- Email: [email protected]
CH2 and CH3 domain deleted IgG1 paraproteins inhibit differently Fc receptor mediated binding and cytolysis
Immunology Letters, 12 (1986) 307-312
Elsevier Imlet 749
CH2 A N D CH3 D O M A I N D E L E T E D IgG1 PARAPROTEINS INHIBIT DIFFERENTLY Fc RECEPTOR MEDIATED BINDING A N D CYTOLYSIS Gabriella S/kRMAY l, Roy JEFFERIS 2 and J~inos GERGELY 1 ~Dept. of Immunology of L. EOtvOs University, GOd 2131, Jdvorka S. 14, Hungary, and 2Dept. of Immunology, The University of Birmingham, Birmingham B15 2TJ, U.K.
(Received 31 January 1986) (Accepted 5 February 1986)
I. Summary Domain deleted paraproteins are suitable tools to study the interaction between IgG domains and Fc receptor (FcR) binding sites. The effect of the C3,2 or C73 domain deleted paraproteins was compared on antibody dependent cellular cytotoxicity (ADCC) and on FcR mediated rosette formation. The C72 domain deleted paraprotein (TIM) had no significant effect on lymphocyte or monocyte mediated ADCC, while the C3,3 domain deleted paraprotein (SIZ) inhibited both types of cytotoxicity. FcR dependent rosette formation was also inhibited by SIZ but TIM was ineffective. The data further confirm our previous results suggesting a significant role of C72 domain in the transfer of killing signal in ADCC and that of C73 domain in the high affinity binding to lymphocyte FcR.
2. Introduction Several cell types express surface receptors which bind human IgG through interactions with Key words: Fc receptor binding sites - lgG domains - anti-
body dependent cellular cytotoxicity Abbreviations: FcR, Fc receptors; ADCC, antibody depen-
dent cellular cytotoxicity;PBMC, peripheral blood mononuclear cells; EA, erythrocytescoated with antibody; DEAE, diethylaminomethyl;SDS, sodium dodecylsulphate; PAGE, polyacrylamidegel electrophoresis
defined sites of the Fc region (reviewed in [1, 2]). A subpopulation of the FcR positive cells function as effectors in the ADCC [3]. Operationally these are designated as K cells. Within this heterogeneous population [4, 5] some are lymphocytes, others are monocytes [6, 7]. K cells exhibit a high association constant for IgGl and IgG3 subclass proteins. Accordingly, these IgG molecules mediate ADCC, while IgG2 and IgG4 proteins are inactive. The precise role of FcR in activating the lytic mechanism which results in the destruction of target cells in ADCC is not known. Activation of the lysis can be inhibited by the ligands, monomeric IgG and Fc [8, 9]. The contribution of the C72 and C73 domains in the binding to FcR and the functional role in ADCC has been studied extensively [8-16]. We have previously demonstrated that both the C72 and C3,3 domains of the sensitizing antibody are necessary for obtaining an optimal ADCC effect with lymphocytes. We have proposed a model in which two binding sites are necessary and have different impact. FcR/C73 domain interaction contributes to the close contact between effector and target ceils, while the FcR/C72 is crucial for triggering the lytic process for which the linear sequence Lys 2 7 4 - A r g 301 of the C72 domain seems to be responsible. In contrast, monocyte mediated ADCC depended only on one high avidity binding site of FcR interacting with the C72 domain of IgG [14-161. These results are consistent with other recent reports; however, earlier studies claimed that the 307
C'73 domain does have a functional role in monocyte mediated A D C C [17--19]. We have further investigated the role of (7-},2 and C~.3 domains in the interaction between effector and target cells in monocyte and lymphocyte mediated A D C C systems. Our results confirm and extend the hypothesis that the C~2 domain triggers the lyric function in A D C C mediated both by lymphocytes and by monocytes. The results suggest that lymphocyte FcR has a high affinity binding site interacting with the C-,,3 domain and a lower affinity binding site reacting with the C')2 domain of the antigen complexed lgG. 3. Materials and Methods
3.1. Separation of lymphocytes Lymphocytes from the blood of healthy donors were prepared according to the method of B6yum [20]. The peripheral blood mononuclear cell (PBMC) suspension was used in the antibody dependent cytotoxic assay. Typically, it contained 1 5 - 2 5 % monocytes as indicated by latex phagocytosis. For the assay with lymphocytes, monocytes were depleted by iron phagocytosis and plastic adherence. The resulting suspension contained less than 2% monocytes. 3.2. Antibody dependent cellular O'totoxicio'
