Journal of Clinical Laboratory Analysis 5:114-120 (1991)

Molecular Forms of Soluble Human T Lymphocyte Receptor for Sheep Erythrocytes in Serum and Saliva

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, Brazil; 2Disciplina de Imunologia, Escola Paulista de Medicina, SBo Paulo, Brazil

Using a specific serum anti-soluble T lym- phocytes receptor for sheep erythrocytes (E) and SDS-PAGE, we detected radioactive bands of molecular weight 58,000 in immu- noprecipitates of supernatant of heated hu- man lymphocytes (SHL), in the supernatant of PHA stimulated lymphocyte cultures (SLC), normal human serum (NHS), and serum from cancer and uremia patients, labelled with 1311. By Sephadex G-200 chromatography, in addi- tion to this fraction, we detected molecules of molecular weight higher than 150,000 which interact with the anti-soluble receptor serum (anti-Rs), in serum from cancer and uremia patients. These molecules were de- tected in NHS or SHL after concentration or by prolonged exposure of SDS-PAGE with some labelled and immunoprecipitated SHL samples.

The soluble receptors of molecular weights

58,000 (Rsl) and more than 150,000 (Rs2) were fully identical when analyzed by immu- nodiffusion with anti-Rs serum. When sub- mitted to immunoelectrophoresis, Rsl showed electrophoretic migration similar to that of albumin, while Rs2 showed a pattern close to that of alpha2-globulin. However, Rs2 did not show antigenic relationship with IgM and was not an immune complex with IgG. Even though the presence of Rs in human saliva has not yet been reported, molecules that interact with anti-Rs serum have been de- tected in human saliva and are fully identical to molecules found in supernatant of heated human T lymphocytes and NHS. The Rs molecules present in human saliva have a molecular weight and electrophoretic migra- tion similar to those of Rsl from SLC and from human serum and have no antigenic relationship with human albumin.

Kev words: E-receDtors. T IvmDhocvtes. CD2. radioactive bands

INTRODUCTION

Human T lymphocytes spontaneously interact with sheep erythrocytes (E) through a T lymphocyte membrane receptor, forming rosettes (1 -3). The molecules involved are CD2 (T11) (4) and the E surface antigen TI ITS (5 ) . The natural CD2 ligand in man is the LFA-3 antigen (6), which is present in all cell lines capable of acting as accessory cells (7). With the aid of monoclonal antibodies, CD2 antigens have been biochemically defined as polypeptides of molecular weights from 40,000 to 58,000 (4,8- 11); this heterogeneity in molecu- lar weight is due to variations in the side branches of carbo- hydrates ( 1 1). Even though many studies have been conducted on the role of CD2 at the T lymphocyte membrane level, lit- tle is known about the soluble receptor present in human serum (12,13). Mendes et al. (14) after immunizing sheep with autol- O ~ O U S erythrocytes sensitized with soluble receptors (Rs) pres- ent in supernatant of heated lymphocytes, obtained polyclonal antibodies specific to the E receptor. When these anti-Rs sera were used in electro-immunodiffusion (“rocket electropho- resis”), an increase in Rs levels was detected in the serum of patients with cancer, uremia, and leprosy (1.5). Increased

0 1991 Wiley-Liss, Inc.

CD2 levels have also been observed by other authors, in the serum of patients with cancer, using anti-CD2 monoclonal antibodies (1 3). The study of these soluble receptors is impor- tant if we consider that they have an immunosuppressive effect as evaluated by inhibition of blastogenic response (16). In the present study, we have characterized the structures rec- ognized by anti-Rs polyclonal heterologous serum which are present in heated solutes of human lymphocytes, in normal human serum and in serum from cancer and uremia patients with high Rs levels. Even though the presence of soluble E receptors in human saliva had not yet been reported before, molecules interacting with anti-Rs serum have been detected in human saliva and have been shown by immunodiffusion

