Multispecific lymphoid cell surface receptors - Europe PMC

2 downloads 0 Views 904KB Size Report
Dec 27, 1976 - We would like to thank Dr. W. H. Konigsberg for his help and support of this project. This work was supported by Grants AI-08614,. GM-12607 ...
Proc. Natl. Acad. Sci. USA Vol. 74, No. 3, pp. 1224-1228, March 1977

Immunology

Multispecific lymphoid cell surface receptors (immunoglobulins/polyfunctional antibodies)

MARK J. CZAJA*, FRANK F. RICHARDS*, AND JANOS M. VARGAt Departments of *Internal Medicine and tDermatology, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06510

Communicated by Aaron B. Lerner, December 27, 1976

ABSTRACT When mice are sequentially immunized with two antigens to give an oligoclonal "double-binding" antibody response, there is a concomitant increase of "double-binding" cell surface receptors on their splenic lymphocytes. Competition studies suggest that the capacity to bind the two ligands, bovine pancreatic ribonuclease (EC 3.1.4.22) and a 2,4-dinitrophenyl (Dnp) derivative, is a function of the same molecules. In ribonuclease-primed mice, an early response to bovine gamma globulin containing an average of 60 Dnp groups per molecule is the appearance of an increasing number of cells bearing surface receptors bindin both ribonuclease and Dnp. Later, these double-binding celL are diluted by cells that bind Dnp, but not ribonuclease. The analogous phenomenon is observed when the two antigens are used in reverse order. While other reports suggest that there may be several different receptors in relatively undifferentiated cells from unimmunized mice, it seems likely that cells committed to antibody production carry a predominant multispecific cell surface immunoglobulin receptor. The antibody combining region is known to be multispecific in the sense that structurally diverse antigens can bind to different sites within the antibody combining region (1-4). Such multiple binding is also functional because production of a single immunoglobulin species binding two diverse antigenic determinants may be induced by either of these two determinants (5). A simple, but experimentally untested, extension of this idea is that the antigen receptor on the surface of the bone-marrow-derived (B) lymphocyte contains the same combining region as the immunoglobulin secreted by the cell. Thus, an antibody-producing cell precursor (APCP) would respond to two diverse antigens by virtue of the interaction of two determinants with a single polyfunctional B-cell receptor. However, DeLuca et al. (6) have recently presented persuasive evidence that in unimmunized mice, 30-50% of all splenic lymphocytes bind more than one antigen and that the receptors for the two antigens employed were located on different receptor molecules of the cell surface. Similar conclusions arise from the work of Liacopoulos et al. (7). These authors suggest that APCPs may exhibit a stage of differentiation where several different types of receptors are present on the APCP and at which time the cell has the potential to produce several antibodies. The final immunoglobulin secreted by the cell would depend on the receptor that is stimulated. Thus, the observed polyfunctionality of antibody-producing cells could be mediated either by several monofunctional APCP receptors or by single APCP receptor molecules that are polyfunctional. We here present evidence that under conditions during which immunoglobulins binding both 2,4-dinitrophenyl (Dnp) groups and bovine pancreatic ribonuclease (EC 3.1.4.22) Abbreviations: ABC, antigen-binding cell; APCP, antibody-producing cell precursor; Dnp, 2,4-dinitrophenyl radical; Dnp6oBGG, bovine gamma globulin containing 60 Dnp radicals per molecule (average); HBBS, Hanks' balanced buffer solution; PBS, 0.1 M sodium phosphate buffer at pH 7.4/0.15 M NaCl; RNase, bovine pancreatic ribonuclease; TL-Dnp, L-tyrosyl-(N'-2,4-dinitrophenyl)-L-lysine hydrochloride.

are induced, there is an associated specific increase of cell surface receptors that also bind Dnp and ribonuclease competitively. It therefore seems likely that splenic lymphocytes do contain polyfunctional cell surface receptors. We do not rule out the possibility that a single cell may contain more than one type of polyfunctional immunoglobulin receptor.