(ADCC) Lymphocyte mediated cytotoxicity was performed as described earlier [21]. Briefly, washed R h + human erythrocytes were treated with papain (1%) at 20°C for 10 min, labelled with 400 p,Ci = 14.8 MBq sodium 5~chromate (Amersham, U.K.), and sensitized with anti-D lgG 250 ~g/ml (OHVI, Hungary). The assay was carried out in U-shaped wells (Cooke microtiter plate). Each sample was tested in triplicate. Effector cells (PBMC) were added to the targets at different ratios. The plates were centrifuged at 2 0 0 x g for 5 rain, then incubated overnight at 37 °C in 5% CO,2, atmosphere. Monocyte mediated lysis was performed as described before, but the target erythrocytes were not pretreated with papain [6]. The results were calculated according to the formula: 308
specific 5~Cr release experimental release - spontaneous release total incorporated activity
x I00
3.3. EA rosette assay H u m a n erythrocytes ( O + , R h + ) were sensitized by anti-D lgG globulin fraction (OHVI) at 37°C for 60 rain, then the 2% washed suspension was added to an equal volume of 5x10%ml lymphocytes. After 10 min incubation at 37'~( the samples were centrifuged at low speed, then stored overnight at 4°C. The number of rosette forming cells was estimated by surveying at least 200 lymphocytes in each sample. In the rosette inhibition experiments different amounts of immunoglobulins were added to the lymphocytes, the samples were incubated at room temperature for 20 rain, then the sensitized erythrocytes were added. 3.4. Preparation of lgG and fragments lgG or IgGl myeloma proteins were isolated from serum by DEAE cellulose chromatograph1,'. The IgG fraction was digested by papain. The undigested IgG was removed by gel fihration on a Sephadex GI00 column, then the Fab and Fc fragments were separated by DEAE cellulose chromatography. To obtain the pFc' fragment the IgG fraction was digested by pepsin (2 mg/100 mg protein), then the digest was fractionated on a Sepharose 6B column. 3.5. Preparation of domain deleted paraproteins Protein TIM was isolated from a myeIoma serum, as the breakthrough fraction from DEAE celhdose equilibrated and eluted with 0.01 M phosphate buffer (pH 7.0). Normal polyclona[ IgG was removed by passage over a Sepharose 4 B - p r o t e i n A column to which protein T I M does not bind. Protein SIZ was similarly isolated from a myeloma serum. Normal IgG was eluted from DEAE cellulose with 0.01 M phosphate buffer (pH 7.0) and protein SIZ by a stepwise increase of the molarity to 0.03 M. S D S - P A G E analysis demonstrated the presence of the low MT paraprotein and transferrin only [17].
4. Results
deleted ]gG
C73
LY 50
(TIM) and C73 deleted (SIZ) paraprotein on A D C C mediated by lymphocytes or monocytes 100/zl effector cells at a concentration of 3 × 106/ml were incubated with different amounts of paraproteins or IgG1. The cytotoxicity effected on papain treated or untreated target cells was determined (Fig. 1). IgG1 and CH3 domain deleted paraprotein (SIZ) inhibited both the lymphocyte and monocyte mediated ADCC. 1.2/~g of IgGl and 6/zg of SIZ inhibited 50O7o of the lymphocyte mediated ADCC, while in the monocyte mediated ADCC about 25 times less IgG1 and five times less SIZ was needed to obtain similar levels of inhibition. The CH2 domain deleted TIM had no effect on monocyte mediated ADCC and slightly inhibited the lymphocyte mediated system.
/.
/
4.1. Comparison of the effect of C'y2 deleted 40
/'
10
u)
/
X
//5
C 1
o o b
/'
/
/"
30
20
]gG1
C 7 2 deleted ]gG
,/X
3
h40 X 40 30
/
/
X
lO f
& AW
1 1001 A
g a
control
1 Effec'~cr
3
1
3
t e r g e t c~bo
Fig. 2. Comparison of ADCC activity of lymphocytes (LY) and monocytes (MO) in the presence of different amounts of [gGs at 3:1 and hi. effector:target ceil ratio. One typica~ experiment out of four. Control (x), 10 #g ( • ), 2 #g ( • ), 0.4 ~g ( • ) and 0.08 ~tg ( • ) IgG.
40 20 0 08
~oc
3
04
2
10 ~g
~-----~coo~o,
~
Figure 2 illustrates the effect of IgG1, SIZ and TIM on lymphocyte and monocyte mediated ADCC, respectively, at 3:1 and 1:1 effector to target cell ratio. 4.2. Comparison of the effect of IgGl and its
40
i
0 08
J
0 4
--
2
10 )Jg
Fig. I. Inhibition o f lymphocyte (A) and monocyte (B) mediated ADCC by C'y2 ( • ) and C'73 ( • ) domain deleted paraproteins or by IgGl ( • ) . The cytotoxic assay was carried out in the presence of different amounts of IgGs. The results were calculated taking the values of untreated control samples as 100%. Mean _+ SD of three independent experiments.
fragments on A D C C mediated by lymphocytes or monocytes In order to compare the effect of domain deleted paraproteins and that of the IgG fragments on lymphocyte or monocyte mediated ADCC, the effector cells were preincubated with 20/zg of IgG, Fab, Fc or pFc'. 20/zg IgG1 or Fc completely blocked the monocyte mediated effect, while it only partially inhibited the lymphocyte mediated ADCC. pFc fragment did not change monocyte ADCC, while inhibiting slightly the lymphocyte mediated killing (Fig. 3). 309
S'
Monocytes
Lymphocytes
deleted SIZ and IgG1 inhibited EA rosette formation with a similar efficiency, while the CH2 domain deleted TIM had no effect (Fig. 4).
