In vivo behavior of monoclonal antibodies
Eur. J. Immunol. 1991. 21: 943-950
Nasim Yousafo, Jonathan C. Howardo and Bryan D. WilliamsO+ Department of Rheumatology, University Hospital of Waleso, Cardiff and Department of Immunology, Agricultural and Food Research Council Institute of Animal Physiology and Genetics Researcho, Babraham, Cambridge
Targeting behavior of rat monoclonal IgG antibodies in vivo: role of antibody isotype, specificity and the target cell antigen density* The studies described in this report were designed to investigate factors that could influence the behavior of erythrocytes following their interaction with monoclonal antibodies (mAb) in a fully homologous experimental opsonization system in vivo.The clearance profiles and tissue distribution of target erythrocytes were examined in both normal and decomplemented rats preinjected with rat IgG2, or IgG2b mAb directed against the same or different sites on RTIAa, the classical class I major histocompatibility complex antigen of the DA rat. Complement played a major role in augmenting the clearance and promoting hepatic sequestration of target erythrocytes in rats preinjected with IgG2, mAb directed against the S site. In contrast, an intact complement system was not an essential requirement for erythrocyte clearance when S site-specific IgGZb mAb were used. With each antibody tested, (DA x PVG)F1 cells, expressing about half as much antigen, were removed significantly slower than DA erythrocytes, this finding being more pronounced when the animals had been preinjected with mAb of the IgG2, isotype. A comparison of the tissue distribution of DA and (DA x PVG)F1 erythrocytes indicated that hepatic uptake was greater for target cells expressing higher antigen density. A considerable degree of heterogeneity was observed in the in vivo behavior of the target erythrocytes with three groups of IgG2b mAb that recognized different sites on the class I molecule. The S site-specific IgG2b mAb were much more efficient in the hepatic Fc receptor-mediated clearance system than were the P site-directed mAb of the same subclass. Our results suggest that antibody specificity may also be a contributory factor, in addition to antibody isotype and target cell antigen density, in determining the fate of target cells in vivo.
1 Introduction The potential use of mAb as specific therapeutic agents has attracted widespread interest. There is now considerable evidence which suggests that passively administered mAb can modify or alter immune responsiveness and disease processes in both humans [l-41, and experimental animals [5-91. It is also anticipated that human mAb against the Rhesus antigen D may replace the polyclonal anti-D used in the prophylaxis of hemolytic disease of the newborn [lo]. However, it is already apparent from the literature that there are marked differences in the effectiveness of treatment with different mAb even when they are directed against the same antigen [3, 5, 11-14]. The effector mechanisms responsible for the outcome of an efficient mAb serotherapy remain unclear. Although the relevance of antibody isotype in the effectiveness of mAb therapy has been recognized in several studies [3, 15, 161,
[I 90311
*
943
This work was supported by a grant from the Arthritis and Rheumatism Council of Great Britain and in part by USPHS Project Grant CA 34913.
Correspondence: Nasim Yousaf, Department of Immunology, University College and Middlesex School of Medicine, Arthur Stanley House, 40-50 Tottenham Street, London W1P 9PG, GB Abbreviation: CVF Cobra venom factor 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991
there are suggestions that antibody specificity and the ability of some antibodies to interact with the complement system may also influence the therapeutic efficacy of mAb in experimental animals [ ll] . More recently, the therapeutic potential of an antibody has been correlated with the antigen density on the target cell surface [17]. It is therefore clearly important to establish the factors that determine the efficiency of mAb serotherapy. If cell depletion is desirable to achieve optimum therapeutic effects [2, 3, lo], then it would be important to maximize the effector mechanism for eliminating the target cells in vivo. It is likely that opsonization of the target cells and their subsequent removal by the mononuclear phagocyte system or their destruction by C-dependent lysis may be critical events in determining the therapeutic usefulness of mAb selected for immunotherapy. The Fc region-dependent effector functions of both human and rat Ig have been examined in several in virro experimental systems. Studies with human anti-Rh(D) mAb not only indicate the differences in the ability of different IgG isotypes to promote erythrocyte binding to M@ [18], but also suggest that mAb of the same subclass can vary significantly in their functional activities [19, 201. Similarly, a comparison of a wide range of rat mAb of different isotypes has revealed that rat IgG2, is considerably less efficient than IgG2b in mediating the effector functions such as C lq binding, C activation [21-241, antibody-dependent cell-mediated cytotoxicity [25], and M@ FcR binding [26]. However, an analysis of the functional properties of rat mAb in vivo, under fully homologous conditions, has been lacking.
+
014-2980/91/0404-0943$3.50 .25/0
944
Eur. J. Immunol. 1991. 21: 943-950
N. Yousaf, J. C. Howard and B. D. Williams
The experiments presented in this report were undertaken in an attempt to define the variables likely to be associated with efficient mAb serotherapy. The expression of MHC class I antigens on rat erythrocytes and the availability of a large number of rat mAb directed against various epitopes on the rat RTIAa class I antigen [27, 281, allowed us to investigate the factors which may influence the targeting behavior of mAb in vivo. In our experimental system, we examined the clearance profiles and tissue distribution of target erythrocytes following their infusion into normal and decomplemented rats preinjected with mAb of the IgG2, or IgGZb subclasses directed against the same or different sites on the RTIAa molecule. Since the RTIAa antigen level on erythrocytes from (DA x PVG)F1 rats is 50% as compared to that on the DA erythrocytes, we were able to investigate the effects of reducing the target cell antigen density for each of the 19 different mAb used in our study.
(Na2Wr04, The Radiochemical Centre, Amersham, GB) essentially as described previously [31], except that 3.7 MBq (100 pCi) of 99Tcand 50 pCi (1.85 MBq) of W r were used, respectively, to label 100 p1 of packed (DA x PVG)Fl erythrocytes and DA erythrocytes. Following the labeling procedure the cells were washed three times and resuspended in saline at a concentration of 10% (v/v) . 2.5 In vivo clearance of target erythrocytes
The procedure adopted for measuring the clearance of radiolabeled rat erythrocytes in vivo was essentially as described previously [31]. Briefly, equal volumes of the DA and (DA x PVG)Fl cell suspensions (lo%, v/v), labeled with different isotopes, were first mixed and 200 pl of this mixture were then injected into the tail vein of PVG rats. Blood samples (20 pl) were taken from the tail at various 2 Materials and methods time intervals and the radioactivity associated with the cells and plasma was measured separately in an LKB (Bromma, 2.1 Animals Sweden) Compu Gamma. The animals were killed 90 min after injection of the cells and the radioactivity within Male rats of the PVG (RTIC) and DA (RTla) strains various organs was determined. Appropriate corrections weighing 180-200 g were obtained from Bantin and King- were made for the decay of radioisotopes and for the man Ltd. (Hull, GB). (DA x PVG)Fl rats were either bred cross-over between channels when two isotopes were used in the Department of Immunology (AFRC Institute of in the same animal. The survival of radiolabeled erythroAnimal Physiology and Genetics Research, Cambridge, cytes after their injection into normal rats has been GB), or purchased from Bantin and Kingman Ltd. Splen- reported in earlier studies [31]. mAb (1 or 2 ml) were ectomized or sham-splenectomized animals were used administered (i.v.) into rats 15-20 min prior to the injection 4 weeks after surgery. of radiolabeled erythrocytes, and the in vivo behavior of the target cells was then assessed. Preliminary experiments revealed no major differences between these two doses, and 2.2 Decomplementation of animals with cobra venom it was assumed, therefore, that effective saturation in vivo factor could be achieved with 1 ml of antibody. However, some differences seen in these studies apparently related to the Cobra 'enom factor (CVF) was purified from nub variation between individual animals since a lower dose (Sigma Company Ltd*9 GB) (0.5 ml) was equally effective for several of the mAb according to the method of Lachmann and Hobart [20]. used. Rats were decomplemented by the i.v. injection of 30 U of CVF24 h prior to the experiment.The C3 levels at the time of study were < 5% of that seen in normal rat plasma. 2.6 ~ ~of data ~ l ~ ~ i 2.3 mAb Rat mAb were directed against RTIAa, the MHC class I antigen of the DA rat. The derivation of these mAb and their in v i m binding characteristics have been reported elsewhere [27, 28, 301. In the present study, mAb of the IgG2, and IgG2b subclasses with specificity for the same or different sites on the RTIAd molecule were used. All mAb were in the form of tissue culture SN containing 10% FCS and 0.1% sodium azide, with specific antibody concentration being about 20 pg/ml. These reagents were dialyzed extensively against PBS, passed through 0.22-pm Millipore filter (Millipore S.A., Molsheim, France), and then stored in appropriate aliquots at -20 "C until required. Before their use in the clearance studies any insoluble material was removed by centrifugation at 1600 x g.
The data were expressed as the percentage of radioactivity remaining in circulation at various times after the injection of the cells.The radioactivity present at 1 min was taken to be 100%. The initial half-clearance time (Tln), the time taken for 50% of the radiolabeled cells to leave the circulation, was calculated for each animal, and where appropriate, linear regression analysis was applied for its assessment. The significance of differences observed between the various groups was analyzed using Student's r-test.
