1 chain and that 4C7 reacts with a 380-kDa laminin α5 chain. Immunohistochemistry with the hLN-α1G4/G5 antibody and 4C7 revealed that the two antibodies stained human kidney, developing and dystrophic muscle in distinct patterns. Our data indicate that the previously reported expression patterns in developing, adult, and dystrophic human muscle tissues with 4C7 should be re-interpreted as an expression of laminin α5 chain. Our data are also consistent with earlier work in mouse, indicating that laminin α1 is largely an epithelial laminin chain not present in developing or dystrophic muscle fibers. Cellular interactions with the extracellular matrix have been implied to be important for several stages of muscle development (1von der Mark K. Öcalan M. Differentiation. 1989; 40: 150-157Crossref PubMed Scopus (109) Google Scholar, 2Jaffredo T. Horwitz A.F. Buck C.A. Rong P.M. Dieteren-Lievre F. Development (Camb.). 1988; 103: 431-446PubMed Google Scholar, 3Menko S. Boettiger D. Cell. 1987; 51: 51-57Abstract Full Text PDF PubMed Scopus (310) Google Scholar, 4Gullberg D. Ekblom P. Int. J. Dev. Biol. 1995; 39: 845-854PubMed Google Scholar). An intact linkage to the surrounding basement membrane has been demonstrated to be of importance also for muscle stability in the adult stage (5Campbell K.P. Cell. 1995; 80: 675-679Abstract Full Text PDF PubMed Scopus (762) Google Scholar, 6Worton R. Science. 1995; 270: 755-756Crossref PubMed Scopus (180) Google Scholar). During regeneration events following muscle damage, the basement membrane acts as a scaffold for the generation of new muscle fibers (7Bischoff R. Development (Camb.). 1990; 109: 943-952PubMed Google Scholar, 8Hughes S.M. Blau H.M. Nature. 1990; 345: 350-353Crossref PubMed Scopus (170) Google Scholar). It is thus important to understand the molecular composition of basement membranes in muscle. Laminin-2, with the chain composition α2, β1, γ1, is present in the muscle lineage from early stages of development in the mouse (9Schuler F. Sorokin L. J. Cell Sci. 1995; 108: 3795-3805Crossref PubMed Google Scholar, 10Velling T. Collo G. Sorokin L. Durbeej M. Zhang H.Y. Gullberg D. Dev. Dyn. 1996; 207: 355-371Crossref PubMed Scopus (60) Google Scholar) and is apparently the major laminin isoform in adult muscle basement membranes (11Ehrig K. Leivo I. Argraves W.S. Ruoslahti E. Engvall E. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 3264-3268Crossref PubMed Scopus (322) Google Scholar). The finding that genetic defects affecting laminin α2 can cause muscular dystrophy has highlighted the importance of laminin-2 for the structural integrity of muscle (12Allamand V. Sunada Y. Salih M. Straub V. Ozo C. Al-Turaiki M. Akbar M. Kolo T. Colognato H. Zhang X. Sorokin L. Yurchenco P. Tryggvason K. Campbell K. Hum. Mol. Genet. 1997; 6: 747-752Crossref PubMed Scopus (111) Google Scholar, 13Helbling-Leclerc A. Zhang X. Topaloglu H. Cruaud C. Tesson F. Weissenbach J. Tome F.M. Schwartz K. Fardeau M. Tryggvason K. Guicheney P. Nat. Genet. 1995; 11: 216-218Crossref PubMed Scopus (572) Google Scholar). Molecular compensation in certain forms of muscular dystrophies by increased expression of laminin chains may decrease the severity of the diseases. Some evidence for this has been obtained in immunohistochemistry studies with the antibody 4C7, which is one monoclonal antibody from a panel of antibodies raised against human placental laminins (14Engvall E. Davis G.E. Dickerson K. Ruoslahti E. Varon S. Manthorpe M. J. Cell Biol. 1986; 103: 2457-2465Crossref PubMed Scopus (263) Google Scholar, 15Engvall E. Earwicker D. Haaparanta T. Ruoslahti E. Sanes J.R. Cell Regul. 