50
5C
40
4C
b
30
3C
5. Discussion
5
2o
2C
The role of the C3,2 and C'~3 domains in the binding of IgG to different cell types and their role in the mediation of FcR dependent cytotoxicity (ADCC) has been widely studied [8-16]. Spiegelberg et al. [8] suggested that interaction of FcR with both C72 and C-~3 domains is essential for optimal ADCC. They also suggested that the C3,3 domain expresses a high affinity binding site for FcR on human K cells while interaction with a lower affinity C3,2 domain binding site is necessary to trigger the lytic mechanism. Klein et al. [9] suggested a major role for the C3'3 domain and a significant role for the C3,2 domain of human IgG1 in binding to the FcR of activated mouse T cells. Inhibition of ADCC by synthetic peptides representing linear sequences of IgG1 and by monoclonal antibodies specific for epitopes expressed on C3,2 or C3,3 domains led us to postulate that the C72 domain mediates the signal for lysis in A D C C by both monocytes and lymphocytes; the C3,3 domain being responsible for the binding and close contact of the effector lymphocytes and target cells [14-16]. These findings and their interpretations are in agreement with previous reports [3, 8-10]. However, since McAb specific for the C73 or the inter-C72/C73 domain regions did not inhibit monocyte mediated A D C C we assumed that the C3,3 domain has no role in this system. This agrees with the data of Woof et al. [17] who showed that neither the pFc' fragment nor a C3,2 domain deleted IgG1 paraprotein (TIM) inhibited the binding of IgG1 to the FcR of monocytes. Similarly, a C-~3 domain deleted paraprotein (SIZ) failed to inhibit, suggesting that intact Fc was essential for expression of the site recognized by monocyte FcR. In contrast, Matre and T6nder [13] have concluded that both placental and monocyte FcR recognize the C3,3 domain of IgG. In the present study inhibition of monocyte
u _u
~ ~o
1C
o_ v'l
9 "q 13
OT 110 Target
i 110
1 ~3 effector
rcltlO
Fig. 3. Effect of [gG and its fragments on ADCC mediated by lymphocytesor monocytes. Control (x), lymphocyteswere preincubated with 20 #g of lgG1 (•), Fc ( • ), pFc (•) of Fab ( • ) fragments.
4.3. Effect of the domain deleted paraproteins
SIZ and TIM on EA rosette formation of lymphocytes 2.5 × 105 monocyte-depleted effector cells were incubated with different amounts of immunoglobulins. The rosette formation of these cells with human erythrocytes sensitized with anti-D IgG is illustrated in Fig. 4. CH3 domain
12 control
i
10
8
L W
6
O
4
2
0'08
0'4
2'0
10~0 # g
Fig. 4. EA rosette inhibition by C23 domain ( • ) but not by C'~2 domain deleted ( • ) paraprotein as compared with the IgGl (•) treated samples. Lymphocyteswere preincubated with different amounts of lgGs for 30 min, then EA were added. The dashed lines show the _+ SD of the control. Mean + SD of 6 experiments. 310
and lymphocyte mediated ADCC by the paraproteins SIZ and TIM has been evaluated and compared with intact IgG Fab, Fc and pFc' fragments. Intact IgG, C3,3 domain deleted paraprotein and the Fc fragment abrogated monocyte ADCC while the C3'2 domain deleted protein TIM and the pFc' fragment had no effect. This finding also suggests that the C73 domain does not participate in monocyte ADCC. Lymphocyte mediated ADCC was inhibited by IgG, paraprotein SIZ and Fc while pFc' and paraprotein TIM inhibited only weakly. These results are in agreement with our previous findings, suggesting that both C3,2 and C3'3 domains are involved in lymphocyte ADCC, while the C3,2 domain alone is involved in monocyte ADCC. Monocyte ADCC was more sensitive to inhibition than lymphocyte ADCC, requiring concentrations of SIZ and IgG 5- and 25-fold lower, respectively, to give the same degree of inhibition. These differences can be explained by differences in the FcR binding constants of the two cell types. Human monocyte FcR binds monomer IgG1 and IgG3 with a binding affinity (Kas s 5x10 s M -1) that is three orders of magnitude greater than for lymphocyte FcR [19]. Inhibition of monocyte ADCC by IgG Fc or the C3'3 deleted protein SIZ suggests that only the C3,2 domain is recognized by monocyte FcR. The absence of inhibition with protein TIM lacking the C73 domain suggests it has no role in monocyte FcR recognition. Woof et al. [17] concluded also that the C73 domain does not participate in monocyte FcR binding; they and others have proposed that the monocyte FcR binds either to the C72 domain alone or to the interdomain region [17, 18, 22]. In the study of Woof et al. [17] the C~3 deleted protein SIZ did not inhibit the binding of IgG to monocyte FcR. There are several possible explanations for the apparent discrepancy between these and our results. These inhibition systems are different. Woof et al. [17] looked for inhibition of binding of soluble monomer IgG to monocyte FcR whereas we have tested the inhibition of ADCC where the competing ligand is IgG bound to human erythrocytes. The sensitivi-
