Human interleukin 2

Journal oflmrnunologicalMethods, 97 (1987) 215-220 Elsevier

215

JIM04250

H u m a n interleukin 2 Detection at the picomolar level by sandwich enzyme immunoassay B. Ferrua, C. Aussel and M. F e h l m a n n Unitd 210 INSERM, Facultd de Mddecine, Laboratoire d'Im~nunologie, Chemin de Valombrose, Nice, France (Received 25 September 1986, revised received 5 November 1986, accepted 11 November 1986)

Human interleukin 2 was detected at the pM level by a simple sequential sandwich enzyme immunoassay. The lymphokine to be assayed was first extracted from supernatants of mitogen-activated Jurkat leukemic T cells or peripheral blood lymphocytes using anti-recombinant interleukin 2 rabbit IgG insolubilized onto polystyrene microtiter plates and was revealed by an anti-interleukin 2 FatY fragment conjugated to peroxidase. The whole method could be performed within 8 h and allowed the measurement of interleukin 2 irrespective of its degree of glycosylation. Among the currently used mitogens, only ConA at a concentration above 10/xg/ml interfered with the assay. The method was carefully compared to the reference bioassay and was found to be only 3-5 times less sensitive. Key words: Interleukin-2; Sandwich enzyme immunoassay; ELISA

Introduction

Human T cell growth factor (MW 1550017000) or interleukin 2 (IL-2) is a soluble sialoglycoprotein that promotes in vitro and in vivo mitogen and antigen-induced T cell proliferation (Morgan et al., 1976). This lymphokine has also been shown to act on other lymphoid cells including B cells (Teranishi et al., 1984), and activated or natural killer ceils (Moretta et al., 1984), to induce or enhance gamma-interferon production (Farrar et al., 1981; Pearlstein et al., 1983) and to potentiate in vivo the action of tumor-specific T cells (Hefeneider et al., 1983). For these reasons, the immunotherapeutic potential of IL-2 has been investigated in patients with

Correspondence to: B. Ferrua, U 210 INSERM, Laboratoire d'Immunologie, Facult6 de Mrdecine, 06034 Nice Cedex, France.

immunodeficiencies (Flomberg et al., 1983) and malignancies (Heberman, 1984). To date, the only method capable of measuring the extremely low levels of IL-2 present in cell supernatants has been a bioassay (Gillis et al., 1978), using IL-2-dependent cytotoxic T cell lines (CTLL). Unfortunately, this method is time-consuming and requires the maintenance of target cells. In addition, it is subject to various forms of interference including mycoplasma contamination, the occurrence of cytotoxic agents masking the IL-2 activity and possible presence of residual mitogenic substances mimicking the IL-2 effect on certain target cells. There has been a need to develop simple, specific yet sensitive, methods for quantitating IL-2 in biological samples and recently immunoassays employing enzyme- or radio-labelled reagents have been described. However, most of these methods fail to match the exquisite sensitivity of the bioassay (Gehman et al., 1984; Ohike et al.,

0022-1759/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

216 1986) or are still relatively complicated to set up and to perform because of the use of radiolabelled and immunopurified reagents (Cardenas et al., 1986). In the present study, human IL-2 was detected at the pM level using a very simple and convenient sequential sandwich enzyme immunoassay method. Anti-human recombinant IL-2 rabbit IgG physically absorbed onto a microtiter plate was used as the capture antibody and the corresponding Fab' fragment conjugated to horseradish peroxidase by means of a maleimide derivative was selected as the tracer. The detection limit of the method approaches that of the bioassay, permitting the recognition of recombinant IL-2 as well as glycosylated IL-2 from Jurkat T cells or peripheral blood lymphocytes. Of the mitogens tested only high concentrations of ConA gave false reactions.

Materials and methods

Flat bottom polystyrene microtiter plates (96 wells) Immulon II were obtained from Dynatech (Denkendorf, F.R.G.). Bovine serum albumin (BSA) fraction V, horseradish peroxidase (HRP), grade I 250 U / m g , PHA, streptomycin, penicillin and glutamine were from Boehringer Mannheim (Mannheim, F.R.G.). Phorbol myristic acetate (PMA), o-phenylenediamine dihydrochloride (OPD) and sodium pyruvate were purchased from Sigma (St. Louis, MI). The coupling reagent, succinimidyl 4-(N-maleimidomethyl) cyclohexane carboxylate was obtained from Pierce, Rockford (IL). RPMI 1640 medium was from Flow (McLean, VI) and fetal calf serum (FCS) was from Eurobio (France). HRP-labelled anti-rabbit IgG sheep Fab' was prepared following the method of Ishikawa et al. (1983). Human recombinant IL-2 Human recombinant IL2 (RIL-2) was kindly provided by Biogen Medical Research (Geneva, Switzerland). According to the manufacturer, this material was at least 97% pure RIL-2 with a specific activity of 13 U / n g . Production of IL-2 from cell cultures Natural human IL-2 was derived from a T