3 Results 3.1 In vivo behavior of target erythrocytes with S site-specific IgGh mAb
S site-specific mAb of the IgGz, isotype removed radiolabeled DA and ( D A x PVG)Fl erythrocytes from the circulation of normal rats in a non-linear fashion. Typical Rat erythrocytes were radiolabeled with 99Tc(Na299Tc04, clearance profiles with two different IgG2, mAb (JY1/174 University Hospital of Wales, Cardiff, GB) or 51Cr and JY 1/163) are shown in Fig. 1.The clearance profiles of
2.4 Labeling of erythrocytes with V
c and 51Cr
~
Eur. J. Immunol. 1991. 21: 943-950
In vivo behavior of monoclonal antibodies
945
Table 1. In vivo behavior of DA erythrocytes in normal and CVF-treated rats preinjected with various mAb Normal rats') % Iniected radioactivityd) Splken Liver (in)
Antibody nh)
S site IgGh JY11174 JY11184 W15S JY3/109 JY11163 S site IgGlb JY3l223 N1/98 JYll232 JY3l208 P site IgGZb JYm3 R3/13 WlOP
4
13.4 f 1.2 18.7 f 3.7 7.2 f 1.4 17.4 f 2.4 20.9 f 0.5
54.9 f 2.2 41.8 f 4.4 42.7 f 1.0 52.8 f 2.5 29.8 f 4.9
3 3 3 3 3
51.7 f 13.0 44.7f 8.7 70.0f 4.0 54.7f 9.9 90.0 rf: 10.0
42.6 f 4.0 34.0f5.6 35.5 f 2.1 41.0f2.1 29.0 f 2.5
9.9k 1.1 12.6 k 4.1 12.0 f 2.2 10.3f 2.1 5.5 f 0.5
13.6 f 0.6 19.3 f 1.6 21.9 f 1.6 24.5 f 5.1
32.0 f 1.3 33.9 f 0.8 35.7 f 1.6 39.3 f 5.6
44.4 f 1.5 31.1 f 1.6 35.8 f 2.3 26.0 f 8.0
5 7 5 4
15.0f 13.4f 15.6f 21.0f
3.6 1.5 2.4 5.2
41.0f4.7 37.4f2.9 39.0f5.2 41.1 f 8 . 3
23.9 f 4.7 27.5 f 3.2 30.2 +- 4.9 19.1 f 8.6
4 7 2
15.1 f 3.8 23.6f 1.1
30.7 f 3.4 19.4 f 0.7 28.0
3 7 3
22.0f 2.1
30.8
37.0 f 2.8 38.1 f 1.7 43.8
31.8f 2.8 40.7f 5.9
60.1 f 3 . 8 48.6f3.4 60.1 f 3 . 8
17.5 f 5.3 4.4 f 0.4 4.3 f 0.6
4 3 3 6
25.0 f 3.9 14.3 f 1.9 15.7 f 2.8
44.2 f 8.7 38.6 f 5.3 41.9 rf: 9.4
15.8 f 4.5 34.3 f 6.4 29.5 f 9.9
4 3 3
28.lf 3.4 17.2f 1.7 16.8f 1.2
61.9f4.1 62.0f 1.4 53.9f5.1
7.5 f 1.8 13.1 f 3.7 21.8 f 6.3
2.0 f 0.2
5.5 rf: 0.2
6 4 4 5
8 5
4 .
None a) b) c) d)
6.5 rf: 0.3 11.9 f 1.9 11.2 f 2.3 12.8 f 2.4 45.5 f 8.4
5
Q site IgGzb JY1/99 JY 11116 YR1/100
CVF-treated ratsa) Tlnc) YO Injected radioactivityd) (min) Spleen Liver
nb)
TIR')
Data are given as mean values & SE for each group. The number of animals in each group. The initial half-clearance time (Tin) is expressed in minutes. Localization within the spleen and liver is expressed as the percentage of the radioactivity administered.
the target cells with other three IgG2, mAb (not shown) were intermediate between the extreme examples shown in Fig. 1. The results obtained in these experiments are summarized in Table 1 (for DA cells) and Table 2 [for (DA x PVG)Fl cells]. In general, DA cells were cleared rapidly with an initial Tl0 of less than 13 min (Fig. 1 and Table 1). Although the overall clearance profiles of these
cells were somewhat variable with the various IgGz, antibodies employed, the importance of C in their in vivo clearance is demonstrated clearly in Fig. 1. C depletion not only led to a change in the clearance profile, but it also significantly delayed the clearance rate of the target cells with all the S site-specific IgGh mAb examined (Table 1). Cell lysis did not occur to any significant degree in normal
JV 1/174
M
a n 8
. \
o
10
2d
b
i
rb
2;
3.2
do
48
de
ek
,
J 7s
o wn)
Figure I. Clearance profiles of DA and (DAxPVG)F, erythrocvtes in normal (-) and CVF-treated (---) rats preinjectedwiths site-specificIgG2, antibodies.
a
Ib
24
j2
io
is
de
04
;2
I
Eur. J. Immunol. 1991. 21: 943-950
N. Yousaf, J. C. Howard and B. D. Williams
946
Table 2. In vivo behavior of (DA x PVG)FI erythrocytes in normal and CVF-treated rats preinjected with various mAb
Antibodv
Normal ratsa) % Iniected radioactivityd) Tld) Liver (min) Spl;en
IlhJ
S site IgGh JY11174 JY11184 R2/15S JY3/109 JY11163 S site IgGZh JY3/223 JYlN8 JYlD32 JY3R08
4 5 4 4
56.7 f 14.7 77.2 f 11.2 >WJ 54.6+ 4.4
4
> W'
CVF-treated ratsa) % Injected radioactivityd) (min) Spleen Liver
TI#
nh)
13.8 f 0.4 16.6 f 3.4 4.3 f 1.5 19.4 f 3.0 13.5 f 1.6
20.2 f 3.2 11.3 f 1.3 14.9 f 2.0 16.1 f 1.5 6.3 f 1.1
3 2 3 3 2
> 9oe) > 9oe) > W) > W) > W)
18.8 f 4.1 20.8 15.7 f 1.9 19.6 f 1.8 10.0
4.2 f 0.4 3.8 4.2 f 0.2 4.2 f 0.2 3.6
17.5 f 1.3 11.9f0.9 15.3 f 2.0 6.4 f 2.4
4 4
29.9 f 4.6 23.2 f 2.4 24.5 f 2.7 43.8 f 8.5
39.8 f 2.7 39.4 f 1.3 39.0 f 3.8 37.4 f 2.9
6.2 f 0.6 9.2 f 1.6 12.1 f 3.1 6.0 f 2.4
5 7 5 4
26.9f 36.6f 34.1 f 48.2f
7.5
40.2f 1.3 35.8 f 1.6 37.4 f 1.9 40.5 f 1.2
P site IgGZb JYm3 RW3 R2/10P
4 4
29.1 f 8.7 46.3f 4.1 57.0
45.6 f 1.7 39.4 f 1.6 37.4
9.6 f 1.1 6.7 f 0.7 5.9
3 2 3
33.0 f 2.6 55.0 75.5 f 16.4
48.5 f 3.5
5.4 f 0.7
38.2 32.1 f 6.6
3.0 3.6 k 0.7
0 site lgG2h JY 1/99 JY11116 YRlllOO
4 3
47.8f 3.8 25.7f 2.4 M.Of 2.6
31.5 f 2.6 46.3 f 3.1 42.2 f 2.4
4.9 f 0.9 11.1 f 3.6 9.1 f 3.8
4
50.0f 8.8 25.0 f 2.5 31.2 f 0.4
38.0 f 3.4 54.4 f 4.8 45.9 f 3.1
3.4 f 0.3 4.0 f 0.7 5.8f 1.6
a) b) c) d) e)
2
3
0.8 2.0 2.0
6 5
3 3
Results are shown as mean values k SE for each group. The number of animals in each group. The initial half-clearance time is expressed in minutes. Localization within the spleen and liver is expressed as the percentage of the radioactivity administered. The radioactivity remaining in circulation at 90 min, expressed as percent 1-min counts, ranged from 51% to 69% for these groups of animals.
rats, there being no increase in the level of free isotope detected in plasma above that seen in the control animals which had received saline.
Specific accumulation of radioactivity was seen in the spleen and the liver, and distribution of the target erythrocytes within these tissues is summarized in Table 1. The
uptake of DA cells in normal rats preinjected with JY1/174, JY1/184, R2/15S and JY3/109 was predominantly hepatic, although significant localization also occurred in the spleen with most of the S site-specific IgG2, mAb. In CVF-treated rats, however, spleen was the primary site of accumulation of the target cells with each of the five IgG2, mAb used (Table 1).
M Cbll.
--I J
I
1
, 0
8
16
24
32
40
48
56
64
72
0
T i m (mh)
8
16
24
32
40
48
56
64
72
Figure2. Clearance profiles of DA and (DA x PVG)F, erythrocytes in normal (-) and CVF-treated (---) animals that had been passively infused with S site-directed mAb of the IgG2h isotype.