ty of the systems may differ and there could be a distinction between C3'2 domain structures responsible for binding and mediating triggering of the lysis. The higher concentrations of protein SIZ, compared to intact IgG, required to give the same degree of A D C C inhibition suggest that the monocyte FcR binding site in the C3,2 domain of SIZ may be conformationally altered, relative to intact IgG, resulting in a lower binding affinity. In the intact molecule the C3,3 domain may contribute to maintaining the conformation of the C3,2 domain. Removal of the C3,3 domain, resulting in some loss of conformation, is consistent with data describing C'1,2 as a "soft" domain [23]. We demonstrated previously that McAb specific for epitopes of the C3'3 domain inhibited both lymphocyte ADCC and the binding of antibody sensitized erythrocytes to lymphocyte FcR. On the other hand, McAb recognizing C3,2 epitopes inhibited ADCC but not binding. The McAb that inhibited lymphocyte ADCC recognized epitopes expressed by the C3,2 deleted protein TIM or the C3,3 deleted protein SIZ [24]. We now demonstrate that while protein SIZ inhibited EA rosette formation (binding) similarly to intact IgG, protein TIM gave no inhibition. These data can be rationalized when the different modes of interaction are considered. The C~3 specific McAb may inhibit binding by direct competition for the binding site or by steric inhibition following its binding to an epitope in the same topographical region as the binding site. However, in the present assay, protein TIM is competing with the C3,3 domain of IgG as presented when bound to the erythrocyte membrane, effectively, in aggregated form. Lymphocyte FcR has a higher binding affinity for complexed than monomer IgG [I]. The inhibition of lymphocyte FcR binding by the C3,3 deleted protein SIZ appears to contradict this explanation since the FcR/C3,2 interaction is thought to be of low affinity [8, 9]. However, it is possible that while the affinity is low the dissociation constant is also low and the bound SIZ sterically blocks interaction between the lymphocyte FcR and the C3'3 domain of erythrocyte bound IgG. 311
These and previous data demonstrate a critical role f o r t h e C3,2 d o m a i n o f I g G in t r i g g e r i n g t h e lytic m a c h i n e r y o f b o t h l y m p h o c y t e a n d m o n o cyte m e d i a t e d A D C C .
References [1] Dickler, H. B. (1976) Adv. lmmunol. 24, 167-241. [2] Unkeless, J. C., Fleit, H. and Mellman, I. S. (1981) Adv. Immunol. 31, 247-270. [3] Michaelsen, T. E., Wisloff, F. and Natvig, J. B. (1975) Scand. J. Immunol. 4, 71-78. [4] MacDonald, H. R., Bonnard, G. D., Sardat, B. and Zavodnik, S. A. (1975) Scand. J. Immunol. 4, 487. [5] Parillo, J. E. and Fauci, A. S. (1977) Immunology 33, 839. [6] Yust, I., Frish, B. and Goldsher, N. (1980) Eur. J. Immunol. 10, 127 [7] McCarley, D. L., Shah, V. O. and Weiner, R. S. (1983) J. Immunol. 131, 1780. [8] Spiegelberg, H. k., Pertmann, H. and Perlmann, P. (1976) J. lmmunol. 116, 1464-1471. [9] Klein, M., Neauport-Sautes, C., Ellerson, J. R. and Fridman, W. H. (1977) J. lmmunol. 119, 1077-1088. [10] Mackennon, I. C. M., Connell, G. E. and Gotch, F. M. (1974) Immunology 26, 303-310.
312
[11] Wisloff, E, Michaelsen, T. E. and Froland, S. S. (1974) Scand. J. Immunol, 3, 29. [12] Barnett-Foster, D. E., Sjoquist, J. and Painter, R. H. (1982) Mol. Immunol. 19, 407. [13] Matre, R. and T6nder, O. (1984) Scand. J. Immunol. 19, 75-78. [14] Sgtrmay, G., Benczur, M., Patrfinyi, Gy., Klein, E., Kahn, S., Stanworth, D. and Gergely, J. (1984) Mol. Immunol. 21, 43-45. [15] Sarmay, G., Jefferis, R., Klein, E., Benczur, M. and Gergel~, J. 11985) Eur. J. Immunol., in press. [16] Gergely, J., S~indor, M., Sarmay, G. and Uher, E (1985) Trans. Biochem. Soc., in press. [17] Woof, J. M., NikJaafar, M. I., Jefferis, R. and Burton, D. (1984) Mol. Immunol. 21, 523-527. [18] Ratcliff, A. and Stanworth, D. R. (19821 Immunol. kett. 4, 215. [19] Burton, D. R. (1985) Mol. lmmunol. 22, 161-206. [20] B6yum, A. (1968) Scand. 3. Clin. Lab. Invest. 21, Suppl., 97-112. [21] Benczur, M., Gy6rffy, G., Garam, T., Varga, M., Medgyesi, G. A., S~.ndor, M. and Petr~myi, G. (1979) Immunobiology 156, 320-329. [22] Dorrington, K. J. and Klein, M. H. (1982) Mol. Immunol. 19, 1215. [23] Deisenhofer, J. 11981) Biochemistry 20, 2361-2370. [24] NikJaafar, M. I., kowe, J. A., king, N. R. and Jefferis, R. (1984) Mol. lmmunol. 21, 137.