1990; 1: 731-740Crossref PubMed Scopus (297) Google Scholar). These antibodies were generated prior to the current knowledge about the existence of multiple laminin isoforms. The 4C7 antibody does not react with the laminin α2 chain but has been considered to react with the human α1 chain (14Engvall E. Davis G.E. Dickerson K. Ruoslahti E. Varon S. Manthorpe M. J. Cell Biol. 1986; 103: 2457-2465Crossref PubMed Scopus (263) Google Scholar, 15Engvall E. Earwicker D. Haaparanta T. Ruoslahti E. Sanes J.R. Cell Regul. 1990; 1: 731-740Crossref PubMed Scopus (297) Google Scholar). The antibody, commercially available under different names, has been widely used to detect human α1 chain (previously called A chain) both in muscle and non-muscle tissues (14Engvall E. Davis G.E. Dickerson K. Ruoslahti E. Varon S. Manthorpe M. J. Cell Biol. 1986; 103: 2457-2465Crossref PubMed Scopus (263) Google Scholar,16Korhonen M. Virtanen I. J. Histochem. Cytochem. 1997; 45: 569-581Crossref PubMed Scopus (30) Google Scholar, 17Vachon P.H. Beaulieu J.F. Am. J. Physiol. 1995; 268: G857-G867PubMed Google Scholar, 18DeArcangelis A. Neuville P. Boukamel R. Lefebre O. Kedinger M. Simon-Assmann P. J. Cell Biol. 1996; 133: 417-430Crossref PubMed Scopus (133) Google Scholar, 19Virtanen I. Laitinen L. Korhonen M. J. Histochem. Cytochem. 1995; 43: 621-628Crossref PubMed Scopus (62) Google Scholar, 20Sanes J.R. Engvall E. Butkowski R. Hunter D.D. J. Cell Biol. 1990; 111: 1685-1699Crossref PubMed Scopus (504) Google Scholar). The 4C7 antigen has been detected in basement membranes of normal muscle, and in blood vessels in muscle tissue (21Sewry C.A. Chevallay M. Tome F.M. Histochem. J. 1995; 27: 497-504Crossref PubMed Scopus (42) Google Scholar), and increased expression of it in muscular dystrophies has been documented in numerous reports (22Sewry C.A. Philpot J. Mahony D. Wilson L.A. Muntoni F. Dubowitz V. Neuromuscul. Disord. 1995; 5: 307-316Abstract Full Text PDF PubMed Scopus (103) Google Scholar, 23Mundegar R.R. von Oertzen J. Zierz S. Muscle & Nerve. 1995; 18: 992-999Crossref PubMed Scopus (18) Google Scholar, 24Sunada Y. Edgar T.S. Lotz B.P. Rust R.S. Campbell K.P. Neurology. 1995; 45: 2084-2089Crossref PubMed Scopus (114) Google Scholar). It thus seemed reasonable to suggest molecular compensation by α1 chain in muscular dystrophies (22Sewry C.A. Philpot J. Mahony D. Wilson L.A. Muntoni F. Dubowitz V. Neuromuscul. Disord. 1995; 5: 307-316Abstract Full Text PDF PubMed Scopus (103) Google Scholar, 23Mundegar R.R. von Oertzen J. Zierz S. Muscle & Nerve. 1995; 18: 992-999Crossref PubMed Scopus (18) Google Scholar, 24Sunada Y. Edgar T.S. Lotz B.P. Rust R.S. Campbell K.P. Neurology. 1995; 45: 2084-2089Crossref PubMed Scopus (114) Google Scholar), particularly since many reports convincingly have shown that laminin-1 (α1, β1, γ1) can stimulate proliferation, motility, and development of muscle cells in vitro (1von der Mark K. Öcalan M. Differentiation. 