In vivo behavior of monoclonal antibodies
Eur. J. Immunol. 1991. 21: 943-950
The effect of lower antigen density on the surface of the target cells and their subsequent behavior in vivo was examined by the administration of (DA x PVG)F1 cells into animals that had received the S site-directed IgG2, mAb, and the data from both normal and CVF-treated rats are shown in Fig. 1 and Table2. With each antibody, (DA x PVG)FI erythrocytes were always cleared at a much slower rate than DA cells. Once again, C depletion markedly influenced the clearance profiles of (DA x PVG)Fl cells and their subsequent localization within the spleen and liver. In CVF-treated rats spleen was the only site of localization of these cells (Table 2). 3.2 In vivo behavior of target erythrocytes with S site-specific IgGa mAb
The clearance of DA and (DA x PVG)Fl erythrocytes by IgGzb mAb directed against the S site (Fig. 2,Tables 1 and 2) contrasted strongly with that shown above for IgG2, mAb. In normal rats the clearance of both DA and (DA x PVG)FI cells was more linear (Fig. 2). Furthermore, with mAb of the IgG2b isotype, the presence of an intact C system was not essential for erythrocyte clearance. Indeed, in the case of JY1/98 antibody, the initial Tln was achieved at a somewhat faster rate for both DA and (DA x PVG)Fl erythrocytes in CVF-treated rats as compared with the normal group (Fig. 2). A similar change in the clearance rate was also seen with JY1/232 mAb (Table 2). The splenic and hepatic localization of DA erythrocytes in both normal and CVF-treated rats preinjected with various S site-specific IgG2b mAb is shown in Table 1. In both groups, these cells accumulated in the spleen as well as the liver, and in general, C depletion did not produce any major changes in the level of radioactivity within these tissues. The splenic accumulation of cells was much more marked than that seen with IgG2, mAb directed against the S site. Fig. 2 also demonstrates the influence of reducing the antigen density on the target erythrocytes in both normal and CVF-treated rats that had been preinjected with different S site-directed IgGZb mAb. The clearance of (DA x PVG)Fl erythrocytes from the circulation of normal rats was significantly prolonged (Table 2), the initial T m with each antibody being approximately twice as long as that seen for the DA cells (Table 1). With all the S sitespecific IgG2b mAb, spleen was the predominant site of sequestration of (DA x PVG)Fl erythrocytes in both normal and CVF-treated animals (Table 2). Similar results were obtained with other IgG2b mAb (JYU75.1, JY3/50, JY1/134, JY3/141) directed against the Ssite (data not shown).
results show that antibody specificity is also a significant contributor to the pattern of erythrocyte clearance in vivo. For IgG2b antibodies directed against the P and Q epitopes, the clearance of DA erythrocytes was more consistently splenic than for IgG2b mAb against the S epitope. This effect was further enhanced in CVF-treated animals, where a striking excess of splenic clearance was apparent. There was variation in the extent of hepatic clearance, both in normal and CVF-treated rats, with JY2/73 of the P-site antibodies, and YR1/100 of the Q-site antibodies showing substantially more hepatic clearance than the others. In decomplemented animals, R3/13 and R2/10P were virtually inactive in promoting hepatic clearance. This result was confirmed in splenectomized, CVF-treated rats, where DA erythrocytes were essentially not cleared by these two antibodies (Fig. 3). In contrast, IgG2b antibodies specific for the S site could still mediate erythrocyte clearance in these rats at a significant rate. In general, for IgG2b mAb directed against the P and Q epitopes, significant hepatic clearance in CVF-treated animals was correlated with a short initial Tin, and antibodies like JY2/73 (P-site) and YR1/100 (Q-site) had clearance behavior closer to the typical IgG2b mAb specific for the S site. Individual variation between mAb competitive for the same target epitopes was a consistent feature of these experiments. Fig. 4 shows an example for the clearance of DA and (DA x PVG)Fl erythrocytes in normal and CVFtreated rats by the two Q-site specific mAb, JY1/99 and JY1/116, where microheterogeneity within this group of antibodies is known [28, 301. Each antibody presented a consistent profile of erythrocyte clearance, and the data with this group of mAb is summarized in Table 1 (for DA cells) and Table 2 [for (DA x PVG)F1 cells]. Since all antibodies were assessed at various doses to ensure effective saturation in vivo, the differences in the amount of erythrocyte-bound antibody could not account for the differences observed between mAb of the same isotype. Interestingly, a comparison of the initial Tln obtained for both types of target erythrocytes in decomplemented
I33113 R2l10P JY31208 JY 1198 0 JY 11232 A JY31223
0
0 A
+
L
10
3.3 Role of antibody specificity in the targeting behavior Of 1gGB mAb The previous experiments showed the importance of antibody isotype in determining the fate of target erythrocytes opsonized by mAb in vivo. In the next series of experiments we used a variety of IgG2b mAb directed against different epitopes on the RTIAd molecule. The
947
I-,,,,, 0
10
20
30
40
SO
1
80
70
d0
00
Tlmdmln)
Figure 3. Comparison of the clearance profiles of DA erythrocytes following the administration of the P site-specific (R3/13, WlOP) and the S site-directed (JY3D08, JY1/98, JYlD32, JY3D23) IgGzb mAb into decomplemented splenectomized rats. The data shown for different mAb is representative of at least two separate experiments.
N.Yousaf, J. C. Howard and B. D. Williams
948
Eur. J. Immunol. 1991. 21: 943-950 animals preinjected with either JY1/98 or JY1/232 (S site IgG2b mAb) when compared with the normal groups that had received the same antibodies (Fig. 2). Furthermore, C depletion had no effect on the tissue distribution of cells opsonized by S site IgG2b antibodies but it clearly influenced the tissue distribution of DA erythrocytes when IgG2, mAb had been used (Table l).These findings suggest that IgG2, mAb activated homologous C in vivo more efficiently than did mAb of the IgG2b isotype. This result contrasts with in virro systems using heterologous C which have shown that rat IgG2b is considerably more efficient than IgG2, in binding and activating the first component of C and inducing cell lysis [21,22,24].Variations in the ability of rat IgG isotypes to fix C of different species have been noted [32-341. However, no systematic study of homologous C activation by rat mAb in vivo has been reported. Recently, the lytic ability of rat mAb has been correlated with the nature of the antigen recognized on the surface of nucleated cells [35], but it is not clear how antigendependent effects are related to C activation.
c
o
a
16
24
32
40
48
56
64
72
Time (min)
Figure 4. Clearance profiles of DA and (DA x PVG)F1 erythrocytes in normal (-) and CVF-treated (---) animals that had been preinjected with the Q site-specific mAb of the IgG2bsubclass (JY1/99, JY11116).
animals preinjected with various IgG2b mAb revealed that the clearance rates of the (DA x PVG)Fl cells (Table 2) with the S site-specific mAb (JY1/98, JY1/232, JY3/223) were comparable to that observed for the DA cells when either R3/13 (P site-specific antibody), or JY1/99 (Q sitespecific antibody) had been used (Table 1).
4 Discussion Three important points have emerged from the results presented in this report. Our findings suggest that the antibody isotype, antibody specificity and the antigen density on the surface of the target erythrocytes influence the fate of the target cells in vivo. The striking differences seen in the in vivo behavior of the IgG2, and IgG2b isotypes are of considerable interest. In general, the initial clearance of DA erythrocytes in normal animals preinjected with S site-specific IgG2, mAb was rapid. Although a degree of heterogeneity was observed in the overall clearance profiles of the target erythrocytes with the various mAb in this group, one important feature to emerge from these studies was the demonstration that C played a major role in augmenting the clearance and promoting hepatic sequestration of the DA cells with each of the five IgG2, mAb examined. In marked contrast to this pattern, an intact C system was not an essential requirement for the clearance of target erythrocytes when IgG2b mAb specific for the same site were used. We also noted that the overall clearance of the target cells was somewhat faster in decomplemented
A comparison of the results obtained for both types of target erythrocytes (DA cells, Table 1; (DA x PVG)F1 cells, Table 2) in different groups of rats preinjected with various IgG2b mAb suggests that the site to which the mAb bind on the RTIAa molecule is also an important factor in determining their behavior in vivo. In contrast to the targeting behavior of the S site-specific mAb, DA erythrocytes were cleared exclusively by the spleen in CVF-treated rats preinjected with either R3/13, R2/10P (both P-site IgG2b mAb), or JYlJ99 (IgG2b) which recognized the Q site on the class I antigen (Table 1). These findings indicated that some IgGZb mAb when bound to the target cells were recognized poorly by the liver when an intact C system was not available. This view was reinforced by the contrasting results obtained in decomplemented splenectomized animals preinjected with either S or P site-directed mAb of the IgG2b isotype (Fig. 3). One possible interpretation of these observations is that differences in antibody specificity are likely to be responsible for the differences in the recognition pattern of the target cells by the hepatic Kupffer cells. It seems that this type of discrimination in the recognition of different IgG2b mAb can also occur within the spleen. Interestingly, we noticed that in CVF-treated rats the initial clearance rates of the (DA x PVG)Fl cells (Table 2) with either the S site-specific mAb (JY1/98, JYlJ232, JY3/223), or Q site-directed mAb (YR1/100, JY1/116) were comparable to that seen for the DA cells (Table 1) when either R3/13 or R2/10P (P site-specific mAb), or JY1/99 (Q sitespecific antibody) had similarly been used in decomplemented animals. The results from this comparison provide further support for the idea that the in vivo behavior of the target cells may be influenced by differences in antibody specificity. The observation that different rat mAb of the same isotype varied considerably in their targeting behavior in vivo, is compatable with the in virro experiments which have demonstrated differences in the efficiency of human IgGl anti-Rh(D) mAb to elicit erythrocyte lysis by human lymphocyte effector cells [19, 201. Additional studies with the Q site-specific IgG2b mAb (Fig. 4, Tables 1 and 2), seem to indicate that the type of variation we have observed in the present study could also result from microheterogeneity within a group of mAb that were competitive for the same site [28, 301. Our in vivo findings are in agreement with the in vitro studies which have
Eur. J. lmmunol. 1991. 21: 943-950
suggested that an appropriate orientation of the antibody molecules on the target cell surface is an important factor for an efficient antibody-dependent cell-mediated cytotoxicity activity elicited by human killer lymphocytes [36].We cannot, however, exclude the possibilities that the variability seen in the in vivo behavior of different mAb reflects differences in the affinity of individual antibodies, or variation in the level of glycosylation and side-chain oligosaccharide heterogeneity in the Fc region of the antibody [37-401. The use of (DA x PVG)Fl erythrocytes, which express half the number of antigenic determinants, allowed a direct comparison of the effects of antigen density on the surface of the target cells and their subsequent behavior in vivo. With each antibody, the (DA x PVG)F, erythrocytes were removed at a slower rate than the DA cells. However, in contrast to the behavior of IgGzt, antibodies in both normal and CVF-treated rats (Fig. 2 and Table 2), mAb of the IgG2, isotype were clearly much less efficient in removing (DA x PVG)F1 erythrocytes from the circulation, even in the presence of an intact C system (Fig. 1and Table 2).The reasons for these observed differences between the two isotypes are not clear but they could partially be attributed to isotype-related structural differences such as the hinge region length and the degree of segmental flexibility [41-431. Considerable differences in the opsonizing efficiency of human IgG3 and IgGl mAb [anti-Rh(D)] to mediate erythrocyte binding to M@ FcR have also been noted [MI. The mechanisms which lead to the removal of target erythrocytes from the circulation most likely involve their recognition by the M@ within the spleen and liver. In our previous studies we have shown that the hepatic clearance of rat erythrocytes coated with a rat IgG2, mAb is exclusively dependent upon C3 receptors [44]. However, the contrasting behavior of the S site-specific IgG2, and IgG2h mAb seen in decomplemented rats (Figs. 1 and 2, Table l), suggests that the hepatic Kupffer cell FcR recognized an antibody of the IgGZhsubclass more efficiently than the IgGfa, whereas splenic M@ FcR mediated the clearance of target erythrocytes with either isotype although IgG2h was more effective. The data presented in this report and our previous findings on the FcR-mediated clearance system within the spleen [45,46] and the liver [47], are in partial agreement with in vitro experiments which indicate that rat splenic M@ do not bind mAb of the IgG2, subclass well but do express a high-affinity receptor for the IgG2b isotype [26, 481. The unique structure of the microenvironment within the splenic red pulp may facilitate an interaction between erythrocyte-bound IgG2, and low-affinity FcR for this IgG isotype on the surface of splenic M a . Although the significance of antibody isotype in the effectiveness of mAb serotherapy has been noted in both animal studies [ll,15,161 and in the treatment of lymphoid malignancies in humans [3], our experimental system has enabled us to demonstrate that not only do the two rat IgG isotypes behave very differently in vivo but also that the antigen density on the target cells is an important factor in determining their subsequent fate in vivo. Furthermore, our studies emphasize that different mAb of the same isotype can vary considerably in their targeting pattern in
In vivo behavior of monoclonal antibodies
049
vivo and this heterogeneity may relate to differences in antibody specificity. Received November 7, 1990.