Elsevier Imlet 749
CH2 A N D CH3 D O M A I N D E L E T E D IgG1 PARAPROTEINS INHIBIT DIFFERENTLY Fc RECEPTOR MEDIATED BINDING A N D CYTOLYSIS Gabriella S/kRMAY l, Roy JEFFERIS 2 and J~inos GERGELY 1 ~Dept. of Immunology of L. EOtvOs University, GOd 2131, Jdvorka S. 14, Hungary, and 2Dept. of Immunology, The University of Birmingham, Birmingham B15 2TJ, U.K.
(Received 31 January 1986) (Accepted 5 February 1986)
I. Summary Domain deleted paraproteins are suitable tools to study the interaction between IgG domains and Fc receptor (FcR) binding sites. The effect of the C3,2 or C73 domain deleted paraproteins was compared on antibody dependent cellular cytotoxicity (ADCC) and on FcR mediated rosette formation. The C72 domain deleted paraprotein (TIM) had no significant effect on lymphocyte or monocyte mediated ADCC, while the C3,3 domain deleted paraprotein (SIZ) inhibited both types of cytotoxicity. FcR dependent rosette formation was also inhibited by SIZ but TIM was ineffective. The data further confirm our previous results suggesting a significant role of C72 domain in the transfer of killing signal in ADCC and that of C73 domain in the high affinity binding to lymphocyte FcR.
2. Introduction Several cell types express surface receptors which bind human IgG through interactions with Key words: Fc receptor binding sites - lgG domains - anti-
body dependent cellular cytotoxicity Abbreviations: FcR, Fc receptors; ADCC, antibody depen-
dent cellular cytotoxicity;PBMC, peripheral blood mononuclear cells; EA, erythrocytescoated with antibody; DEAE, diethylaminomethyl;SDS, sodium dodecylsulphate; PAGE, polyacrylamidegel electrophoresis
defined sites of the Fc region (reviewed in [1, 2]). A subpopulation of the FcR positive cells function as effectors in the ADCC [3]. Operationally these are designated as K cells. Within this heterogeneous population [4, 5] some are lymphocytes, others are monocytes [6, 7]. K cells exhibit a high association constant for IgGl and IgG3 subclass proteins. Accordingly, these IgG molecules mediate ADCC, while IgG2 and IgG4 proteins are inactive. The precise role of FcR in activating the lytic mechanism which results in the destruction of target cells in ADCC is not known. Activation of the lysis can be inhibited by the ligands, monomeric IgG and Fc [8, 9]. The contribution of the C72 and C73 domains in the binding to FcR and the functional role in ADCC has been studied extensively [8-16]. We have previously demonstrated that both the C72 and C3,3 domains of the sensitizing antibody are necessary for obtaining an optimal ADCC effect with lymphocytes. We have proposed a model in which two binding sites are necessary and have different impact. FcR/C73 domain interaction contributes to the close contact between effector and target ceils, while the FcR/C72 is crucial for triggering the lytic process for which the linear sequence Lys 2 7 4 - A r g 301 of the C72 domain seems to be responsible. In contrast, monocyte mediated ADCC depended only on one high avidity binding site of FcR interacting with the C72 domain of IgG [14-161. These results are consistent with other recent reports; however, earlier studies claimed that the 307
C'73 domain does have a functional role in monocyte mediated A D C C [17--19]. We have further investigated the role of (7-},2 and C~.3 domains in the interaction between effector and target cells in monocyte and lymphocyte mediated A D C C systems. Our results confirm and extend the hypothesis that the C~2 domain triggers the lyric function in A D C C mediated both by lymphocytes and by monocytes. The results suggest that lymphocyte FcR has a high affinity binding site interacting with the C-,,3 domain and a lower affinity binding site reacting with the C')2 domain of the antigen complexed lgG. 3. Materials and Methods
3.1. Separation of lymphocytes Lymphocytes from the blood of healthy donors were prepared according to the method of B6yum [20]. The peripheral blood mononuclear cell (PBMC) suspension was used in the antibody dependent cytotoxic assay. Typically, it contained 1 5 - 2 5 % monocytes as indicated by latex phagocytosis. For the assay with lymphocytes, monocytes were depleted by iron phagocytosis and plastic adherence. The resulting suspension contained less than 2% monocytes. 3.2. Antibody dependent cellular O'totoxicio'
(ADCC) Lymphocyte mediated cytotoxicity was performed as described earlier [21]. Briefly, washed R h + human erythrocytes were treated with papain (1%) at 20°C for 10 min, labelled with 400 p,Ci = 14.8 MBq sodium 5~chromate (Amersham, U.K.), and sensitized with anti-D lgG 250 ~g/ml (OHVI, Hungary). The assay was carried out in U-shaped wells (Cooke microtiter plate). Each sample was tested in triplicate. Effector cells (PBMC) were added to the targets at different ratios. The plates were centrifuged at 2 0 0 x g for 5 rain, then incubated overnight at 37 °C in 5% CO,2, atmosphere. Monocyte mediated lysis was performed as described before, but the target erythrocytes were not pretreated with papain [6]. The results were calculated according to the formula: 308