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

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

Work supported by CNPq, CONCITECIPR, and CPGiUEL

Molecular Forms of E Receptor 11 5

Chromatography on Sephadex G-200

Three-milliliter samples of the NHS pool, of the serum pool from patients with cancer and uremia, of SHSL, and of the normal human saliva pool were applied to Sephadex (3-200. The column (2 X 60 cm) was equilibrated with 0.15 M Tris-NaCI buffer, pH 8.0.

in gel to be identical to components of supernatant of heated lymphocytes and normal human serum. Thus, another pur- pose of the present study was to compare a possible soluble E receptor present in normal human saliva with soluble E receptors from other sources.

MATERIALS AND METHODS

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

A peripheral lymphocyte suspension obtained by treating whole blood with powdered iron and by centrifugation in Ficoll-Hypaque (density 1.076) solution was adjusted to the concentration of 1 x lo7 lymphocytesiml in Hanks' solution, heated for 1 hour in a water bath at 45°C and centrifuged, yielding SHPL. To obtain SHSL, spleens obtained from cadav- eric donors for organ transplantation were decapsulated and minced in Hanks' solution. The cell suspension obtained by filtration through a sieve and gauze was processed as described for SHPL.

Human Serum

Normal human serum was obtained from peripheral blood of normal donors, and serum from cancer patients was obtained from the Cancer Institute of Londrina, Brazil. Cancer patients were selected on the basis of soluble receptor (Rs) levels in serum, regardless of histological type, stage, or treatment. Serum samples from uremic patients with high Rs levels were obtained from the Evangelical Hospital, Londrina, Brazil. Pools of 5 sera were used in the fractionation experiments on Sephadex G-200, and individual serum samples in the remain- ing experiments.

Preparation of Anti-Rs Sheep Serum

The procedure was that recommended by Mendes et al. (14) with some modifications. Autologous E (0.5%) were incu- bated with SHPL ( 1 .O ml, viv) for 18 hours at 4"C, washed with ice-cold Hanks' solution, and resuspended in 0.5 ml Hanks' solution. Sensitized E emulsified in complete Freund's adjuvant (viv) were innoculated subcutaneously into a sheep at several sites, with a 30-day interval between inoculations. A third subcutaneous inoculation was then performed with incomplete Freund's adjuvant and the animal was submitted to test bleeding 15 days thereafter.

Immunization of Rabbits With Normal Human Serum and Human Albumin

Normal human serum (NHS) and albumin obtained by NHS precipitation with ammonium sulfate were inoculated into rab- bits subcutaneously (albumin fraction) or intramuscularly (NHS). The first and second inoculum were prepared in com- plete Freund's adjuvant for the albumin fraction and alumen precipitated for NHS.

Double lmmunodiffusion (Ouchterlony)

The procedure was carried out on glass slides (2.5 X 7.5 or 5.0 x 7.5 cm) covered with a 3.0 mm high layer of 1% agar in 0.85% NaCl containing 0.02% sodium azide.

lmmunoelectrophoresis

The procedure was carried out on glass slides covered with 1% agarose in hippurate buffer. Six slides (2.5 X 7.5 cm) were submitted to electrophoresis using hippurate buffer, pH 8.8 (17) at 150 volts and 30 mA. After the run, anti-Rs serum was added to the channel and the preparation was incubated for 24-48 hours in a humid chamber.

lmmunoprecipitation

The lymphocyte suspension (200 pl, at a concentration of 1 x lo7 cellsiml), 200 pl of human serum, SHL, or serum fraction samples were labelled with '"1 by the chloramine method. The lymphocyte suspensions were washed, resus- pended in 200 pl Hanks' solution, heated at 45°C for 1 hour, and centrifuged. The samples were dialyzed and incubated with an equal volume (100 pl) of preimmune sheep serum at 37°C for 30 minutes, transferred to a tube with a 100 pl sedi- ment of a 10% Sruphylococcus aureus (LF strain) suspen- sion, incubated at 28°C for 15 minutes, and centrifuged. The supernatant was divided into two portions: 0. I5 M PBS (viv) was added to one portion, and anti-Rs serum (viv) to the other. Both portions were incubated again with Staphylococcus uureus, and, after washing by centrifugation, the cell pellet was treated with I30 pl of sample buffer containing mercap- toethanol. This cell suspension was boiled for 3 minutes, centrifuged, and the supernatant obtained was submitted to polyacrylamide gel electrophoresis.

