Int. J. Cancer: 50,811-816 (1992) 0 1992 Wiley-Liss, Inc.
Publicationof the International Union Against Cancer Publication de I’Union Internationale Contre le Cancer
AN IDIOTYPIC REPLICA OF CARCINOEMBRYONIC ANTIGEN INDUCING CELLULAR AND HUMORAL RESPONSES DIRECTED AGAINST HUMAN COLORECTAL TUMOURS L.G. DURRANT’’~, G.W.L. DENTON’,E. JACOBS’, M. MEE’, R. MOSS’, E.B. AUSTIN’,R.W. BALDWIN’,J.D. HARDCASTLE* and R.A. ROBINS’ ’CRC Laboratories, University of Nottingham, Nottingham, N G 7 2RD; and ‘Department of Surgery, University of Nottingham, UK A monoclonal anti-idiotypic antibody (anti-Id MAb) 708 (I&), which inhibitedthe binding of NCRC23 (IgG,) MAb to CEA and prevented radiolabelled CEA from binding to MAb NCRC23, was produced. No recognition of 3 other anti-CEA antibodies, 3 other IgG, or 2 IgM MAbs was observed with this anti-idiotypic antibody. When an immunoblotting technique was used, 708 anti-Id MAb failed to bind to isolated heavy or light chains of MAb NCRC23, whereas binding was observed with intact antibody. Mouse, rat and human lymphocytes (in vitro) were immunized with 708 anti-Id MAb and the resultant Ab3 antibodies all inhibited binding of labelled 708 anti-Id MAb to MAb NCRC23 and also reacted with CEA, showing that 708 anti-Id MAb induced anti-CEA antibody responses. Similarly, mice immunized with 708 anti-Id MAb could be restimulatedin vitro with either CEA or tumour cells expressing CEA which induced specific T-cell proliferative responses. Human tumour-infiltrating lymphocytesisolated from colorectal tumours or peripheral blood T cells from cancer patients were stimulated in vitro with 708 anti-Id MAb or an irrelevant IgG,, antibody. Six days later both sets of lymphocyteswere restimulatedwith CEA, and lymphocytes primed to 708 anti-Id MAb proliferated in response to CEA. These results suggest that 708 anti-Id M A b can act as an idiotypic replica of CEA and stimulate cellular and humoral anti-CEA immune responses. It is therefore of great interest as an idiotypic vaccine against colorectal cancer.
There is considerable interest in the development of antiidiotypic antibodies for use as vaccines against a number of diseases, including cancer. In the normal course of events, tumour growth and expansion may continue for want of an immune response, either because the tumour-associated antigens are recognized as self or due to repression of anti-tumour immune responses. Anti-idiotypic antibodies may induce a primary, i.e. an anti-anti-idiotypic, response against tumour antigen, or they may select or amplify a pre-existing immune repertoire by up-regulating a normally repressed response. Induction of an immune response by antibodies acting as “network antigens” is attractive, as anti-idiotypic antibodies may activate part of the immunological repertoire that is not susceptible to stimulation by antigen. This may act via activation of “silent clones” (Bona et al., 1981; Stein and Soederstroem, 1984; Schick et aL, 1987), or provide selective pressure to promote the induction of an idiotype for which the appropriate germ-line genes are not present (Moser et al., 1983). Thus, the possibility of activating responses not inducible by antigen makes idiotypic intervention very attractive for tumour immunot herapy. CEA is a complex glycoprotein (M, 180,000)found in 90% of colorectal cancers and 77% of gastric cancers and in an average of 87 ? 5% of epithelial cells per lesion (Durrant et al., 1989). It is therefore an appropriate target antigen for idiotypic immunotherapy. Various glycoproteins, immunologically related to CEA, are reported to be present on normal tissues. These include normal cross-reacting antigen (NCA) which has a tissue distribution similar to that of CEA but is also found on normal granulocytes and macrophages; NCA-2 and normal foetal antigen (NFA-2) which are found in meconium and adult faeces; and biliary glycoprotein I which is present in normal bowel (Rogers, 1983). A murine MAb, NCRC23, has been produced which binds to a CEA-specific epitope (Price et
al., 1987) and shows minimal cross-reactivity with normal tissues. It has also been used for the radioimmunodetection of tumours in colorectal cancer patients without any adverse reaction (data not shown). This antibody has been used to raise anti-idiotypic antibodies for immunotherapy of colorectal cancer. MATERIAL AND METHODS
