Connective Tissue Research, 1991, Val. 25, pp. 265-219 Reprints available directly from the publisher Photocopying permitted by license only 0 1991 Gordon and Breach Science Publishers S.A. Printed in the United States of Amenca

POLYCLONAL ANTIBODIES TO TROPOELASTIN AND THE SPECIFIC DETECTION AND

MEASUREMENT OF TROPOELASTIN IN VZTRO

IAN W. PROSSER,? LOREN A. WHITEHOUSE,t WILLIAM C. PARKS,$ MONA STAHLE-BACKDAHL,t ALEKSANDER HINEK,? PYONG WOO PARK,§

and ROBERT I? MECHAM,?§

tRespirutory and Critical Care and *Dermatology Divisions, Department of Medicine, Jewish Hospital at Washington University Medical Center, St. Louis, Missouri, USA

§Department of Cell Biology and Physiology, Washington University School of Medicine, S t . Louis, Missouri, USA

(Received May I S , 1990; in revised form August 13, 1990; accepted August 27, 1990)

Because tropoelastin is difficult to purify, most antibodies to elastin are raised against the insoluble form of the molecule. While these antibodies cross-react with tropoelastin, antigenic differences between insoluble and soluble elastin suggest that antibodies raised directly against tropoelastin might provide a more sensitive and specific reagent for evaluating tropoelastin production in elastin-producing systems. Using an improved method for purifying tropoelastin from tissue culture explants, we describe the generation and characterization of an antibody to bovine tropoelastin. This antibody was used to develop a sensitive, direct-binding immunoassay capable of quantifying small levels of tropoelastin in conditioned medium from cultured cells. This assay takes advantage of the propensity of tropoelastin to adsorb to vinyl microtiter plates, even in the presence of serum proteins. This property, in combination with the increased sensitivity obtained using antibodies to tropoelastin, provides for a direct-binding immunoassay that detects nanogram quantities of tropoelastin directly in cell culture medium, avoiding sample preparation steps that result in extensive loss of tropoelastin. In addition, this direct-binding assay is ten- to 30-fold more sensitive than the existing competitive ELISA assays.

KEYWORDS: elastin, tropoelastin, immunoassay, ELISA, polyclonal antibodies

Elastic fibers are a cons'ituent of many mammalian connective tissues, providing structural integrity and reversible extensibility or deformability to tissues such as major arterial vessels, lung and skin. Their physiological importance lies in the elastomeric properties of elastin, the protein which forms the functional component of the fiber. Elastin is synthesized and secreted as a soluble, single-chain protein, tropoelastin, which is rapidly cross-linked in the extracellular matrix into insoluble elastin by the modification and covalent crosslinking of lysine side chains.]-3

The development of immunoassays that measure tropoelastin production has contributed greatly to understanding elastin biosynthesis. Because tropoelastin is difficult to purify in large quantities, these immunoassays have used antisera prepared against solubilized forms of mature, cross-linked elastin. Radioimmunoassay (RIA) and enzyme-linked immuno- assay (ELISA) techniques have been widely used,4-6 and the results are usually expressed ~ ~

Addressfor correspondence: Dr. Robert Mecham, Jewish Hospital, Pulmonary Research, 216 South Kingshigh- way, St. Louis, MO 63110, USA

265

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

266 I . W. PROSSER et al.

relative to a solubilized-elastin standard. Because antisera to solubilized, mature elastin may have a different affinity for tropoelastin, the calculated values for tropoelastin in samples may be inaccurate. More precise values, however, may be obtained if tropoelastin is used as a competitive standard.

In addition to problems with precise quantitation of tropoelastin, these assays present a number of technical difficulties. RIA, although sensitive, requires the preparation of [125I]- labeled, solubilized-elastin, and competitive binding assays are subject to interference from non-specific inhibition of antibody binding by serum components in culture medium.6.7 Microplate ELISA avoids the use of hazardous radioactive reagents with limited shelf-life, but published protocols4~5~* utilize a competitive technique which requires extensive sample preparation and is also affected by serum components.

To eliminate these difficulties, we have prepared a highly specific polyclonal antiserum against tropoelastin purified from organ cultures of fetal bovine ligamentum nuchae. This antiserum was used to develop a direct-binding ELISA which allows the rapid and direct measurement of nanogram quantities of tropoelastin in cell culture medium containing calf serum. Assay results were validated by comparison with measurements from previous assays, with estimates of tropoelastin biosynthesis by immunoprecipitation and by measure- ment of steady-state cellular tropoelastin mRNA levels.

EXPERIMENTAL PROCEDURES

Materials

Vinyl 96-well microtiter plates (#2595) were from Costar, Cambridge, MA. Peroxidase- labeled goat anti-rabbit and goat anti-mouse immunoglobulin, and ABTS enzyme substrate solution were obtained from Kirkegaard and Perry Laboratories Inc., Gaithersburg, MD. [3,4,5-3H]-Leucine, [~L-~~SI-UTP, ENHANCE were from New England Nuclear-Dupont, Boston, MA and IgGSorb was from The Enzyme Center Inc., Boston, MA. HRP Color Development Solution was from Bio Rad, Richmond, CA. pGEM-4Z was purchased from Promega, Madison, WI and plastic tissue culture chamber slides were from Lab Tek Division, Miles Laboratories Inc., Naperville, IL. Insoluble elastin and human leukocyte elastase were from Elastin Products Co., St. Louis, MO and all other reagents were from Sigma, St. Louis, MO.

