Development of an Enzyme-Linked Immunosorbent Assay for Goldfish ...

BIOLOGY

OF REPRODUCFION

40, 68-73

Development

OLIVIER

(1989)

of an Enzyme-Linked lmmunosorbent for Goldfish Gonadotropin

KAH,2’3

iEsus

pj’,2 and Laboratoire

de

NUNEZ RODRIGUEZ,3 BRETON4

BERNARD

Physiologie

des

UA Avenues

Cellulaires2

France

de

Avenue

Biologie des

Talence, Laboratoire

ANDR#{201}CALAS,2

CNRS 339 des Facult#{233}s

Talence, Laboratoire

Interactions

Assay

Marine3

Facult#{233}s France

and de Physiologie des Poissons4lNRA Campus de Beaulieu Rennes,

France

ABSTRACT An enzyme-linked immwzosorbent assay (ELISA) for goldfish gonadotropin (GTH) was developed with the intent of devising a simple, reliable and nonradioisotopic assay for the measurement of GTH in goldfish biological samples. In this assay, soluble GTH of the standards or samples competes with carp GTH (cGTH) immobilized on a solid support (96-well microplate) for the fixation on antibodies to the 1-subunit of carp gonadotropin. The immobilized antigen-antibody complexes are then revealed by the peroxidase-antiperoxidase (PAP) technique. After revelation of the peroxidase activity, the absorbance value of each well is measured with a microplate reader. The cGTH concentration used for coating the wells is 2 nglml and the final dilution of the specific antibody is 1:80,000. The assay can be performed within 24 h and can be used over a range of 0.125-4 ng/ml. At about 50% binding, the intra- and interassay coefficients of variation are 5% and 9% respectively. The displacement cur.’es generated by goldfish plasma or pituitary perifusion fractions were strictly parallel to the standard cGTH. In addition, the stimulation by salmon gonadotropinreleasing hormone of pituitary fractions perjfused in vitro caused an immediate increase in the GTH measured in the collected fractions, strongly reinforcing the assumption that this assay indeed measures

GTH.

INTRODUCTION

0Th levels in fish. Homologous RIAs have, for instance, been developed for carp (Breton et al., 1971; Kobayashi et al., 1985), different salmonids (Crim et a!., 1973; Billard et al., 1976; Suzuki et al., 1988a), and catfish (Goos et al., 1986), but heterologous systems have been also employed to measure GTH levels in goldfish (Breton et aL, 1972; Crim et a!., 1976; Hontela and Peter, 1978) and in eels (Dufour et al., 1983) using the carp RIA. Fish pituitary glands have been thought for a long time to secrete only one 0TH; most of these assays were developed against a single 0Th molecule, frequently referred to as Con A-Il or maturational 0TH (Idler and Ng, 1979). Recently, Suzuki et a!. (1988b) have presented evidence for the existence of two chemically distinct OTHs, 0Th I and 0TH II, purified from

The purification of pituitary gonadotropins (OTHs) from different teleost species and the development of specific, sensitive, and reliable methods for their measurement was a prerequisite for our understanding of the hormonal interactions underlying fish reproduction. Since the introduction of radioimmunoassays (RIAs, Yallow and Berson, 1959), this methodology has been the only one used by different laboratories to measure

Accepted March 15, 1989. Received December 21, 1988. tReprint requests: Di. 0. Kah, Laboratoire Compar&, UA CNRS 651, Parc Valrose, 06034

de Physiologic Cellulaire Nice Cedex, France.

