An indirect enzymoimmunological assay for hyaluronidase

Journal oflmmunologieal Methods, 104 (1987) 223-229 Elsevier

223

JIM04533

An indirect enzymoimmunological assay for hyaluronidase Bertrand Delpech, Philippe Bertrand and Claude Chauzy Laboratory of Irnmunochemistry, Centre Henri-Becquerel, Rue d'Amiens, 76000 Rouen, France (Received 9 February 1987, revised received 15 April 1987, accepted 23 June 1987)

The use of a hyaluronic acid-binding proteoglycan (hyaluronectin) as a probe for the detection of hyaluronic acid has facilitated the development of an indirect enzymo-immunological assay for hyaluronidase. Plastic microtest ELISA plates were coated with hyaluronic acid. Incubation with hyaluronidase led to the destruction of insolubilized hyaluronic acid in proportion to the hyaluronidase concentration of samples. Residual hyaluronic acid was assayed by its capacity to bind immune complexes made up of hyaluronectin supplemented with alkaline phosphatase-conjugated anti-hyaluronectin antibodies. The technique was very sensitive and permitted the detection of as little as 10 -1° NFU of bovine testicular hyaluronidase. Hyaluronidase was detected by this technique in human sera, bee venom and culture medium of human hepatoma cell lines. Key words: ELISA; Hyaluronidase; Hyaluronectin; Hyaluronic acid

Introduction

Many techniques have been proposed for the assay of hyaluronidases, a group of enzymes which catalyse the degradation of hyaluronic acid and chondroitin sulphate inside (lysosomal hyaluronidase) our outside the body (testicular, venom or bacterial hyaluronidase). Assay techniques rely on viscosity reduction of HA solutions (Dorfman, 1948), or on turbidity-reducing activity in the presence of albumin (Dorfman and Ott, 1948), or on the increase of the reducing power of an HA solution due to the formation of oligosaccharides (Meyer et al., 1941). More sensitive techniques

Correspondence to: B. Delpech, Laboratory of Immunochemistry, Centre Henri-Becquerel, Rue d'Amiens, 76000 Rouen, France. Abbreviations: HA, hyaluronic acid; HN, hyaluronectin; IC, HN-anti-HN immune complexes; TRU, turbidity reducing unit; NFU, national formulary unit; PBS, phosphate-buffered saline; BSA, bovine serum albumin.

have also been described such as the spectrophotometric method of Benchetrit et al. (1977) which can detect 5 x 10 -5 NFU. Richman and Baer (1980) set up a convenient plate assay capable of detecting 0.3 x 10 -3 NFU, whereas Fiszer-Szafarz (1984) described a gel electrophoresis technique which could detect 2 x 10 -3 NFU. To analyse the interaction between HA and hyaluronectin (a proteoglycan which binds specifically and reversibly to HA) (Delpech and Halavent, 1981; Bertrand and Delpech, 1985; Delpech et al., 1986) we set up an indirect enzymoimmunological assay for HA on HA-coated plastic microtest plates (Delpech et al., 1985). We observed that incubation of hyaluronidase in the wells of plates completely suppressed the HN binding capacity of HA-coated plates, whereas other enzymes had no effect. This prompted us to use HA-coated plates as a solid substrate for the assay of hyaluronidase. The HN-binding capacity of plates was measured by adding HN-anti-HN immune complexes to the wells, a method which was also used to detect

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

224

tissue HA (Girard et al., 1984, 1986). The technique detects as little as 10 -a° NFU and permits the assay of hyaluronidase activity in human sera, bee venom and human hepatoma cell culture media.