Electrophoresis on a Polyacrylamide Gel Gradient With SDS (SDS-PAGE)

Electrophoresis was carried out on a linear gradient of 7.5% to 15% polyacrylamide gel (8.5 cm high and 1 mm thick) containing individual wells in the stacking gel. Labelled and inimunoprecipitated samples ( S O pl) and 10 pl of unlabelled protein molecular weight standards (bovine albumin, oval- bumin, carbonic anhydrase, phosphorylase B, and a-galac- tosidase) were run in individual wells. The electrophoretic run was performed on Tris-glycine buffer, pH 8 . 2 , at 125 V. The gel was then stained with Coomassie blue, destained, dried, and exposed to X-ray film.

116 ltano et at.

Chromatography on Sephadex G-200

Analysis of the chromatographic eluates of serum from patients with cancer or uremia showed the presence of two fractions capable of interacting with anti-Rs serum: one, which was close to albumin, was denoted Rs 1; and the other, which corresponded to the region with molecular weight higher than 150,000 (preceding the IgG peak), was called Rs2. The Rs2 fraction was detected in NHS after tenfold concentration of the sample. A positive fraction of elution similar to that of serum Rsl was detected in SHL, and a positive fraction was detected at the end of the exclusion volume only in one Sam- ple after concentration (Fig. 2). a

Fig. 1. Determination of the molecular weight of soluble receptor by SDS- PAGE autoradiography in iininunoprecipi~ates from supernatant of heated lymphocytes a) obtained from labelled lymphocytes (1 and 2), SHL (3 and 4), supernatant of lymphocyte cultures stimulated with PHA (5 and 6); b) NHS (1 and 2), serum from a patient with cancer (3 and 4), and serum from a uremic patient (5 and 6). labelled with ' " I . The immunoprecipitates were obtained from samples incubated with preimmune sheep serum. The samples were then transferred to a tube containing Sruphdococcus (iurrus-A sediment. After incubation, the supernatant was divided into two portions. PBS was added to one portion ( I , 3, and 5), and anti-Rs sheep serum to the other and both were again incubated with a Sruphy/ococcus uureus-A suspension. The cell pellets were treated with sample buffer containing a reducing agent and the supernatants obtained were submitted to electrophoresis. M = Molecular weight.

RESULTS

Determination of Rs Molecular Weight by SDS-PAG E

Samples of a SHL immunoprecipitate obtained from "'I- labelled peripheral lymphocytes, or from an SHL immunopre- cipitate, or from the supernatant of a "lI-labelled lymphocyte culture (SLC) showed a radioactive band of molecular weight 58,000 when submitted to SDS-PAGE. The band was absent in the control. In some SHL and SLC samples, a band of molecular weight higher than 150,000 was observed. SDS- PAGE of immunoprecipitates from NHS samples and from sera from cancer or uremic patients showed only a radioactive band of molecular weight of approximately 58,000 (Fig. 1).

Analysis of Rsl and Rs2

Analysis of the Rs 1 and Rs2 fractions from SHSL by chro- matography on Sephadex G-200 with anti-Rs serum showed a line of total identity by immunodiffusion in gel. The same result was observed for the Rsl and Rs2 fractions of NHS. A cross-precipitation line was obtained by the same proce- dure when the Rs2 fraction from serum and anti-IgM and anti-Rs serum were used, indicating that Rs2 is not related to IgM. To rule out the possibility that Rs2 was bound to IgG, the Rs2 fraction was passed through a Sepharose-Protein A column. When the bound and unbound fractions were tested by the Ouchterlony procedure in relation to the anti-Rs serum, precipitation was observed only in the unbound fraction.