Monoclonal antibodies NCRC23 (IgG,) and NCRC24 (IgG,,) are MAbs which recognize CEA-specific epitopes (Price et al., 1987). NCRC16 (IgG,) and NCRC17 (IgG,) are MAbs which recognize both CEA and NCA (Price et al., 1987). H65 and NCRC34 are IgG, MAbs which recognize CD5 and cytokeratin 18, respectively. C14 and NCRC30 are IgM antibodies and NCRC37 is an IgG, antibody; all of these recognize C14gp200 colorectal-tumour-associated antigen (Price et al., 1986). 381, an IgG, MAb used as the control antibody, does not bind to CEA. 730 is an anti-idiotypic mouse MAb recognizing the binding site of 791T/36 anti-tumour MAb (Doran et al., 1990). Immunization protocols and MAb production BAL,B/c mice were immunized by S.C. injection of 50 pg of MAb NCRC23 coupled to solubilized keyhole-limpet haemocyanin by glutaraldehyde fixation, and emulsified in Freund’s complete adjuvant. Animals were boosted 15 and 28 days after the primary immunization with 50 pg of the same immunogen in Freund’s incomplete adjuvant. On day 59, 4 days prior to fusion, animals were boosted with 50 pg of the same immunogen by i.v. injection. Splenocytes were fused with the mouse myeloma NSO. Culture supernatants were screened by ELISA for the presence of anti-idiotypes and positive hybridomas were cloned 3 times by limiting dilution. Isotypes were determined by means of a kit from Binding Site, Birmingham, UK. MAbs were purified from culture supernatants by affinity chromatography on protein A/Sepharose (Pharmacia, Uppsala, Sweden). Sandwich idiotype ELISA This assay was used to detect anti-idiotype antibodies in culture supernatant and to determine the specificity of the purified anti-idiotypic MAb. MAb NCRC23 or control antibodies were diluted to 5 pg/mI in PBS and added (50 pl/well) to polyvinyl chloride flat-bottomed microtitre plates. After overnight incubation at 4”C, wells were blocked with 50 p1 of BSA (1% in PBS) for 1 hr at room temperature. Culture supernatants or purified antibodies were added (50 pl/well) and incubated at room temperature for 1 hr. After washing, the biotinylated antibodies (identical to those used to coat the plate) were added at a concentration of 25 ng/well. After further washing, avidin peroxidase (BRL, Gaithersburg, MD)
’TOwhom correspondence and reprint requests should be sent. Received: October 12,1991
812
DURRANT ET AL.
was added and the plates were left for 1 hr at room temperature. After 5 washings in PBS/O.O5% Tween 20, ABTS substrate was added in H,O, and the optical density was read after 10 min at 405 nm.
Direct idiotype ELISA Microtitre plates were coated with MAb NCRC23 or other anti-CEA or IgG, MAbs at 5 p,g/ml overnight and blocked with BSA. The biotinylated 708 anti-Id MAb was added at 500 ng/ml in the presence of 0-5,000 ng/ml of unlabelled 708 anti-Id MAb for 1 hr at room temperature. After washing, avidin peroxidase was added and the assay developed. Inhibition of binding of 708 anti-Id MAb to MAb NCRC23 by anti-708 anti-id MAb antibodies Microtitre plates were coated with MAb NCRC23 and blocked with BSA. Biotinylated 708 anti-Id MAb (50 ng/well) was pre-incubated for 1 hr at room temperature with dilutions of a rat serum pool (1/3-1/30,000), a mouse serum pool (1/3-1/300,000) or human tissue culture supernatant (neat to 1/loo) before addition to the MAb NCRC23-coated plates for a further 1 hr at room temperature. After washing, avidin peroxidase was added and the assay developed as above. Direct CEA-bindingELISA CEA (10 pg/ml) was dried on to microtitre plates at 37°C overnight and blocked with BSA. Rat and mouse sera and supernatant from in vitro culture of human B cells were added, and after 1 hr at room temperature, bound antibody was detected with either rabbit anti-mouse peroxidase (Dakopatts, Copenhagen, Denmark), goat anti-rat peroxidase (Caltag, Leicester, UK) or goat anti-human peroxidase (Sigma, Poole, UK). The assays were then developed. To determine whether the anti-708 anti-Id MAbs recognized the same epitope as MAb NCRC23, rat and mouse sera and human supernatants were added to CEA-coated plates and then biotinylated MAb NCRC23 was added. After 1 hr at room temperature, the assay was developed with avidin peroxidase.
Flow cytometnc competition assays A modification of the assay of Robins et al. (1986) was used to measure inhibition of MAb NCRC23 or other anti-CEA MAbs binding to CEA-coated polystyrene beads by 708 anti-Id MAb. Briefly, MAb NCRC23 or other anti-CEA antibodies labelled with fluorescein isothiocyanate (FITC) were preincubated with 708 anti-Id MAb for 30 min at 4°C. The antibody mixture was then added to lo5CEA-coated beads and incubated for a further 30 min at 4°C. The level of FITClabelled antibody binding to CEA beads was determined by flow cytometric measurements of fluorescence intensity on a FACS IV.