MATERIALS AND METHODS Immunization. Five- to seven-week-old BALB/cN mice were injected subcutaneously with 100 ,ug of bovine gamma globulin containing an average of 60 Dnp groups per molecule (Dnp6oBGG) or ribonuclease (RNase; Worthington 12 times recrystallized) in complete Freund's Adjuvant and were boosted 18 days later with 100 ,ug of the same or with the heterologous antigen subcutaneously. Unimmunized control mice formed group A. Immunized mice were divided into six groups: B, primed with RNase, but not boosted; C, primed with RNase, boosted with RNase; D, primed with Dnp6oBGG, but not boosted; E, primed with Dnp6oBGG, boosted with Dnp6oBGG; F, primed with RNase, boosted with Dnp6OBGG; G, primed with Dnp6oBGG, boosted with RNase. Antigens. L-Tyrosyl-(Nl-2,4-dinitrophenyl)-L-lysine hydrochloride (TL-Dnp) and its radioiodinated derivative 125ITL-Dnp were synthesized by a previously reported method (5). The specific radioactivity of 125I-TL Dnp was 4.5 X 105 dpm/mol. DnproBGG was prepared by the method of Eisen (8). 25I-labeled RNase 125I-RNase was prepared by Ihe method of Hunter and Greenwood (9) and had a specific radioactivity of 1 X 104 Ci/mol. Preparation, Radiolabeling, and Counting of Spleen Cell Preparations. Because cell surface receptors are lost rapidly at 370 and are stable at 00, all phases prior to cell fixation were carried out at 00. Individual mouse spleens were teased apart in 2.0 ml of Hanks' balanced buffer solution (HBBS). The suspension was then gently homogenized in a syringe and layered carefully on top of 1.5 ml of Lymphoprep in a 5 ml tube and subsequently centrifuged at 1000 X g for 30 min in an IEC centrifuge. The lymphocyte fraction was resuspended in 10 ml HBSS and centrifuged at 1000 rpm for 5 min. The pellet was resuspended in 200 ,l HBSS and then applied to the polylysine-coated microscope slide and was incubated for 20 min at 0° without haptens. Control cell suspensions were in addition incubated with 20 ml of Earle's minimal essential culture medium containing 5% fetal calf serum for 30 min at 370 prior to the Lymphoprep step, to remove any passively adsorbed antibodies (10). Because this step did not alter results, it was subsequently omitted. The microscope slides were coated with polylysine prior to the application of the cells (11). With this technique the cells were attached to the slide with minimum damage, and sticking of radioiodinated antigens to the glass slide was reduced. In different experiments, numbered as the columns in Table 1, the cells attached to the polylysine-glass slides were incubated for 60 min at 00 with a 50 ,l volume al1224

Proc. Natl. Acad. Sci. USA 74 (1977)

Immunology: Czaja et al. cli

oq

Lo

5

C

C

co

O3)

>e

C

CD

+1

OO

~

0

oLo

oZ

0. 4-~

14)

Lo0~

E-

Lo

co

+

LO

LO

LO~

Cl

L-

LO

Lo

0~ oq

L-

0)

t-

Cl1

0-C oo

"0

LD

o

+ co

oq

C

l

.-

OLo

E0Co

C1C C~~~~~~~~~~o

Hu

*a

t

t

r

oo

0C

COe

O

o

"0

C

a

1

Cl

Co

'

ct Y

LoC~ ~

~

) 0C0

0)

c

O00

C)Co

eE-,

~0

0

C

00~

0

0

000

0~C

o0

+

)

C C 3 )r 0)M

OC

0

00

Coi

CoCY)o r-

Q Qo

0D

0

04 LO

LOI

.° a)

X

LO)

ClO

0~~~~~~~~~C)0

co

"0

4) C>

oo

"0 o

Co:

0+0I

0Y

c0

'ti

~~~~o

0+I-

0.

~~ CC

4)

C13 V-.

S4:3

C

r

T!Cl

0)

CoD

W

$D0 C

lr

0)

LO

0)

Lo

Cl

Q0

Co

co

.c

Lo

co

CY)

3

.) o

==

-4 z 0 ~~~~~~ ~~~~

la .0

L:,t

~ ~~C

O~ c

+I

+I

+1

oo

to

~~~0

r-I

r-I

." 4

Co

a>

tootoo

z CZ

0.