leukemic line Jurkat kindly supplied by Dr. Schmitt-Verhulst, Centre d'Immunologie (Marseille-Luminy, France) and subcloned by limiting dilution. The cells were cultured in RPMI 1640 medium supplemented with 5% FCS, 50 I U / m l penicillin, 50 /*g/ml streptomycin, 200 /~M glutamine, 200 ~tM sodium pyruvate, 10 /,M 2mercaptoethanol at 37°C in 5% CO 2 humidified atmosphere and usually stimulated with 100/*g/ml PHA and 10 n g / m l PMA. Natural IL-2 was also obtained from peripheral blood lymphocytes (PBL) from pooled healthy donors following the method of Korner et al. (1986). In both cases, IL-2 levels ranged from 1.5 to 10 n g / m l (19.5-130 U/ml). Bioassay for IL-2 The biological activity of IL-2 was determined by measuring its ability to induce the proliferation of an IL-2-dependent murine cytotoxic T lymphocyte line (CTLL2) kindly provided by Dr. Pierres, Centre d'Immunologie (Marseille-Luminy, France). Briefly, serial dilutions (0.1 ml volume) of IL-2 in RPMI-supplemented medium were mixed in a microtiter plate with 0.1 ml of medium containing 5000 CTLL2 for 24 h at 37°C in a 5% CO 2 humidified atmosphere. The cultures were then pulsed for 4 h with 50 #1 of medium containing 1 /,Ci of [3H]thymidine, CEA (Saclay, France). The cells were washed using a cell harvester Titertek, Flow (McLean, VI) and counted by a standard scintillation technique. Anti-RIL-2 rabbit antiserum One rabbit was immunized at 2 weeks intervals with 20 /,g of RIL-2 dissolved in 0.2 ml of saline and emulsified with complete Freund's adjuvant. Subcutaneous and multisite injections were performed in the back and the animal was bled from the marginal ear vein 4 days after each injection. The different antiserum batches were kept frozen until testing. Substrate solution The color development solution consisted in 0.1 M phosphate citrate buffer pH 5.5, hydrogen peroxide 0.02%, OPD 3 m g / m l (final pH 5.0) and was freshly prepared in a disposable vessel protected from light.

217

Measurement of antiserum titer Anti-IL-2 activity was determined by an ELISA method. RIL-2 (40 ng) in 0.1 ml of 0.1 M phosphate buffer pH 7.2 (PB) was distributed into the wells of a microtiter plate and incubated overnight at room temperature (RT). After unbound antigen was aspirated off, the wells were rinsed four times with buffer and saturated with PB 1% BSA (PBBSA) for I h at room temperature. Duplicate 0.1 ml aliquots of the different antiserum batches diluted in PB-BSA were delivered into the IL-2coated wells and incubated for 90 min at room temperature. After washing with PB 0.1% Tween 20, the wells were incubated for another 90 min with 0.1 ml of HRP-labelled anti-rabbit IgG sheep Fab' diluted in PB-BSA buffer, washed as described above and filled with 0.1 ml of substrate solution. The enzymatic reaction proceeded for 30 min at RT in the dark and was quenched with 0.1 ml of 2 N HzSO 4. Absorbances were read at 492 nm in an automatic plate reader (MR 600, Dynatech). Anti-RIL-2 rabbit IgG From the anti-RIL-2 bleeding showing the highest titer as measured by the ELISA procedure, an IgG fraction was prepared with caprylic acid according to the method of Steinbuch et al. (1970). This fraction was selected for the coating step or further purified by D E 52 cellulose (Whatman, England) chromatography before peptic digestion. HRP-labelled anti-RIL-2 rabbit Fab' The Fab' fragment, prepared from the repurifled anti-RIL-2 rabbit IgG by peptic digestion and 2-mercaptoethylamine reduction was coupled to H R P by means of succinimidyl 4-(N-maleimido methyl) cyclohexane-l-carboxylate following the method of Ishikawa et al. (1983) without modification. After coupling, the Fab'-HRP conjugatecontaining peak was isolated by ACA 44 Ultrogel gel filtration (IBF, France), supplemented with 1% BSA and 0.01% thimerosal, mixed with 1 vol. glycerol and stored aliquoted at - 2 0 ° C . Under these conditions no appreciable loss of immunoreactivity or enzymatic activity was observed within 3 months.