Inhibition of binding of CEA by anti-idiotype Plates were coated with MAb NCRC23 and blocked with BSA. 708 anti-Id MAb (10 p,g/ml) was added to half the plate and then 30-1,000 ng of radiolabelled lZ5ICEA were added to 708 anti-Id MAb blocked and unblocked wells. Radioactivity was counted on a gamma counter. Mouse T-cell assays Mice were immunized with 708 anti-Id MAb which had been coupled to KLH by glutaraldehyde fixation and emulsified in Freund’s complete adjuvant. Two weeks later, they were re-injected with the same immunogen in Freund’s incomplete adjuvant; 2 weeks after this, splenocytes were obtained. They were plated into round-bottomed microtitre plates in RPMI 1640 containing 5 x M mercaptoethanol and 10% FCS at 5 x lo4 cells/well. To quadruplicate sets of wells were added previously irradiated (5,000 rad) CEA-positive (MKN45) and CEA-negative (791T) cells at a density of lo4 cells/well, or
CEA antigen (10 pg/ml), or epithelial mucin antigen (10 pgiml) as a negative control. To separate quadruplicate wells were added 708 anti-Id MAb or irrelevant IgG,, MAb at 1-10,000 ng/well. Cells were incubated for 5 days at 37°C and then pulsed for 24 hr with 1 pCi of [3H] thymidine. Radioactivity was counted on a liquid scintillation counter.
Human-tumour-infiltrating lymphocyte assay Colorectal turnour was finally minced and disaggregated (Durrant et al., 1986). A single-cell suspension was sorted for infiltrating (lymphocytes, monocytes and fibroblasts: TLA) and tumour cells were sorted by the MAb magnetic-bead technique (Durrant, 1991). Alternatively, peripheral blood lymphocytes (PBL) were separated on lymphocyte separation medium (Durrant et al., 1984). Either the lymphocytes were stimulated directly, or T cells were purified on magnetic beads (Dynal, Oslo, Norway) coated with an anti-CD2 MAb, then the remaining cells were irradiated and used as antigen-presenting cells. Cells were then stimulated with either 708 anti-Id MAb or IgG,, myeloma antibody at 10 pg/ml, challenged after 6 days with CEA (10 pg/ml) or a control antigen (human epithelial mucin, 10 pg/ml), left for a further 5 days, pulsed with [3H] thymidine (1 pCi/well) and counted on a liquid scintillation counter. Immunoblotting of M b NCRC23 and 708 anti-Id MA6 MAb NCRC23 and 708 anti-Id MAb were reduced in 5% mercaptoethanol, and separated heavy and light chains were resolved in a 7.5% polyacrylamide, Phastgel (Pharmacia, Milton Keynes, UK). Non-reduced antibody was also run in this gel. Protein was transferred to nitrocellulose by Western blotting using a Phast system. Blots were blocked with 1%BSA in PBS containing 0.05% Tween-20. Biotinylated 708 anti-Id MAb was added to the MAb NCRC23 blot and biotinylated MAb NCRC23 to the 708 anti-Id MAb blot. The blots were then developed with avidin peroxidase and 0.4% AEC and H,O,(Austin et al., 1989). Statistics All results were analysed for significance by Student’s t-test. RESULTS
Generation of anti-idiotype Mitbs Out of 183 hybridoma supernatants screened, 8 were positive according to MAb NCRC23 sandwich idiotype ELISA. After cloning, an IgG,, MAb, 708, was obtained. To assess the specificity of this MAb, it was tested in sandwich idiotype ELISA against 3 other IgG, MAbs and 2 IgM antibodies. 708 anti-Id MAb could only bridge MAb NCRC23 antibody (Table I). An irrelevant anti-idiotypic MAb, 730, failed to bridge MAb NCRC23 (Table I). Similarly, specificity was also tested by direct idiotype ELISA (Fig. 1). Biotinylated 708 anti-Id MAb TABLE I - ELISA TO DETECT WHETHER THE ANTI-IDIOTYPIC MAb 708 CAN BRIDGE MAbs
MAb
NCRC23 H65 NCRC37 NCRC34 C14 NCRC30
708 MAb
Absorbance % SE (405 nm) 730 anti-Id MAb
0.812 +- 0.17 0.161 +- 0.026 0.217 & 0.01 0.105 & 0.01 0.175 ? 0.013 0.129 0.014
0.114 2 0.01 NDZ 0.101 2 0.02 ND ND 0.141 k 0.01
Antigen’
1.24
k 0.08 ND 1.115 2 0.06 0.650 2 0.02 2.247 2 0.154 2.667 t 0.119
‘Antigen was included as a positive control to show that the MAbs can bridge. The antigen for NCRC23 was CEA; that for NCRC37, C14 and NCRC30 was C14gp200; and that for NCRC34 was cytokeratin 18.-’ND, not determined.
813
INDUCTION OF ANTI-CEA IMMUNE RESPONSES
TABLE 111- INHIBITION OF BINDING OF CEA TO MAb NCRC23 BY 708 ANTI-IDIOTYPIC MAb
0.8 1
Concentration of "'1 CEA (ng)
E 0.4 3
{
\
I !