1989; 40: 150-157Crossref PubMed Scopus (109) Google Scholar, 25Echtermeyer F. Schober S. Pöschl H. von der Mark H. von der Mark K. J. Biol. Chem. 1996; 271: 2071-2075Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 26Goodman S.L. Risse G. von der Mark K. J. Cell Biol. 1989; 109: 799-809Crossref PubMed Scopus (158) Google Scholar). Nevertheless, we and others have failed to detect laminin α1 chain in developing mouse muscle tissue (9Schuler F. Sorokin L. J. Cell Sci. 1995; 108: 3795-3805Crossref PubMed Google Scholar, 27Klein G. Ekblom M. Fecker L. Timpl R. Ekblom P. Development (Camb.). 1990; 110: 823-837PubMed Google Scholar, 28Tiger C.-F. Gullberg D. Muscle & Nerve. 1997; (in press)PubMed Google Scholar), and no increased expression of this chain was seen in dystrophic mouse muscle (28Tiger C.-F. Gullberg D. Muscle & Nerve. 1997; (in press)PubMed Google Scholar). Furthermore, comparing the staining pattern in non-muscle tissue of 4C7 in human, rat, and hamster with the pattern seen in mouse and rat with other antibodies, there is a discrepancy. In non-neural tissues of mouse and rat the α1 chain is largely confined to epithelial basement membranes (27Klein G. Ekblom M. Fecker L. Timpl R. Ekblom P. Development (Camb.). 1990; 110: 823-837PubMed Google Scholar, 29Ekblom M. Klein G. Mugrauer G. Fecker L. Deutzmann R. Timpl R. Ekblom P. Cell. 1990; 60: 337-346Abstract Full Text PDF PubMed Scopus (196) Google Scholar, 30Durbeej M. Fecker L. Hjalt T. Zhang H.Y. Salmivirta K. Klein G. Timpl R. Sorokin L. Ebendal T. Ekblom P. Ekblom M. Matrix Biol. 1996; 15: 397-413Crossref PubMed Scopus (91) Google Scholar), but the 4C7 antigen is widely distributed in developing and adult human, rat, and hamster tissues (15Engvall E. Earwicker D. Haaparanta T. Ruoslahti E. Sanes J.R. Cell Regul. 1990; 1: 731-740Crossref PubMed Scopus (297) Google Scholar, 19Virtanen I. Laitinen L. Korhonen M. J. Histochem. Cytochem. 1995; 43: 621-628Crossref PubMed Scopus (62) Google Scholar, 20Sanes J.R. Engvall E. Butkowski R. Hunter D.D. J. Cell Biol. 1990; 111: 1685-1699Crossref PubMed Scopus (504) Google Scholar). The staining pattern of 4C7 in human tissues is also in disagreement with the distribution of the α1 mRNA in human tissues (31Vuolteenaho R. Nissinen M. Saino K. Byers M. Eddy R. Hirvonen H. Shows T.B. Sariola H. Engvall E. Tryggvason K. J. Cell Biol. 1994; 124: 381-394Crossref PubMed Scopus (245) Google Scholar, 32Nissinen M. Vuolteenaho R. Boot-Handford R. Kallunki P. Tryggvason K. Biochem. J. 1991; 276: 369-379Crossref PubMed Scopus (100) Google Scholar). Antibody 4C7 might thus detect some other α laminin chain, but this proposal is speculative (33Ekblom P. Curr. Opin. Cell Biol. 1996; 8: 700-706Crossref PubMed Scopus (77) Google Scholar, 34Miner J.H. Patton B.L. Lentz S.I. Gilbert D.J. Snider W.D. Jenkins N.A. Copeland N.G. Sanes J.R. J. Cell Biol. 1997; 137: 685-701Crossref PubMed Scopus (584) Google Scholar) and has not been rigorously tested. Currently, five different laminin α chains have been described (35Timpl R. Curr. Opin. Cell Biol. 1996; 8: 618-624Crossref PubMed Scopus (554) Google Scholar) and 4C7 could potentially detect any of these or might detect several α chains. In a cell line, 4C7 immunoprecipitated a large chain in the 400-kDa range together with 200-kDa chains, but the nature of the 400-kDa chain was not studied (15Engvall E. Earwicker D. Haaparanta T. Ruoslahti E. Sanes J.R. Cell Regul. 