.=

's: Q

cY

4)

0.Lo

CY)

r4

*

Lo

LO

CoD

ulO

CD

0

.0 _.

C

°

.C

0.

as

0

0

to 0

LO~

Lo

Co

4) CC3

D

0q

4-

Cl

A

LO

I

H

=

*

0.4S

14

0

H--

CoD

co

Co

r-

r!C m0

-i

'-!

r

co

C

U

vLC

LO

0)

r--

+1

03

~t

q 0

m

0

0)

0c

!' C

Cl'T

+1

C

to

to

a) +I'

r

c

-.

0C

0)(

o

c

"E

+ O

0o

Q C

CoLO

0-

0i

LOCo r

!

0y)

LoCl

0C

Co)

.

.

Co

Co

0

C9

iO

00

b

w

E

0)

a cl!

Q

co

4)

'4-

Cl

C)

5

CD

01 e

c

-4

Co)Co )' ;1

uLO

lp~~~~~~~~~~~~~~e

Lq

z

o = w

C4

C-

0

lo00

CoS

0*Es

ze

Pz

0.S= "0C O i..i C.) E4 -4 r, CL) + .o ,v :j .= 4J. .

.

r. 0

E E

0-4

0gj ~ ~ "cU)r ~ ~

N

~

~

~

0

.4

&O 0 -4 .

>. = .0 *_4

-I.-

-9

1225

1226

Immunology: Czaja et al.

iquot containing [in phosphate-buffered saline (PBS)]: 1, 1 X 106 dpm of 125I-RNase (0.1 mM); 2, 1 X 106 dpm of 125I-RNase plus RNase (1 mM); 3, 1 X 106 dpm of 125I-RNase plus TL-Dnp (1 mM); 4, 1 X 106 dpm of 125I-TL-Dnp (0.1 mM); 5, 1 X 106 dpm of 125I-TL-Dnp plus TL-Dnp (1 mM); 6, 1 X 106 dpm of 12I-TL-Dnp plus RNase (1 mM). After incubation the cells were washed three times with PBS, fixed with 1% (wt/vol) glutaraldehyde in PBS, air dried, and coated with an ultra-thin layer of nuclear track emulsion by a bubble technique that eliminates radioactive decay tracks beyond the apparent cell boundary (12). Eight days later the slides were developed and the cells were stained directly with Giemsa. Antigen-binding cells (ABC) were counted by light microscopy; the adopted criterion for an ABC was the presence of five or more silver grains within the apparent cell boundary. On each slide 2000 cells were counted.