1000 300 100 30
0.2
n
4:
Binding of '''1 CEA to NCRC23 (cpm in the presenceof:
2
SE)
~
Nothing
kG*t' (500 ne)
708 (500 ng)
55,883 f 4073 56,821 k 3871 2,795 t 191 20.222 t 361 29.416 f 581 1.243 +- 58 '488 f 5 81621 & 208 71636 f 174 2.027 f 19 1.981 2 40 214 t 13 186 t 46 194 2 16 209 f 73
0.0 1
1
I708
10 antibody]
100
1000
(pg/ml)
FIGURE 1 - Binding of 708 anti-Id MAb to 4 anti-CEA MAbs, NCRC23 (W),NCRC24 (A),NCRC16 ( 0 )and NCRC17 (O), and to another IgG, MAb NCRC37 (A). Binding of 730 anti-Id MAb to NCRC23 (0).Error bars are covered by symbols. TABLE I1 - INHIBlTION OF BINDING OF ANTI-CEA MAbs TO CEA BY 708 ANTI-IDIOTYPIC MAb
Concentration of 708 anti-Id MAb (ILgW
166 55 18 6
2 Concentration of 730 anti-Id MAb (uelml)
166
%binding ? SE NCRC23 (20 uelml)
NCRC24 (20 uelml)
NCRC17 (20 ueirnl)
2 t 0.05 3 k 0.01 86 t 0.07 105 f 1.0 101 f 1.1
56 k 0.1 86 f 3.0 96 f 1.6 99 k 2.0 94 +- 2.0
101 f 4.0 100 t 2.0 99 f 1.0 103 t 2.0 101 t 2.0
NCRC23 (20ILgiml)
18 6
96 f 1 101 f 2 104 t 5 97 t 1
2
99 f 1
55
could bind to plates coated with MAb NCRC23 but not to another CEA-specific MAb, NCRC24, or to 2 CEA/NCAreactive MAbs, NCRC16 and NCRC17, or to another IgG, antibody. 730 anti-Id MAb failed to bind to MAb NCRC23. Binding of biotinylated 708 anti-Id MAb to MAb NCRC23 was inhibited by unlabelled 708 anti-Id MAb. If an anti-idiotype MAb (Ab2) is capable of mimicking antigen, it must recognize idiotypic determinants within the antigen-combining site of Abl and therefore inhibit binding of Abl to antigen and binding of antigen to Abl. 708 anti-Id MAb inhibited the binding of MAb NCRC23 to CEA (Table 11) but had no effect on the binding of MAb NCRC17, and only at very high concentrations did it inhibit binding of NCRC24 to CEA. Equimolar concentrations of 708 anti-Id MAb and MAb NCRC23 resulted in 86% inhibition of binding. 730 anti-Id MAb failed to prevent MAb NCRC23 from binding to CEA. 708 anti-Id MAb (500 ng) also inhibited binding of radiolabelled CEA to MAb NCRC23-coated microtitre plates (Table 111) whereas 730 anti-Id had no effect.
Immunoblottingof 708 anti-Id MAb and MAb NCRC23 Idiotypes can be present on the isolated heavy and light chains of immunoglobulins or result from the association of these 2 chains. After reduction, the heavy and light chains of 708 anti-Id MAb and MAb NCRC23 antibodies were separated electrophoretically and transferred onto nitrocellulose. 708 anti-Id MAb bound to intact, but not to isolated, heavy and light chains of MAb NCRC23, and MAb NCRC23 bound to
FIGURE 2 - Immunoblots of (A-E) 708 anti-Id MAb and (E-J) MAb NCRC23. Blots (A) and (D) are of non-reduced 708 anti-Id MAb developed with anti-mouse antiserum and biotinylated NCRC23, respectively. Immunoblots (B,C,E) are of reduced 708 anti-Id MAb developed with anti-mouse antiserum, anti-kappa light chain and biotinylated NCRC23, respectively. Immunoblots (F) and (I) are of non-reduced MAb NCRC23 developed with anti-mouse antiserum and biotinylated 708 anti-Id MAb, respectively. Immunoblots (G,H,J) are of reduced NCRC23 developed with anti-mouse antiserum, anti-kappa light chain and biotinylated 708 anti-Id MAb, respectively. Molecular weights are shown to the left. intact, but not to isolated, heavy and light chains of 708 anti-Id MAb (Fig. 2).