1990; 1: 731-740Crossref PubMed Scopus (297) Google Scholar). To clarify the discrepancies in laminin α1 distribution in mouse and human tissues, we made antibodies to the recombinant E3 domain of human laminin α1 and compared the specificity of these antibodies with that of 4C7. Immunoprecipitation of laminins from cell lines producing varying amounts of either α1 or α5 mRNA allowed a precise distinction of antibody specificity. Furthermore, we compared the distribution of laminin α1 chain and the 4C7 antigen in developing human muscle and in dystrophic human muscle tissue. Since the distribution of the laminin α1 chain in mouse kidney has been well described, we also analyzed the expression patterns in human kidney. A 1180-base pair-long fragment from the 3′-end of human laminin α1 chain (nucleotide residues 8140–9320) corresponding to the E3 region (carboxyl-terminal globular domain G4-G5) was amplified by PCR 1The abbreviations used are: PCR, polymerase chain reaction; kb, kilobase(s); PAGE, polyacrylamide gel electrophoresis; DMD, Duchenne muscular dystrophy. from a 4.5-kb laminin α1 cDNA sequence ((36), clone number 7 supplied by E. Engvall The Burnham Institute, La Jolla Cancer Research Center) using AmpliTaq® (Perkin-Elmer). The primers were modified to include restriction sites for NotI and NheI to facilitate cloning into the expression vector. Primer sequences were as follows: forward primer, 5′ GCC CCG CTA GCT CCC GAT GCA GAG GAC AGC A 3′; reverse primer, TCA GTT GCG GCC GCT CAG GAC TCG GTC CCA GG. The obtained PCR product was ligated into a TA-vector (PCR II™, Invitrogen) for sequence confirmation. Sequencing was performed with a Pharmacia T7 Sequencing™ kit (Pharmacia Biotech Inc.). The sequenced PCR product was cleaved with NotI/NheI and inserted into the episomal pCEP-Pu vector (which is a modified pCEP4 (Invitrogen) vector, provided by E. Pöschl Institute of Experimental Medicine, Friedrich-Alexander-University, Erlangen, Germany). The insertions sites were sequenced prior to transfection. 106 human embryonic kidney cells 293 EBNA (Invitrogen, Catalog number R-620-07) were stably transfected with 15 μg of hLN-α1G4/G5 in pCEP-Pu using lipofectAMINE™ reagent (Life Technologies, Inc.), according the instructions from the manufacturer. Transfected cells were selected in 2 μg/ml puromycin and 0.25 mg/ml G418 (Life Technologies, Inc.) and the medium from cells grown under serum-free conditions was analyzed for recombinant protein by SDS-PAGE. Purified protein was separated on a 10% SDS-PAGE under reducing conditions, visualized with Coomassie Brilliant Blue, excised, and digested “in-gel” with trypsin according to Ref. 37Rosenfeld J. Capdevielle J. Guillemot J.C. Ferrara P. Anal. Biochem. 1992; 203: 173-179Crossref PubMed Scopus (1132) Google Scholar. Liberated peptides were further analyzed as described in Ref. 38McCourt P.A.G. Ek B. Forsberg N. Gustafson S. J. Biol. Chem. 1994; 269: 30081-30084Abstract Full Text PDF PubMed Google Scholar. One peptide (SPQVQSFDFS) was analyzed and found to be identical with amino acids 3048–3057 in Ref. 36Haaparanta T. Uitto J. Ruoslahti E. Engvall E. Matrix. 