et

68

ELISA

FOR GOLDFISH

chum salmon pituitaries. On the basis of subunit structure, 0Th 1 is more related to the luteinizing hormone (LH) of tetrapods and 0TH II to follicle-stimulating hormone (FSH) (Itoh et a!., 1988). 0Th I also corresponds to the first 0Th purified in fish on the basis of its ability to induce intrafollicular oocyte maturation in vitro (Jalabert et a!., 1974; Breton et a!., 1976). The present work refers to the GTh previously purified from carp and therefore to the 0Th I of Suzuki et a!. (1988b). Initially, enzyme-linked immunosorbent assays (ELISA) were developed to measure the titer and specificity of antisera. However, more recently, their use has been extended to the detection and quantification of various antigens including mammalian pituitary hormones (Ishikawa et a!., 1982; Yochitake et a!., 1982; Signorella and Hymer, 1984; Farrington and Hymer, 1987; Spearow and Trost, 1987). According to these studies, ELISAs are reproducible and sensitive, and fulfill the criteria of valid assays. Among other advantages, this methodology does not require the use of radioactivity, which is often a limiting factor. For this reason, we have attempted to develop and validate a “cold” competitive assay by using an alternative method based on the ELISA principle. Recently, a competitive ELISA for mammalian LH was introduced by Spearow and Trost (1987). In this assay, the specific anti-LH antibodies are coated on a microplate and competition takes place between peroxidase-labeled LH and the free LH of the samples. A sandwich immunoassay involving enzyme-labeled antibodies also has been developed for human LH (Chow et a!., 1985). The strategy employed in our assay is slightly different, since free 0TH of the standards or samples will compete with immobilized 0Th coated on a 96-well microplate for the fixation on antibodies against the subunit of carp 0Th (3-c0TH). The immobilized antigen-antibody complexes are then reacted using the peroxidase-antiperoxidase (PAP) method. Peroxidase activity is revealed by a chromogen and the absorbance values of the wells are measured on a microplate reader. MATERIALS

AND

METHODS

Materials

The standard cOTH was prepared procedure described for salmon -0Th

according (Breton,

to the 1981),

GONADOTROPIN

69

and the antibodies against 3-cGTh were raised in rabbits (Breton et al., 1983). The swine anti-rabbit immunoglobulin 0 (IgO) and the rabbit PAP complexes were from Dakopatts (Glostrup, Denmark). Ortho-phenylenediamine, bovine serum albumin (BSA, grade V), and Tween 20 were obtained from Sigma Chemical Company (St. Louis, MO). Norma! swine serum was obtained from the local abattoir. The microplates routinely used for the assay were Nunc immunoplates, Type II (Intermed, Denmark). The optical density as read on a Titertek plate reader.

ELISA

Procedure

The assay involves 9 principa! steps separated, when indicated, by a wash cycle (3 washes each time) carried out as follows: The content of the plate was discarded by inversion, and each well was filled with 0.01 M phosphate buffer (pH 7.4) containing 0.05% Tween 20 (PB-’!’). After 2 mm, the contents were discarded and the plate was struck against a filter paper before being filled again. Step 1: Coating. The stock solution of cGTH (1 j.Lg/ ml) was diluted to 2 ng/m! in carbonate buffer (0.05 M; pH 9.6) and distributed 200 .tlJwell in all wells of the plate except for the first column, whose wells received 200 t1 of a 2 ng/ml solution of BSA in the same buffer. The plates were incubated overnight under slight agitation (14-16 h) at 4#{176}C. Step 2 (concurrent with Step I). Tubes containing either dilutions of the samples to be measured or different concentrations of the standard cGTh were incubated (v/v) with antiserum to f3-cGTh (1:80,000 final) overnight (14-16 h) at 4#{176}C. Dilutions were made in PBT containing 2% norma! swine serum (PB-T-NSS). Standard cOTh was serially diluted (factor of 2) from 7.8 to 0.0625 ng/ml final. Step 3: Saturation. After inversion of the plate, all wells received 200 p.1 of PB-T-NSS and were incubated 30 mm at 37#{176}C. Wash

cycle.

antibody. Two-hundred microliters of the mixture prepared in Step 2 was added to the wells. The first two columns received only the specific antibody at the final dilution (1:80,000). The plate was incubated for 2 h at 37#{176}C. Wash cycle. Step 5: Secondary antibody. All wells received 200 p.1 swine anti-rabbit 1g0 diluted 1:5000 in PB-T-NSS and were incubated for I h at 37#{176}C. This secondary antibody is used in excess so that only one site will Step

4: Primary

KAH

70 react with the primary antibody, leaving free for binding to PAP complexes. Wash

the other site

cycle.

complexes. All wells received 200 p.! PAP complexes diluted 1:5000 in PB-T-NSS and were incubated for 45 mm at 37#{176}C. Step

Wash

6: PAP

cycle.