Characterization of antisera generated by immunization with anti-708 anti-Id MAb Mice and rats were immunized with 708 anti-Id MAb or an IgG,, MAb linked to KLH in the presence of either Freund's adjuvant or the saporin adjuvant Quil A (Superfos Biosector als Vedbaek Denmark). Human peripheral blood lymphocytes were stimulated in vitro with 708 anti-Id MAb or IgG,, MAb. The resultant antisera or culture supernatants were screened for inhibition of binding of 708 anti-Id MAb to MAb NCRC23 and for direct binding to CEA. Antiserum from animals immunized with anti-Id MAb 708 KLH in the presence of either Freund's or Quil A adjuvant gave, in each case, significantly greater inhibition of binding of biotinylated 708 anti-Id MAb to MAb NCRC23 (Fig 3a,b) and higher binding to CEA (Fig. 4a,b) than did antiserum from animals immunized with an irrelevant IgG,, MAb. The antiserum from animals immunized with anti-Id MAb 708/KLH in Freund's adjuvant was superior, with end-point titres for binding to CEA of 1/1,000 to 1/3,000 for the mouse serum and 1/100 to 1/300 for the rat serum. The rat serum failed to bind to normal colon antigen (NCA), whereas weak binding was detected in the 708 anti-Id MAb Freund's adjuvant antiserum (titre 1/100), the IgG, MAb Freund's adjuvant antiserum (titre 1/10) and the 708 anti-Id MAb Quil A antiserum (titre
814
DURRANT ET AL. 1207
a
r'"
V
".la
.
0001
,001
.01
.1
/**
I
0
1
,0001
,001
.01
1
1
DILUTION OF MOUSE SERUM (1TTITRE)
DILUTION OF MOUSE SERUM (1TTITRE)
b
0.0 ,0001
,001
.01
1
1
,0001
0
DILUTION OF RAT SERUM (1ITITRE)
120 110
1
c 0.71
T
T-T
100
CI
,001
.01
1
1
DILUTION OF RAT SERUM (1TTITRE)
o,61
c
0.5
90
0
f 80 s
70 0.3 60
0.2
50
1
3 9 27 TISSUE CULTURE SUPERNATANT (1TTITRE)
81
1
DILUTION
2
6 F TISSUE
4
8
16
32
CULTURE SUPERNATANT FROM B CELLS
FIGURE 3 - Binding of 708 anti-Id MAb to MAb NCRC23 in the presence of (a) 708 anti-Id MAb immune mouse serum, (b) 708 anti-Id MAb immune rat serum and (c) 708 anti-Id-MAbstimulated human-B-cell supernatant. Animals were immunized with 708 anti-Id MAb in Freund's adjuvant (W), 708 anti-Id MAb in Quil A adjuvant (O), IgG, in Freund's adjuvant ( 0 )or IgG,, in Quil A (0).In (c), the human-B-cell supernatants were stimulated with 0 (a)or 10 (0)ng/ml of 708 anti-Id MAb or 10 ng/ml irrelevant IgG,, MAb (H). In a and b the error bars are covered by the symbols.
FIGURE 4 - Binding of (a) 708 anti-Id MAb immune mouse serum, (b) 708 anti-Id MAb immune rat serum and (c) 708 anti-Id-MAb-stimulated human B-cell tissue culture supernatant to CEA. Animals were immunized with 708 anti-Id MAb in Freund's adjuvant (W), 708 anti-Id MAb in Quil A adjuvant (O), IgG, in Freund's adjuvant ( 0 )or IgG,, in Quil A (0).The human B cells were stimulated in vitro with either 0 (a), 1(A), 10 (M)or 100 ( 0 )ng/ml708 anti-Id MAb or 10 ng/ml irrelevant IgG MAb (A). The error bars are covered by the symbols. *p > O.OOl,zb**p > 0.01.
1/ 10). Similarly, culture supernatants from human peripheralblood mononuclear cells which had been stimulated in vitro with 708 anti-Id MAb contained antibodies which inhibited 708 anti-Id MAb binding to MAb NCRC23 (Fig. 3c) and bound directly to CEA (Fig. 4c). Unstimulated B-cell cultures of B cells stimulated with an irrelevant IgG,, MAb produced no such antibodies (Fig. 3c, 4c).