RESULTS Experiments described here establish that lymphoid cells competitively bind both a Dnp-based hapten, TL-Dnp, and RNase to their surfaces.f The competitive nature of the binding suggests that these two substances bind to the same surface molecules. Such TL-Dnp + RNase binding surface molecules exist in the lymphocytes of unimmunized mice. They are more plentiful in mice immunized with Dnp6oBGG and in mice immunized with RNase. However, a major expansion of the cell clones containing surface molecules binding both TL-Dnp and RNase occurs in mice immunized with Dnp60BGG followed by RNase or vice versa. Such double-binding molecules are retained on the cell surface under conditions in which any passively adsorbed antibodies are shed. Table 1 demonstrates the existence of polyfunctional ABC surface molecules. In unimmunized animals Dnp-binding cells occur with a frequency of approximately 85/104 cells (Table 1; immunization schedule A, column 4, referred to as A4 data). This high frequency has been found by other workers (see ref. 6 for references) and may reflect the fact that many clones of antibody-producing cells produce immunoglobulins complementary to Dnp. The frequency of RNase-binding ABCs (Table 1, Al data) is somewhat higher and probably represents clones of cells producing antibodies directed against a number of surface determinants of RNase. Very few of the ABCs of nonimmunized animals bind both TL-Dnp and RNase, i.e., the inhibition by heterologous antigen is very low (Table 1, A5 and A6 data). When mice are immunized and boosted with the same antigen, the cell compartment binding that antigen increases in size, but the relative number of double-binding ABCs remains very low. Homologous antigen in 10-fold excess reduces the number of ABCs to approximately 10% in cells from unimmunized mice, indicating that ABC surface molecules are saturatable with an average binding constant Ko of 1 X 106 M-1 or greater. When the compartment is expanded, there is evidence that saturation is no longer complete. The likeliest cause is that, along with high-affinity ABCs, low-affinity ABCs are also expanded, and they are not saturated at 1 X 10-4 M hapten concentrations. The most striking finding is the large relative expansion of double-binding ABCs when mice are immunized with Dnp6oBGG followed by RNase or vice versa (Table 1, F3, G3 data and F6, G6 data; and Fig. 1). The pattern of expansion was t Previous experiments have shown that TL-Dnp behaves or resembles e-Dnp-lysine in binding to anti-Dnp antibodies. The antigenic determinant TL-Dnp and determinants on RNase have previously been shown to bind competitively to the same combining region of certain polyfunctional antibodies (5, 14).

Proc. Natl. Acad. Sci. USA 74 (1977)

'boosting" with

A,

37 %

// %

vt oS

//,O

Dnpj0BGG

F~ /

Unimmun ized mice

/

0 V ,A _0

2

i'boosting"

with

RRNase

5% 66 %

FIG. 1. Antigen-binding cells (ABCs) and double-binding (RNase + Dnp) ABCs from BALB/c mouse spleens after various immunization procedures. The area of each circle is proportional to the total number of antigen-binding cells at day 10 after the last immunization procedure. The hatched sectors and the figures under the sectors refer to the total proportion of ABCs that are RNase + Dnp double-binding cells. Data from which this figure is drawn are given in Table 1.

interesting. Mice were killed at 3, 7, 10, and 23 days after boosting with the heterologous antigen. If the mouse was primed with RNase and boosted with Dnp6OBGG, then early after boosting (3 days) the priming (RNase) response predominated. There were relatively large numbers of RNase-binding cells, but only about half of all RNase-binding cells were TLDnp-RNase double-binders. However, in the same population, almost all of the TL-Dnp-binding cells (90%) were RNase-Dnp double-binders. At 10 days, the percentage of RNase-binding cells that were RNase-Dnp double-binders rose to 71%, suggesting that the clones of double-binding cells had expanded to their maximum. At 23 days these cells were about the same in number or perhaps had begun to decline somewhat (63%). When the percentages of Dnp-binding cells in these populations that are also Dnp-RNase double-binders are estimated, the percentage falls from 90% at 3 days to 15% at 23 days (Fig. 2 and Table 1, F6 data). This decline shows that a "primary" anti-Dnp response arose later after boosting and that a declining percentage of cells are Dnp-RNase double-binders. Thus, at the level of the cell surface "receptor," the earliest consequence of double immunization with a pair of crossreactive antigens is that early after heterologous boosting there is a selective increase in ABCs binding both antigens. Later, when the "primary response" to the boost becomes apparent, RNase-Dnp doublebinders become only a small portion of the Dnp-binding cells. This phenomenon mirrors exactly the pattern of appearance of double-binding antibodies in rabbits (5) and mice (unpublished) and suggests strongly that there is a direct relationship between the appearance of double-binding receptors and the secretion of double-binding antibody. When the order of immunization is reversed (Dnp6oBGG followed by RNase) the equivalent phenomenon is observed (Table 1, G3 data). Control Experiments. To check that the appearance of double-binding ABCs was a consequence of double immunization, control experiments checking the primary and secondary responses of both Dnp6oBGG and RNase were carried out. As shown in Table 1, C1-6 and E1-6 data, these did not give rise to large populations of double-binding cells. Because other studies have shown in both rabbits and mice that the sequential

Proc. Natl. Acad. Sci. USA 74 (1977)

Immuno'logy: Czaia et al.