Materials and methods

Materials Plastic microtest plates for enzymoimmunological assays were obtained from Nunc (Poly Labo Block, Strasbourg, France) and Dynatech (Fresnes, France). Absorbances were read on a Titertek multiskan instrument (Flow laboratories, 92 Puteaux, France). Incubations were performed in a Thermomix water bath (Roucaire, 78 Vrlizy, France) or in cuvettes maintained at 37 °C in an air cabinet (Jouan, 44 Saint Herblain, France). Hyaluronic acid (grade I), bovine testicular hyaluronidase (290 NFU/mg), calf intestinal alkaline phosphatase, bee venom (grade I), Streptomyces griseus protease (4.4 U/mg), bovine liver fl-glucuronidase (9 x 103 U/mg), and Clostridium histolyticum collagenase (type I, 150 U / m g ) were purchased from Sigma Chemical Co. (Saint Louis, MO). Stock HA solution was made with hyaluronic acid from Fluka (Buchs, Switzerland). Streptomyces hyaluronidase (2000 T R U / m g ) was from Calbiochem (San Diego, CA). Proteus vulgaris chondroitinase ABC was from Seikagaku Kogyo (Tokyo, Japan); porcine pancreatic trypsin (3200 U/mg) from Industrie Biologique Fran~aise (92 Villeneuve-la-Garenne, France); porcine pancreas elastase (8.7 U/mg) from Millipore (Freehold, N J). Leech hyaluronidase (Orgelase) was a gift from Biopharma (Swansea, U.K.). Proteins were assayed with the Bio-Rad protein kit (Munich, F.R.G.). Human hepatoma cell lines were cultivated in RPMI 1640 culture medium supplemented with 10% foetal calf serum (Biopro, 68 Mulhouse, France) for 2 days, then in serum-free medium for 2 days. Culture flasks were obtained from CML (77 Nemours, France).

Methods Indirect immunological assay for hyaluronidase on plastic microtest plates Plastic microtest plates were coated with HA (100 mg/1 in 0.1 M sodium bicarbonate) and left overnight at 4 ° C. Coated plates were rinsed with distilled water before use to eliminate ions. Samples for assay were diluted in the appropriate buffer and incubated in plates in duplicate. After incubation with hyaluronidase samples, wells were rinsed with PBS or with deionized water three times and incubated for at least 4 h at 37 °C with HN immune complexes conjugated with alkaline phosphatase diluted in 0.1 M sodium phosphate at pH 7. Alternatively, incubations were performed according to a two-step procedure. HN (150 ng/ml in PBS supplemented with BSA) was incubated in wells for 4 h. The wells were rinsed and incubated with the diluted conjugated antibodies overnight. After incubation with immune complexes or with antibodies, wells were rinsed three times with PBS, then incubated with substrate (p-nitrophenyl phosphate 1 mg/ml in 1 M diethanolamine with 1 mM MgC12 at pH 9.8) at 37°C for 1-2 h and absorbances were recorded. The activity was calculated according to the suppression of IC binding, S = (1 - A / A m ) × 100 where A m is the maximum absorbance at 405 nm in the absence of hyaluronidase, and A the absorbance measured for samples. Preparation and titration of immune complexes Alkaline phosphatase-conjugated anti-HN-purified antibodies were diluted 1/500 in 0.1 M phosphate containing BSA 1 g/1 and sodium azide 0.25 g/1 at pH 7, and supplemented (v/v) with dilutions of purified HN in the same buffer. Immune complexes were incubated on an HA-coated plate for 18 h at 37°C and the phosphatase activity bound to the plate was measured as described. The solution which gave the highest absorbance at 405 nm was selected as a probe to detect HA in plates. The purification of HN and antibodies, conjugation of antibodies and the enzymoimmunological assay for HA in solution were each performed as described previously (Delpech et al., 1985).