1991; 11: 151-160Crossref PubMed Scopus (60) Google Scholar. For the generation of antibodies to human laminin α1, medium was collected from confluent 293 EBNA hLN-α1G4/G5 cells under serum-free conditions and supplemented with 1 mm benzamidine, 1 mm EDTA, 1 mm N-ethylmaleimide. Collected medium was diluted 1:2 in water, passed over a 10-ml DEAE-Sepharose® Fast Flow (Pharmacia Biotech Inc.) column, serially connected to a 5-ml Hi trap Heparin-Sepharose® column (Pharmacia Biotech Inc.). The DEAE column was disconnected, and the heparin column was washed in 0.1 m NaCl, 20 mmTris-HCl, pH 8.0, prior to eluting in 0.3 m NaCl in 20 mm Tris-HCl, pH 8.0. Peak fractions containing recombinant protein were concentrated on a second 1-ml Hi trap Heparin-Sepharose column, and the resulting peak fraction was used for immunizations of two rabbits, using 50-μg injections intramuscularly at intervals of 3 weeks. For immunohistochemistry, the antibodies were affinity-purified on the recombinant protein as described in Ref. 39Rose O. Rohwedel J. Reinhardt S. Bachmann M. Cramer M. Rotter M. Wobus A. Starzinski-Powitz A. Dev. Dyn. 1994; 201: 245-249Crossref PubMed Scopus (77) Google Scholar prior to staining. The polyclonal antibody to human laminin α5 was generated as follows: an extract from human placenta was purified by affinity chromatography on a laminin β1 chain antibody (Ab 545) as described in Ref. 40Champliaud M. Lunstrum G. Rousselle P. Nishiyama T. Keene D. Burgeson R. J. Cell Biol. 1996; 132: 1189-1198Crossref PubMed Scopus (228) Google Scholar. A major 380-kDa purified protein band on SDS-PAGE was cleaved with trypsin and microsequencing of resulting peptides revealed laminin α5 sequences (see “Results”). The original 380-kDa SDS-PAGE band was used to generate the rabbit polyclonal antibodies to laminin α5 chain. The monoclonal antibody recognizing laminin β1/γ1 chains (clone 4C12.8) was obtained from Immunotech. The polyclonal antibody to intact mouse laminin-1 was from Sigma (L9393). The 4C7 antibody (sold under the name mAb 1924) was from Chemicon. To visualize the proximal tubules a polyclonal antibody specific for a brush border antigen of proximal tubules was used (41Ekblom P. Miettinen A. Saxén L. Dev. Biol. 1980; 74: 263-274Crossref PubMed Scopus (42) Google Scholar). JAR cells (human choriocarcinoma cells ATCC No HTB-144), RD (rhabdomyosarcoma ATCC No CCL-136), WWCS-1 (Wilm's tumor cell line (42Talts J.F. Aufderheide E. Sorokin L. Ocklind G. Mattson R. Ekblom P. Int. J. Cancer. 1993; 54: 868-874Crossref PubMed Scopus (23) Google Scholar)), and G6 (cloned primary human fetal myoblasts (43Jin P. Farmer K. Ringertz N.R. Sejersen T. Differentiation. 1993; 54: 47-54Crossref PubMed Google Scholar)) were grown in Dulbecco's modified Eagle's medium under standard conditions. Cells were labeled overnight in in the presence of 25 μCi/ml [35S]methionine/cysteine (pro-Mix 35S cell labeling mix (Amersham Corp.)). Medium was collected from cells, centrifuged, and supplemented with protease inhibitors (1 mm benzamidine, 1 mm EDTA, 1 mm N-ethylmaleimide). The centrifuged medium was processed for immunoprecipitation as described (44Eng H. Herrenknecht K. Semb H. Starzinski-Powitz A. Ringertz N. Gullberg D. Differentiation. 1997; 61: 169-176Crossref PubMed Scopus (13) Google Scholar). For Western blotting, conditioned medium was collected from JAR cells. Medium was passed over a Ricinus communis agglutinin I-agarose column (Vector Laboratories), washed extensively in phosphate-buffered saline, and eluted with 0.5 md(+)-galactose (Sigma). Eluted proteins were directly used for immunoprecipitation. Immunoprecipitated proteins were solubilized in SDS-PAGE sample buffer and resolved on a 5% SDS-PAGE gel under reducing conditions. Separated proteins were transferred to nitrocellulose membranes in