Journal of Clinical Laboratory Analysis 5 1 62-167 (1991)

Purification of Soluble Human T Lymphocyte Receptors for Sheep Erythrocytes

Eiko N. Itano,' Mario A. Ono,' Mari Sumigawa,' leda M. Longo,2 Nayla C. Moura,* and Nelson F. Mendes2

' Departamento de Patologia Geral, Centro de Ciencias Biologicas, Universidade Estadual de Londrina, Parana, and 'Disciplina de Imunologia, Escola Paulista de Medicina, Sdo Paulo, Brazil

Using a polyclonal heterologous anti-solu- ble E-receptor serum, we identified mole- cules of molecular weight circa 58,000 and 150,000.

The soluble receptor molecule with molec- ular weight of approximately 58,000 (Rsl) was initially purified from supernatant of heated lymphocytes through chromatography on Sephadex G-200 and/or DEAE-cellulose. The soluble receptor molecule with molecu- lar weight of approximately 150,000 (Rs2) is detected at high levels in the serum of pa-

tients with cancer and uremia. Rsl and Rs2 present in serum from cancer patients were purified by chromatography on Sephadex G-200 and by affinity chromatography using anti-Rsl IgG.

13' I-labelled supernatant of heated lym- phocytes binds to sheep erythrocytes and the elution and analysis of the molecules ad- sorbed showed bands of molecular weights approximately 58,000 and 150,000, confirm- ing the receptor activity of these molecules.

Key words: E-receptors, CD2, chromatography

INTRODUCTION

Some of the fundamental studies on the human T lympho- cyte receptor for sheep erythrocytes, describing the phenom- enon of rosette formation and the characterization and quantitation of this receptor in a soluble form, were devel- oped in this laboratory (1 -7).

Human T lymphocyte membrane receptors for sheep eryth- rocytes (E), known as CD2, detected by monoclonal antibodies, have molecular weights from 40,000 to 58,000 (8-12). This heterogeneity in molecular weight is due to variation in the side branches of carbohydrates ( 1 2). This structure contains at least three different epitopes, denoted T1 1 . 1 , which is the E binding site present in all T lymphocytes and thymocytes; TI 1.2, the tissue distribution of which is identical to that of TI 1 . 1 , and T. 11.3, which is present only in activated T lym- phocytes ( I 1). The CD2 antigen has been purified by affinity chromatography, using monoclonal antibodies. Electropho- resis of the purified and '251-labelled material has demon- strated the presence of several bands. Plunkett et al. in 1986 obtained bands of molecular weights of 50,000,54,000, and 58,000, and after prolonged autoradiography, bands of mo- lecular weights of 32,000,36,000, and 44,000 (13).

Brown et al., in 1987, obtained more evident bands of mo- lecular weights 52,000, 54,000, and 56,000 and after more prolonged autoradiography bands of molecular weights of ap- proximately 22,000 and 100,000 (12).

Even though many studies are available on T lymphocyte membrane CD2, little is known about soluble E receptors

0 1991 Wiley-Liss, Inc.

present in human serum ( 5 ) . In this study, anti-soluble re- ceptor (anti-Rs) serum was obtained by immunizing sheep with autologous E coated with Rs (14). Previous studies from our group have demonstrated that Rs levels are elevated in cancer, uremia and leprosy patients (7). This anti-Rs serum recognizes molecules of molecular weights of approximately 58,000 and 150,000, the latter being detected in high levels in the serum of patients with cancer and uremia (15). The purpose of this study was to purify these soluble E-receptors.

MATERIALS AND METHODS

Supernatant of Heated Peripheral Lymphocytes (SHPL) and of Heated Spleen Lymphocytes (SHSL) (5)

A peripheral lymphocyte suspension was obtained by treat- ing whole blood with powdered iron and by centrifuging in Ficoll-Hypaque solution of 1.076 density. The final cell sus- pension was adjusted to 1 X lo7 lymphocytes/ml in Hanks', then heated for 1 h in a water bath at 45"C, and centrifuged, yielding SHPL. To obtain SHSL, spleens obtained from ca- daveric donors for organ transplants were decapsulated and

Received August 17, 1990; accepted August 23, 1990.