Purification of Bovine Ligamentum Nuchae Tropoelastin

Tropoelastin was prepared from fetal calf ligamentum nuchae by a modification of a previously described method.9 The ligamentum nuchae was removed at a local slaughter house from a late gestation (200-270 day) bovine fetus and placed immediately in Dul- becco’s modified Eagle’s medium (DMEM)-10% calf serum with 2X antibiotics. In the laboratory, adherent tissue was removed and the ligament finely minced with a McIlwain tissue chopper. The minced tissue was incubated overnight at 37°C in DMEM supplemented with 100 pg/ml p-aminopropionitrile fumerate (BAPN), 50 pg/ml penicillamine, 80 pg/ml gentamycin, 1 ng/ml dexamethasone, 30 mM N-2-hydroxy-ethylpiprazine-N-2-ethane- sulfonic acid (HEPES) buffer, pH 7.4, and 3% calf serum. After the incubation, the tissue

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN 261

was extensively washed with ice cold water to remove serum components and homogenized, using a Brinkman Polytron, in 0.5 N acetic acid containing 2.5 p.g/ml pepstatin A. Following overnight stirring at 4°C in 250 ml polyethylene centrifuge bottles, insoluble material was removed by centrifugation at 16,OOOxg for 30 min at 4°C and discarded. Tropoelastin was extracted from the supernatant by the sequential addition on ice of 1.5 volumes of propanol and 2.5 volumes of butanol. 10 The precipitate that formed during this step was removed by centrifugation and the supernatant was flash evaporated to dryness on a rotary evaporator at 40°C. The residue was suspended in a minimal volume of ice-cold, dilute acetic acid (0.25% vlv), and tropoelastin was precipitated with 3 volumes of ice-cold acetone. The precipitated tropoelastin was collected by centrifugation (50,000 x g for 30 min), dissolved in water and lyophilized. Purity of the isolated tropoelastin was assessed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and amino acid analysis.

Solubilization of Insoluble Elastin

Bovine welastin was prepared by oxalic acid treatment of ligamentum nuchae insoluble elastin prepared with hot sodium hydroxide.5.11 Elastin peptides were obtained using a modification of the procedure of Senior et al. 12 Human leukocyte elastase was added to a suspension of bovine ligamentum nuchae insoluble elastin ( 1 part elastase to 100 parts elastin, w/w) in phosphate buffered saline, pH 8.0, and allowed to incubate at 37°C overnight. The reaction mixture was passed over a Trasylol-Sepharose column to remove active enzyme and then absorbed with antibody to human neutrophil elastase coupled to Affi-Gel 10.

Cell Culture

Fetal calf ligamentum nuchae fibroblast (FCL) and aortic smooth muscle cell (SMC) cultures were established from tissue explants as previously described. 13 Fetal bovine chondroblasts (FBC) were enzymatically dissociated from ear cartilage. 14315 Cultures were maintained in DMEM containing high glucose and high bicarbonate supplemented with antibiotics, nonessential amino acids, and 5% or 10% (vlv) calf serum. First or second passage cells were used in all experiments. A2058 melanoma cells16 were maintained as above.

Production of Antiserum to Tropoelastin

New Zealand White rabbits were immunized by multiple subcutaneous injections in the back with a total of 5 Fg of purified bovine tropoelastin in 500 cl.1 phosphate-buffered saline (PBS), pH 7.4, emulsified in an equal volume of Freund’s complete adjuvant. Booster injections (5 p.8) were given in incomplete Freund’s adjuvant at one month intervals and animals were bled by puncture of the central ear artery 8-10 days after boosting.

To determine the titer of the tropoelastin antiserum, vinyl 96-well microtiter plates were acid washed by soaking for 30 min in 4 N sulfuric acid, rinsed with water and air dried before use. The plates were then coated overnight at 4°C with tropoelastin (20 ng in 100 p1 per well) dissolved in borate buffered saline (BBS). Coated plates were stored dry in a humidified chamber at 4°C for up to one month without noticeable change in antibody-binding capacity.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

268 I. W. PROSSER et al.

Before use, tropoelastin-coated plates were washed three times with 0.1% Tween-20 in PBS (PBS-Tween), then incubated for 2 h at 37°C with PBS-Tween containing 1 mg/ml ovalbumin (100 pl/well) to block non-specific binding. After washing the plate three times with PBS- Tween, serial dilutions of tropoelastin antisera (100 $/well) in the same buffer were added in triplicate to the plate and incubated for 1.5 h at 37°C. Pre-immune rabbit serum and saline controls were included on each plate. Plates were washed as before and incubated for 1.5 h at 37°C with peroxidase-labeled, goat anti-rabbit immunoglobulin (GAR-HRP) diluted 1: lo00 in PBS-Tween (100 pl/well), washed again, and 100 pl of freshly prepared ABTS enzyme substrate solution added to each well. Color was allowed to develop for 20-30 min and the reaction was stopped by the addition of 50 p1 of 10 mM sodium a i d e to each well. Reaction product was quantified spectrophotometrically at a wavelength of 405 mm. Results were read as positive if absorbance was at least twice that of the highest dilution of preimmune serum wells.