a Trans-Blot cell (Bio-Rad). Membranes were incubated with primary antibody, washed in TBS + 0.05% Tween 20, followed by peroxidase-coupled sheep anti-rabbit IgG (Amersham) and developed using the ECL system (Amersham). Serial sections, 5–8-μm thick, of muscle biopsies of boys referred for diagnostic purpose and shown to lack dystrophin, of muscle samples of human fetuses with a gestation age of 22 weeks, and of biopsies of human kidneys were cut in a Reichert Jung cryostat at −25 °C. The sections were collected on individual slides as well as on the same slide (one sample of human muscular dystrophy, one of human fetal muscle, and one of human kidney) to allow a direct comparison of staining intensity in the three different types of samples. The staining procedure was carried out as described in Ref. 28Tiger C.-F. Gullberg D. Muscle & Nerve. 1997; (in press)PubMed Google Scholar. Total RNA was isolated using Qiagen RNeasy midi kit according to the manufacturer's instructions. Northern blotting was performed as described (28Tiger C.-F. Gullberg D. Muscle & Nerve. 1997; (in press)PubMed Google Scholar). For laminin α1, the 1180-base pair-long fragment used for recombinant laminin expression was used as a probe. For laminin α5 a reverse transcription PCR amplified mouse cDNA (nucleotides 290–891) was obtained from newborn mouse kidney total RNA and used as a probe as described (30Durbeej M. Fecker L. Hjalt T. Zhang H.Y. Salmivirta K. Klein G. Timpl R. Sorokin L. Ebendal T. Ekblom P. Ekblom M. Matrix Biol. 1996; 15: 397-413Crossref PubMed Scopus (91) Google Scholar). In addition a 1.3-kb human EST clone (accession W67855) was obtained from the Integrated Molecular Analysis of Genomes and their Expression (I.M.A.G.E.) consortium (I.M.A.G.E. consortium Clone ID:342926, United Kingdom Human Genome Mapping Project Resource Center, Hinxton, Cambridge, United Kingdom) and was sequenced from the vector T7 site and in a span of 360 nucleotides found to show 73% identity to mouse laminin α5 chain (nucleotides 9874–10233) (45Miner J.H. Lewis R.M. Sanes J.R. J. Biol. Chem. 1995; 270: 28523-28526Abstract Full Text Full Text PDF PubMed Scopus (224) Google Scholar) and to be identical with a partial cDNA sequence for human laminin α5 in nucleotides 1903–2262 (46Durkin M.E. Loechel F. Mattei M.-G. Gilpin B.J. Albrechtsen R. Wewer U. FEBS Lett. 1997; 411: 296-300Crossref PubMed Scopus (47) Google Scholar). The opposite end of the EST clone was found to be identical to the untranslated end of the partial human laminin α5 cDNA sequence (nucleotides 2891–3125) (46Durkin M.E. Loechel F. Mattei M.-G. Gilpin B.J. Albrechtsen R. Wewer U. FEBS Lett. 1997; 411: 296-300Crossref PubMed Scopus (47) Google Scholar). The 1.3-kb fragment was excised with NotI/EcoRI and used as a probe. To obtain reagents specific for human laminin α1 for immunohistochemistry on human tissues, we stably expressed cDNA coding for the E3 region (carboxyl-terminal globular domains G4-G5) of laminin α1 episomally in 293 EBNA cells. Recombinant protein (hLN-α1G4/G5) displayed an estimated molecular mass of 45 kDa under nonreducing conditions, shifted mobility to 55 kDa upon reduction, and bound heparin-Sepharose (Fig. 1). Following purification on DEAE- and heparin-Sepharose the hLN-α1G4/G5 (Fig. 1) was used as an immunogen to generate polyclonal antibodies. These antibodies are further characterized below. We also generated an antibody to human laminin α5 chain. The polyclonal rabbit antibody, anti-hLNα5, was raised to a 380-kDa SDS-PAGE band purified from human placenta by immunoaffinity chromatography on anti-laminin β1 IgG (data not shown). Amino acid sequencing of tryptic peptides obtained from the 380-kDa band revealed two peptides, which were identified in the deduced amino acid sequence from a recently identified partial human laminin α5 cDNA sequence (46Durkin M.E. Loechel F. Mattei M.