225

Results With the conjugated antibodies used (which contained 300 ng Ig/ml) the optimum concentration of H N in immune complex solution was 300-400 n g / m l which was a little more than two molecules of antigen per antibody molecule (Fig. 1). Hyperconcentration of H N reduced antibody binding, due to competition for insolubilized HA between immune complexes and free HN. The absorbances which were measured when wells had been incubated (before the addition of immune complexes) with formate or acetate buffers (i.e., the buffers used in the hyaluronidase assay) were lower than the absorbances obtained after preincubation in wells with PBS. Therefore it was necessary to use slightly longer substrate incubation times than in the indirect enzymoimmunological assay for HA (Delpech et al., 1985). The sensitivity of the assay was not improved by the two-step procedure when H N was incubated in the coated plate before the addition of antibodies. The decrease of IC binding to HA-coated plates was due to hyaluronidase activity alone. It was observed only after incubation with samples containing either hyaluronidase or chondroitinase and at pH and salt concentrations appropriate for these enzymes. HA was not significantly desorbed by the buffers alone even after a 24 h incubation. In comparison, hyaluronidases were able to

.< 1

./.7

\.

~ o.

~

G

L

\ ~)

J 5L2

1

oT°~.

/

i! tl . . . .

<

1o

.j/

0.5

solubilize HA from wells within a few minutes (Fig. 2). Longer incubation led to a decrease of HA in the solution, which was due to further degradation of solubihzed HA. Heating of samples (e.g., 5 0 ° C for 5 min for serum hyaluronidase) or protease digestion of samples suppressed hyaluronidase activity. Proteases alone, as well as fl-glucuronidase, did not alter the IC binding to HA-coated plates. Furthermore, IC were not degraded during the incubation in wells previously treated with hyaluronidase and could bind later to undegraded HA-coated wells. Results were the same whether IC were preincubated with fully digested HA-coated wells or not preincubated, showing that the decrease in alkaline phosphatase activity was not due to degradation of a component from the IC-alkaline phosphatase complex. Using this technique hyaluronidase activity was measured under a variety of conditions. The plate HA digestion increased as a function of time. Bovine testicular activity was maximum at p H 4.5, whereas for human serum hyaluronidase the optimum pH was 3.5 (Fig. 3). Bovine testicular hyaluronidase was inactive at the concentration used at pH 3 and below, whereas inactivation was less pronounced at alkaline pH. There was an absolute

, 128

, 32

¢

| 8

, 2 tLN TitP('

Fig. 1. HN-anti-HN immune complex titration. The HN stock solution was 3 m g / l . For details see m e t h o d s section.

Min u t ~,~

6~)

Fig. 2. Effect of streptomyces hyaluronidase (1 T R U / m l ) on HA-coated plate. (O) Left scale: immune complex binding to the plate decreases after incubation of wells with hyaluronidase. (C)) Right scale: HA is solubilized by hyaluronidase. Samples were heated at 1 0 0 ° C for 15 rain to inactivate hyaluronidase before HA measurement. (A) Spontaneous release in the absence of hyaluronidase.

226

l/~}

lo0t

o"~l..-°~

/

\

i// \

!~'~0

\

o

0

\ o.1

2

3

4

~

{,

7 pH

Fig. 3. O p t i m u m pH for hyaluronidases. Incubation was for 1 h at 3 7 ° C . (e) 1 N F U / m l bovine testicular hyaluronidase. The pH was adjusted by mixing 0.1 M acetic acid with 0.1 M sodium acetate. Both solutions contained 0.05 M NaCI. (zx) H u m a n serum was diluted 1 / 2 0 in 0.1 M citrate buffer containing 0.25 M NaCI. ( O ) H u m a n serum was diluted 1 / 2 0 in 0.1 M acetate containing 0.25 M NaC1. In both cases pH adjustment was made by mixing alkaline and acidic solutions. Suppression %: see methods section.

requirement for cations since bovine testicular hyaluronidase was not active in double-distilled water or in diluted HC1 at p H 4.5 in the absence of NaC1 or of KC1 (Fig. 4). In 0.1 M acetate buffer at p H 4.5 the bovine testicular hyaluronidase was active in the absence of NaC1 and the addition of NaC1 increased the activity by only 20%. However, an excess of NaC1 inactivated the enzyme with a sharp decrease in activity when the NaC1 concentration was above 0.20 M. H u m a n serum hyaluronidase activity was abolished between p H 4 and 5 depending on the buffer used (Fig. 3). The activity was salt dependent and the m a x i m u m activity at p H 3.5 was obtained using 0.25 M NaC1 in formate or citrate buffer. The human serum activity decreased sharply with NaC1 concentrations above 0.4 M (Fig. 5). The activity was not dialysable and was