Immunoblot Analysis of Tropoelastin

Tropoelastin was transferred from SDS-PAGE gels to nitrocellulose at 0.2 A for 90 min at 4°C. All subsequent steps were carried out at room temperature. Protein blots were washed for 2 h with 0.1% Tween-20 in Tris-buffered saline (TBS-Tween: 10 mM Tris, 150 mh4 NaCl, pH 7.4) containing 5% (w/v) non-fat dry milk (BLOTTO) to block unbound sites. The nitrocellulose was then washed twice with TBS-Tween, 0.5% BLOTTO (15 midwash) on an orbital shaker and incubated for 1 h with the tropoelastin antibody at a dilution of 1:300. The blot was then washed twice with TBS-Tween, 0.5% BLOTTO (15 midwash) and incubated for 1 h with GAR-HRP diluted 1500 in TBS-Tween, 0.5% BLOTTO. After washing two times with TBS-Tween (15 min/wash), freshly made HRP Color Development Solution (4- chloronaphthol) was added.

Direct-binding ELISA for Quantitation of Tropoelastin in Cell Culture Medium

A stock standard solution was made by dissolving purified bovine tropoelastin at a concen- tration of 100 pglml in DMEM containing 5% calf serum and 30 mM HEPES. Before use, the tropoelastin stock solution was diluted to a final concentration of 320 ng/ml in tissue culture medium identical to that used for the cell type to be studied. Serial dilutions were then made in the microtiter plate to yield a range of 32 to 1 ng per well in a volume of 100 ~ 1 . For determining tropoelastin production by cultured cells, HEPES buffer was added to conditioned medium to a final concentration of 30 mM and 100 p1 aliquots were added directly to each well of the microtiter plate. The plate was then incubated for 3 h at 37°C to allow the tropoelastin to adsorb, then washed three times with PBS-Tween. Tropoelastin antibody diluted in the same buffer was added to the plate and incubated for 1.5 h at 37°C. Pre-immune rabbit serum and saline controls were included on each plate. Plates were washed three times with PBS-Tween, incubated for 1.5 h at 37°C with peroxidase-labeled GAR-HRP, washed three times with PBS-Tween, and developed with diamino-benzidine. Optimal concentrations of tropoelastin antibody (generally 1:2000) and GAR-HRP anti- body (1500) were determined by checkerboard titration which yielded a standard curve that is linear in the range 1-16 ng/well with a maximal optical density between 0.7 and 1.1.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN 269

Radiolabeling and Immunoprecipitation of Tropoelastin

SMC cultures, 2-3 days post-confluence, were incubated overnight in leucine-deficient DMEM containing 25 pCi/ml L-[3,4,5-3H] leucine (147 Ci/mmol) and 5% calf serum. Aliquots were removed for determination of tropoelastin levels by ELISA and total protein incorporation by precipitation with trichloroacetic acid (TCA). Protease inhibitors (final concentrations: 5 mM EDTA, 5 mM benzamidine, 1 mM eaminocaproic acid, and 1 mM N-ethylmaleimide) were added and the media was dialyzed against dilute acetic acid (0.25% v/v) to remove unincorporated radiolabel. Dialyzed media were lyophilized and resus- pended in 100 mM 2(N-morpholino)ethanesulfonic acid (MES) buffer, pH 6, containing 150 mM NaCl and enzyme inhibitors. Samples were adjusted to the same concentration of immunoprecipitated radioactivity, brought to 900 ~1 with MES buffer and preabsorbed with 30 p1 preimmune rabbit serum and 200 p1 of a 10% IgGSorb for 3 h at 4°C. Following centrifugation, tropoelastin antibody (12 p1) was added to the supernatant and incubation was continued overnight at 4°C. Immune complexes were precipitated with 50 pl IgGSorb suspension. Precipitated material was washed with MES buffer containing 1% Triton X-100, suspended in reducing sample buffer, and analyzed by SDS-PAGE using high resolution, 0.45 mm thick, 7.5%-12% gradient gels.9.17 For fluorography, gels were impregnated with ENHANCE and exposed to Kodak XAR-5 film at -70°C.

RNA Isolation and Northern Hybridization

Medium was removed from confluent cultures of FCL-240 fibroblasts and the cell layer was washed twice with PBS. Total RNA was isolated from cells by a modification of the differential precipitation procedure outlined by MacDonald et al. 18 Monolayer cultures were rinsed twice with PBS (10 mM potassium phosphate, 150 mM NaCl, pH 7.4), and cells were lysed in situ with 4 M guanidine thiocyanate as described.19 RNA in the cell extracts was precipitated with addition of 0.05 volume of 2 M potassium acetate, pH 5.5,0.08 volume of 1 M acetic acid and 0.7 volume of 100% ethanol and by incubating for 18 h at -20°C. Precipitated material was collected by centrifuging for 20 min at 10,000 X g, dissolved in 7.5 M guanidine hydrochloride containing 10 mM dithiothreitol and 10 mM EDTA. RNA was precipitated with addition of 0.1 volume of 2 M potassium acetate, pH 5.5 , and 0.5 volume of 100% ethanol followed by incubation for 18 h at -20°C. After centrifugation, super- natants were discarded and pellets were washed once with 70% ethanol, dried under vacuum, dissolved in 10 mM EDTA, and extracted once with 3 volumes of a 4: 1 mixture of chloroform and 1-butanol. After two back-extractions with 10 mM EDTA, RNA in the combined aqueous phases was precipitated with 2 volumes of 4.5 M sodium acetate, pH 7, for 18 h at -20°C. RNA was collected by centrifugation, dried under vacuum, dissolved in RNase-free water, and precipitated with 0.1 volume of 3 M sodium acetate, pH 5.2, and 2.2 volumes of 95% ethanol for 2 h at -20°C. Purified RNA was collected by centrifugation, dried and dissolved and stored in water. Purified RNA was denatured in 1 M formaldehyde in the presence of 50% formamide and 50 ng/pl ethidium bromide, resolved by electrophoresis though 1% agarose and transferred passively to nitrocellulose. Hybridization and wash conditions were as described20 except tropoelastin mRNA was detected with nick-translated T66, a 500-base pair exon-specific cDNA of the bovine elastin gene.*' Relative tropoelastin mRNA levels were determined by densitometric scanning of autoradiographs.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