All wells received 200 p.1 of a solution of orthophenylenediamine 0.05% in citratephosphate buffer (0.1 M/ pH 5) containing 0.025% of 30% hydrogen peroxide and were incubated for 20 mm in the dark at 20#{176}C. The solution was prepared just before use and kept in the dark. Preparing the revelation solution just before use and keeping the plates in the dark during the revelation is critical since the orthophenylenediamine solution is unstable, even if kept frozen for prolonged periods (over 1 mo). Step 8. The reaction was stopped by adding 50 p.1 of 2 M sulfuric acid in each well. Step 9. The absorbance value of each well was measured at 492 nm using a plate reader (Multiscan, Titertek), and the results were entered in a microcomputer (Apple lie or McIntosh Plus) for the mathematical processing and statistical analysis using a program developed in our laboratory. All standards and samples measurements were performed in duplicate or triplicate. Step

Analysis

7: Revelation.

of the

Results

Analysis of the results were performed according to Signorella and Hymer (1984). Briefly, a standard curve was obtained by plotting the ratio B/B0 versus the logarithm of the corresponding dose, where B was the mean absorbance value at some GTH dose minus the mean of the 0% values (BSA-coated wells) and B0 the mean absorbance of the 100% values (specific antibody only) minus the mean absorbance of the 0% values. A least-squares linear regression of the logit-log transformation was then performed on the linear part of the standard curve. The resulting equation was used to obtain the concentration of a given sample from its absorbance value. Pituitary

Perifusion

Adult goldfish (25-50 g) were obtained from a natural pond near Bordeaux and were kept in running tap water under a l6L:8D photoregime. The pituitary gland was rapidly dissected and immersed in cold Hank’s

ET AL. buffered saline solution (HBSS) prepared according to MacKenzie et al. (1984). Osmolarity of the perifusion medium was adjusted to 180-300 mOsm with manitol or sucrose. Under a dissecting microscope, the neurointermediate lobe was discarded and the rest of the gland was cut into small fragments. The fragments were divided equally and suspended in two seringue barrels containing 1.5 ml preswollen Biogel P-2 (Bio-Rad Laboratories, Richmond, CA). The bottoms of the seringues were then tapped with the plungers and washed for 3 h with HBSS before collection of the fractions every 2.5 mm. The constant flow rate was 0.3 ml/min. The temperature was maintained at 17#{176}C. After the equilibration period and the collection of 10 fractions. 10-mm pulses of salmon gonadotropin-releasing hormone (sGnRH, 100 nM) were given via a three-way valve. Between two pulses, samples were washed for 45 mm. The 0Th content of the fractions, diluted 20 times in PB-T-NSS, was then measured by using the ELISA procedure described above. RESULTS

A necessary condition for the suitability of such assays is that the coating of proteins is reproducible within the same plate. To assess this point, all wells of a plate were coated with cGTh at 2 ng/ml and incubated overnight. After the saturation step, the specific antibody, diluted 1:80,000, was added to all wells. The rest of the procedure was carried out as described under Material and Methods. The resulting coefficient of variation was found to be 4.6%, demonstrating that the coating step and the rest of the procedure are reasonably reproducible within the same plate. The optimum cOTh concentration used for the coating in Step 1 and the primary antibody dilution were determined by coating different cOTH concentrations (from 0.5 ng/ml to 16 ng/ml; factor of 2) in the wells of the columns and by filling the wells of the lines with serial dilutions of antibody (from 1:20,000 to 1: 160,000; factor of 2). The other steps were conducted as described under Material and Methods. The results shown in Figure 1 demonstrate that several combinations gave satisfactory optic densities between 1.0 and 1.5. The combination routinely used for the assay was 2 ng/ml for the coating of cOTh and 1:80,000 for the final dilution of the primary antibody. Figure 2 shows that changing the dilution of the secondary antibody (Step 5) did not critically modify the final absorbance value. For this reason and to save

ELISA

FOR GOLDFISH

GONADOTROPIN

71

BIB0

O.D.