0.3 (NaCl) N

0.~~

Fig. 4. O p t i m u m NaCI concentration for the bovine testicular hyaluronidase. HA-coated plates were rinsed with deionized water before use. Incubation was for 1 h at 37 o C. Hyaluronidase was 1 N F U / m l in (e) 0.1 M acetate at p H 4.5; ( O ) diluted HC1 at pH 4.5 or (v) in distilled water at pH 7.

0.1

0.5 (NaCI)M

1

Fig. 5. O p t i m u m NaCI concentration for the h u m a n serum hyaluronidase. HA-coated plates were rinsed with deionized water before use. Incubation was for 1 h at 3 7 ° C . Before incubation h u m a n serum was dialysed against buffer without NaC1 then diluted: (O) 1 / 2 0 in 0.1 M citrate at pH 3.5; ( O ) 1 / 2 0 in 0.1 M glycine HCI at pH 3.5; and (zx) 1 / 4 0 in 0.1 M acetate at pH 3.6.

227

1oo

•

5O

12

11

lb

~ -log (NFU/nd)

Fig. 6. Sensitivity of the assay. Sensitivity was measured with bovine testicular hyaluronidase incubated for 24 h at 37 o C. A significant activity was found from 10-10 N F U / m l .

destroyed by protease digestion or by heating at 50 ° C for 5 min. The technique permitted the detection of as little as 10 -1° N F U when the enzyme was incubated for 24 h at 37 ° C in 0.1 M acetate 0.05 M NaC1 buffer supplemented with BSA 1 g/1 and TABLE I H Y A L U R O N I D A S E ACTIVITY OF BIOLOGICAL SAMPLES Hyaluronidase activity was measured by reference to bovine testicular hyaluronidase (I), in bee venom (II), pooled human serum (III), and human hepatoma cell culture HepG2 medium (IV). Incubation was for 2 h at 37 o C. Bee venom was diluted in the same buffer as bovine testicular hyaluronidase. Human serum and human hepatoma cell culture medium were diluted in 0.2 M formate 0.25 M NaCl at p H 3.5. The culture of 87x106 hepatoma cells was made in 100 ml serum free medium for 48 h. Samples

Proteins mg/ml

NFU/ml

NFU/mg

I II III IV

1 1 72 0.67

290 680 2.6 0.2

290 680 0.04 0.30

sodium azide 0.25 g/1 at pH 4.5 (Fig. 6). Using this technique we were able to assay hyaluronidase in different preparations from the leech, bee venom and streptomyces and found their parameters to be close to those of bovine testicular hyaluronidase with an optimum pH between 4 and 5 and activity which was relatively stable at alkaline pH. Conversely the hyaluronidase found in hepatoma cell culture medium was inactive above pH 4. An estimation of activities was made using bovine testicular hyaluronidase as a standard and the results are presented in Table I. A pooled human serum was estimated in three different plates (16 wells per plate) with a standard of bovine testicular enzyme at a concentration which gave a linear response. The serum was diluted 10 -3. The interassay variation was 8.3% whereas the intra-assay variations were above 10% and reached 19.8% in one plate.