270 I. W. PROSSER et al.

Immunohistological Localization of Extracellular Elastin

Cultured FBC-150 cells were washed with Tris-buffered saline to remove serum proteins and fixed with 0.5% glutaraldehyde-0.5% paraformaldehyde. Reactive aldehydes were blocked with 0.15 M glycine and the samples were washed with TBS, postfixed with 1% osmium tetroxide in cacodylate buffer, dehydrated, and embedded in Epon resin.22 Thin sections were stained with uranyl acetate and lead citrate; immunogold staining was as described.23

RESULTS AND DISCUSSION

Purification of Tropoelastin

Purification of tropoelastin was by a modification of the methodology developed earlier in our laboratory.9 By evaluating the recovery at each step of the isolation procedure, we discovered substantial loses of tropoelastin during dialysis (see below) and when tropoelas- tin-containing solutions were exposed to unsiliconized glass. We therefore modified the isolation protocol to avoid dialysis and have substantially increased the overall yield of purified protein. SDS-PAGE of the isolated product (Fig. 1) showed the characteristic bovine tropoelastin isoforms which migrate at a molecular weight of approximately 65 m a . The multiple lower molecular weight bands which migrate below tropoelastin may be degrada- tion products that arise during extraction and purification.9 The amino acid composition of the purified product (Table I) is typical of pure tropoelastin and agrees with the composition predicted from the bovine cDNA.21J4

Specificity of Tropoelastin Antiserum

Direct-binding ELISA of serum from rabbits injected with purified tropoelastin showed maximal titers at 6 to 8 months following the initial immunization. No reactivity above background was seen when the antiserum was incubated with wells coated with bovine collagen types I, I1 and IV (coated at 100 pg/well), laminin purified from Engelbreth-Holm- Swarm tumor (50 pg/well), hyaluronic acid or chondroitin sulfate proteoglycan from bovine nasal cartilage (100 pg/well) or with culture medium containing 1-10% calf serum (data not shown).

Immunoblot analysis demonstrated specific binding of the tropoelastin antiserum to authentic tropoelastin and its characteristic lower molecular weight degradation products (Fig. 1). Comparison of the tropoelastin antiserum reactivity with a mouse monoclonal antibody to a cell-recognition sequence of elastin25 showed no qualitative differences between the two antibody preparations (not shown). Use of tropoelastin antibody for ultrastructural immunolocalization demonstrated highly specific labeling of elastin, with antibody-coated gold particles localizing exclusively to the amorphous component of elastic fibers (Fig. 2).

Cross-reactivity of Tropoelastin Antibody with Mature Elastin

Conversion of tropoelastin to insoluble elastin involves covalent modification of lysine side chains to form multifunctional crosslinks.2 As a result, insoluble elastin contains potential

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN

A B

27 1

RGURE 1 SDS-PAGE and immunoblot of bovine tropoelastin. Purified tropoelastin was resolved using 7-12% gradient SDS-gels and visualized by silver staining (A) or transferred to nitrocellulose and reacted with tropoelas- tin antiserum (B). Immunoreactive bands were visualized with peroxidase-labeled second antibody.

antigenic determinants that are different from its soluble precursor.26 To investigate how the tropoelastin antiserum reacts with the mature protein, plates coated with 10 pg/well bovine ligamentum nuchae a-elastin were incubated with the tropoelastin antibody and binding was compared to wells coated with tropoelastin. The tropoelastin antibody bound with high specificity to a-elastin, exhibiting a similar limiting dilution (approximately 1:25,000) to that observed when reacted against plates coated with the immunizing preparation of tropoelastin. This high level of cross-reactivity was confirmed by demonstrating that an antiserum to bovine a-elastin26 bound at a similar titer to plates coated with tropoelastin (20 ng/well) as to plates coated with bovine a-elastin. It should be noted, however, that the amount of a-elastin needed to produce an equivalent titer was 500 times greater than tropoelastin (10 pg/ml compared to 20 ng/ml).