Anti-RIL-2 solid phase Caprylic acid-purified IgG was treated at p H 2.5 following the procedure of Miyai et al. (1981) and then neutrafized and adjusted to 27.4 # g / m l with PB. The wells of a microtiter plate were filled with 0.2 ml of this solution for 3 h at RT. The coating solution was removed by aspiration and the wells were washed four times with buffer. Finally, the plate was filled with PB and stored with a plastic sealer at 4°C until use. RIL-2 standard RIL-2 (0.13 m g / m l ) was diluted 1 : 3 with 1% ( v / v ) acetic acid and then adjusted to 100 n g / m l with PB 0.2% BSA. This stock calibrator solution was kept at - 2 0 ° C and RIL-2 standards ranging from 0 to 10 n g / m l (0-130 U / m l ) were daily prepared in RPMI-supplemented medium. The value of each standard was verified using the BRMP IL-2 reference preparation (Biological Resources Branch, Frederick, MD). Enzyme immunoassay procedure The anti-IL-2-coated plate was quickly rinsed twice with PB and duplicate 0.1 ml aliquots of standards or unknowns were dispensed into the wells followed by 0.05 ml of 0.2 M Tris-HC1 buffer pH 8.2, 2% BSA. The plate was sealed and incubated for 5-16 h at RT. Unbound material was then aspirated off and the plate thoroughly washed with running cold tap water. HRP-labelled anti-RIL-2 rabbit FalY (600 ng in 0.15 ml of PB-BSA buffer) was added and incubated for 90 min at RT. The plate was washed as described above and substrate solution (0.15 ml) was added. After a 30 min incubation at RT in the dark, the reaction was stopped with 0.1 ml of 2 N HzSO 4 and the absorbances read at 492 nm with the automatic plate reader.

Results

Anti-RIL-2 antiserum The results of the ELISA test developed for the detection of the anti-RIL-2 antibodies are detailed in Fig. 1. The antiserum titer markedly increased after the third injection and reached a plateau between 40 and 60 days of immunization. At this

218 E

A492 nm

C

100-

// //

"4"

iii

I/

]I

15-

I]

X3

S

7S

//

-

Y

iI

D

III

I

I

II I

I

10-

I// 50-

/I I

v5 E

05-

25-

//

~ /

32

/ i

0

10

t

t . - - - 4 - - ~ - - - 4-,-.-+- -,-+-.- - , - - f - ~ 20

30

40

50

60

70

80

90

Fig. 1. Time-course response of the anti-RIL-2 rabbit as measured by ELISA. The different antiserum batches were tested at 10 3 ( I I ) and i0 -4 (O 0) dilutions on IL-2coated wells (full lines) or BSA-coated wells (dotted lines). The arrows indicate the different booster injections.

stage, antibodies were still detectable at the 10 6 dilution. After this period, in spite of additional boosts, the antiserum titer declined.

Comparison of enzyme immunoassay and bioassay The average dose-response curve for the RIL-2 bioassay and calibration curve for the RIL-2 enzyme immunoassay are depicted in Fig. 2. Under the conditions used, the bioassay exhibited half maximal [3H]thymidine incorporation at 0.5 ng/rnl of RIL-2. The threshold sensitivity calculated as the lowest concentration of RIL-2 giving a thymidine incorporation distinguishable with a 99% confidence from that of standard 0 was 0.008 ng/ml. The ELISA test was only 3-5 times less sensitive with a half maximum absorbance for 1.5 n g / m l (19.5 U / m l ) of RIL-2 and a detection limit lying at 0.04 n g / m l (2.58 pM, 0.52 U / m l ) . In addition the standard ELISA curve was much more reproducible than that of the bioassay which showed extensive variation.

Specificity of the ELISA test Standard curves constructed with RIL-2 or di-

/

I

ii

I

/i ' / / •

/ /

// /

X

yU I

Days alter" /he 1st ~nJechon

i II

I

I

+

o+

IIII

00049

I

0.019

I

/

I

0 078

031

I

I 25

I

I

t

5

10

RIL 2 ng/ml

Fig. 2. Dose-response curve (H H) for the RIL-2 bioassay (mean of five experiments--1 SD) and calibration curve (© ©) for the RIL-2 ELISA (mean of seven experiments_+l SD). Results were expressed as % of maximal [3H]thymidine incorporation (ranging from 50000 to 75000 cpm) or % of absorbance of the highest calibrator (ranging from 1.4 to 1.65).

lutions of supernatant cultures from mitogenactivated Jurkat leukemic T cell line and PBL (Fig. 3) were parallel, thus suggesting that the polyclonal rabbit antibody recognized IL-2 from various sources and differing only in their degree of glycosylation. In addition PMA and PHA used as mitogens did not interfere in the assay (Table I) whereas relative high concentrations of ConA (above 10 /~g/ml) did mimick the effect of IL-2. However, this undesired effect could be easily prevented by adding an excess of a-methyl mannoside in the first assay step.