Determination of the molecular weight of Rsl and Rs2 immunoprecipitates by SDS-PAGE followed by reduction only showed one band of molecular weight of approximately 58,000, which was absent in the control, both with the Rsl and the Rs2 fractions (Fig. 5). The Rsl fraction from serum, when submitted to immunoelectrophoresis using anti-Rs serum showed a precipitation line in the albumin region. When submitted to the same procedure, the Rs2 fraction from serum showed a precipitation line in the alpha 2-globulin region (Fig. 4). The Rsl fraction from SHSL also showed a precipitation line in the albumin region. However, when serum samples were submitted to immunoelectrophoresis, no precipitation line was detected in NHS, though Rs2 was detected in the sera of patients with cancer and uremia (Fig. 4). The slow migration of Rs2 obtained from SHL was observed by anal- ysis of eluates of PAGE gels without SDS.

Comparative Study of Rs From Saliva, Serum, and SHL

Analysis of saliva concentrated 5-fold with anti Rs serum and the other sources of Rs such as SHPL, SHSL, NHS (con- centrated fourfold), and supematant of stimulated lympho- cyte cultures (concentrated fourfold) showed a line of total identity (Fig. 3). '311-labelled human saliva immunoprecipi- tated and analyzed by SDS-PAGE showed a band of molecular weight 58,000, i.e., a band of molecular weight equivalent

Molecular Forms of E Receptor 11 7

O D 1

0.0

7.4

Fig. 2. Spectrophotometric profile at 280 nm of the fractions obtained by chromatography on Sephadex (3-200 of a pool of normal human sera (a), a pool of sera from cancer patients (b), a pool of sera from uremic patients (c), a supernatant of heated spleen lymphocytes (SHSL) (d), SHSL concen- trated ten-fold ( e ) , and a pool of normal human saliva concentrated fourfold (0. Fractions (2.5 ml) were collected and analyzed by a&ar gel precipitation (Ouchterlony) using sheep anti-soluble T lymphocyte serum specific for E

to that of Rsl from SHL and serum (Fig. 5 ) . Analysis of fractions obtained by chromatography of fourfold concentrated saliva pool samples on Sephadex (3-200 only showed one pos- itive fraction which interacted with anti-Rs serum (Fig. 2). The Rs2 fraction was not detected by chromatography of saliva even after tenfold fraction concentration. The positive frac- tion was equivalent to Rs 1 from SHL and serum in terms of elution volume and, when submitted to immunoelectropho- resis, it showed a precipitation line with anti-Rs serum in the albumin region (Fig. 4), i.e., similar to Rsl from SHL and serum. Crossed lines were observed when fourfold con- centrated saliva was submitted to the Ouchterlony test with anti-Rs serum, anti-albumin fraction serum and human albu- min (95% pure grade, Instituto Santa Catarina Hemoderivados e Reagentes Ltda.).

DISCUSSION

In the present study, we prepared a heterologous anti-Rs serum by the method of Mendes (14), which formed a line of total identity with the original anti-Rs serum obtained by Mendes when analyzed with SHL.

After labelling and immunoprecipitation with anti-Rs serum and Staflhyfococcus mireus pol yacrylamide gel electrophore- sis showed a radioactive band of molecular weight of approx- imately 58,000, denoted Rs 1 , in the supernatant of heated

and the positive fractions are indicated as RsI and R,2 (a-c). The Rs2 fraction of normal human serum (a) was positive after ? 10-fold fraction concentration. Only the RsI fraction was detected in SHL (d) and normal human saliva (e). After concentration of the SHSL sample and of the frac- tions collected, both the Rsl and Rs2 fractions were detected (e) . O.D. = optical density at 280 nm. 150 kD (human IgG). Molecular weight ot'69,OOO (human albumin).

peripheral lymphocytes (SHPL), in the supernatant of a lymph- ocyte culture stimulated with PHA, in normal human serum (NHS), and in the serum of patients with cancer or uremia.