5 References 1 Levy, R. and Miller, R. A., Fed. Proc. 1983. 42: 2650. 2 Waldmann, T. A., Goldman, C. K., Bongiovanni, K. E , Sharrow, S. O., Davey, M. P., Cease, K. B., Greenberg, S. J. and Longo, D. L., Blood 1988. 72: 1805. 3 Dyer, M. J. S., Hale, G., Hayhoe, F. G . J. and Waldmann, H., Blood 1989. 73: 1431. 4 Ortho Multicenter Transplant Study Group, N. Engl. J. Med. 1985. 313: 337. 5 Adelman, N. E.,Watling, D. L. and McDevitt, H. O., J. Exp. Med. 1983. 158: 1350. 6 Boitard, C., Michie, S., Senurier, P., Butcher, G. W., Larkins. A. I? and McDevitt, H. O., Proc. Natl. Acad. Sci. USA 1985.82: 6627. 7 Wofsy, D. and Seaman, W. E., J. Exp. Med. 1985. 161: 378. 8 Waldor, M. K., Sriram, S., Hardy, R., Herzenberg, L. A., Herzenberg, L. A., Lanier, L., Lim, M. and Steinman, L., Science 1985. 227: 415. 9 Acha-Orbea, H., Mitchell, D. J.,Timmermann, L.,Wraith, D. C.,Tausch, G. S.,Waldor, M. K., Zamvil, S. S., McDevitt, H. 0. and Steinman, L., Cell 1988. 54: 263. 10 Hughes-Jones, N. C., Blood Rev. 1989. 3: 53. 11 Kelley,V. E., Gaulton, G. N., Strom,T. B., J. Immunol. 1987. 138: 277 1. 12 Kelley, V. E., Gaulton, G. N., Hattori, M., Ikegami, H., Eisenbarth, G. and Strom, T. B., J. Immunol. 1988. 140: 59. 13 Ferrara, J. L. M., Marion, A., McIntyre, J. F., Murphy, G . F. and Burakoff, S. J., J. Immunol. 1986. 137: 1874. 14 Granstein, R. D., Goulston, C. and Gaulton, G., J. Immunol. 1986. 136: 898. 15 Cobbold, S. I?, Jayasuriya, A., Nash, A., Prospero,T. D. and Waldmann, H., Nature 1984. 312: 548. 16 Waldor, M. K., Mitchell, D., Kipps,T. J., Herzenberg, L. A. and Steinman, L., J. Immunol. 1987. 139: 3660. 17 Kummer, U., Thierfelder, S. and Mysliwietz, J., Eur. J. Irnmunol. 1990. 20: 107. 18 Wiener, E., Atwal, A.,Thompson, K. M., Melamed. M. D., Gorick, B. and Hughes-Jones, N. C., Immunology 1987. 62: 401. 19 Armstrong, S. S.,Wiener, E., Garner, S. F., Urbaniak, S. J. and Contreras, M., Br. J. Haematol. 1987. 66: 257. 20 Kumpel, B. M., Leader, K. A., Merry, A. H., Hadley, A. G.. Poole, G. D., Blancher, A., Goossens, D., Hughes-Jones, N. C. and Bradley, B. A., Eur. J. Immunol. 1989. 19: 2283. 21 Hughes-Jones, N. C., Gorick, B. D. and Howard, J. C.. Eur. J. Immunol. 1983. 13: 635. 22 Bindon, C. I., Hale, G., Clark, M. and Waldmann, H., Transplantation 1985. 40: 538. 23 Bindon, C. I., Hale, G., Hughes-Jones, N. C., Gorick, B. and Waldmann, H., Mol. Immunol. 1987. 24: 587. 24 Bruggemann, M., Teale, C., Clark, M., Bindon, C. and Waldmann, H., J. Immunol. 1989. 142: 3145. 25 Hale, G., Clark, M. and Waldmann, H., J. Immunol. 1985.134: 3056. 26 Denham, S., Barfoot, R. and Jackson, E., Immunology 1987. 62: 69. 27 Howard, J. C., Butcher, G. W., Galfrk, G., Milstein, C. and Milstein, C. P., Immunol. Rev. 1979. 47: 139. 28 Diamond, A. G., Larkin, A. €!,Wright, B., Ellis, S.T., Butcher, G. W. and Howard, J. C., Eur. J. Immunol. 1984. 14: 405. 29 Lachmann, P. J. and Hobart, M. J., in Weir, D. M. (Ed.), Handbook of Experimental Immunology, 3rd Edit., vol. 1, Blackwell Scientific Publications, Oxford 1978, chapter 5A.
950
N. Yousaf, J. C. Howard and B. D. Williams
30 Diamond, A. G., Butcher, G. W. and Howard, J. C., J. Immunol. 1984. 132: 1169. 31 Yousaf, N., Howard, J. C. and Williams, B. D., Immunology 1986. 59: 75. 32 Medgyesi, G. A., Fust, J.. Gergely, J. and Bazin, H., Immunochemistry 1978. 15: 125. 33 Fust, G., Medgyesi, G . A.. Bazin, H. and Gergely, J., Immunol. Lett. 1980. I : 249. 34 Medgyesi, G. A., Miklos, K., Kulics, J., Fiist, G., Gergely, J. and Bazin, H., Immunology 1981. 43: 171. 35 Bindon, C. I., Hale, G. and Waldmann, H., Eur. J. Irnmunol. 1988. 18: 1507. 36 Christiaansen, J. E., Burnside. S. S. and Sears, D. W., J. Immunol. 1987. 138: 2236. 37 Nose, M. and Wigzell, H., Proc. Natl. Acad. Sci. USA 1983.80: 6632. 38 Leatherbarrow, R. J., Rademacher,T.W., Dwek, R. A.,Woof, J. M., Clark, A., Burton, D. R., Richardson, N. and Feinstein, A.. Mol. Immunol. 1985. 22: 407.
Eur. J. Immunol. 1991. 21: 943-950 39 Walker, M. R., Lund, J.,Thompson, K. M. and Jefferis, R., Biochem. J. 1989. 259: 347. 40 Rothman, R. J. ,Warren, L.,Vliegenthart, J. F. G. and Hard, K. J., Biochemistry 1989. 28: 1317. 41 Oi, V. T., Vuong, T. M., Hardy, R., Reidler, J., Dangl, J., Herzenberg, L. A. and Stryer, L., Nature 1984. 307: 136. 42 Dangl, J. L., Wensel., T. G., Momson, S. L., Stryer, L., Herzenberg, L. A. and Oi,V. T., EMBO J. 1988. 7: 1989. 43 Schneider, W. I?, Wensel, T. G., Stryer, L. and Oi,V. T., Proc. Natl. Acad. Sci. USA 1988. 85: 2509. 44 Yousaf, N., Howard, J. C. and Williams, B. D., Immunology 1988. 64: 193. 45 Yousaf, N., Howard, J. C. and Williams, B. D., Immunology 1986. 59: 81. 46 Yousaf, N., Howard, J. C. and Williams, B. D., Clin. Exp. Immunol. 1986. 66: 654. 47 Yousaf, N., Howard, J. C. and Williams, B. D., Clin. Exp. Immunol. 1989. 78: 218. 48 Denham, S., Barfoot, R. and Sills, J., Immunology 1990. 69: 329.