negative human tumour-cell lines (Table IV). The splenocytes gave a significantly higher proliferative response to the CEApositive than to the CEA-negative cell lines and to CEA antigen compared with breast tumour mucin. The splenocytes also proliferated in response to re-challenge with 708 anti-Id MAb and also to IgG,, (Table IV). Human tumour-infiltrating lymphocytes (TILs) were separated on magnetic beads coated with an anti-CD2 MAb from an enzymatically disaggregated single-cell suspension of colorectal tumour cells. These lymphocytes were stimulated in vitro in the presence of irradiated antigen-presenting cells with
Cellular immune responses induced by 708 anti-Id MAb Mice primed in viva with anti-Id MAb 708/KLH were screened for proliferation against CEA-positive and CEA-
INDUCTION OF ANTI-CI?A IMMUNE RESPONSES
815
leads to the induction of anti-tumour effects, resulting in suppression of tumour growth. In a syngeneic rat system, tumour suppression correlates with induction of a cellular rather than an antibody response (Dunn et al., 1987). Incorporation of ['HI thymidine (cpm ? SD) Three approaches to anti-idiotypic therapy of human cancer In vitro challenge 708 anti-Id IgG, have been reported. Colorectal patients have been treated primed primed (Herlyn et al., 1987) with heterologous (goat) anti-idiotypic 1,191 t 188 923 2 196 791T Cells (lo4) antibodies raised against the MAb 17-1A. Melanoma patients 961 t 227 MKN45 Cells (lo4) 2,274 2 182 have been treated with a mouse anti-idiotypic MAb which mimics the high-molecular-weight human melanoma antigen 708 anti-Id MAb 10 pg/ml 4,847 t 436 4,833 2 48 (Mittleman et al., 1990). Both of these trials revealed some 1 pg/ml 6,993 2 649 ND favourable responses although they were aimed at inducing IgG,, MAb humoral rather than cellular immune responses, and both 7,659 2 54 5,516 t 528 10 pg/ml induced anti-species antibodies. Our own trial in colorectal 6,567 2 118 ND 1 pg/ml cancer patients with a human anti-idiotypic MAb, which 2,745 2 149 274 t 64 CEA recognizes the binding site of the anti-tumour MAb 791T/36, 240 2 5 EMA 657 2 92 showed cellular anti-tumour immune responses in 5 out of 6 patients and high levels of plasma IL-2 in 4 out of 6 patients ND. not determined. (Robins et al., 1991). No antibody responses were observed, and there was no associated toxicity. Two of the patients had stable disease, for a period of 15 months in one case, and were alive and well at the time of writing. The aim of this study was to investigate other tumourx 6 associated antigens as potential targets for idiotypic immunoW n therapy. CEA is a promising target as it is a homotypic z 5 adhesion molecule widely expressed on colorectal tumours. Cancer patients are almost always tolerant to this antigen but 5 4 idiotypic immunization has been shown to break self-tolerance i= in other systems. In this communication we describe the production and characterization of an anti-idiotypic MAb 2 3 W which acts as a network antigen (Kohler et al., 1989). LL 708 anti-Id MAb reacts specifically with NCRC23 MAb and A 2 can inhibit binding of this antibody to CEA and of CEA to this 0 U antibody. The binding of 708 anti-Id MAb to MAb NCRC23 n 1 requires an association of the heavy and light chains, as immunoblotting of separated chains destroys recognition. n " Analysis of the Ab3 response in animals immunized with 708 TLA-1 TLA-2 PEL-2 PEL-3 anti-Id MAb clearly shows that Ab3 antibodies which prevent 708 anti-Id MAb from binding to MAb NCRC23 and which FIGURE 5 - Proliferative response of human-tumour-infiltrating lymphocytes to 708 anti-Id MAb and CEA antigen. Lymphocytes bind directly to CEA are produced. Similarly, immunization of were treated with 708 anti-Id MAb (m) or IgG,, (0) and after 6 human lymphocytes in vitro with 708 anti-Id MAb resulted in days they were rechallenged with CEA for a further 7 days. human antibodies which recognized both 708 anti-Id MAb and Alternatively,lymphocytes were primed with 708 anti-Id MAb (m) CEA antigen. These results provide strong evidence that the or IgG,, (B) and challenged 6 days later with human epithelial former can act as a replica of a B-cell epitope on CEA. antigen. Several other research groups have also identified CEA as a potential target for idiotypic immunotherapy. Monestier et al. either 708 anti-Id MAb or an irrelevant IgG, antibody and (1989) produced 2 murine MAbs which recognized determithen 6 days later they were challenged with CEA (Fig. 5). Only nants within the antibody-combining site of a CEA-specific TILs which had been primed with 708 anti-Id MAb prolifer- MAb. Kohler's group have recently produced a mouse MAb to ated in response to challenge with CEA. Peripheral blood the combining site of a CEA-specific mouse MAb which lymphocytes were also isolated from 2 cancer patients. TILs induces anti-CEA antibodies in mice (Bhattacharya-chatterwere also isolated from the tumour of 1 of these patients for jee et aL, 1990). This anti-anti-idiotypic mouse anti-serum has direct comparison. Either lymphocytes or T cells, primed with a fine specificity which is similar, but not identical, to that of 708 anti-Id MAb in the presence of antigen-presenting cells, the original anti-CEA mouse MAb. responded with enhanced proliferation to a second challenge 708 anti-Id MAb is, however, unique in that it is the first with CEA antigen. The T cells failed to respond if they had anti-idiotypic replica of CEA which can not only prime mouse been primed with IgG, or if they had been primed with 708 splenocytes to proliferate in response to CEA and CEAanti-Id MAb and then challenged with an irrelevant antigen expressing human tumour cells but can also prime humansuch as epithelial much antigen. tumour-infiltrating lymphocytes and peripheral-blood T cells from cancer patients to proliferate in vitro specifically to CEA antigen. The epitopes recognized by T cells are small linear DISCUSSION peptides which are easily presented by major histocompatibilAnti-tumour responses induced by anti-idiotypic manipula- ity molecules. Even though 708 anti-Id MAb only binds to tion have been demonstrated against both murine- (Forstrom intact MAb NCRC23, the T-cell epitope which stimulates cell et aL, 1983; Lee et aL, 1986) and rat-tumour-associated proliferation may be a linear determinant expressed on the antigens. It has been reported (Kennedyet al., 1985; Raychaud- heavy or light chain. This has been shown for an anti-idiotypic huri et al., 1987) that stimulation of anti-idiotypic responses antibody which mimics mycobacterial protein. Ab2 only bound TABLE N - PROLIFEKATION OFlN WVO-PKIMED 708 ANI'I-ID .MAb AND IgG,, MAh MOUSE SPLENOCYTES IN RESPONSE TO CEA-NEGAIIVE (791TJ, CEA-POSITIVE (MKN45J TUhlOUR CELL I.ISES, CEA, EPITHEI.IAL MUCIV. 7U8 AKTI-ID MAh AND IeG,, MAh
71 T
816
DURRANT ET AL.