Address reprint requests to Eiko N. Itano, Departamento de Patologia Geral, CCB, Universidade Estadual de Londrina, 8605 1 Londrina, Parana, Brad .

Research supported by CNPq, CONCITEC/PR, and CPG/UEL.

Purification of Soluble E-Receptors 163

anti-Rsl rabbit IgG. After several washes, elution was per- formed with 0.05 M sodium citrate buffer, pH 2.8. Fractions of 2.0 ml were collected into tubes containing 60 pl of 1 M Tris-HCI buffer, pH 8.5, and analysed with a spectrophoto- meter at 280 nm. The fractions with the highest absorbances at 280 nm were mixed, and the resulting pool was dialysed in 0.15 M PBS, concentrated, and analysed by PAGE.

minced in Hanks’ solution. The cell suspension obtained by filtration through a sieve and through gauze was processed as described for SHPL.

Serum of Cancer Patients

Serum samples provided by the clinical laboratory of the Cancer Institute of Londrina (Parana, Brazil) were selected in terms of high Rs levels without considering histological type, stage or treatment. A pool of 5 sera was used in frac- tionation experiments using chromatography on Sephadex G- 200. The sera were divided into aliquots and stored at - 20°C.

Chromatography on Sephadex G-200

Three milliliter samples from the serum pool from cancer patients and SHSL were applied to a Sephadex G-200 col- umn (2 x 60 cm) balanced with 0.15 M Tris-NaC1, pH 8.0. Fractions (2.5 ml) were collected with an automatic fraction collector, read in a spectrophotometer at 280 nm, and sub- mitted to Ouchterlony’s double immunodiffusion in gel.

Chromatography on DEAE Cellulose

Approximately 1 .5 ml of the SHSL fraction (=0.6 mg/ml) were dialysed against 0.01 M ammonium acetate buffer, pH 8.0, and then applied to a DEAE cellulose column (2 X 50 cm). Elution was performed with a 0.01 to 3.0 M gradient of the ammonium acetate buffer, pH 8.0, followed by reading in a spectrophotometer at 280 nm. The fractions were dialysed, lyophilized, dissolved in 0.5 ml phosphate buffered solution (PBS), and analyzed by polyacrylamide gel electro- phoresis and by gel diffusion with anti-Rs serum.

Double lmmunodiffusion in Gel (Ouchterlony)

The procedure was carried out on a glass slide covered with 1% agar. The samples were applied to wells in the gel and maintained in a humid chamber at room temperature for 24-48 h. The gel was washed, dried, and stained.

Preparation of Anti-Rsl Rabbit IgG

Anti-Rsl serum was obtained by inoculating a rabbit sub- cutaneously with 0.4 ml of the Rsl fraction of SHSLG0.6 mg/ml). The first two immunizations were performed with Freund’s complete adjuvant (v/v), and the third was performed with incomplete Freund’s adjuvant (viv) at 12 day intervals. After test bleeding, the hyperimmune serum was passed through a Sepharose-protein A column.

Affinity Chromatography With Anti-Rsl Rabbit IgG

Anti-Rsl rabbit IgG was coupled to Sepharose 4B CnBr according to the Pharmacia manual. Approximately 0.8 ml of the Rsl and Rs2 fractions of serum from a cancer patient was applied separately to a column packed with 3.5 ml Sepharose-

lmmunoprecipitation

Samples labelled with I3’I by the chloramine T method were incubated with an equal volume (100 11.1) of pre- immune serum at 37°C for 30 min, transferred to a tube con- taining a 100 pl sediment of a 10% Staphylococcus aureus (LF strain) suspension, and incubated at 28°C for 15 min. After centrifugation, the supernatant was divided into two portions. Phosphate buffered saline (PBS) (v/v) was added to one, and anti-Rs serum (v/v) to the other. Both were incu- bated again at 28°C with 200 p1 10% Staphylococcus aureus for 15 min. After several washes with 0.05% NET buffer, the cell pellet was treated once with 130 pl sample buffer, boiled, centrifuged, and submitted to polyacrylarnide gel electro- phoresis.