This result is in accordance with the molecular weight of CD2 which has been reported to range from 40,000 to 58,000, depending on the variation in the lateral carbohydrate branches (8). It is possible that Rsl detected by anti-Rs serum is CD2, since, in addition to the similarity in molecular weight, an increase in both Rs (1 5 ) and CD2 ( 13) levels has been reported in the serum of cancer patients.

In addition to this fraction of molecular weight of approx- imately 58,000, in some concentrated SHL samples analyzed by chromatography on Sephadex (3-200 or by prolonged auto- radiography and polyacrylamide gel electrophoresis of labelled SLH immunoprecipitates, another fraction of high molecu- lar weight (more than 150,000) was detected, called Rs2. Similarly, the reactivity of the Rs2 fraction of NHS with anti- Rs serum was detected in a fraction obtained by chromatog- raphy on Sephadex G-200 only after sample concentration. In contrast, in the serum of patients with cancer or uremia with high Rs levels, the Rs2 fraction was detected directly without previous concentration. This may indicate that Rs2 contributes to the increase in Rs in these patients.

The Rs2 band was not identified by SDS-PAGE in the immunoprecipitate of human serum or in the Rs 1 fraction of

118 ltano et al.

a b

a

d

C

Fig. 3. Immunodiffusion: a) Rs2 fraction (NHS Rs2) and Rsl (NHS R,I) fraction of normal human serum obtained by fractionation of normal human serum on Sephadex G-200 and between the Rs2 fraction (SHSL 1) and the RS1 fraction (SHSL 11) of the supernatant of heated spleen lymphocytes (SHSL) obtained by SHSL fractionation on Sephadex (3-200. NHS I = Rs2 fraction of NHS. b) Anti-SRBC receptor sheep serum (A) and anti-human IgM sheep serum (C), with the Rs2 fraction obtained by fractionation of serum from a patient with cancer on Sephadex (3-200 (B). c) Normal human saliva concentrated fivefold ( I ), supernatant of heated thymus lymphocytes (2), supernatant of heated spleen lymphocytes (3), supernatant of a lym- phocyte culture stimulated with PHA (4), normal human serum concentrated five-fold ( S ) , with anti-R, sheep serum (A). d) Anti-albumin fraction rab- bit serum (A) with anti-Rs sheep serum (B), using purified human albumin (C) and human saliva concentrated fivefold (D) as antigens.

serum from patients with cancer, but the presence of a high molecular weight band (Rs2) in some samples of SHL immunoprecipitates suggests that at least part of this band consists of molecules that do not present disulfide bonds.

Even though Rs2 shows electrophoretic migration (alpha 2-globulin) and a molecular weight of more than 150,000, no antigenic relationship with IgM was observed when Rs2 was analyzed with anti-IgM serum. Also, this is not an immune complex with IgG since the Rs2 fraction was passed through a Sepharose-Protein A column and the reactivity with anti-Rs serum was not removed. The fact that the L and H chains of Ig were not observed by immunoprecipitation with anti-Rs serum and by SDS-PAGE in any preparation contain- ing receptors rules out the possibility of the involvement of any immunoglobulin class. Maugeri (1 8) also demonstrated that Rs is not C-reactive protein, acid alpha-1 glycoprotein, ceruloplasmin, transferrin, or alpha-2-macroglobulin. Anal- ysis of saliva samples by gel filtration, SDS-PAGE, and immu- noelectrophoresis demonstrated the presence of a molecule of molecular weight of approximately 58,000 which reacted with anti-Rs serum, being therefore similar to the Rsl of SHL and serum.