In contrast to the reactivity of the tropoelastin antibody towards a-elastin, no detectable antibody binding was detected in plates coated with 10 pg/well of elastin peptides derived from an overnight digestion of insoluble elastin with neutrophil elastase. Minimal cross- reactivity was evident when the coating concentration was increased to 100 pg/well, but the

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

272 I. W. PROSSER et al.

TABLE I

Amino acid composition of purified bovine tropoelastin compared with the composition predicted from the full length bovine tropoelastin cDNA.t

ResiduesllOOO

Purified bovine Predicted from tropoelastin tropoelastin cDNA

CYS 2 3 HYP 17 Asx 8 4 Thr 11 11 Ser 11 10 Glx 19 14 Pro 107 121 (Pro + Hyp) GlY 309 318 Ala 192 205 Val 127 130 Met 0 0 Ile 26 26 Leu 65 57 TYr 8 10 Phe 33 31 HY 1 0 0 His 0 0 LY s 53 52 Arg 10 8

+From Yeh et ~ 1 . 2 1 Values do not include the signal peptide.

limiting dilution (1:200) was several orders of magnitude lower than that found with tropoelastin as the coating antigen. Bovine a-elastin antiserum also had minimal cross- reactivity with elastase-derived peptides (titer 1:400). These findings suggest that the extensive hydrolysis associated with elastase digestion of insoluble elastin destroys shared epitopes which are predominantly responsible for the antigenicity of tropoelastin and a-elastin.

Binding of Tropoelastin to Multiwell Dishes in the Presence of Serum

During our initial characterization of the tropoelastin antiserum, we noted that purified tropoelastin bound rapidly and irreversible to the wells of plastic microtiter plates and that binding was only slightly reduced in the presence of large amounts of blocking protein. Upon further investigation, it was found that tropoelastin could bind to the plastic even in the presence of 10% fetal calf serum. To investigate this property more completely, doubling dilutions of known concentrations (as determined by amino acid analysis) of pure tropoelas- tin were diluted in HEPES-buffered DMEM with differing concentrations of calf serum. These samples were then added directly to acid washed microtiter plates, incubated for 3 hr at 37"C, and the amount of bound tropoelastin was determined with tropoelastin antiserum and GAR-HRP as described in Methods. The highest sensitivity was obtained in the absence of calf serum (Fig. 3A), although nonspecific binding of primary or secondary antibodies to residual binding sites on the plastic resulted in high background levels. The dilution curve in the absence of serum was linear from 10 to 80 ng/ml, with saturation of binding sites

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN 273

FIGURE 2 Immunogold labeling of pulmonary artery SMC cultures with tropoelastin antiserum. Shown is the specific localization of 5 nm tropoelastin antibody-coated gold particles to elastin fibers in the matrix of the cultured cells. Magnification = 54,ooOX.

for tropoelastin at about 160 ng/ml. Non-specific antibody binding could be completely blocked by the addition of 2.5% calf serum, but with some reduction in sensitivity as evident by an extension of the linear range to 20 to 160 ng/ml. Increasing the calf serum concentra- tion to 5% produced a further minimal reduction in sensitivity, and no further decrease was noted in tropoelastin binding from 5% to a maximum of 10% calf serum. Calibration curves remained linear in the range from 20 to 160 ng/ml, with linear regression coefficients typically greater than 0.96. In addition, pretreatment of microtiter plates by washing in 4 N sulfuric acid enhanced tropoelastin binding and reduced plate-to-plate variability.

The effect of pH on binding of tropoelastin to microtiter plates was investigated by diluting known amounts of tropoelastin in DMEM containing 5% calf serum and changing the pH by the addition of sodium phosphate buffers. After the final pH of each solution was verified, a standard curve was established for each pH value. Tropoelastin binding was slightly greater at acid than at alkaline pH (Fig. 3B), but observed differences were negligible. In addition, no significant difference in tropoelastin binding was found when comparing DMEM containing 5% calf serum buffered at neutral pH with phosphate, HEPES or bicarbonate.

The propensity of tropoelastin to bind to microtiter plates, even when at low concentra- tions in a complex protein mixture, may be due in part to its unusual molecular properties. Tropoelastin is an extremely hydrophobic protein,* and it is likely that hydrophobic

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

214 I. W. PROSSER et al.

1.5

2 1.0 k n s

0.5

0.0

* Noserum -8- 2.5% Serum * 5.01Serum -0- 101Serum

0 5 10 15 20 Tropoelastin (ng/100 ~ 1 )

pH6.35 pH6.50 * pH7.85

..

0 5 10 15 20 Tropoelastin (ng/100p1)

FIGURE 3 Effect of calf serum and pH on tropoelastin binding to microtiter wells. Panel A: Purified tropoelastin was serially diluted with DMEM alone or DMEM containing various concentrations of calf serum up to 10%. The samples were incubated on microtiter plates as described in Materials and Methods and bound tropoelastin was detected with tropoelastin antiserum and peroxidase-labeled second antibody. Panel B: Tropoelastin was diluted in DMEM-5% calf serum and the pH was adjusted by the addition of sodium phosphate buffers. Bound tropoelastin was determined as above.

interactions are largely responsible for its adsorption to the vinyl surface. This possibility agrees with the observation that substantial losses of tropoelastin occur during purification, most likely due to adsorption of protein to dialysis membranes and to the wall of extraction vessels (see below).