1,00

0,80

1:20000 1:40000 1:80000 #{149} 1:160000

2

-_

0,60

0,40

0,20

0,00 0

,01

1

,1

10

GTH (ng/mI) FIG. 1. Effects of the carp gonadotropin (GTH) concentration used for coating the wells and of the dilution of the primary antibody on the final optic density (0.D.). All further steps were conducted as described under Materials

ng/mI

Dose FIG. 3. A compilation of 16 independent ratio B/B0 versus the dose (ng/ml). Vertical

displacement curves indicating bars represent standard error.

the

and Methods.

reagents, a dilution of 1:5000 was routinely used for the assay. For Step 6, PAP dilutions from 1:1250 to 1: 10,000 gave similar optic densities as shown on Figure 2. A dilution of 1:5000 of this reagent was chosen as standard condition for the assay. Figure 3 shows a typical standard curve obtained under routine conditions. A statistical analysis using the F-test, performed on 16 standard curves, demonstrated that there was no statistical difference between them. Despite this reproducibility, a standard curve was performed on every plate. An F-test for the linearity of regression was used to test the validity of the logit transformation. To reach a correlation coefficient over

O.D.

2nd

antibody

dilution

0.99, the highest dose (7.8 ng/rnl) and the lowest dose (0.0625 ng/ml) had to be skipped, resulting in a displacement curve ranging from 3.9 ng/ml to 0.125 ng/ ml. Samples of plasma and pituitary perifusion fractions were assayed at different dilutions (factor 2) to test for parallelism. Dilutions of plasma samples and anterior lobe perifusion fractions generated curves of similar shape compared to the standard cGTH. After logit transformation, the parallelism was tested by covariant analysis (Fig. 4). The sensitivity of the assay, defined as the lowest reliable value of the standard curve, was 0.125 ng/ml,

S.

4. o

Standard

#{149} Perifusion

0 -j

cOTH fraction

#{149} Plasma

2

0#{149}

-2

-4

log 1/64 0,125

1:1250

1:2500

1:5000

1:10000

1:20000

1:40000

PAP dilution FIG. 2. Effects of the dilution of the secondary antibody and of the peroxidas -mtiperoxidase (PAP) complexes on the fmal optic density (01).). The carp gonadotropin concentration used for coating was 2 ng/ml and the dilution of the primary antibody was 1:80,000.

I;16 0,5

1/4

dilution or dose ng/ml

FIG. 4. Logit/log transformation showing the parallelism between a fraction of a pituitary perifusion (100 tVwell) and a goldfish plasma sample (100 tI/ well) compared to reference carp gonadotropin (cGTH). The resulting equations were as follows: standard cGT}1: y = -0.0063 - I .2366x, R = 0.99; perifusionfraction:y=-0.9046l.2140x,R= 1.00;plasma:y= 1.39501.2951x,R = 0.99.

KAH

72 5 E

0

3

2

1 0

#{182}io sGnRH

20

3#{176}?

sGnRH

FIG. 5. Effects of 10-mm pulses of salmon (sGnRH) (100 nM) on the carp gonadotropin fragments perifused in vitro.

40

‘o

50 Fraction

number

gonadotropin-releasing hormone (GTH) released by anterior lobe

although lower values such as 60 pg/mi still resulted in a significant displacement. The intraassay variation calculated by measuring 10 times the same sample within the same assay was found to be 5% around 50% binding. The interassay variation calculated by measuring the same sample in 9 independent assays was shown to be 9% around 50% binding. If the assay indeed measures GTH levels in goldfish biological samples, one would expect that in vitro stimulation of perifused gonadotropes by OnRH would increase the basal 0TH level measured in the perifusates. Accordingly, Figure 5 shows that repeated stimulations of perifused 0TH cells by sGnRH (100 nM) cause immediate and sharp increases of the 0Th levels. DISCUSSION In conclusion, the present 0TH enzymoimmunoassay was developed with the objective of devising a specific, sensitive, reliable, rapid, and nonradioisotopic assay. Compared with the techniques of Spearow and Trost (1987) or Chow et al. (1985) for LH measurement, the present strategy does not require hormone or antibody labeling by an enzyme and therefore does not involve any purification step. On the other hand, the method does not appear to consume more reagents than the already existing RIM for fish gonadotropins. In terms of specificity, displacement curves generated by plasma dilutions or by anterior lobe perifusion fractions were strictly parallel to the standard curve. Antibodies against -cGTh have been shown to crossreact with total cOTH, but not with a-cGTH. Inversely, an antiserum against a-cGTh does not recognize 3cGTH (Breton et al., 1983). Therefore, antibodies against -cGTh does not cross-react with thyroid-slim-