Assay precision Three IL-2-containing samples, kept frozen and with values spanning the useful assay range were assayed 10 times in the same series or on seven different occasions including different batches of anti-RIL-2-coated plates. Within-run and day-today reproducibility was good (Table II) with coefficients of variation ranging from 4.4% to 9.3%.

219 5UPERNATANT DILUTION A 4 9 2 nm

1:64

1!32

I

I

1:16 I

1:8

1:4

I

I

1:2 I

1:1 I

T A B L E II W I T H I N - R U N A N D D A Y - T O - D A Y PRECISION OF IL-2 ELISA Number of determinations

Mean (ng/ml)

Standard deviation (ng/ml)

Coefficient of variation (%)

Within run

10 10 10

0.42 2.3 6.8

0.031 0.13 0.30

7.4 5.6 4.4

D a y to day

7 7 7

0.48 2.1 7.4

0.045 0.15 0.51

9.3 7.1 6.8

1.5-

1.0-

0.5-

././/

T A B L E III

:,...j./'/ 0

dd

0.0049

I

0.078

i

0 _31

A P P L I C A T I O N OF T H E IL-2 ELISA TEST TO M E A S U R E IL-2 R E L E A S E F R O M T H E J U R K A T CELL L I N E ON SEVEN D I F F E R E N T OCCASIONS I

1.25

i

1

5

10

Mitogens

RI L 2 n o / m l

Fig. 3. Parallelism experiment. Representation of curves drawn with RIL-2 (O 0) and dilutions of supernatants from mitogen-activated Jurkat cell line (~, A) or PBL (It II). Dilutions were prepared in RPMI-supplemented medium.

PHA 1-100 ~g/ml

ELISA IL-2 (ng/ml) N.D. Bioassay IL-2 (ng/ml) N.D.

a

P M A P H A 100 1-100 l~g/ml n g / m l + P M A 10 ng/ml N.D.

4.0-1.7-0.4-0.4-2.5-4.0-3.5

N.D.

4.5-1.3-0.2-0.3-3.2-4.5-4.0

a Not detectable.

Application of the IL-2 ELISA test IL-2 released from the Jurkat cell line activated with PHA a n d / o r PMA was quantified both by ELISA and bioassay (Table III). As already published, (Wiskocil et al., 1985) PHA or PMA alone TABLE I E F F E C T OF M I T O G E N S O N IL-2 ELISA Mitogen PMA (ng/ml)

Equivalent RIL-2 measured ( n g / m l ) 0.1 1.0 i0.0

< 0.04 < 0.04 < 0.04

PHA (~g/ml)

1.0 10.0 100.0

< 0.04 < 0.04 < 0.04

ConA (~g/ml)

5.0 10.0 20.0 50.0

< 0.04 0.05 0.3 18.0

at currently used concentrations were unable to produce measurable amounts of IL-2 whereas high levels of the lymphokine were detected on seven different occasions in the supernatants of cells activated with both compounds. In addition, IL-2 values measured by both techniques were similar.

Discussion

The major difficulty in developing a radio- or enzyme immunoassay method for IL-2 measurement which could replace the complex bioassay is to achieve a sufficiently low detection limit. Among the various available immunoassay designs, the sandwich protocol has been demonstrated to be the most sensitive (Ekins et al., 1984) especially when a large excess of high avidity first and second antibody are present in the reaction vessel.