-G. Gilpin B.J. Albrechtsen R. Wewer U. FEBS Lett. 1997; 411: 296-300Crossref PubMed Scopus (47) Google Scholar) (TableI).Table IAmino acid sequences of tryptic peptides obtained from the 380-kDa band were identified as human laminin α5, based on the deduced amino acid sequence from a partial human laminin α5 cDNA (46Durkin M.E. Loechel F. Mattei M.-G. Gilpin B.J. Albrechtsen R. Wewer U. FEBS Lett. 1997; 411: 296-300Crossref PubMed Scopus (47) Google Scholar)Human laminin-α5 partialKFYLQGPEPEPGQGTED cDNA residue #21‖‖‖‖‖‖‖‖ ‖‖‖ ‖‖‖Tryptic peptide FYLQGPEPDPGQ-TEDHuman laminin-α5 partialRQATGDYMGVSLR cDNA residue #46‖‖‖‖‖‖‖‖‖‖‖‖Tryptic peptideQATGDYMGVSLR Open table in a new tab To characterize the antibodies we had generated we tested a number of cell lines in Northern blotting for their expression of laminin α1 mRNA. As a comparison we also probed for laminin α5 mRNA. Laminin α1 mRNA was only detected in JAR cells, whereas laminin α5 mRNA was detected at moderate levels in JAR cells and RD cells and and at lower levels in WCCS-1 cells (Fig.2 A). The size of laminin α1 mRNA was estimated to approximately 10 kb, and the laminin α5 mRNA, as detected by both laminin α5 probes, was distinctly larger. The size was in agreement with the previously reported size of 11–12 kb in mouse (45Miner J.H. Lewis R.M. Sanes J.R. J. Biol. Chem. 1995; 270: 28523-28526Abstract Full Text Full Text PDF PubMed Scopus (224) Google Scholar). When laminins were immunoprecipitated with an antibody to laminin β/γ chains, two α chain bands with molecular masses of 400 and 380 kDa were observed on SDS-PAGE from JAR cells (Fig. 2 B). In contrast, the hLN-α1G4/G5 antibodies only precipitated a complex containing 400- and 200–220-kDa bands. The monoclonal antibody 4C7 has been suggested to recognize laminin α1 chain. However, when medium from metabolically labeled JAR cells was immunoprecipitated with anti-hLN-α1G4/G5 and 4C7 in parallel, different laminin α chain bands were obtained. Whereas anti-hLN-α1G4/G5 precipitated the 400-kDa band, 4C7 precipitated a 380-kDa band (Fig. 2 B). In mouse the laminin α5 chain has been reported to have an molecular mass of 380 kDa (47Sorokin L. Frieser M. Pausch F. Kröger S. Ohage E. Deutzmann R. Dev. Biol. 1997; (in press)PubMed Google Scholar). From JAR cells anti-hLNα5 precipitated a 380-kDa band (Fig. 2 B). From RD cells lacking reactivity with anti-hLN-α1G4/G5, antibodies to laminin β/γ chains and anti-hLNα5 still precipitated a laminin with a molecular mass of the α chain of 380 kDa. A 380 kDa α chain band was also precipitated from RD and G6 cells with 4C7 (data not shown). WWCS-1 cells, shown in Northern to lack laminin α1 mRNA and to express low levels of laminin α5 mRNA, only precipitated visible β/γ chain complexes in immunoprecipitation under the conditions used. Western blotting of JAR cell medium and proteins immunoprecipitated from this medium with antibodies