Fig. 4. Analysis of electrophoretic migration by immunoelectrophoresis using anti-soluble receptor sheep serum (anti-R,) and anti-normal human serum rabbit serum (anti-NHS). a: Rsl fraction (1,2) and Rs2 fraction (3,4) obtained by fractionation of serum from a patient with cancer ( I ) on Sephadex G-200. b: Serum from a patient with cancer (l), human saliva concentrated fivefold (2), supernatant of heated lymphocytes (3), and Rsl fraction of human saliva (4).

The Rs 1 molecule has electrophoretic migration and molecular weight similar to those of albumin, which is the most cationic protein in plasma and therefore has the greatest facility in binding to the cell surface, and both Rsl and albumin are detected on the surface of human lymphocytes (19). However, the fact that human saliva analyzed with anti- Rs serum and anti-albumin serum showed a crossed line by immunodiffusion in gel rules out the possibility that Rsl is albumin.

The similarity of saliva Rsl and SHL and serum Rsl in terms of molecular weight, electrophoretic migration, and the common antigenic determinants recognized by anti-Rs serum suggests that soluble E receptors are present in human saliva. Although it was not possible to detect Rs2 in samples of nor- mal human saliva or of saliva from cancer patients, even when the samples were concentrated tenfold, Rs2 may have been present since it was detected in all other preparations ana- lyzed. The lack of detection of this molecular form in saliva may be explained by its degradation by enzymes of by its presence in very small amounts in saliva.

Molecular Forms of E Receptor 11 9

1 2 M W

b

Fig. 5. Determination of the molecular weight of the soluble receptor by autoradiography and SDS-PAGE. a: R,2 fraction immunoprecipitate (1,2,5-10) and Rs I fraction immunoprecipitate (3,4). b: Human saliva immunoprecipitate (1,Z). The immunoprecipitates were obtained from labelled samples incubated with preimmune sheep serum. Samples were then trans- ferred to a tube containing Sruph$ococcus uureus-A sediment. After incu-

Molecules of distinct molecular weights (14,000, 42,000, and 220,000) involved in rosette formation between T lym- phocytes and E have been detected on the E surface and of erythrocytes from other species (20).

Preliminary data obtained by us in studies currently under- way have demonstrated that both Rsl and Rs2 have the abil- ity to interact with E. Thus, more than one molecular form of E receptor must participate in rosette formation between T lymphocytes and E. Also, the multiple, and often opposite, CD2 activities reported in the literature may be due to the existence of more than one form of E receptor. The lack of detection of this molecular form by other investigators may be explained by its low concentration in normal preparations or sera, or may possibly be due to the restricted recognition by monoclonal antibodies, whereas the polyclonal recogni- tion system is wider.

REFERENCES

I . Lay WH. Mendes NF, Bianco C , Nussenzweig V: Binding of sheep red blood cells to a large population of human lymphocytes. Nufure 230531-532, 1971.

2. Mendes NF. Tolnai MEA, Silveira NPA, Metzgar RS, Gilbertsen R: Technical aspects of rosette tests used to detect human complement recep- tor Cp) and sheep erythrocyte binding (T) lymphocytes, J Immunol I I1:860-867, 1973.

3 . Mendes NF: Immunological identification of human lymphoid cell pop- ulations. Lynpholop 10:85-93, 1977.

4. Meuer SC, Hussey RE, Fabbi M , Fox D. Acuto 0. Fitzgerald KA, Hogdon JC, Protentis JP, Schlossman SF, Reinherz EL: An alternative pathway of T cell activation: a functional role for the SO Kd T I I sheep erythrocyte receptor protein. Ce/l36:897-906. 1984.