Discussion

An indirect enzymoimmunological assay was developed for hyaluronidase, taking advantage of the loss of HN binding by HA-coated plates after digestion of the plates by hyaluronidase. The suppression of the plate's capacity to bind HN was due to hyaluronidase activity since: (i) it resulted in immediate solubilization of HA fragments in wells; (ii) it was pH and salt dependent; (iii) it was abrogated by sample heating or protease treatment; (iv) preincubation of wells with other enzymes, especially with proteases, did not alter the binding capacity of the plate for HN even at high concentrations. Only chondroitinase ABC was able to destroy HA, an observation which is in agreement with its known properties (Yamagata et al., 1968). Moreover, the results obtained concerning pH and salt dependence of bovine testicular hyaluronidase and of serum hyaluronidase were concordant with previous reports (Meyer et al., 1941; Bollet et al., 1963; De Salegui and Pigman, 1967; Platt, 1967; Yang et al., 1975; Yamada et al., 1977; Gold, 1982; Fiszer-Szafarz, 1984). The assay is very simple since it is reduced to three fillings of wells with (i) the test solution, (ii) immune complexes, which are very stable and can be stored at 4 ° C for months without loss of activity, and (iii)

228

alkaline phosphatase substrate. The use of immune complexes provides a simple three-step technique. A four-step technique involving incubation with test sample followed by incubation with HN then with conjugated antibodies and finally with the substrate is longer and does not improve the sensitivity. A two-step procedure using HN conjugated with alkaline phosphatase, similar to the use of enzyme-linked hyaluronate-binding proteoglycan for the detection of tissue HA (RippeUino et al., 1986) is faster but needs a larger amount of the hyaluronate binding component (unpublished). Furthermore, labelling with the specific antibody is spontaneous and the preparation of conjugated antibodies is fast and simple. An indirect enzymoimmunological assay was previously described for HA, in which incubation of HN with samples containing HA led to a decrease of HN binding to the plates as a function of the HA content of the samples. It is clear that the presence of a hyaluronidase active at pH 7 could give a false positive result for HA and lead to misinterpretation. The protease digestion and heating of samples avoid such misinterpretation, as these procedures destroy hyaluronidase activity. Results obtained with preparations containing hyaluronidase were in agreement with the values given in the literature. Inter- and intra-assay reproducibilities were good, provided that a standard was used on every plate and the incubation time long enough (2 h) to minimize differences between the opposite ends of plates. Previous results have suggested a role for hyaluronidase in cancer invasiveness (Fiszer-Szafarz and Gullino, 1970; Fiszer-Szafarz, 1981). On the other hand it has also been shown that the serum hyaluronidase activity is reduced compared to normal serum activity (Fiszer-Szafarz, 1968; Chakraborti et al., 1982) and that hyaluronidase has a protective role in experimental tumors (Pawlowski et al., 1979). The simple, fast and sensitive technique described here should be valuable in the further study of lysosomal hyaluronidase.

Acknowledgements This work was supported by the University of Rouen and the F~d6ration des Centres de Lutte contre le Cancer.

References Benchetrit, L.C., Pahuja, S.L., Gray, E.D. and Edstrom, R.D. (1977) A sensitive method for the assay of hyaluronidase activity. Anal. Biochem. 79, 431. Bertrand, P. and Delpech, B. (1985) Interaction of hyaluronectin with hyaluronic acid oligosaccharides. J. Neurochem.