ET AL. ulating hormone (TSH), which is not available for carp. Concerning the other pituitary hormones, it has not been possible to test the cross-reactivity since neither carp growth hormone nor prolactin (PRL) are available. The anti-cGTh does not cross-react with salmon PRL (P. Prunet, personal communication), salmon growth hormone (P.Y. Le Bail, personal communication), or salmon TSH (B. Breton, unpublished data), but this can be related to the important immunological speciesspecificity existing for a given hormone between fish. Nevertheless, this assay was shown to detect changes in the GTH levels after sOnRH stimulation strongly supporting the fact that it indeed measures 0TH. Concerning its sensitivity, this ELISA appears to be at least equivalent to the RIA for fish gonadotropins since levels as low as 60 pg/mi can be detected. From previous studies (Signorella and Hymer, 1984) and our own preliminary experiments, prolonged preincubation of the standards or samples with the specific antibody prior to adding the mixture to the cOTH-coated wells results in a greater sensitivity than immediately adding both reagents to the wells. It must be pointed out that the sensitivity was considerably increased by using the PAP method, which proved to be superior to the streptavidin-biotin or the indirect immunoperoxidase methods. This method allows improvement of the sensitivity of the assay and the notable reduction in the concentration of cGl’H used for the coating step, resulting in a reduced consumption of reagents. If sensitivity is not critical, the different incubation periods can be reduced, resulting in a shorter procedure. Beside the major advantage of being free of radioactivity handling and wastes, this assay provides a number of other positive aspects. First, the whole procedure can be performed in less than 24 h including counting time and result analysis. This is of obvious interest, especially for in vitro work since it is possible to perform an entire experiment in 2 days. Indeed, most RIA procedures currently in use for fish gonadotropins necessitate several days (Breton et al., 1972; Crim et al., 1976; Hontela and Peter, 1978; Dufour et a!., 1983). To be free of radioactivity and to use only stable reagents allows one to set up an assay in a very short period of time. Therefore, it is not necessary to have a large number of samples before running the ELISA. This is particularly valuable for small research groups, where only a limited amount of samples have to be measured. Another advantage of ELISA is the reduction in cost in terms of both equipment and reagent con-

ELISA

FOR GOLDFISH

sumption, which should allow more research groups to set up their own assay. In addition, the increasing popularity of ELISA in biology has recently led to the development of devices allowing these assays to be automated. REFERENCES Bilard

R, Richard M, Breton B, 1976. Stimulation do Ia s#{233}cr6tion gonadotrope hypophysaire apths castration chez Ia true arc-en-ciel; variation de Ia rponse an cours du cycle reproducteur. CR Acad Sci 283:171-74 Breton B, 1981. A study of salmon pituitary gonadotropin dissociation. (len Comp Endocrinol 45:147-52 Breton B, Billard R, Jalabert B, Khan 0, 1972. Dosage radioimmunologique des gonadotropines plamaliques chez Carassiuz auratus an cours do nycuimre et pendant I’ovulation. Gen Camp Endocrinol 18:4.63-68 Breton B, Jalabert, Reinaud P, 1976. Purification of gonadoiropin from rainbow trout (Salmo gairdneri Richardson) pituitary glands. Ann BioI Mint Biochim