Polyacrylamide Gel Electrophoresis Without SDS (PAGE)

Samples of 30-50 ~l (a mixture of 0.1 ml sample, 0.5 ml glycerol, and 0.05 bromophenol blue) were submitted to poly- acrylamide gel electrophoresis in tubes, applying 3.0 mA per tube. After the electrophoretic run, the gel was stained for protein with 2.5% Coomassie blue and for glycoprotein with the Schiff reagent ( 16).

Polyacrylamide Gel Electrophoresis With SDS (SDS-PAGE)

Labelled and immunoprecipitated samples (SO 11.1) and 10 p1 of molecular weight protein standards (bovine albu- min, ovalbumin, carbonic anhydrase, phosphorylase b, and P-galactosidase) were run in individual wells on a plate with a 7.5- 15% polyacrylamide gradient. The electrophore- tic run was performed with Tris-glycine buffer, pH 8.2, at 125 V. After staining with Coomassie blue, the gel was vacuum dried and exposed to X-ray film, followed by devel- oping and fixing.

Preparation of E-Rs Eluates

Sheep erythrocytes (E) were first fixed on a flat-bottomed plate by the Cobbold and Waldmann method (1981) with some modifications. Fifty microliters of poly-L-lysine (mo- lecular weight of 70,000- 150,000, Sigma Chemical Co., St. Louis, MO) were pipetted into each well and incubated at room temperature for 1 h. After washing with PBS, 50 p1 of 1% E were pipetted and incubated at 4°C for 30 min. After

164 ltano et al.

2 0-

L5-

10-

0.5-

RESULTS

Purification of SHSL Rsl : Preparation and Analysis of Anti-Rsl Serum

The Rsl fraction was purified by SHSL fractionation on Sephadex G- 200 (Fig. 1) alone or combined with DEAE- cellulose (Fig. 2). Considering that, in both cases, PAGE anal- ysis of the fractions showed only one band stained for protein with electrophoretic migration similar to that of the bromo- phenol blue stain (Fig. 3), the Sephadex G-200 fraction was used to obtain anti-Rs 1 serum. The anti-Rs 1 thus obtained was submitted to immunodiffusion in gel against SHL and compared with anti-Rs serum, showing a line of total identity.

In order to determine whether anti-Rs 1 serum was recog- nizing the Rsl fraction of molecular weight 58,000, labelled SHPL was immunoprecipitated with anti-Rs serum and anti- Rs 1 serum and submitted to SDS-PAGE. Radioactive bands of molecular weight of 58,000 and higher than 150,000 were observed in both immunoprecipitates (Fig. 4).

0 . D 4

n

- Fractions

Fig. 1. Spectrophotometric profile at 280 nm of fractions obtained from I .0 ml of tenfold concentrated supernatant of heated spleen lymphocytes by chromatography on Sephadex (3-200. The fractions (2.5 ml) were concen- trated fivefold and analyzed by precipitation in gel (Ouchterlony) using sheep anti-soluble T lymphocyte E receptor serum, and the positive fractions are denoted RsI and Rs2. O.D., optical density at 280 nm.

removal of the excessive E with PBS, blockade was performed with PBS containing 2% normal human serum (NHS) ab- sorbed with E. After washing with PBS containing 0.2% NHS absorbed with E, 20 pl of ”‘I-labelled sample plus 20 p1 Hanks’ solution were pipetted. Twenty microliters of the sam- ple plus 20 ~1 of anti-Rs serum were pipetted as control. After incubation at 4°C for 4 h, each well was washed twice with PBS containing 0.2% NHS absorbed with E. After removal of PBS by inversion, 100 p1 Triton X-100 were pipetted and the eluate containing the receptor was concentrated with ice- cold ethanol (viv), incubated at - 20°C for 18 h, and centri- fuged. The precipitate was diluted with sample buffer and analyzed by SDS-PAGE.