specific 5~Cr release experimental release - spontaneous release total incorporated activity
x I00
3.3. EA rosette assay H u m a n erythrocytes ( O + , R h + ) were sensitized by anti-D lgG globulin fraction (OHVI) at 37°C for 60 rain, then the 2% washed suspension was added to an equal volume of 5x10%ml lymphocytes. After 10 min incubation at 37'~( the samples were centrifuged at low speed, then stored overnight at 4°C. The number of rosette forming cells was estimated by surveying at least 200 lymphocytes in each sample. In the rosette inhibition experiments different amounts of immunoglobulins were added to the lymphocytes, the samples were incubated at room temperature for 20 rain, then the sensitized erythrocytes were added. 3.4. Preparation of lgG and fragments lgG or IgGl myeloma proteins were isolated from serum by DEAE cellulose chromatograph1,'. The IgG fraction was digested by papain. The undigested IgG was removed by gel fihration on a Sephadex GI00 column, then the Fab and Fc fragments were separated by DEAE cellulose chromatography. To obtain the pFc' fragment the IgG fraction was digested by pepsin (2 mg/100 mg protein), then the digest was fractionated on a Sepharose 6B column. 3.5. Preparation of domain deleted paraproteins Protein TIM was isolated from a myeIoma serum, as the breakthrough fraction from DEAE celhdose equilibrated and eluted with 0.01 M phosphate buffer (pH 7.0). Normal polyclona[ IgG was removed by passage over a Sepharose 4 B - p r o t e i n A column to which protein T I M does not bind. Protein SIZ was similarly isolated from a myeloma serum. Normal IgG was eluted from DEAE cellulose with 0.01 M phosphate buffer (pH 7.0) and protein SIZ by a stepwise increase of the molarity to 0.03 M. S D S - P A G E analysis demonstrated the presence of the low MT paraprotein and transferrin only [17].
4. Results
deleted ]gG
C73
LY 50
(TIM) and C73 deleted (SIZ) paraprotein on A D C C mediated by lymphocytes or monocytes 100/zl effector cells at a concentration of 3 × 106/ml were incubated with different amounts of paraproteins or IgG1. The cytotoxicity effected on papain treated or untreated target cells was determined (Fig. 1). IgG1 and CH3 domain deleted paraprotein (SIZ) inhibited both the lymphocyte and monocyte mediated ADCC. 1.2/~g of IgGl and 6/zg of SIZ inhibited 50O7o of the lymphocyte mediated ADCC, while in the monocyte mediated ADCC about 25 times less IgG1 and five times less SIZ was needed to obtain similar levels of inhibition. The CH2 domain deleted TIM had no effect on monocyte mediated ADCC and slightly inhibited the lymphocyte mediated system.
/.
/
4.1. Comparison of the effect of C'y2 deleted 40
/'
10
u)
/
X
//5
C 1
o o b
/'
/
/"
30
20
]gG1
C 7 2 deleted ]gG
,/X
3
h40 X 40 30
/
/
X
lO f
& AW
1 1001 A
g a
control
1 Effec'~cr
3
1
3
t e r g e t c~bo
Fig. 2. Comparison of ADCC activity of lymphocytes (LY) and monocytes (MO) in the presence of different amounts of [gGs at 3:1 and hi. effector:target ceil ratio. One typica~ experiment out of four. Control (x), 10 #g ( • ), 2 #g ( • ), 0.4 ~g ( • ) and 0.08 ~tg ( • ) IgG.
40 20 0 08
~oc
3
04
2
10 ~g
~-----~coo~o,
~
Figure 2 illustrates the effect of IgG1, SIZ and TIM on lymphocyte and monocyte mediated ADCC, respectively, at 3:1 and 1:1 effector to target cell ratio. 4.2. Comparison of the effect of IgGl and its
40
i
0 08
J
0 4
--
2
10 )Jg
Fig. I. Inhibition o f lymphocyte (A) and monocyte (B) mediated ADCC by C'y2 ( • ) and C'73 ( • ) domain deleted paraproteins or by IgGl ( • ) . The cytotoxic assay was carried out in the presence of different amounts of IgGs. The results were calculated taking the values of untreated control samples as 100%. Mean _+ SD of three independent experiments.
fragments on A D C C mediated by lymphocytes or monocytes In order to compare the effect of domain deleted paraproteins and that of the IgG fragments on lymphocyte or monocyte mediated ADCC, the effector cells were preincubated with 20/zg of IgG, Fab, Fc or pFc'. 20/zg IgG1 or Fc completely blocked the monocyte mediated effect, while it only partially inhibited the lymphocyte mediated ADCC. pFc fragment did not change monocyte ADCC, while inhibiting slightly the lymphocyte mediated killing (Fig. 3). 309
S'
Monocytes
Lymphocytes
deleted SIZ and IgG1 inhibited EA rosette formation with a similar efficiency, while the CH2 domain deleted TIM had no effect (Fig. 4).