to intact Abl (Rees et al., 1987) but a specific T-cell response could be stimulated by heavy chain alone (Rees and Lamb, 1990). Similarly a putative T-cell epitope has been identified in the CDR2 domain of an anti-idiotypic MAb 2F10 which induces protection against transfer and growth of L1210/GZL tumour (Raychaudhuri et al., 1990).
708 anti-Id MAb appears to be suitable for immunotherapy of gastrointestinal cancer. As it is a mouse antibody, it may be necessary to immunize with antibody fragments or genetically cloned Fab to avoid anti-mouse humoral responses which may interfere with repeated immunizations. Alternatively, the antibody could be humanized (Reichmann et aL, 1988).
REFERENCES
AUSTIN,E.B., ROBINS,R.A., DURRANT,L.G., PRICE, M.R. and BALDWIN, R.W., Human monoclonal anti-idiotypic antibody to the tumour-associated antibody 791T/36. Immunology, 67,525-530 (1989). BHATTACHARYA-CHATTERJEE, M., MUKERJEE, S., BIDDLE,W.. FOON, K.A. and KOHLER,H., Murine monoclonal anti-idiotype antibody as a potential network antigen for human carcinoembryonic antigen. J. Immunol., 145,2758 (1990). BONA,C.A., HERBER-KATz, E. and PAUL,W.E., Idiotype-anti-idiotype regulation. I. Immunization with a Levan binding myeloma protein leads to the appearance of auto-anti-(anti-idiotype) antibodies and to the activation of silent c1ones.J. ap. Med., 153,951-961 (1981). DORAN,M., DURRANT, L.G., AUSTIN,E.B., BALDWIN,R.W. and ROBINS,R.A., A syngeneic anti-idiotypic antibody to an anti-tumour antibody which induces cellular and humoral anti-tumour responses. Brit. J. Cancer, 62,500 (1990). DUNN,C.P.L.., JOHNSON, C.A., STYLES,J.M., PEASE,S.S. and DEAN, C.J., Vaccination with syngeneic monoclonal anti-idiotype protects against a tumour challenge. Immunology, 60,181-186 (1987). DURRANT, L.G., Rapid and efficient method for separating tumour infiltrating cells and tumour cells from colorectal tumours. J. immunol. Meth., (1991). (In press). DURRANT, L..G., PARKER,M., KENWORTHY, N. and TAYLOR,G.M., Characterisation of a human x mouse T cell hybridoma and identification of a clone secreting and binding interleukin-2. Immunology, 52, 119-124 (1984). DURRANT, L.G., ROBINS,R.A., ARMITAGE, N.C., HARDCASTLE, J.D. and BALDWIN, R.W., Association of antigen expression and DNA ploidy in human colorectal tumours. Cancer Rex, 46, 3543-3549 (1986). DURRANT, L.G., ROBINS,R.A. and BALDWIN, R.W., Flow cytometric screening of monoclonal antibodies for drug or toxin targeting to human cancer. J. nat. Cancer Inst., 81,688-691 (1989). I. and FORSTROM, J.W., NELSON,K.A., NEPOM,G.T., HELLSTROM, HELLSTROM. K.E., Immunization to a syngeneic sarcoma by a monoclonal auto-anti-idiotypic antibody. Nature (Lond.), 303, 627-629 (1983). HERLYN, D., WETTENDORF, M., SCHMOLL, E., ILIOPOULOS, D., SCHEDEL,I.. DREIKHAUSEN, V., RUB, R., ROSS,A.H., JAKSCHE, H., SCRIBA, H., Anti-idiotype immunization of cancer paM. and KOPROWSKI, tients: modulation of the immune response. Proc. nat. Acad. Sci. (Wash.), 84,8055-8059 (1987). KENNEDY, R.C., DREESMAN, G.R., BUTEL,J.S. and LANFORD,R.E., Suppression of an in vivo tumour formation induced by SV40transformed cells in mice receiving anti-idiotypic antibodies. J. exp. Med., 161,1432-1436 (1985). T., SRINIVASAN, S., KAVERI,S., MORKOHLER,H., KIEBER-EMMONS, ROW, W.J.W., MULLER, S., KANG,C-Y. and RAYCHAUDHURI, S., Short analytical review: revised immune network concepts. Clin. Immunol. Immunopatliol., 52,104-116 (1989). LEE,V.K., I~ELLSTROM,K.E. and NEPOM,G.T., Idiotypic interactions in immune responses to tumour associated antigens. Biochim. biophys. Acta, 965,127-128 (1986). T., YANG,H., YAMADA, M., MITTLEMAN, A., CHEN,Z.J., KAGESHITA, BASKIND, P.,GOLDBERG, N., Puccio, C . , AHMED,T., ARLIN,2. and S., Active specific immunotherapy in patients with melaFERRONE, noma. J. chi. Invest., 86,2136-2144 (1990).