Eur. J. Immunol. 1991. 21: 943-950
Nasim Yousafo, Jonathan C. Howardo and Bryan D. WilliamsO+ Department of Rheumatology, University Hospital of Waleso, Cardiff and Department of Immunology, Agricultural and Food Research Council Institute of Animal Physiology and Genetics Researcho, Babraham, Cambridge
Targeting behavior of rat monoclonal IgG antibodies in vivo: role of antibody isotype, specificity and the target cell antigen density* The studies described in this report were designed to investigate factors that could influence the behavior of erythrocytes following their interaction with monoclonal antibodies (mAb) in a fully homologous experimental opsonization system in vivo.The clearance profiles and tissue distribution of target erythrocytes were examined in both normal and decomplemented rats preinjected with rat IgG2, or IgG2b mAb directed against the same or different sites on RTIAa, the classical class I major histocompatibility complex antigen of the DA rat. Complement played a major role in augmenting the clearance and promoting hepatic sequestration of target erythrocytes in rats preinjected with IgG2, mAb directed against the S site. In contrast, an intact complement system was not an essential requirement for erythrocyte clearance when S site-specific IgGZb mAb were used. With each antibody tested, (DA x PVG)F1 cells, expressing about half as much antigen, were removed significantly slower than DA erythrocytes, this finding being more pronounced when the animals had been preinjected with mAb of the IgG2, isotype. A comparison of the tissue distribution of DA and (DA x PVG)F1 erythrocytes indicated that hepatic uptake was greater for target cells expressing higher antigen density. A considerable degree of heterogeneity was observed in the in vivo behavior of the target erythrocytes with three groups of IgG2b mAb that recognized different sites on the class I molecule. The S site-specific IgG2b mAb were much more efficient in the hepatic Fc receptor-mediated clearance system than were the P site-directed mAb of the same subclass. Our results suggest that antibody specificity may also be a contributory factor, in addition to antibody isotype and target cell antigen density, in determining the fate of target cells in vivo.
1 Introduction The potential use of mAb as specific therapeutic agents has attracted widespread interest. There is now considerable evidence which suggests that passively administered mAb can modify or alter immune responsiveness and disease processes in both humans [l-41, and experimental animals [5-91. It is also anticipated that human mAb against the Rhesus antigen D may replace the polyclonal anti-D used in the prophylaxis of hemolytic disease of the newborn [lo]. However, it is already apparent from the literature that there are marked differences in the effectiveness of treatment with different mAb even when they are directed against the same antigen [3, 5, 11-14]. The effector mechanisms responsible for the outcome of an efficient mAb serotherapy remain unclear. Although the relevance of antibody isotype in the effectiveness of mAb therapy has been recognized in several studies [3, 15, 161,
[I 90311
*
943
This work was supported by a grant from the Arthritis and Rheumatism Council of Great Britain and in part by USPHS Project Grant CA 34913.
Correspondence: Nasim Yousaf, Department of Immunology, University College and Middlesex School of Medicine, Arthur Stanley House, 40-50 Tottenham Street, London W1P 9PG, GB Abbreviation: CVF Cobra venom factor 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991
there are suggestions that antibody specificity and the ability of some antibodies to interact with the complement system may also influence the therapeutic efficacy of mAb in experimental animals [ ll] . More recently, the therapeutic potential of an antibody has been correlated with the antigen density on the target cell surface [17]. It is therefore clearly important to establish the factors that determine the efficiency of mAb serotherapy. If cell depletion is desirable to achieve optimum therapeutic effects [2, 3, lo], then it would be important to maximize the effector mechanism for eliminating the target cells in vivo. It is likely that opsonization of the target cells and their subsequent removal by the mononuclear phagocyte system or their destruction by C-dependent lysis may be critical events in determining the therapeutic usefulness of mAb selected for immunotherapy. The Fc region-dependent effector functions of both human and rat Ig have been examined in several in virro experimental systems. Studies with human anti-Rh(D) mAb not only indicate the differences in the ability of different IgG isotypes to promote erythrocyte binding to M@ [18], but also suggest that mAb of the same subclass can vary significantly in their functional activities [19, 201. Similarly, a comparison of a wide range of rat mAb of different isotypes has revealed that rat IgG2, is considerably less efficient than IgG2b in mediating the effector functions such as C lq binding, C activation [21-241, antibody-dependent cell-mediated cytotoxicity [25], and M@ FcR binding [26]. However, an analysis of the functional properties of rat mAb in vivo, under fully homologous conditions, has been lacking.
+
014-2980/91/0404-0943$3.50 .25/0
944
Eur. J. Immunol. 1991. 21: 943-950
N. Yousaf, J. C. Howard and B. D. Williams
The experiments presented in this report were undertaken in an attempt to define the variables likely to be associated with efficient mAb serotherapy. The expression of MHC class I antigens on rat erythrocytes and the availability of a large number of rat mAb directed against various epitopes on the rat RTIAa class I antigen [27, 281, allowed us to investigate the factors which may influence the targeting behavior of mAb in vivo. In our experimental system, we examined the clearance profiles and tissue distribution of target erythrocytes following their infusion into normal and decomplemented rats preinjected with mAb of the IgG2, or IgGZb subclasses directed against the same or different sites on the RTIAa molecule. Since the RTIAa antigen level on erythrocytes from (DA x PVG)F1 rats is 50% as compared to that on the DA erythrocytes, we were able to investigate the effects of reducing the target cell antigen density for each of the 19 different mAb used in our study.
(Na2Wr04, The Radiochemical Centre, Amersham, GB) essentially as described previously [31], except that 3.7 MBq (100 pCi) of 99Tcand 50 pCi (1.85 MBq) of W r were used, respectively, to label 100 p1 of packed (DA x PVG)Fl erythrocytes and DA erythrocytes. Following the labeling procedure the cells were washed three times and resuspended in saline at a concentration of 10% (v/v) . 2.5 In vivo clearance of target erythrocytes
The procedure adopted for measuring the clearance of radiolabeled rat erythrocytes in vivo was essentially as described previously [31]. Briefly, equal volumes of the DA and (DA x PVG)Fl cell suspensions (lo%, v/v), labeled with different isotopes, were first mixed and 200 pl of this mixture were then injected into the tail vein of PVG rats. Blood samples (20 pl) were taken from the tail at various 2 Materials and methods time intervals and the radioactivity associated with the cells and plasma was measured separately in an LKB (Bromma, 2.1 Animals Sweden) Compu Gamma. The animals were killed 90 min after injection of the cells and the radioactivity within Male rats of the PVG (RTIC) and DA (RTla) strains various organs was determined. Appropriate corrections weighing 180-200 g were obtained from Bantin and King- were made for the decay of radioisotopes and for the man Ltd. (Hull, GB). (DA x PVG)Fl rats were either bred cross-over between channels when two isotopes were used in the Department of Immunology (AFRC Institute of in the same animal. The survival of radiolabeled erythroAnimal Physiology and Genetics Research, Cambridge, cytes after their injection into normal rats has been GB), or purchased from Bantin and Kingman Ltd. Splen- reported in earlier studies [31]. mAb (1 or 2 ml) were ectomized or sham-splenectomized animals were used administered (i.v.) into rats 15-20 min prior to the injection 4 weeks after surgery. of radiolabeled erythrocytes, and the in vivo behavior of the target cells was then assessed. Preliminary experiments revealed no major differences between these two doses, and 2.2 Decomplementation of animals with cobra venom it was assumed, therefore, that effective saturation in vivo factor could be achieved with 1 ml of antibody. However, some differences seen in these studies apparently related to the Cobra 'enom factor (CVF) was purified from nub variation between individual animals since a lower dose (Sigma Company Ltd*9 GB) (0.5 ml) was equally effective for several of the mAb according to the method of Lachmann and Hobart [20]. used. Rats were decomplemented by the i.v. injection of 30 U of CVF24 h prior to the experiment.The C3 levels at the time of study were < 5% of that seen in normal rat plasma. 2.6 ~ ~of data ~ l ~ ~ i 2.3 mAb Rat mAb were directed against RTIAa, the MHC class I antigen of the DA rat. The derivation of these mAb and their in v i m binding characteristics have been reported elsewhere [27, 28, 301. In the present study, mAb of the IgG2, and IgG2b subclasses with specificity for the same or different sites on the RTIAd molecule were used. All mAb were in the form of tissue culture SN containing 10% FCS and 0.1% sodium azide, with specific antibody concentration being about 20 pg/ml. These reagents were dialyzed extensively against PBS, passed through 0.22-pm Millipore filter (Millipore S.A., Molsheim, France), and then stored in appropriate aliquots at -20 "C until required. Before their use in the clearance studies any insoluble material was removed by centrifugation at 1600 x g.
The data were expressed as the percentage of radioactivity remaining in circulation at various times after the injection of the cells.The radioactivity present at 1 min was taken to be 100%. The initial half-clearance time (Tln), the time taken for 50% of the radiolabeled cells to leave the circulation, was calculated for each animal, and where appropriate, linear regression analysis was applied for its assessment. The significance of differences observed between the various groups was analyzed using Student's r-test.