to laminin β/γ chains, and subsequently blotted with antibodies to hLN-α1G4/G5 (Fig. 2 C, lane a), revealed strong reactivity with the 400-kDa band. Anti-hLNα5 reacted weakly with a 380-kDa band in the material precipitated by laminin β/γ chains (lane b). As shown inC the material immunoprecipitated from JAR cells with the 4C7 antibody did not react with hLN-α1G4/G5 antibodies (lane c), whereas Western blotting with anti-hLNα5 resulted in reactivity with the 380-kDa band (lane d). We also performed silver staining to independently illustrate the size difference between the laminin α chains precipitated by the two antibodies. Silver staining of proteins immunoprecipitated with anti-hLN-α1G4/G5 revealed a distinct 400-kDa band together with 200–220-kDa bands (lane e), whereas silver staining of 4C7 reactive material revealed the 380-kDa band in addition to 200–220-kDa bands (lane g). The 200–220-kDa bands were recognized in Western blotting by a polyclonal antibody to mouse laminin α1β1γ1 chains (data not shown). When human adult kidney was stained with affinity-purified antibodies to human laminin α1 (anti-hLN-α1G4/G5) and human laminin α5 (4C7), contrasting staining patterns were observed. Anti-hLN-α1G4/G5 selectively stained a subset of proximal tubuli, whereas 4C7 stained proximal and distal tubuli, glomerular basement membranes, and blood vessels (Fig.3, A–C). In agreement with previously reported data, 4C7 stained muscle fibers in addition to blood vessels in human fetal muscle (Fig.4 A). In biopsy material from a dystrophic DMD boy, 4C7 stained basement membranes of muscle fibers, blood vessels, and somewhat more intensely groups of small diameter regenerating muscle fibers (Fig. 4 B). In contrast, the hLN-α1G4/G5 antibody did not specifically stain either developing or dystrophic muscle tissue (Fig. 4, C and D).Figure 4Detection of laminin α1 and laminin α5 in human skeletal muscle. Immunofluorescence with 4C7 on cross-sections of human fetal muscle revealed the presence of laminin α5 in muscle fiber basement membranes and blood vessels (A), whereas a parallel section lacked detectable levels of human laminin α1 (C). In DMD patient sections 4C7 stained larger blood vessels, capillaries, and around small myofibers (B), whereas laminin α1 was not detected in a parallel section (D). Bar: 100 μm.View Large Image Figure ViewerDownload Hi-res image Download (PPT) In some forms of muscular dystrophy the primary defect is a disturbed linkage between the muscle fiber and the basement membrane (5Campbell K.P. Cell. 1995; 80: 675-679Abstract Full Text PDF PubMed Scopus (762) Google Scholar, 6Worton R. Science. 1995; 270: 755-756Crossref PubMed Scopus (180) Google Scholar). This leads to muscle degeneration but also to a regeneration event where satellite cells are activated, replicate, and fuse to form new myofibers. During this process the basement membrane is used for migration and as a scaffold for the formation of new fibers (7Bischoff R. Development (Camb.). 1990; 109: 943-952PubMed Google Scholar, 8Hughes S.M. Blau H.M. Nature. 1990; 345: 350-353Crossref PubMed Scopus (170) Google Scholar). Little is known about the basement membrane components made during regeneration. Laminins (35Timpl R. Curr. Opin. Cell Biol. 1996; 8: 618-624Crossref PubMed Scopus (554) Google Scholar) and collagen IV (48Ninomiya Y.M. Kagawa M. Iyama K. Naito I. Kishoro Y
Referência(s)