Direct-binding Assay for Tropoelastin Production in vitro

Having established that purified tropoelastin binds to microwell dishes in the presence of serum, we next determined whether a direct-binding assay could be used to evaluate tropoelastin production by cultured cells. Confluent cultures (100 cm2 culture dishes) of Fl3C-150 and A2058 cells were incubated for 24 h with DMEM (5 ml/plate) supplemented with 5% calf serum and 100 pg/ml BAPN. DMEM-5% calf serum incubated for 24 hat 37°C in a sterile dish which contained no cells was used as a control. Conditioned and control media were diluted with unconditioned DMEM-5% calf serum, 10 kl of 1 M HEPES was

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN 21 5

1.5

1 .o -

0.5 -

added per ml of medium, and the samples were added directly to microtiter wells. After a 3 h incubation at 37"C, the plates were washed and developed with tropoelastin antibody as described above. For a standard curve, known amounts of purified tropoelastin diluted in DMEM-5% calf serum were included on each plate. Figure 4 demonstrates the linearity of the standard curve as well as linear values for serial dilutions of medium samples from FBC-150 cells. No immunoreactive material was observed in control medium or in medium derived from A-2058 cells, a cell line which does not produce elastin, thus confirming the specificity of the assay for tropoelastin measurement.

The ability to accurately determine elastin levels in unprocessed samples is useful since appreciable losses of tropoelastin may occur during sample preparation. For example, to determine whether dialysis adversely affects tropoelastin recovery, medium (10 ml) was harvested from confluent bovine pulmonary artery SMC cultures (100 cm* dishes) after 3 days incubation, enzyme inhibitors were added, and the sample was split into two portions. One portion was stored at -7O"C, and the other was dialyzed at 4°C over 48 h against four 8 1 changes of 0.5 N acetic acid. The samples were then lyophilized and dissolved in water to their original volume. Tropoelastin levels were determined by direct-binding ELISA. These results showed that the dialyzed sample contained approximately 25% as much immuno- reactive tropoelastin as the undialyzed, frozen sample, suggesting that extensive losses of tropoelastin occurred during dialysis, despite the presence of serum and protease inhibitors. The loss of tropoelastin was variable from experiment to experiment, with little repro- ducibility evident even in the same culture system.

0 0.2

0.0 0 6 TIl 1:20 1:40 1:w 1:1a

Dilution 0

0

0. 0 . 0 1 . r ' I . I . I ' 8 . I . I . , . I .

Verification of ELISA Results by Immunoprecipitation and Northern Analysis

To verify that the values measured by the direct-binding ELISA were indicative of tropoelas- tin synthesis by cultured cells, results obtained by this method were compared with assessment of total medium tropoelastin using an immunoprecipitation assay. Tropoelastin synthesis and secretion follow monophasic kinetics with negligible intracellular sequestra-

4 0

FIGURE 4 Binding of tropoelastin in cell culture medium to microtiter plates. Shown is a typical standard curve generated with purified tropoelastin dissolved in DMEM-10% calf serum and developed with tropoelastin antibody. Serial dilutions of culture medium from FBC-150 cells is shown in the insert.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

216 I. W. PROSSER et al.

tion or degradation.27.28 Therefore, measurement of medium tropoelastin in the presence of BAPN, which prevents cross-linking of tropoelastin, allows a sensitive and specific estimate of total elastin production by cultured cells.

Bovine pulmonary artery SMC were incubated with [3H]-leucine as described under Materials and Methods, and total protein incorporation was estimated by TCA precipitation. To provide a range of concentrations of newly-synthesized tropoelastin, culture medium was sampled from different flasks at various times of incubation and at different degrees of confluency. Tropoelastin levels in aliquots of labeled medium were quantified by ELISA, and the results were expressed as nanograms of tropoelastin per 105 TCA-precipitable counts per minute. Additional aliquots containing an equal number of TCA-precipitable counts were immunoprecipitated with tropoelastin antibody and analyzed by SDS-PAGE. The total immunoprecipitated material was loaded onto the gel and the labeled species were detected by fluorography. Protein precipitated by tropoelastin antibody resolved as a tight multimer at approximately 65 kD (Fig. 5). These species, which co-migrated with authentic tropoelas- tin, were not precipitated by pre-immune rabbit serum. Estimation of the relative amount of tropoelastin precipitated from each sample of culture medium was performed by scanning densitometry and was in agreement with ELISA data from similar aliquots.

FIGURE 5 Correlation of tropoelastin estimations in labeled smooth muscle medium by ELISA and immu- noprecipitation techniques. Medium was harvested at 12 h (B and C), 24 h (D and E), and 36 h (G and H) from smooth muscle cultures labeled with [3H]-leucine and precipitated with tropoelastin antibody. Analysis by SDS- PAGE and Auorography demonstrates the characteristic tropoelastin multimers running slightly faster than bovine serum albumin. B, D and G: tropoelastin antibody. C, E and H: pre-immune rabbit serum. A and F: [W]-labeled protein standards include phosphorylase b (93 m a ) , bovine serum albumin (68 m a ) , and egg albumin (45 kDa) (arrows, top to bottom). The table at the bottom of the figure compares tropwlastin values determined by ELISA at the 12, 24, and 36 h time points with tropoelastin levels estimated by scanning the immunoprecipitated product in B , D, and G . Figures in parenthesis give percent increase in tropoelastin concentration relative to the 24 h time point.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN 277