50
5C
40
4C
b
30
3C
5. Discussion
5
2o
2C
The role of the C3,2 and C'~3 domains in the binding of IgG to different cell types and their role in the mediation of FcR dependent cytotoxicity (ADCC) has been widely studied [8-16]. Spiegelberg et al. [8] suggested that interaction of FcR with both C72 and C-~3 domains is essential for optimal ADCC. They also suggested that the C3,3 domain expresses a high affinity binding site for FcR on human K cells while interaction with a lower affinity C3,2 domain binding site is necessary to trigger the lytic mechanism. Klein et al. [9] suggested a major role for the C3'3 domain and a significant role for the C3,2 domain of human IgG1 in binding to the FcR of activated mouse T cells. Inhibition of ADCC by synthetic peptides representing linear sequences of IgG1 and by monoclonal antibodies specific for epitopes expressed on C3,2 or C3,3 domains led us to postulate that the C72 domain mediates the signal for lysis in A D C C by both monocytes and lymphocytes; the C3,3 domain being responsible for the binding and close contact of the effector lymphocytes and target cells [14-16]. These findings and their interpretations are in agreement with previous reports [3, 8-10]. However, since McAb specific for the C73 or the inter-C72/C73 domain regions did not inhibit monocyte mediated A D C C we assumed that the C3,3 domain has no role in this system. This agrees with the data of Woof et al. [17] who showed that neither the pFc' fragment nor a C3,2 domain deleted IgG1 paraprotein (TIM) inhibited the binding of IgG1 to the FcR of monocytes. Similarly, a C-~3 domain deleted paraprotein (SIZ) failed to inhibit, suggesting that intact Fc was essential for expression of the site recognized by monocyte FcR. In contrast, Matre and T6nder [13] have concluded that both placental and monocyte FcR recognize the C3,3 domain of IgG. In the present study inhibition of monocyte
u _u
~ ~o
1C
o_ v'l
9 "q 13
OT 110 Target
i 110
1 ~3 effector
rcltlO
Fig. 3. Effect of [gG and its fragments on ADCC mediated by lymphocytesor monocytes. Control (x), lymphocyteswere preincubated with 20 #g of lgG1 (•), Fc ( • ), pFc (•) of Fab ( • ) fragments.
4.3. Effect of the domain deleted paraproteins
SIZ and TIM on EA rosette formation of lymphocytes 2.5 × 105 monocyte-depleted effector cells were incubated with different amounts of immunoglobulins. The rosette formation of these cells with human erythrocytes sensitized with anti-D IgG is illustrated in Fig. 4. CH3 domain
12 control
i
10
8
L W
6
O
4
2
0'08
0'4
2'0
10~0 # g
Fig. 4. EA rosette inhibition by C23 domain ( • ) but not by C'~2 domain deleted ( • ) paraprotein as compared with the IgGl (•) treated samples. Lymphocyteswere preincubated with different amounts of lgGs for 30 min, then EA were added. The dashed lines show the _+ SD of the control. Mean + SD of 6 experiments. 310
and lymphocyte mediated ADCC by the paraproteins SIZ and TIM has been evaluated and compared with intact IgG Fab, Fc and pFc' fragments. Intact IgG, C3,3 domain deleted paraprotein and the Fc fragment abrogated monocyte ADCC while the C3'2 domain deleted protein TIM and the pFc' fragment had no effect. This finding also suggests that the C73 domain does not participate in monocyte ADCC. Lymphocyte mediated ADCC was inhibited by IgG, paraprotein SIZ and Fc while pFc' and paraprotein TIM inhibited only weakly. These results are in agreement with our previous findings, suggesting that both C3,2 and C3'3 domains are involved in lymphocyte ADCC, while the C3,2 domain alone is involved in monocyte ADCC. Monocyte ADCC was more sensitive to inhibition than lymphocyte ADCC, requiring concentrations of SIZ and IgG 5- and 25-fold lower, respectively, to give the same degree of inhibition. These differences can be explained by differences in the FcR binding constants of the two cell types. Human monocyte FcR binds monomer IgG1 and IgG3 with a binding affinity (Kas s 5x10 s M -1) that is three orders of magnitude greater than for lymphocyte FcR [19]. Inhibition of monocyte ADCC by IgG Fc or the C3'3 deleted protein SIZ suggests that only the C3,2 domain is recognized by monocyte FcR. The absence of inhibition with protein TIM lacking the C73 domain suggests it has no role in monocyte FcR recognition. Woof et al. [17] concluded also that the C73 domain does not participate in monocyte FcR binding; they and others have proposed that the monocyte FcR binds either to the C72 domain alone or to the interdomain region [17, 18, 22]. In the study of Woof et al. [17] the C~3 deleted protein SIZ did not inhibit the binding of IgG to monocyte FcR. There are several possible explanations for the apparent discrepancy between these and our results. These inhibition systems are different. Woof et al. [17] looked for inhibition of binding of soluble monomer IgG to monocyte FcR whereas we have tested the inhibition of ADCC where the competing ligand is IgG bound to human erythrocytes. The sensitivi-