3 Results 3.1 In vivo behavior of target erythrocytes with S site-specific IgGh mAb
S site-specific mAb of the IgGz, isotype removed radiolabeled DA and ( D A x PVG)Fl erythrocytes from the circulation of normal rats in a non-linear fashion. Typical Rat erythrocytes were radiolabeled with 99Tc(Na299Tc04, clearance profiles with two different IgG2, mAb (JY1/174 University Hospital of Wales, Cardiff, GB) or 51Cr and JY 1/163) are shown in Fig. 1.The clearance profiles of
2.4 Labeling of erythrocytes with V
c and 51Cr
~
Eur. J. Immunol. 1991. 21: 943-950
In vivo behavior of monoclonal antibodies
945
Table 1. In vivo behavior of DA erythrocytes in normal and CVF-treated rats preinjected with various mAb Normal rats') % Iniected radioactivityd) Splken Liver (in)
Antibody nh)
S site IgGh JY11174 JY11184 W15S JY3/109 JY11163 S site IgGlb JY3l223 N1/98 JYll232 JY3l208 P site IgGZb JYm3 R3/13 WlOP
4
13.4 f 1.2 18.7 f 3.7 7.2 f 1.4 17.4 f 2.4 20.9 f 0.5
54.9 f 2.2 41.8 f 4.4 42.7 f 1.0 52.8 f 2.5 29.8 f 4.9
3 3 3 3 3
51.7 f 13.0 44.7f 8.7 70.0f 4.0 54.7f 9.9 90.0 rf: 10.0
42.6 f 4.0 34.0f5.6 35.5 f 2.1 41.0f2.1 29.0 f 2.5
9.9k 1.1 12.6 k 4.1 12.0 f 2.2 10.3f 2.1 5.5 f 0.5
13.6 f 0.6 19.3 f 1.6 21.9 f 1.6 24.5 f 5.1
32.0 f 1.3 33.9 f 0.8 35.7 f 1.6 39.3 f 5.6
44.4 f 1.5 31.1 f 1.6 35.8 f 2.3 26.0 f 8.0
5 7 5 4
15.0f 13.4f 15.6f 21.0f
3.6 1.5 2.4 5.2
41.0f4.7 37.4f2.9 39.0f5.2 41.1 f 8 . 3
23.9 f 4.7 27.5 f 3.2 30.2 +- 4.9 19.1 f 8.6
4 7 2
15.1 f 3.8 23.6f 1.1
30.7 f 3.4 19.4 f 0.7 28.0
3 7 3
22.0f 2.1
30.8
37.0 f 2.8 38.1 f 1.7 43.8
31.8f 2.8 40.7f 5.9
60.1 f 3 . 8 48.6f3.4 60.1 f 3 . 8
17.5 f 5.3 4.4 f 0.4 4.3 f 0.6
4 3 3 6
25.0 f 3.9 14.3 f 1.9 15.7 f 2.8
44.2 f 8.7 38.6 f 5.3 41.9 rf: 9.4
15.8 f 4.5 34.3 f 6.4 29.5 f 9.9
4 3 3
28.lf 3.4 17.2f 1.7 16.8f 1.2
61.9f4.1 62.0f 1.4 53.9f5.1
7.5 f 1.8 13.1 f 3.7 21.8 f 6.3
2.0 f 0.2
5.5 rf: 0.2
6 4 4 5
8 5
4 .
None a) b) c) d)
6.5 rf: 0.3 11.9 f 1.9 11.2 f 2.3 12.8 f 2.4 45.5 f 8.4
5
Q site IgGzb JY1/99 JY 11116 YR1/100
CVF-treated ratsa) Tlnc) YO Injected radioactivityd) (min) Spleen Liver
nb)
TIR')
Data are given as mean values & SE for each group. The number of animals in each group. The initial half-clearance time (Tin) is expressed in minutes. Localization within the spleen and liver is expressed as the percentage of the radioactivity administered.
the target cells with other three IgG2, mAb (not shown) were intermediate between the extreme examples shown in Fig. 1. The results obtained in these experiments are summarized in Table 1 (for DA cells) and Table 2 [for (DA x PVG)Fl cells]. In general, DA cells were cleared rapidly with an initial Tl0 of less than 13 min (Fig. 1 and Table 1). Although the overall clearance profiles of these
cells were somewhat variable with the various IgGz, antibodies employed, the importance of C in their in vivo clearance is demonstrated clearly in Fig. 1. C depletion not only led to a change in the clearance profile, but it also significantly delayed the clearance rate of the target cells with all the S site-specific IgGh mAb examined (Table 1). Cell lysis did not occur to any significant degree in normal
JV 1/174
M
a n 8
. \
o
10
2d
b
i
rb
2;
3.2
do
48
de
ek
,
J 7s
o wn)
Figure I. Clearance profiles of DA and (DAxPVG)F, erythrocvtes in normal (-) and CVF-treated (---) rats preinjectedwiths site-specificIgG2, antibodies.
a
Ib
24
j2
io
is
de
04
;2
I
Eur. J. Immunol. 1991. 21: 943-950
N. Yousaf, J. C. Howard and B. D. Williams
946
Table 2. In vivo behavior of (DA x PVG)FI erythrocytes in normal and CVF-treated rats preinjected with various mAb
Antibodv
Normal ratsa) % Iniected radioactivityd) Tld) Liver (min) Spl;en
IlhJ
S site IgGh JY11174 JY11184 R2/15S JY3/109 JY11163 S site IgGZh JY3/223 JYlN8 JYlD32 JY3R08
4 5 4 4
56.7 f 14.7 77.2 f 11.2 >WJ 54.6+ 4.4
4
> W'
CVF-treated ratsa) % Injected radioactivityd) (min) Spleen Liver
TI#
nh)
13.8 f 0.4 16.6 f 3.4 4.3 f 1.5 19.4 f 3.0 13.5 f 1.6
20.2 f 3.2 11.3 f 1.3 14.9 f 2.0 16.1 f 1.5 6.3 f 1.1
3 2 3 3 2
> 9oe) > 9oe) > W) > W) > W)
18.8 f 4.1 20.8 15.7 f 1.9 19.6 f 1.8 10.0
4.2 f 0.4 3.8 4.2 f 0.2 4.2 f 0.2 3.6
17.5 f 1.3 11.9f0.9 15.3 f 2.0 6.4 f 2.4
4 4
29.9 f 4.6 23.2 f 2.4 24.5 f 2.7 43.8 f 8.5
39.8 f 2.7 39.4 f 1.3 39.0 f 3.8 37.4 f 2.9
6.2 f 0.6 9.2 f 1.6 12.1 f 3.1 6.0 f 2.4
5 7 5 4
26.9f 36.6f 34.1 f 48.2f
7.5
40.2f 1.3 35.8 f 1.6 37.4 f 1.9 40.5 f 1.2
P site IgGZb JYm3 RW3 R2/10P
4 4
29.1 f 8.7 46.3f 4.1 57.0
45.6 f 1.7 39.4 f 1.6 37.4
9.6 f 1.1 6.7 f 0.7 5.9
3 2 3
33.0 f 2.6 55.0 75.5 f 16.4
48.5 f 3.5
5.4 f 0.7
38.2 32.1 f 6.6
3.0 3.6 k 0.7
0 site lgG2h JY 1/99 JY11116 YRlllOO
4 3
47.8f 3.8 25.7f 2.4 M.Of 2.6
31.5 f 2.6 46.3 f 3.1 42.2 f 2.4
4.9 f 0.9 11.1 f 3.6 9.1 f 3.8
4
50.0f 8.8 25.0 f 2.5 31.2 f 0.4
38.0 f 3.4 54.4 f 4.8 45.9 f 3.1
3.4 f 0.3 4.0 f 0.7 5.8f 1.6
a) b) c) d) e)
2
3
0.8 2.0 2.0
6 5
3 3
Results are shown as mean values k SE for each group. The number of animals in each group. The initial half-clearance time is expressed in minutes. Localization within the spleen and liver is expressed as the percentage of the radioactivity administered. The radioactivity remaining in circulation at 90 min, expressed as percent 1-min counts, ranged from 51% to 69% for these groups of animals.
rats, there being no increase in the level of free isotope detected in plasma above that seen in the control animals which had received saline.
Specific accumulation of radioactivity was seen in the spleen and the liver, and distribution of the target erythrocytes within these tissues is summarized in Table 1. The
uptake of DA cells in normal rats preinjected with JY1/174, JY1/184, R2/15S and JY3/109 was predominantly hepatic, although significant localization also occurred in the spleen with most of the S site-specific IgG2, mAb. In CVF-treated rats, however, spleen was the primary site of accumulation of the target cells with each of the five IgG2, mAb used (Table 1).
M Cbll.
--I J
I
1
, 0
8
16
24
32
40
48
56
64
72
0
T i m (mh)
8
16
24
32
40
48
56
64
72
Figure2. Clearance profiles of DA and (DA x PVG)F, erythrocytes in normal (-) and CVF-treated (---) animals that had been passively infused with S site-directed mAb of the IgG2h isotype.