The validity of the direct-binding assay was also evident when ELISA values were compared to tropoelastin mRNA concentrations. Correlation of steady-state and functional tropoelastin mRNA levels (as estimated by cell-free translation) with tropoelastin produc- tion in a number of elastin-producing tissues indicates that elastin biosynthesis is controlled at a pretranslational level 9 Therefore, estimates of steady-state tropoelastin mRNA con- centrations correlate closely with tropoelastin released into the medium. We have recently observed29 that tropoelastin expression in FCL fibroblasts is markedly reduced by exposure to 1,25-dihydroxyvitamin D, [1,25(0H), D,], the major hydroxylated form of vitamin D,. Because vitamin D can modulate elastin production in vitro, we used this reagent to examine the relationship between tropoelastin synthesis, as measured by direct-binding ELISA, and steady-state tropoelastin &A levels as estimated by Northern hybridization. Confluent cultures of FCL-240 cells were treated for 48 h with 100 kg/ml BAPN with or without 10-7 M 1,25(OH), D,. Medium was removed and processed for direct-binding ELISA, and RNA was extracted from the cell layer. Equal amounts of total RNA (5 Fg) from each sample was resolved by electrophoresis through a formaldehyde-agarose gel, transferred passively to nitrocellulose and hybridized with a [32P]-labeled exon-specific tropoelastin cDNA probe. Consistent with earlier studies, treatment with 1,25(OH), D, resulted in about a 6-fold decrease in tropoelastin levels in medium of ligament fibroblasts (Fig. 6B). Similarly, steady-state levels of tropoelastin &A were reduced 6.5 fold (Fig. 6A, B) verifying that the direct-binding ELISA can accurately determine physiologic changes in tropoelastin expression.

FIGURE 6 Confluent cultures of FCL-240 fibroblasts were treated for 48 h with 10-7 M 1,25(OH), D, dissolved in 100% ethanol. Control cells received the same volume of 100% ethanol (0.5% final concentration), and both groups were treated with 100 p g h l BAPN. Medium was removed and processed for direct-binding ELISA, and RNA was isolated from the cell layers. A: Autoradiogram of blot hybridized with the [)2P]-labeled elastin probe T66. Total RNA (5 pg) from control (lane 1) and 1,25(0H), D,-treated cells (lane 2) was analyzed by Northern hybridization as described under Material and Methods. The migration of tropoelastin mRNA (3.5 kb) was compared to 18 and 28s rRNA subunits. Autoradiographic exposure was for 36 h with an intensifying screen. Equivalency of RNA loading (5 f ig per lane) was assessed by ethidium bromide fluorescence of rRNA subunits which demonstrated no difference among the samples (not shown). B: Tropoelastin mRNA bands in the autoradio- gram in panel A were scanned using a densitometer. The results, in arbitrary units, are presented in this histogram (solid bars) and demonstrate a 6.5-fold decrease in the steady-state levels of tropoelastin mRNA in FCL-240 cells treated with 1,25(0H), D,. Tropoelastin protein levels in the medium, expressed as nanograms per microgram DNA (open bars), were reduced 5.7-fold in the 1,25(0H), D,-treated cells.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

278 I. W. PROSSER et al.

Sensitivity and Advantages of Direct-binding ELISA

The direct-binding ELISA described in this report is an order of magnitude more sensitive than our previously described competitive-binding assay.6 In the competitive ELISA, tropoelastin values could only be determined in medium samples that had been extensively dialyzed and concentrated by lyophilization and resolubilization. The sensitivity range of the direct-binding assay (i.e., the linear portion of the standard curve) corresponds closely to the range of tropoelastin concentrations achieved in cell culture medium after 2-3 days incubation. In fetal bovine pulmonary artery SMC and fetal bovine ligamentum nuchae fibroblast cultures, tropoelastin could be estimated directly in medium within the range of sensitivity of the assay. Fetal bovine chondroblast cultures produced much larger amounts of elastin, requiring dilution of the medium. Dilution series performed with such media showed good linearity within the useable range of sensitivity (Fig. 4). Because the sample requirement for the assay is small (100 pYwell), serial samples can be taken from the same culture to follow the accumulation of tropoelastin in the medium, allowing temporal analysis of elastin biosynthesis. Measurement of elastin production by a small number of cells was also possible.

ACKNOWLEDGMENTS

We thank Dr. Joel Rosenbloom, University of Pennsylvania, for providing the elastin cDNA T66. Technical assistance was given by Clarina Tisdale, Lisa Mecham, Gertrude Crump, and Noreen Person. The excellent secretarial assistance of Terese Hall is gratefully acknowledged. This work was supported by NIH grants HL26499 and HL41926 to R.I?M. and HL41040 to W.C.I?

REFERENCES

1. Partridge, S. M. (1962) Elastin. Adv. Prof. Chem. 17:227-302. 2. Franzblau, C. (1971) Elastin. Comprehensive Biochem. 26~659-712. 3. Paz, M. A., Keith, D. A., and Gallop, I? M. (1982) Elastin isolation and cross-linking. Meth. Enzymol. 82:

571-587. 4. Giro, M. G., Hill, K. E., Sandberg, L. B., and Davidson, J. M. (1984) Quantitation of elastin production in

cultured vascular smooth muscle cells by a sensitive and specific enzyme-linked immunoassay. Collagen Rel. Res. 4:21-34.