1227

2.0

100

No of ABCs 100 cells

% of ABCs that bind both Dnp 50 and RNase

10

IA.

100 .

'.0

No of ABCs 1 00 cells

% of ABCs that bind 50 both Dnp and RNase

10 I

3

7

10

I 1.0

23

Days FIG. 2. The proportion of double-binding (RNase + Dnp) ABCs as a function of time after heterologous boosting. (A) Spleen ABCs from BALB/cN mice primed with Dnp60BGG and boosted with RNase (immunization schedule given in Materials and Methods section): (a) 1251-TL-Dnp ABCs/100 cells counted; (0) 1251-TL-Dnp + RNase-binding ABCs/100 cells counted. (B) Spleen ABCs from BALB/cN mice primed with RNase and boosted with Dnp6oBGG (immunization schedule given in Materials and Methods section): (v) 125I-RNase ABCs per 100 cells counted; (A ) 1251-RNase + Dnp-binding ABCs per 100 cells counted.

immunization protocol gives rise to double-binding humoral antibodies, control experiments were carried out to check that passive adsorbtion of antibodies did not account for the observed results. When ABCs were subjected to the conditions of Berken and Benacerraf (10) in which adsorbed antibodies are shed from lymphocyte surfaces, ABC numbers remained within 5% of those on untreated cells.

DISCUSSION We find a specific increase in Dnp + RNase double-binding ABC following sequential immunization of BALB/c mice with Dnp6oBGG followed by RNase and vice versa. Under identical conditions double-binding antibodies are produced (5), but passive adsorbtion of secreted antibodies does not appear to be responsible for this phenomenon. Thus, cells producing antibodies with polyfunctional combining regions appear to have the same type of polyfunctional cell surface "receptors." There is evidence compatible with the presence of several different "receptors" on the same APCP (6, 7, 13). This concept does not contradict the data presented here. Thus, it is possible that a single APCP may contain one or more than one type of "polyfunctional" cell surface receptor. We have chosen a small hapten and a relatively large protein antigen pair that is known to bind to the combining region of a single immunoglobulin (5, 14) and we have obtained evidence that a single cell surface molecule binds both of our antigens. Perhaps, had a different pair of antigens been chosen, whose binding sites were located on different molecules, results similar to those of DeLuca et al. (6) might have been obtained. Other authors find no double-binding ABCPs in spleens of unimmunized mice when they incubate a single pair of antigens with cells at concentrations of approximately 10,uM and 0.1 tM,

respectively (15). However, screening experiments show that many (approximately 100 to 200) "randomly" chosen antigen pairs must be tested before a serum is found that contains immunoglobulins binding both members of a single pair with K0 = 105 M-l or higher (3, 4). Since the same may be true for cell surface receptors, neither monospecificity nor multispecificity is shown by such experiments. It has also been shown that antisera produced by sequential immunization of two antigens vary in the number of component immunoglobulins that bind both antigens. Thus, for a pair of determinants such as Dnp and inosine, perhaps only one or two clones of cells produce Dnp + inosine double-binding antibodies (5). Other pairs of antigens such as Dnp and RNase induce a larger number of clones secreting double-binding antibodies. We chose Dnp and RNase because it seemed likely that the double-binding ABCs would occur in large numbers and be easier to demonstrate (14). An early sequel of immunological boosting with two crossreactive heterologous antigens is the specific induction of ABCs bearing receptors for both antigens. This phenomenon resembles the pattern of double-binding antibody production (5). Our experiments do not distinguish between lymphocytes that are predominantly bone-marrow-derived and those that are predominantly thymus-dependent (T). High resolution separation techniques use the presence of immunoglobulins on, or the application of immunoglobulins to, the cell surface and may therefore interfere with cell surface receptors. It seems likely that the ABCs we observe are probably not all immediate precursors of antibody-producing cells. The relatively small expansion of ABCs after primary immunization suggests that only a small proportion are active precursors; the others may be resting "memory" cells and possibly T-cells. Mice primed and challenged with the same antigen also produce double-binding ABCs (Table 1, C3, E6 data and C6,