0. D 1 1 6 M

Fig. 2. Spectrophotometric profile at 280 nm of fractions obtained by chro- matography on DEAE-cellulose by refractionation of the Rsl fraction re- sulting from fractionation of the supernatant of heated spleen lymphocytes on Sephadex (3-200. The fractions ( I .5 ml) were pooled, forming new frac- tions that were designated with the letters A to H. These fractions were con- centrated and analyzed by precipitation in gel (Ouchterlony) using sheep anti-receptor serum. resulting in only one positive fraction, which is marked with E. O.D., optical density at 280 nm.

Fig. 3. Polyacrylamide gel electrophoresis carried out in a tube without SDS, with Coomassie blue staining. B: Rsl fraction interacting with anti- R, serum, obtained by fractionation of the supernatant of heated spleen lym- phocytes on Sephadex G-200. C: Rsl fraction interacting with anti-R, serum, obtained by fractionation of the supernatant of heated spleen lym- phocytes on Sephadex (3-200, followed by refractionation on DEAE cellu- lose. 0, origin; A.B. , bromophenol blue.

Purification of Soluble E-Receptors 165

0. D 1 150 Kd

4 69Kd

I

Fig. 4. Result of SDS-PAGE autoradiography of immunoprecipitates of the supernatant of heated lymphocytes (SHPL= SALP) with anti-Rs rabbit serum (anti-R,I) and anti-Rs sheep serum (anti-R,). The immunoprecipitates were obtained from "'I-labelled samples previously incubated with pre- immune serum (SHPL-PBS). The samples were then transferred to a tube containing Sruphylocuccus uureus-A sediment. After incubation, the super- natants were divided into two portions. PBS was added to one (SHPL-PBS) and anti-R, rabbit serum (anti-R, SHPL) or anti-Rs sheep serum to the other. After incubation with Staphylococcus aurcus the cell pellets were treated with sample buffer containing a reducing agent and the supernatants obtained were submitted to electrophoresis. MW, unlabelled molecular weight standards.

Purification of Rsl and Rs2 From Human Serum

Chromatography on Sephadex (3-200 of a pool of sera from cancer patients yielded two positive fractions capable of interacting with anti-Rs serum: one coincided with albumin ( R s l ) and the other had a molecular weight higher than 150,000 (preceding the IgG peak) (Fig. 5).

The Rsl and Rs2 fractions from the serum of cancer pa- tients were refractionated on Sepharose-anti-Rs l IgG. Poly- acrylamide gel electrophoresis of these fractions resulted in a stronger band of rapid migration similar to that of the bro- mophenol blue stain. An equivalent result was obtained with the Rs2 fraction (Fig. 6).

Analysis of E-Rs Eluates

SDS-PAGE of the E eluate resulted in radioactive bands of molecular weight of 58,000 and higher than 150,000 with the Rs2 fraction of SHSL, and of 58,000 with the Rsl frac- tion of SHSL (Fig. 7).

Fig. 5. Spectrophotometric profile at 280 nm of fractions obtained by chm- matography on Sephadex G-200 of a pool of sera from cancer patients. The fractions were analyzed by precipitation in gel (Ouchterlony) using sheep anti-soluble T lymphocyte E receptor, and the positive fractions are denoted as RS1 and Rs2. O.D. , optical density at 280 nm.

DISCUSSION

The CD2 structure detected on the surface of human T lym- phocytes by different monoclonal antibodies was initially thought to be an E receptor or an antigen closely related to the receptor since these antibodies can block rosette forma- tion with E and identify T lymphocytes. In addition, Kamoun et al., in 1981, reported simultaneous movement of the CD2 antigen and the E receptor when lymphocytes were treated with anti-CD2 monoclonal antibody (anti-9.6).