5. Mecham, R. I? and Lange, G. (1982) Antibodies to insoluble and solubilized elastin. Meth. Enzymol. 82: 744-759.

6. Wrenn, D. S. and Mecham, R. I? (1987) Immunology of elastin. Meth. Enzymol. 144:246-259. 7. Mecham, R. I? and Lange, G. (1980) Measurement by radioimmunoassay of soluble elastins from different

animal species. Connect. Tiss. Res. 7:247-252. 8. Kucich, U., Christner, I?, Lippmann, M., Kimbel, I?, Williams, G., Rosenbloom, J., and Weinbaum, G. (1984)

Utilization of a peroxidase antiperoxidase complex in an enzyme-linked immunosorbent assay of elastin- derived peptides in human plasma. Am. Rev. Respir. Dis. 131:709-713.

9. Wrenn, D. S., Parks, W. C., Whitehouse, L. A,, Crouch, E. C., Kucich, U., Rosenbloom, J., and Mecham, R. I? (1987) Identification of multiple tropoelastins secreted by bovine cells. J. Biol. Chem. 262:2244-2249.

10. Sandberg, L. B. and Wok, T. B. (1982) Production and isolation of soluble elastin from copper-deficient swine. Meth. Enzymol. 82:657-665.

11. Lansing, A. T., Rosenthal, T. B., Alex, M., and Dempsey, E. W. (1952) The structure and characterization of elastic fibres as revealed by elastase and by electron microscopy. Anat. Rec. 114555-575.

12. Senior, R. M., Griffin, G. L., and Mecham, R. I? (1980) Chemotactic activity of elastin-derived peptides. J. Clin. Invest. 66:859-862.

13. Mecham, R. I?, Whitehouse, L. A,, Wrenn, D. S., Parks, W. C., Griffin, G. L., Senior, R. M., Crouch, E. C., Stenmark, K. R., and Voelkel, N. E (1987) Smooth muscle-mediated connective tissue remodeling in pulmonary hypertension. Science 237:423-426.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.

DIRECT ASSAY FOR TROPOELASTIN 279

14. Quintarelli, G., Starcher, B . C., Vocaturo, A,, Di-GianFilippo, E, Gotte, L., and Mecham, R. F! (1979)

15. Mecham, R . I? (1987) Modulation of elastin synthesis: In vitro models. Meth. Enzymol. 144(D):232-246. 16. Todaro, G. J., Fryling, C., and DeLarco, J. E. (1980) Transforming growth factors produced by certain human

tumor cells: polypeptides that interact with epidermal growth factor receptors. Proc. Natl. Acad. Sci. USA

17. Matsudaira, I? T. and Burgess, D. R. (1978) SDS microslab linear gradient polyacrylamide gel electrophoresis. Anal. Biochem. 87:386-396.

18. MacDonald, R . J., Swift, G. H., Przybyla, A. E., and Chirgwin, J. M. (1987) Isolation of RNA using guanidinium salts. Meth. Enzymol. 152:219-227.

19. Parks, W. C., Whitehouse, L. A,, Wu, L. C., and Mecham, R. F! (1988) Terminal differentiation of muchal ligament fibroblasts: Characterization of synthetic properties and responsiveness to external stimuli. Develop. Biol. 129555-564.

20. Parks, W. C., Secrist, H., Wu, L. C., and Mecham, R. I? (1988) Developmental regulation of tropoelastin isoforms. J. Biol. Chem. 263:4416-4423.

21. Yeh, H., Ornstein-Goldstein, N., Indik, Z . , Sheppard, I?, Anderson, N., Rosenbloom, J. C., Cicila, G., Yoon, K., and Rosenbloom, J. (1987) Sequence variation of bovine elastin mRNA due to alternative splicing. Collagen Rel. Res. 7:235-247.

22. Hinek, A., Thyberg, J., Fribexg, U. (1976) Heterotopic allotransplantation of isolated aortic cells. An electron microscopic study. Cell Tiss. Res. 172:59-79.

23. Mecham, R. I?, Hinek, A, , Cleary, E. G., Kucich, U., Lee, S. J., and Rosenbloom, J. (1988) Development of immunoreagents to ciliary zonules that react with protein components of elastic fiber microfibrils and with elastin-producing cells. Biochem. Biophys. Res. Commun. 151:822-826.

24. Raju, K. and Anwar, R. A. (1987) Primary structures of bovine elastin a, b, and c deduced from the sequences of cDNA clones. J. Biol. Chem. 2625755-5762.

25. Wrenn, D. S., Griffin, G. L., Senior, R. M., and Mecham, R. F! (1986) Characterization of biologically active domains on elastin: Identification of a monoclonal antibody to a cell recognition site. Biochemistry 255172- 5176.

26. Mecham, R. F! and Lange, G. (1982) Antigenicity of elastin: Characterization of major antigenic determinants on insoluble elastin. Biochemistry 21:669-673.

27. Saunders, N. A. and Grant, M. E. (1985) The secretion of tropoelastin by chick-embryo artery cells. Biochem. J. 230:217-225.

28. Sephel, G. C. and Davidson, J. M. (1986) Elastin production in human skin fibroblasts and its decline with age. J. Invest. Dermatol. 86:279-285.

29. Parks, W. C., Hinek, A., Rosenbloom, A., and Mecham, R. I? (1989) Down regulation of elastin expression by 1,25 dihydrocholecalciferol-vitamin D3. J. Invest. Derm. 92:497.

Fibrogenesis and biosynthesis of elastin in cartilage. Connect. Tiss. Res. 7:l-19.

77:5258-5262.

Con

nect

Tis

sue

Res

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f T

oron

to o

n 08

/09/

12Fo

r pe

rson

al u

se o

nly.