1228

Immunology: Czaja et al.

ES3 data). As compared with double-binding cells in crossstimulated mice, the number may be either quite small or much larger. This suggests that the secondary response produced by priming and challenging with the same antigen may sometimes be an inefficient evoker of all cell clones binding the antigen, including the double-binding cells. At other times it may be very efficient. It seems that the order of immunization is unimportant for this phenomenon. When the results of DeLuca et al. (6) and Liacopoulos et al. (7) are considered together with the work reported here, a sequence of events is suggested. In unimmunized animals, a relatively large number of APCs are found which are not yet committed to antibody production and which carry on their surface several different immunoglobulin cell surface receptors. This probably occurs concomitantly with an early stage of cell differentiation. In mice immunized to give an oligoclonal response of double-binding antibodies, there is a corresponding increase of one double-binding cell surface receptor. It is possible that later in differentiation the cell committed to antibody production carries predominantly a single polyfunctional receptor. We would like to thank Dr. W. H. Konigsberg for his help and support of this project. This work was supported by Grants AI-08614, GM-12607, and HL-16126 from the U.S. Public Health Service and GB-43482-X from the National Science Foundation. J.M.V. is the recipient of a Research Career Development Award from the Public Health Service.

Proc. Nati. Acad. Sci. USA 74 (1977) 1. Rosenstein, R. W., Musson, R. A., Armstrong, M., Konigsberg, W. H. & Richards, F. F. (1972) Proc. Natl. Acad. Sci. USA 69, 877-881. 2. Richards, F. F. & Konigsberg, W. H. (1973) Immunochemistry

10,545-553. 3. Richards, F. F., Konigsberg, W. H., Rosenstein, R. W. & Varga, J. M. (1975) Science 187, 130-137. 4. Inman, J. K. (1974) in The Immune System, Genes, Receptors, Signals, eds. Sercarz, E. E., Williamson, A. R. & Fox, C. F., (Academic Press, New York), pp. 37-52. 5. Varga, J. M., Konigsberg, W. H. & Richards, F. F. (1973) Proc. Natl. Acad. Scd. USA 70,3269-3274. 6. DeLuca, D., Miller, A. & Sercarz, E. (1975) Cell. Immunol. 18, 255-273. 7. Liacopoulos, P., Couderc, J. & Bleux, C. (1976) Ann. Immunol. (Paris) 127C, 519-530. 8. Eisen, H. N. (1964) Methods Med. Res. 10, 94-102. 9. Hunter, W. M. & Greenwood, F. C. (1962) Nature 194, 495496. 10. Berken, A. & Benacerraf, B. (1966) J. Exp. Med. 123, 119137. 11. Mazia, D., Schatten, G. & Sale, W. (1975) J. Cell Biol. 66, 198-200. 12. Varga, J. M., Saper, M., Lerner, A. M. & Fritsch, P. (1976) J. Supramol. Struct. 4, 45-49. 13. DeLuca, D., Decker, J., Miller, A. & Sercarz, E. (1974) Cell. Immunol. 10, 1-18. 14. Varga, J. M., Lande, S. & Richards, F. F. (1974) J. Immunol. 112, 1565-1570. 15. Julius, M. H., Janeway, C. A., Jr. & Herzenberg, L. A. (1976) Eur. J. Immunol. 6,288-292.