5. Hiinig T, Mitnacht R, Tiefenthaler G , Kohler C. Miyasaka M: TI ITS the cell surface molecule binding to the “erythrocyte receptor” of T

bation. the supernatant was divided into two parts, PBS was added to one part (odd numbers) and anti-Rs sheep serum to the other. Both portions were incubated again with a Sruphdococcus crureus-A suspension. The cell pel- lets were treated with the sample buffer (containing a reducing agent) and the supernatants obtained were submitted to electrophoresis. MW = molec- ular weight.

lymphocytes cellular distribution, purification to homogeneity and bio- chemical properties. EurJlrnmunol 16(2): 1615-1621, 1986.

6. Plunkett ML, Sanders ME, Selvxaj T, Dustin ML, Springer TA: Rosetting of activated human T lymphocytes with autologous erythrocytes. J E.rp Med 165(3):664-676, 1987.

7 . Halvorsen R , Leivestad T, Gaudemack G , Thorsby E: Accessory cell- dependent T-cell activation via Ti-CD3 Involvement of CDZ-LFA-3 inter- actions. ScandJImmurioi 28(3):2??-284, 1988.

8. 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 ExpMed 153:207-212, 1981

9. Verhi W. Greaves MF, Schneider C, Koubek K, Janossy G, Stein H , Kung P, Goldstein G: Monoclonal antibodies OKTl I and OKT I 1 A have pan-T reactivity and block sheep erythrocyte “receptors”. Eur J lmmunol 1 2 3 - 8 6 , 1982.

10. Plunkett ML, Springer TA: Purification and characterization of the lym- phocyte function-associated-2 (LFA-2) molecule. J lmmunol 136( 1 1 ): 4 I8 1-4 187, 1986.

I . 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. EurJ / tn~ii~nol 17: 15-20. 1987.

2. Mendes NF, Saraiva PJ, Santos OBO: Restorative effect of normal human serum. transfer factor and thymosin on the ability of heated human lym- phocytes to form rosettes with sheep erythrocytes. Cell lmmunol 17560-566. 1975.

3. Oh SK, Lafenson D, Morgan Jr AC: Quantification of soluble E-receptor in the serum of patients with various diseases and its accompanying in vitro immunosuppression in neoplasia. Scnnd Jlmmunol22:S 1-60, 1985.

4. Mendes NF, Bernd LAG, Cintra HHA, Mendes CMF, P r a p CL, Moura NC: Human T lymphocyte receptor for sheep erythrocytes: characteriza- tion of a specific antiserum and its application in the detection and quan- titation of the receptor in a soluble form. Celllmmunol 72: 143- 150, 1982.

15. Moura NC, Longo IM. Bernd LAG. Mendes NF: Quantitation of the soluble receptor of human T lymphocytes for sheep erythrocytes by electroimmunodiffusion in the serum of patients with cancer, uremia and leprosy. Experirnfiu 39(3):306-308, 19x3.

16. Musatti CC, Soares VA, Santos LMB. DeLima JJG, Mendes NF: Immu-

120 ltano et al.

nosuppressive effect of soluble E receptors in uremic serum. Chi Irnnzunol lrnmunopathol 14:403-410, 1979.

17. Ambler J , Rodgers M: Two new non-barbiturate buffers for electropho- resis of serum proteins on cellulose acetate membranes. Clin Chem 26(8): 122 I - 1223, 1980.

18 . Maugeri IML: Receptor de linfocitos T humanos para eritrocitos de carneiro: dosagem por imunoeletrodifusio em soro de pacientea corn hanseniase, leucemias e linfomas. Comparasio com dosagem serica de

“proteinas de fase aguda”. PhD thesis-Escola Paulista de Medicina. Sao Paulo, Brazil, 1982.

19. Andersen MM: Leucocyte-associated plasma proteins. Scand J Clin Lab Invest 43:49-59, 1983.

20. Bernard A, Tran HC, Boumsell L: Three different erythrocyte surface molecules are required for spontaneous T cell rosette formation. J lmmunol 139( I): 18-23, 1987.