In vivo behavior of monoclonal antibodies
Eur. J. Immunol. 1991. 21: 943-950
The effect of lower antigen density on the surface of the target cells and their subsequent behavior in vivo was examined by the administration of (DA x PVG)F1 cells into animals that had received the S site-directed IgG2, mAb, and the data from both normal and CVF-treated rats are shown in Fig. 1 and Table2. With each antibody, (DA x PVG)FI erythrocytes were always cleared at a much slower rate than DA cells. Once again, C depletion markedly influenced the clearance profiles of (DA x PVG)Fl cells and their subsequent localization within the spleen and liver. In CVF-treated rats spleen was the only site of localization of these cells (Table 2). 3.2 In vivo behavior of target erythrocytes with S site-specific IgGa mAb
The clearance of DA and (DA x PVG)Fl erythrocytes by IgGzb mAb directed against the S site (Fig. 2,Tables 1 and 2) contrasted strongly with that shown above for IgG2, mAb. In normal rats the clearance of both DA and (DA x PVG)FI cells was more linear (Fig. 2). Furthermore, with mAb of the IgG2b isotype, the presence of an intact C system was not essential for erythrocyte clearance. Indeed, in the case of JY1/98 antibody, the initial Tln was achieved at a somewhat faster rate for both DA and (DA x PVG)Fl erythrocytes in CVF-treated rats as compared with the normal group (Fig. 2). A similar change in the clearance rate was also seen with JY1/232 mAb (Table 2). The splenic and hepatic localization of DA erythrocytes in both normal and CVF-treated rats preinjected with various S site-specific IgG2b mAb is shown in Table 1. In both groups, these cells accumulated in the spleen as well as the liver, and in general, C depletion did not produce any major changes in the level of radioactivity within these tissues. The splenic accumulation of cells was much more marked than that seen with IgG2, mAb directed against the S site. Fig. 2 also demonstrates the influence of reducing the antigen density on the target erythrocytes in both normal and CVF-treated rats that had been preinjected with different S site-directed IgGZb mAb. The clearance of (DA x PVG)Fl erythrocytes from the circulation of normal rats was significantly prolonged (Table 2), the initial T m with each antibody being approximately twice as long as that seen for the DA cells (Table 1). With all the S sitespecific IgG2b mAb, spleen was the predominant site of sequestration of (DA x PVG)Fl erythrocytes in both normal and CVF-treated animals (Table 2). Similar results were obtained with other IgG2b mAb (JYU75.1, JY3/50, JY1/134, JY3/141) directed against the Ssite (data not shown).
results show that antibody specificity is also a significant contributor to the pattern of erythrocyte clearance in vivo. For IgG2b antibodies directed against the P and Q epitopes, the clearance of DA erythrocytes was more consistently splenic than for IgG2b mAb against the S epitope. This effect was further enhanced in CVF-treated animals, where a striking excess of splenic clearance was apparent. There was variation in the extent of hepatic clearance, both in normal and CVF-treated rats, with JY2/73 of the P-site antibodies, and YR1/100 of the Q-site antibodies showing substantially more hepatic clearance than the others. In decomplemented animals, R3/13 and R2/10P were virtually inactive in promoting hepatic clearance. This result was confirmed in splenectomized, CVF-treated rats, where DA erythrocytes were essentially not cleared by these two antibodies (Fig. 3). In contrast, IgG2b antibodies specific for the S site could still mediate erythrocyte clearance in these rats at a significant rate. In general, for IgG2b mAb directed against the P and Q epitopes, significant hepatic clearance in CVF-treated animals was correlated with a short initial Tin, and antibodies like JY2/73 (P-site) and YR1/100 (Q-site) had clearance behavior closer to the typical IgG2b mAb specific for the S site. Individual variation between mAb competitive for the same target epitopes was a consistent feature of these experiments. Fig. 4 shows an example for the clearance of DA and (DA x PVG)Fl erythrocytes in normal and CVFtreated rats by the two Q-site specific mAb, JY1/99 and JY1/116, where microheterogeneity within this group of antibodies is known [28, 301. Each antibody presented a consistent profile of erythrocyte clearance, and the data with this group of mAb is summarized in Table 1 (for DA cells) and Table 2 [for (DA x PVG)F1 cells]. Since all antibodies were assessed at various doses to ensure effective saturation in vivo, the differences in the amount of erythrocyte-bound antibody could not account for the differences observed between mAb of the same isotype. Interestingly, a comparison of the initial Tln obtained for both types of target erythrocytes in decomplemented
I33113 R2l10P JY31208 JY 1198 0 JY 11232 A JY31223
0
0 A
+
L
10
3.3 Role of antibody specificity in the targeting behavior Of 1gGB mAb The previous experiments showed the importance of antibody isotype in determining the fate of target erythrocytes opsonized by mAb in vivo. In the next series of experiments we used a variety of IgG2b mAb directed against different epitopes on the RTIAd molecule. The
947
I-,,,,, 0
10
20
30
40
SO
1
80
70
d0
00
Tlmdmln)
Figure 3. Comparison of the clearance profiles of DA erythrocytes following the administration of the P site-specific (R3/13, WlOP) and the S site-directed (JY3D08, JY1/98, JYlD32, JY3D23) IgGzb mAb into decomplemented splenectomized rats. The data shown for different mAb is representative of at least two separate experiments.
N.Yousaf, J. C. Howard and B. D. Williams
948
Eur. J. Immunol. 1991. 21: 943-950 animals preinjected with either JY1/98 or JY1/232 (S site IgG2b mAb) when compared with the normal groups that had received the same antibodies (Fig. 2). Furthermore, C depletion had no effect on the tissue distribution of cells opsonized by S site IgG2b antibodies but it clearly influenced the tissue distribution of DA erythrocytes when IgG2, mAb had been used (Table l).These findings suggest that IgG2, mAb activated homologous C in vivo more efficiently than did mAb of the IgG2b isotype. This result contrasts with in virro systems using heterologous C which have shown that rat IgG2b is considerably more efficient than IgG2, in binding and activating the first component of C and inducing cell lysis [21,22,24].Variations in the ability of rat IgG isotypes to fix C of different species have been noted [32-341. However, no systematic study of homologous C activation by rat mAb in vivo has been reported. Recently, the lytic ability of rat mAb has been correlated with the nature of the antigen recognized on the surface of nucleated cells [35], but it is not clear how antigendependent effects are related to C activation.
c
o
a
16
24
32
40
48
56
64
72
Time (min)
Figure 4. Clearance profiles of DA and (DA x PVG)F1 erythrocytes in normal (-) and CVF-treated (---) animals that had been preinjected with the Q site-specific mAb of the IgG2bsubclass (JY1/99, JY11116).
animals preinjected with various IgG2b mAb revealed that the clearance rates of the (DA x PVG)Fl cells (Table 2) with the S site-specific mAb (JY1/98, JY1/232, JY3/223) were comparable to that observed for the DA cells when either R3/13 (P site-specific antibody), or JY1/99 (Q sitespecific antibody) had been used (Table 1).
4 Discussion Three important points have emerged from the results presented in this report. Our findings suggest that the antibody isotype, antibody specificity and the antigen density on the surface of the target erythrocytes influence the fate of the target cells in vivo. The striking differences seen in the in vivo behavior of the IgG2, and IgG2b isotypes are of considerable interest. In general, the initial clearance of DA erythrocytes in normal animals preinjected with S site-specific IgG2, mAb was rapid. Although a degree of heterogeneity was observed in the overall clearance profiles of the target erythrocytes with the various mAb in this group, one important feature to emerge from these studies was the demonstration that C played a major role in augmenting the clearance and promoting hepatic sequestration of the DA cells with each of the five IgG2, mAb examined. In marked contrast to this pattern, an intact C system was not an essential requirement for the clearance of target erythrocytes when IgG2b mAb specific for the same site were used. We also noted that the overall clearance of the target cells was somewhat faster in decomplemented
A comparison of the results obtained for both types of target erythrocytes (DA cells, Table 1; (DA x PVG)F1 cells, Table 2) in different groups of rats preinjected with various IgG2b mAb suggests that the site to which the mAb bind on the RTIAa molecule is also an important factor in determining their behavior in vivo. In contrast to the targeting behavior of the S site-specific mAb, DA erythrocytes were cleared exclusively by the spleen in CVF-treated rats preinjected with either R3/13, R2/10P (both P-site IgG2b mAb), or JYlJ99 (IgG2b) which recognized the Q site on the class I antigen (Table 1). These findings indicated that some IgGZb mAb when bound to the target cells were recognized poorly by the liver when an intact C system was not available. This view was reinforced by the contrasting results obtained in decomplemented splenectomized animals preinjected with either S or P site-directed mAb of the IgG2b isotype (Fig. 3). One possible interpretation of these observations is that differences in antibody specificity are likely to be responsible for the differences in the recognition pattern of the target cells by the hepatic Kupffer cells. It seems that this type of discrimination in the recognition of different IgG2b mAb can also occur within the spleen. Interestingly, we noticed that in CVF-treated rats the initial clearance rates of the (DA x PVG)Fl cells (Table 2) with either the S site-specific mAb (JY1/98, JYlJ232, JY3/223), or Q site-directed mAb (YR1/100, JY1/116) were comparable to that seen for the DA cells (Table 1) when either R3/13 or R2/10P (P site-specific mAb), or JY1/99 (Q sitespecific antibody) had similarly been used in decomplemented animals. The results from this comparison provide further support for the idea that the in vivo behavior of the target cells may be influenced by differences in antibody specificity. The observation that different rat mAb of the same isotype varied considerably in their targeting behavior in vivo, is compatable with the in virro experiments which have demonstrated differences in the efficiency of human IgGl anti-Rh(D) mAb to elicit erythrocyte lysis by human lymphocyte effector cells [19, 201. Additional studies with the Q site-specific IgG2b mAb (Fig. 4, Tables 1 and 2), seem to indicate that the type of variation we have observed in the present study could also result from microheterogeneity within a group of mAb that were competitive for the same site [28, 301. Our in vivo findings are in agreement with the in vitro studies which have
Eur. J. lmmunol. 1991. 21: 943-950
suggested that an appropriate orientation of the antibody molecules on the target cell surface is an important factor for an efficient antibody-dependent cell-mediated cytotoxicity activity elicited by human killer lymphocytes [36].We cannot, however, exclude the possibilities that the variability seen in the in vivo behavior of different mAb reflects differences in the affinity of individual antibodies, or variation in the level of glycosylation and side-chain oligosaccharide heterogeneity in the Fc region of the antibody [37-401. The use of (DA x PVG)Fl erythrocytes, which express half the number of antigenic determinants, allowed a direct comparison of the effects of antigen density on the surface of the target cells and their subsequent behavior in vivo. With each antibody, the (DA x PVG)F, erythrocytes were removed at a slower rate than the DA cells. However, in contrast to the behavior of IgGzt, antibodies in both normal and CVF-treated rats (Fig. 2 and Table 2), mAb of the IgG2, isotype were clearly much less efficient in removing (DA x PVG)F1 erythrocytes from the circulation, even in the presence of an intact C system (Fig. 1and Table 2).The reasons for these observed differences between the two isotypes are not clear but they could partially be attributed to isotype-related structural differences such as the hinge region length and the degree of segmental flexibility [41-431. Considerable differences in the opsonizing efficiency of human IgG3 and IgGl mAb [anti-Rh(D)] to mediate erythrocyte binding to M@ FcR have also been noted [MI. The mechanisms which lead to the removal of target erythrocytes from the circulation most likely involve their recognition by the M@ within the spleen and liver. In our previous studies we have shown that the hepatic clearance of rat erythrocytes coated with a rat IgG2, mAb is exclusively dependent upon C3 receptors [44]. However, the contrasting behavior of the S site-specific IgG2, and IgG2h mAb seen in decomplemented rats (Figs. 1 and 2, Table l), suggests that the hepatic Kupffer cell FcR recognized an antibody of the IgGZhsubclass more efficiently than the IgGfa, whereas splenic M@ FcR mediated the clearance of target erythrocytes with either isotype although IgG2h was more effective. The data presented in this report and our previous findings on the FcR-mediated clearance system within the spleen [45,46] and the liver [47], are in partial agreement with in vitro experiments which indicate that rat splenic M@ do not bind mAb of the IgG2, subclass well but do express a high-affinity receptor for the IgG2b isotype [26, 481. The unique structure of the microenvironment within the splenic red pulp may facilitate an interaction between erythrocyte-bound IgG2, and low-affinity FcR for this IgG isotype on the surface of splenic M a . Although the significance of antibody isotype in the effectiveness of mAb serotherapy has been noted in both animal studies [ll,15,161 and in the treatment of lymphoid malignancies in humans [3], our experimental system has enabled us to demonstrate that not only do the two rat IgG isotypes behave very differently in vivo but also that the antigen density on the target cells is an important factor in determining their subsequent fate in vivo. Furthermore, our studies emphasize that different mAb of the same isotype can vary considerably in their targeting pattern in
In vivo behavior of monoclonal antibodies
049
vivo and this heterogeneity may relate to differences in antibody specificity. Received November 7, 1990.