ty of the systems may differ and there could be a distinction between C3'2 domain structures responsible for binding and mediating triggering of the lysis. The higher concentrations of protein SIZ, compared to intact IgG, required to give the same degree of A D C C inhibition suggest that the monocyte FcR binding site in the C3,2 domain of SIZ may be conformationally altered, relative to intact IgG, resulting in a lower binding affinity. In the intact molecule the C3,3 domain may contribute to maintaining the conformation of the C3,2 domain. Removal of the C3,3 domain, resulting in some loss of conformation, is consistent with data describing C'1,2 as a "soft" domain [23]. We demonstrated previously that McAb specific for epitopes of the C3'3 domain inhibited both lymphocyte ADCC and the binding of antibody sensitized erythrocytes to lymphocyte FcR. On the other hand, McAb recognizing C3,2 epitopes inhibited ADCC but not binding. The McAb that inhibited lymphocyte ADCC recognized epitopes expressed by the C3,2 deleted protein TIM or the C3,3 deleted protein SIZ [24]. We now demonstrate that while protein SIZ inhibited EA rosette formation (binding) similarly to intact IgG, protein TIM gave no inhibition. These data can be rationalized when the different modes of interaction are considered. The C~3 specific McAb may inhibit binding by direct competition for the binding site or by steric inhibition following its binding to an epitope in the same topographical region as the binding site. However, in the present assay, protein TIM is competing with the C3,3 domain of IgG as presented when bound to the erythrocyte membrane, effectively, in aggregated form. Lymphocyte FcR has a higher binding affinity for complexed than monomer IgG [I]. The inhibition of lymphocyte FcR binding by the C3,3 deleted protein SIZ appears to contradict this explanation since the FcR/C3,2 interaction is thought to be of low affinity [8, 9]. However, it is possible that while the affinity is low the dissociation constant is also low and the bound SIZ sterically blocks interaction between the lymphocyte FcR and the C3'3 domain of erythrocyte bound IgG. 311
These and previous data demonstrate a critical role f o r t h e C3,2 d o m a i n o f I g G in t r i g g e r i n g t h e lytic m a c h i n e r y o f b o t h l y m p h o c y t e a n d m o n o cyte m e d i a t e d A D C C .
References [1] Dickler, H. B. (1976) Adv. lmmunol. 24, 167-241. [2] Unkeless, J. C., Fleit, H. and Mellman, I. S. (1981) Adv. Immunol. 31, 247-270. [3] Michaelsen, T. E., Wisloff, F. and Natvig, J. B. (1975) Scand. J. Immunol. 4, 71-78. [4] MacDonald, H. R., Bonnard, G. D., Sardat, B. and Zavodnik, S. A. (1975) Scand. J. Immunol. 4, 487. [5] Parillo, J. E. and Fauci, A. S. (1977) Immunology 33, 839. [6] Yust, I., Frish, B. and Goldsher, N. (1980) Eur. J. Immunol. 10, 127 [7] McCarley, D. L., Shah, V. O. and Weiner, R. S. (1983) J. Immunol. 131, 1780. [8] Spiegelberg, H. k., Pertmann, H. and Perlmann, P. (1976) J. lmmunol. 116, 1464-1471. [9] Klein, M., Neauport-Sautes, C., Ellerson, J. R. and Fridman, W. H. (1977) J. lmmunol. 119, 1077-1088. [10] Mackennon, I. C. M., Connell, G. E. and Gotch, F. M. (1974) Immunology 26, 303-310.
312
[11] Wisloff, E, Michaelsen, T. E. and Froland, S. S. (1974) Scand. J. Immunol, 3, 29. [12] Barnett-Foster, D. E., Sjoquist, J. and Painter, R. H. (1982) Mol. Immunol. 19, 407. [13] Matre, R. and T6nder, O. (1984) Scand. J. Immunol. 19, 75-78. [14] Sgtrmay, G., Benczur, M., Patrfinyi, Gy., Klein, E., Kahn, S., Stanworth, D. and Gergely, J. (1984) Mol. Immunol. 21, 43-45. [15] Sarmay, G., Jefferis, R., Klein, E., Benczur, M. and Gergel~, J. 11985) Eur. J. Immunol., in press. [16] Gergely, J., S~indor, M., Sarmay, G. and Uher, E (1985) Trans. Biochem. Soc., in press. [17] Woof, J. M., NikJaafar, M. I., Jefferis, R. and Burton, D. (1984) Mol. Immunol. 21, 523-527. [18] Ratcliff, A. and Stanworth, D. R. (19821 Immunol. kett. 4, 215. [19] Burton, D. R. (1985) Mol. lmmunol. 22, 161-206. [20] B6yum, A. (1968) Scand. 3. Clin. Lab. Invest. 21, Suppl., 97-112. [21] Benczur, M., Gy6rffy, G., Garam, T., Varga, M., Medgyesi, G. A., S~.ndor, M. and Petr~myi, G. (1979) Immunobiology 156, 320-329. [22] Dorrington, K. J. and Klein, M. H. (1982) Mol. Immunol. 19, 1215. [23] Deisenhofer, J. 11981) Biochemistry 20, 2361-2370. [24] NikJaafar, M. I., kowe, J. A., king, N. R. and Jefferis, R. (1984) Mol. lmmunol. 21, 137.