The most concrete evidence that the CD2 structure is an E receptor was obtained after CD2 purification. The direct bind- ing of purified CD2 to the E surface was demonstrated by Selvaraj et al. (1987) by '251 labelling of purified CD2 (17). This binding was inhibited by treating E with purified CD2. After CD2 purification, cDNA for CD2 was obtained. When this cDNA was transferred to COS-1 cells, these cells, which expressed cDNA for CD2, were found to be able to bind to E (18). These data demonstrate that the CD2 antigen is the struc- ture that binds T lymphocytes to E. In this study, anti-Rs heterologous serum was obtained by the method of Mendes et al. (14) by immunizing a sheep with autologous E sensi- tized with Rs. It was then demonstrated that this serum can inhibit rosette formation, is cytotoxic to T lymphocytes, iden- tifies T lymphocytes by indirect immunofluorescence and stimulates blastogenesis of lymphocytes in culture. The se- rum used by us showed a line fully identical to that of the anti-Rs serum obtained by Mendes et al. (14) when analyzed with supernatant from heated lymphocytes (SHL). This anti- Rs serum recognizes molecules of molecular weight of 58,000 (Rs l ) and of molecular weight higher than 150,000 (Rs2) (15). In this study the purification of these receptors was started. The preparation of Rsl from the supernatant of heated lymphocytes was easily purified by filtration on Sephadex (3-200 alone, or in combination with DEAE-

166 ltano et al.

Fig. 6. Polyacrylamide gel electrophoresis carried out in a tube without SDS, with staining for glycoprotein. A: Rsl fraction of serum from a cancer patient, obtained by chromatography on Sephadex G-200 plus affinity chroma- tography (anti-RsI rabbit I&). B: Rs2 fraction of serum from a cancer patient obtained by chromatography on Sephadex G-200 plus affinity chroma- tography (anti-Rsl rabbit IgG). 0, origin; A.B., bromophenol blue.

cellulose, since few proteins exist in this preparation. In con- trast, Rs2 could not be purified from SHL since it was present in low concentration. Serum Rsl and Rs2 were initially pu- rified by chromtography on Sephadex G-200, and then by affinity chromatography with anti-Rsl rabbit IgG, consider- ing that this serum recognizes antigenic determinants in common with anti-Rs serum. When submitted to polyacryl- amide gel electrophoresis without SDS, both the purified Rsl of SHL and the Rsl and Rs2 present in serum showed only one band of electrophoretic migration similar to that of bro- mophenol blue. Considering that Rs2 has the electrophoretic migration of alpha2-gammaglobulin and Rs 1 has migration close to that of albumin (15), the presence of only one band of fast migration in the purified Rs2 fraction may have been due to Rs2 dissociation at low pH.

It is possible that the Rsl detected by anti-Rs serum is CD2, since in addition to the similarity in molecular weight, an increase in Rs levels (7) and in CD2 levels (19) has been reported to occur in the serum of cancer patients. Further- more, analysis of sensitized E eluates confirms the binding of Rsl to E. The molecule Rs2, which has not been described

Fig. 7. SDS-PAGE autoradiography of E eluates sensitized with 1) Rs2 fraction of the supernatant of heated lymphocytes; 2) Rs2 fraction of the supernatant of heated lymphocytes treated with anti-Rs serum; 3) Rsl frac- tion of the supernatant of heated lymphocytes treated with anti-Rs serum; 4) R,l fraction of the supernatant of heated lymphocytes.

in the literature, may have receptor activity, and the lack of detection of this molecular form by other investigators may be explained by its low concentration in normal preparations or sera and also to the restricted recognition by monoclonal antibodies, while our polyclonal serum reacts with multiple epitopes.

ACKNOWLEDGMENTS

We thank Mylene M. Yokoyama and Helena Kaminami for helping with the laboratory work. We are grateful to Dr. Altair J . Mocelin and his group for supplying human spleens from cadaveric donors, to the Cancer Institute of Londrina, and to Dr. Caio M. F. Mendes for supplying Stuphytococcus uureus LF7 strain.

REFERENCES

Lay WH, Mendes NF, Bianco C, Nussenzweig V: Binding of sheep red blood cells to a large population of human lymphocytes. Nature 23053 1-532, 197 1. Silveira NPA, Mendes NF, Tolnai MEA: Tissue localization of two popu- lations of human lymphocytes distinguished by membrane receptors. J lmmunol 108:1456-1460, 1972. Mendes NF, Tolnai MEA, Silveira NPA, Metzgar RS, Gilbertsen R: Technical aspects of rosette test used to detect human complement re- ceptor (B) and sheep erythrocyte binding (T) lymphocytes. J fminunol

Mendes NF, Miki SS, Peixinho Z F Combined detection of human T I Il:860-867, 1973.