Breton

Biophys

16:25-31

B, Kahn G, Burzawa-G#{233}rard E, Billard R. 1971. Dosage radioimmunologique d’une hormone gonadotrope de Carpe (Cyprinus carpio. L). CR Acad Sd 272: 1515-17 Breton B, Zohar Y, Sambroni E. 1983. Dosages radioiminunologiques h#{233}terologuesde gonadoptropincs tie type glycoprot#{233}iques maturantes de Poissons t#{233}l#{233}ost#{233}ens. Aquaculture 40:307-23 Chow SN. Ho-Yuen B. Lee CY, 1985. Applications of monoclonal antibodies in solid-phase immunoassays of human luteinizing hormone. J Appi Biochem 7:114-121 Crim LW, Meyer RK, Donaldson EM, 1973. Radioimmunoassay estimates of plasma gonadotropin levels in the spawning pink salmon. (len Comp Endocrinol 21:69-76 Crim LW, Peter RE, Billard R, 1976. Stimulation of gonadotropin secretion by intraventricular injections of hypothalamic extracts in the goldfish, Carassius auratus. (len Camp Endocrinol 30:77-82 Dufour S, Delerue-Le Belle N, Fontaine YA. 1983. Development of a heterologous radioimmunoassay for eel (Angullia anguilla) gonadotropin. (lea Comp Endocrinol 49:404-13

GONADOTROPIN

73

Farnngton MA. Hymer WC, 1987. An enzyme immunoassay for the growth hormone: applications to the study of growth hormone variants. Life Sd 402479-88 Goos HJTh, Dc Leeuw R, Burzawa.G#{233}rard E. Terlou M. Richter Cii, 1986. Purification of gonadotropic hormone from the pituitary from the african catfish, Clarlas gariepinus (Burchdll), and the development of a homologous radioiinmunoassay. (len Camp Endocrinol 63:162-70 Hositela A, Peter RE, 1978. Daily cycles in serum gonadotropin levels in the goldfish: effects of photoperiode, temperature, and sexual condition. Can J Zool 56:2430-42 Idler DR. Ng TB. 1979. Studies on two types of gonadotropins from both salmon and carp pituitaries. (len Camp Endocnisol 38:421-40 lshikawa K, Narita 0. Noguchi H. Kato K. 1982. Practical enzyme immunoassay for prolactin in human serum. Clin Chim Acts 121:181-87 Itoh H, Suzuki K, Kawauchi H, 1988. The complete amino acid sequences of subunits of two distinct chum salmon GThs. (len Camp Endocrinol 71: 438-51 Jalabert B, Breton B, Billard R, 1974. Dosage biologique des hormones gonedoiropes do poissons par Ic test do maturation in vitro des oocytes tie Truite. Ann Biol Anim Biochem Biophys 14:217-28 Kobayashi M, Aida M, Hanyu 1.1985. Radioimmunoassay for silver carp gonadotropin. Bull Jpn Soc Sci Fish 5 1:1085-91 MacKenzie DS, Gould DR. Peter RE, Rivier J, Vale WW, 1984. Response of superfused goldfish pituitary fragments to mammalian and salmon gonadotropin-relcasing hormones. Life Sci 35:2019-26 Signordlla AP. Hymer WC, 1984. An enzyme-linked immunosorbent assay for rat prolactin. Anal Biochem 136:372-81 Spearow JL, Trost BA, 1987. Development of a sensitive enzyme-linked immunosorbcnt assay for cattle, sheep, rat, and mouse luteinizing hormone. Biol Reprod 37:595-605 Suzuki K. Kanamori A, Nagahama Y, Kawauchi H, l988a Development of salmon GTh I and GTh II radioimmunoassay. Gen Comp Endocrinol 71: 459-67 Suzuki K, Kawauchi H, Nagahama Y, l988b. Isolation and characterization of two distinct gonadoiropins from chum salmon pituitary glands. Gee Camp Endocnnol 71:292-301 Yallow RS, Berson SA, 1959. Assay of plasma insulin in human subjects by immunological methods. Nature (Lond) 184:1648-49 Yoshitakc S. Endo Y, Nakagawa K, Imagawa M, Ohiaki S. Ishikawa E, 1982. A sandwich enzyme immunoassay to human growth hormone in serum using anti-human growth hormone igG-bcta-D-galactosidase conjugates. OinChimActa 125:1-7