45, 434. Bollet, A.J., Bonner, W.H. and Nance, J.L. (1963) The presence of hyaluronidase in various mammalian tissues. J. Biol. Chem. 238, 3522. Chakraborti, A.S., Basu, A. and Mitra, S. (1982) Hyaluronidase activity in murine leukemia & lymphoma. Ind. J. Exp. Biol. 20, 423. Delpech, B. and Halavent, C. (1981) Characterization and purification from human brain of a hyaluronic acid binding glycoprotein, hyaluronectin. J. Neurochem. 36, 855. Delpech, B., Bertrand, P. and Maingonnat, C. (1985) Immunoenzymoassay of the hyaluronic acid-hyaluronectin interaction: application to the detection of hyaluronic acid in serum of normal subjects and cancer patients. Anal. Biochem. 149, 555. Delpech, B., Bertrand, P., Hermelin, B., Delpech, A., Girard, N., Halkin, E. and Chauzy, C. (1986) Hyaluronectin. In: J. Labat-Robert, R. Timpl and U Robert (Eds.), Frontiers of Matrix Biology, vol. 11, Karger, Basel, p. 78. De Salegui, M. and Pigman, W. (1967) The existence of an acid-active hyaluronidase in serum. Arch. Biochem. Biophys. 120, 60. Dorfman, A. (1948) The kinetics of the enzymatic hydrolysis of hyaluronic acid. J. Biol. Chem. 172, 377. Dorfman, A. and Ott, M.L. (1948) A turbidimetric method for the assay of hyaluronidase. J. Biol. Chem., 172, 367. Fiszer-Szafarz, B. (1968) Demonstration of a new hyaluronidase inhibitor in serum of cancer patients. Proc. Soc. Exp. Biol. Med. 129, 300. Fiszer-Szafarz, B. (1981) Acid hydrolases and tumor invasion. Relationship between intracellular distribution of hyaluronidase, cathepsin D and acid phosphatase, and cell proliferation, in rat liver and diethylnitrosamine-induced hepatoma. Biol. Cell 42, 97. Fiszer-Szafarz, B. (1984) Hyaiuronidase polymorphism detected by polyacrylamide gel electrophoresis. Application to hyaluronidases from bacteria, slime molds, bee and snake venoms, bovine testes, rat liver lysosomes, and human serum. Anal. Biochem. 143, 76. Fiszer-Szafarz, B. and Gullino, P.M. (1970) Hyaluronidase activity of normal and neoplastic interstitial fluid. Proc.

Soc. Exp. Biol. Med. 133, 805. Girard, N., Bertrand, P., Delpech, A. and Delpech, B. (1984)

229 Caract6risation tissulaire de l'acide hyaluronique dans le cervelet par l'hyaluronectine. C.R. S6ances Acad. Sci. Ser. III 298, 325. Girard, N., Delpech, A. and Delpech, B. (1986) Characterization of hyaluronic acid on tissue sections with hyaluronectin. J. Histochem. Cytochem. 34, 539. Gold, E.W. (1982) Purification and properties of hyaluronidase from human liver. Biochem. J. 205, 69. Meyer, K., Chaffee, E., Hobby, G.L. and Dawson, M.H. (1941) Hyaluronidases of bacterial and animal origin. J. Exp. Med. 73, 309. Pawlowski, A., Haberman, H.F. and Aravindakshan Menon, I. (1979) The effects of hyaluronidase upon tumor formation in BALB/c mice painted with 7,12-dimethylbenz(a)anthracene. Int. J. Cancer 23, 105. Platt, D. (1967) Glycosaminoglycano-hydrolasen in menschlichen Serum. I. Mitteilung. Blut 15, 274. Richman, P.G. and Baer, H. (1980) A convenient plate assay for the quantitation of hyaluronidase in hymenoptera venoms. Anal. Biochem. 109, 376.

Ripellino, J.A., Klinger, M.M., Margolis, R.U. and Margolis, R.K. (1985) The hyaluronic acid binding region as a specific probe for the localization of hyaluronic acid in tissue sections. Application to chick embryo and rat brain. J. Histochem. Cytochem. 33, 1060. Tolksdorf, S., McCready, M.H., Roy McCullagh, D. and Schwenk, E. (1949) The turbidimetric assay of hyaluronidase. J. Lab. Clin. Med. 34, 74. Yamada, M., Hasegawa, E. and Kanamori, M. (1977) Purification of hyaluronidase from human placenta. J. Biochem., 81,485. Yamagata, T., Saito, H., Habuchi, O. and Suzuki, S. (1968) Purification and properties of bacterial chondroitinases and chondrosulfatases. J. Biol. Chem., 243, 1523. Yang, C.H. and Srivastava, P.N. (1975) Purification and properties of hyaluronidase from bull sperm. J. Biol. Chem., 250, 79.