220

In the present study, an excess of first antibody was ensured by selecting irradiated polystyrene microtiter plates with a high protein-binding capacity. This capacity was improved by pretreating the anti-RIL-2 IgG used for coating under acidic conditions (results not shown). In the same way, HRP-labelled Fab' prepared by the maleimide technique produced low background values and therefore large amounts of labelled antibody could be used in the second step of the test. The advantages of this conjugation procedure have been detailed elsewhere (Ishikawa et al., 1983) and in our hands have been confirmed in various sandwich systems (Ferrua et al., 1985). The IL-2 enzyme immunoassay here described showed an absolute detection limit at 2.58 pM corresponding to 0.26 fmol or 0.004 ng (0.052 U) of material, approaching that of the bioassay. In addition, the ELISA test was much simpler to perform using standard laboratory equipment, more reproducible and less subject to interference than the bioassay. Moreover, it does not demand the maintenance of target ceils and since the reagents are extremely stable it is more adaptable for daily use. Attempts to improve the assay detection limit by using the avidin-biotin system (Vilja et al., 1985), the enzyme amplification procedure with alkaline phosphatase-labelled Fab' (Self, 1985) or the indirect sandwich ELISA with a monoclonal antibody directed against RIL-2 (Cousin et al., 1985) were unsuccessful, each giving unacceptable background absorbances. The test was able to detect unglycosylated RIL-2 as well as various glycosylated forms of natural IL-2 (Gehman et al., 1984) but, unlike the bioassay, could not be applied to the quantitation of mouse IL-2. Contrary to the results of Cardenas et al. (1986) the presence of whole human serum in the first assay step reduced the assay sensitivity only by two-fold. The IL-2 ELISA described here appears to be a valuable alternative to the sophisticated bioassay. Similar assays for other lymphokines could be developed using the same principles. Acknowledgements We wish to thank Drs. L. Schaffar and D. Mary for helpful discussions. The illustration work

of A. G r i m a and C. Minghelli is greatly appreciated.

References Cardenas, J.M., Marshall, P., Henderson, B. and Altman, A. (1986) J. Immunol. Methods 89, 181. Cousin, M.A., Lando, P., Peyrat, M.A. and Soulillou, J.P. (1985) Lymphokine Res. 4, 319. Ekins, R. and Jackson, T. (1984) In: C.A. Bizollon (Ed.), Monoclonal Antibodies and New Trends in Immunoassays (Elsevier, Amsterdam) p. 149. Farrar, W.L., Johnson, H.M. and Farrar, J.J. (1981) J. Immunol. 126, 1120. Ferrua, B., Maiolini, R. and Masseyeff, R. (1985) Ann. Biol. Clin. 4, 591. Flomemberg, N., Welte, K., Mertelsmann, R., Krenan, N.. Ciobanu, N., Venuta, S., Feldman, S., Kruger, G., Kirkpatrick, D., Dupont, B. and O'Reilly, R. (1983) J. Immunol. 130, 2644. Gehman, L.O. and Robb, R.J. (1984) J. Immunol. Methods 74, 39. Gillis, S., Ferm, M.M., Ou, W. and Smith, K.A. (1978) J. Immunol. 120, 2027. Heberman, R.B. (1984) J. Biol. Response Modif. 3,527. Hefeneider, S.H., Conlon, P.J., Henney, C.S. and Gillis. S. (1983) J. Immunol. 130, 222. Ishikawa, E., Imagawa, M., Hashida, S., Yoshitake, S., Hamaguchi, Y. and Uneo, T. (1983) J. Immunoassay 4, 209. Korner, I.J., Dettmer, R., Kopp, J., Gaestel, M. and Malz, W (1986) J. Immunol. Methods 87, 185. Miyai, K., Ishibashi, K. and Kawashima, M. (1981) In: E. Ishikawa, T. Kawai and K. Miyai (Eds.), Enzyme Immunoassay (Igaku-Shoin, Tokyo) p. 186. Mond, J.J., Thompson, C., Schaeffer, M., Finkelman, F.D. and Robb, R.J. (1984) Curr. Top. Microbiol. Immunol., 113, 102. Moretta, A., Pantateo, G., Mingari, M.C., Melioli, G., Moretta, L. and Cerottini, J.C. (1984) Eur. J. Immunol. 14, 121. Morgan, D.A., Ruscetti, F.W. and Gallo, R. (1976) Science 193, 1007. Ohike, Y., Imai, M., Tanaka, E., Mukaida, N., Kasahara, T., Tashibana, K., Miyakawa, Y. and Mayumi, M. (1986) J. Immunol. Methods 87, 245. Pearlstein, K.T., Palladino, M.A., Welte, K. and Vilucek, J. (1983) Cell. Immunol. 80, 1. Self, C.H. (1985) J. Immunol Methods 76, 389. Steinbuch, M., Audran, R. and Pejaudier, L. (1970) C.R. S6ances Soc. Biol. Ses. Fil. 164, 296. Teranishi, T., Hirano, T., Lin, B.H. and Onoue, K. (1984) J. Immunol. 33, 3062. Vilja, P., Kron, K. and Tuohimaa, P. (1985) J. Immunol. Methods 76, 73. Wiskocil, R., Weiss, A., Imboden, J., Kamin-Lewis, R. and Stobo, J. (1985) J. Immunol. 134, 1599.