Purification of Soluble E-Receptors 167

12. Brown MH, Krissansen GW, Totty NF, Sewell WA, Crumpton MJ: Puri- fication and N-terminal amino acid sequence of the human T lympho- cyte CD2 (TI I ) surface antigen. EurJfmmunol 17: 15-20, 1987.

13. Plunkett ML, Springer TA: Purification and characterization of the lym- phocyte function-associated-2 (LFA-2) molecule. J fmmunol 136( 1 I ): 4181-4187, 1986.

14. Mendes NF, Bernd LAG, Cintra HHA, Mendes CFM, Praga C , Moura NC: Human T lymphocyte receptor for sheep erythrocytes: characteri- zation of a specific antiserum and its application in the detection and quan- titation of the receptor in a soluble form. Cell fmmunol72: 143- 150, 1982.

1.5. Itano EN: Formas moleculares do receptor para eritrocitos de carneiro em linficitos T, soro e saliva. PhD thesis. Escola Paulista de Medicina, SHo Paulo, Brazil, 1989.

16. Zacharius RM, Zell TE, Morrison JH, Woodlock JJ: Glycoprotein stain- ing following electrophoresis on acrylamide gels. Anal Biochem 30: 148- 152, 1969.

17. Selvaraj P, Dustin ML, Mitnacht R, Hiinig T, Springer TA, Plunkett ML: Rosetting of human T lymphocytes with sheep and human eryth- rocytes. J fmmunol 138(8):2690-2695, 1987.

18. Sayre H, Chang HC, Hussey RE, Brown NR, Richardson NE, Spagnoli G , Clayton LK, Reinherz EL: Molecular cloning and expression of T I 1 cDNAs reveal a receptor-like structure on human T lymphocytes. Proc NutlAcadSci USA 84(9):2941-2945, 1987.

19. Oh SK, Lapenson D, Morgan Jr AC: Quantification of soluble E- receptor in the serum of patients with various diseases and its accom- panying in vitro immunosuppression in neoplasia. Scand J fmmunol 225-60 , 1985.

and B lymphocytes by rosette formation with sheep erythrocytes and zymosan - C3 complexes. Jfmmunoll13:531-536. 1974.

5 . Mendes NF, Saraiva PJ, Santos OBO: Restorative effect of normal hu- man serum, transfer factor and thymosin on the ability of heated human lymphocytes to form rosettes with sheep erythrocytes. Cell fmmunol

6. Mendes N F Immunological identification of human lymphoid cell popu- lations. Lymphology 10%-93, 1977.

7. Moura NC, Longo IM, Bernd LAG, Mendes NF: Quantitation of the soluble receptor of human lymphocytes for sheep erythrocytes by immunoelectrodiffusion in the serum of patients with cancer, uremia and leprosy. Experienria 39:306-308, 1983.

8. Howard FD, Ledbetter JA, Wong J , Bieber CP, Stinson EB, Herzen- berg LA: Human T lymphocyte differentiation marker defined by mono- clonal antibodies that block E-rosette formation. J fmmunol 126(6): 2117-2122, 1981.

9. Kamoun M. Martin PJ, Hansen JA, Brown MA, Siadak AW. Nowinski RC: Identification of a human T lymphocyte surface protein associated with the E-rosette receptor. J E x p Med 153:207-212, 1981.

10. Verbi W, Greaves MF, Schneider C , Koubek K, Janossy G, Stein H, Kung P, Goldstein G: Monoclonal antibodies OKTl I and OKTl IA have pan-T reactivity and block sheep erythrocyte “receptors”. Eur J fmmunol12:81-86, 1982.

1 1 . Meuer SC, Hussey RE, Fabbi M, Fox D. Acuto 0, Fitzgerald KA, Hodgdon JC, Protentis JP, Schlossman SF, Reinherz EL: An alterna- tive pathway of T cell activation: a functional role for the 50Kd TI I sheep erythrocyte receptor protein. Cell 36:897-906, 1984.

17:560-566, 1975.