5 References 1 Levy, R. and Miller, R. A., Fed. Proc. 1983. 42: 2650. 2 Waldmann, T. A., Goldman, C. K., Bongiovanni, K. E , Sharrow, S. O., Davey, M. P., Cease, K. B., Greenberg, S. J. and Longo, D. L., Blood 1988. 72: 1805. 3 Dyer, M. J. S., Hale, G., Hayhoe, F. G . J. and Waldmann, H., Blood 1989. 73: 1431. 4 Ortho Multicenter Transplant Study Group, N. Engl. J. Med. 1985. 313: 337. 5 Adelman, N. E.,Watling, D. L. and McDevitt, H. O., J. Exp. Med. 1983. 158: 1350. 6 Boitard, C., Michie, S., Senurier, P., Butcher, G. W., Larkins. A. I? and McDevitt, H. O., Proc. Natl. Acad. Sci. USA 1985.82: 6627. 7 Wofsy, D. and Seaman, W. E., J. Exp. Med. 1985. 161: 378. 8 Waldor, M. K., Sriram, S., Hardy, R., Herzenberg, L. A., Herzenberg, L. A., Lanier, L., Lim, M. and Steinman, L., Science 1985. 227: 415. 9 Acha-Orbea, H., Mitchell, D. J.,Timmermann, L.,Wraith, D. C.,Tausch, G. S.,Waldor, M. K., Zamvil, S. S., McDevitt, H. 0. and Steinman, L., Cell 1988. 54: 263. 10 Hughes-Jones, N. C., Blood Rev. 1989. 3: 53. 11 Kelley,V. E., Gaulton, G. N., Strom,T. B., J. Immunol. 1987. 138: 277 1. 12 Kelley, V. E., Gaulton, G. N., Hattori, M., Ikegami, H., Eisenbarth, G. and Strom, T. B., J. Immunol. 1988. 140: 59. 13 Ferrara, J. L. M., Marion, A., McIntyre, J. F., Murphy, G . F. and Burakoff, S. J., J. Immunol. 1986. 137: 1874. 14 Granstein, R. D., Goulston, C. and Gaulton, G., J. Immunol. 1986. 136: 898. 15 Cobbold, S. I?, Jayasuriya, A., Nash, A., Prospero,T. D. and Waldmann, H., Nature 1984. 312: 548. 16 Waldor, M. K., Mitchell, D., Kipps,T. J., Herzenberg, L. A. and Steinman, L., J. Immunol. 1987. 139: 3660. 17 Kummer, U., Thierfelder, S. and Mysliwietz, J., Eur. J. Irnmunol. 1990. 20: 107. 18 Wiener, E., Atwal, A.,Thompson, K. M., Melamed. M. D., Gorick, B. and Hughes-Jones, N. C., Immunology 1987. 62: 401. 19 Armstrong, S. S.,Wiener, E., Garner, S. F., Urbaniak, S. J. and Contreras, M., Br. J. Haematol. 1987. 66: 257. 20 Kumpel, B. M., Leader, K. A., Merry, A. H., Hadley, A. G.. Poole, G. D., Blancher, A., Goossens, D., Hughes-Jones, N. C. and Bradley, B. A., Eur. J. Immunol. 1989. 19: 2283. 21 Hughes-Jones, N. C., Gorick, B. D. and Howard, J. C.. Eur. J. Immunol. 1983. 13: 635. 22 Bindon, C. I., Hale, G., Clark, M. and Waldmann, H., Transplantation 1985. 40: 538. 23 Bindon, C. I., Hale, G., Hughes-Jones, N. C., Gorick, B. and Waldmann, H., Mol. Immunol. 1987. 24: 587. 24 Bruggemann, M., Teale, C., Clark, M., Bindon, C. and Waldmann, H., J. Immunol. 1989. 142: 3145. 25 Hale, G., Clark, M. and Waldmann, H., J. Immunol. 1985.134: 3056. 26 Denham, S., Barfoot, R. and Jackson, E., Immunology 1987. 62: 69. 27 Howard, J. C., Butcher, G. W., Galfrk, G., Milstein, C. and Milstein, C. P., Immunol. Rev. 1979. 47: 139. 28 Diamond, A. G., Larkin, A. €!,Wright, B., Ellis, S.T., Butcher, G. W. and Howard, J. C., Eur. J. Immunol. 1984. 14: 405. 29 Lachmann, P. J. and Hobart, M. J., in Weir, D. M. (Ed.), Handbook of Experimental Immunology, 3rd Edit., vol. 1, Blackwell Scientific Publications, Oxford 1978, chapter 5A.
950
N. Yousaf, J. C. Howard and B. D. Williams
30 Diamond, A. G., Butcher, G. W. and Howard, J. C., J. Immunol. 1984. 132: 1169. 31 Yousaf, N., Howard, J. C. and Williams, B. D., Immunology 1986. 59: 75. 32 Medgyesi, G. A., Fust, J.. Gergely, J. and Bazin, H., Immunochemistry 1978. 15: 125. 33 Fust, G., Medgyesi, G . A.. Bazin, H. and Gergely, J., Immunol. Lett. 1980. I : 249. 34 Medgyesi, G. A., Miklos, K., Kulics, J., Fiist, G., Gergely, J. and Bazin, H., Immunology 1981. 43: 171. 35 Bindon, C. I., Hale, G. and Waldmann, H., Eur. J. Irnmunol. 1988. 18: 1507. 36 Christiaansen, J. E., Burnside. S. S. and Sears, D. W., J. Immunol. 1987. 138: 2236. 37 Nose, M. and Wigzell, H., Proc. Natl. Acad. Sci. USA 1983.80: 6632. 38 Leatherbarrow, R. J., Rademacher,T.W., Dwek, R. A.,Woof, J. M., Clark, A., Burton, D. R., Richardson, N. and Feinstein, A.. Mol. Immunol. 1985. 22: 407.
Eur. J. Immunol. 1991. 21: 943-950 39 Walker, M. R., Lund, J.,Thompson, K. M. and Jefferis, R., Biochem. J. 1989. 259: 347. 40 Rothman, R. J. ,Warren, L.,Vliegenthart, J. F. G. and Hard, K. J., Biochemistry 1989. 28: 1317. 41 Oi, V. T., Vuong, T. M., Hardy, R., Reidler, J., Dangl, J., Herzenberg, L. A. and Stryer, L., Nature 1984. 307: 136. 42 Dangl, J. L., Wensel., T. G., Momson, S. L., Stryer, L., Herzenberg, L. A. and Oi,V. T., EMBO J. 1988. 7: 1989. 43 Schneider, W. I?, Wensel, T. G., Stryer, L. and Oi,V. T., Proc. Natl. Acad. Sci. USA 1988. 85: 2509. 44 Yousaf, N., Howard, J. C. and Williams, B. D., Immunology 1988. 64: 193. 45 Yousaf, N., Howard, J. C. and Williams, B. D., Immunology 1986. 59: 81. 46 Yousaf, N., Howard, J. C. and Williams, B. D., Clin. Exp. Immunol. 1986. 66: 654. 47 Yousaf, N., Howard, J. C. and Williams, B. D., Clin. Exp. Immunol. 1989. 78: 218. 48 Denham, S., Barfoot, R. and Sills, J., Immunology 1990. 69: 329.