MONESTIER, M., DEBBAS,M.E. and GOLDENBERG, D.M., Syngeneic anti-idiotype monoclonal antibodies to murine anti-carcinoembryonic antigen monoclonal antibodies. Cancer Rex, 49,123-127 (1989). M., LAI, A.C.K., NORTON,G., REALE, MORAN,T.M., MONESTIER, M.A., THOMPSON, M.A., SCHULMAN, J.L., RIRLET,R. and BONA,C.A., Characterisation of variable region genes and shared cross-reactive idiotypes of antibodies specific for antigens of various influenza viruses. Viral Immunol., 1,l-12 (1987). MOSER,M., HIERNEUX, L.J. and URBAIN, J., Idiotypic manipulation in mice: BALBic mice can express the cross-reactive idiotype of A/J mice. Proc. nat. Acad. Sci. (Wash.), 80,4474-4479 (1983). PRICE,M.R., EDWARDS, S. and BALDWIN, R.W., Association of the Y hapten with glycoproteins, glycolipids and carcinoembryonic antigen in colorectal carcinoma. Cancer Lett., 33,83-89 (1986). S., JACOBS,E. and BALDWIN, R.W., Mapping PRICE,M.R., EDWARDS, of monoclonal antibody-defined epitope associated with carcinoembryonic antigen, CEA. Cancer Immunol. Immunother., 25,lO-15 (1987). RAYCHAUDHURI, S., KANG,C-Y., KAVERI,S-V., KIEBER-EMMONS, T. and KOHLER,H., Tumor idiotype vaccines. VII. Analysis and correlation of structural, idiotypic and biologic pro erties of protective and non protective Ab2. J. Immunol., 145,760-76y (1990). RAYCHAUDHURI, S., SAEKI,Y., CHEN,J-J, IRIBE,H., FUJI, H. and KOHLER,H., Tumour-specific idiotype vaccines. 11. Analysis of the tumor-related network response induced by the tumor and by internal image antigens (Ab2beta). J. Immunol., 139,271-280 (1987). REES, A.D.M. and LAMB,J.R., Human T-cell recognition of an “internal image” of mycobacterial antigen in an anti-idiotypic antibody. In: A. Osterhaus and F. Uytdehaag (eds.), Idiotypic networks in biology and medicine, pp. 195-201, Elsevier, Amsterdam (1990). REES,A.D.M., SCOGING, A,, DOBSON, N., PRAPUTPIITAYA, K., YOUNG, D., IVANYI,J. and LAME,J.R., T-cell activation by anti-idiotypic antibody: evidence for the internal image. Europ. J. Immunol., 17, 197-201 (1987). REICHMANN, L., CLARKE, M., WALDMANN, H. and WINTER,G., Reshaping human antibodies for therapy. Nature (Lond.), 332, 323327 (1988). ROBINS,R.A., DENTON,G.W.L., HARDCASTLE, J.D., AUSTIN,E.B., BALDWIN, R.W. and DURRANT, L.G., Anti-tumor immune response and interleukin-2 production induced in colorectal cancer patients by immunization with human monoclonal anti-idiotypic antibody. Cancer Res., 51,5425-5429 (1991). ROBINS,R.A., LAXTON,R.R., GARNETT,M., PRICE,M.R. and BALDWIN,R.W., Measurements of tumour-reactive antibody and antibody conjugates by competition, quantitated by flow cytometry. J. immunol. Meth., 90,165-171 (1986). ROGERS,G.T., Carcinoembryonic antigens and related glycoproteins. Molecular aspects and specificity.Biochim. biophys.Acta, 695,227-249 (1983). SCHICK,M.R., DREESMAN, G.R. and KENNEDY, R.C., Induction of an anti-hepatitis B surface antigen response in mice by non-internal image (Ab2p) anti-idiotypic antibodies. J. Immunol., 138, 3419-3424 (1987). STEIN,K. and SOEDERSTROEM, T., Neonatal administration of idiotype or anti-idiotype primes for protection against E. coli K13 infection in mice.J. exp. Med., 160,1001-1007 (1984).
Publicationof the International Union Against Cancer Publication de I’Union Internationale Contre le Cancer
AN IDIOTYPIC REPLICA OF CARCINOEMBRYONIC ANTIGEN INDUCING CELLULAR AND HUMORAL RESPONSES DIRECTED AGAINST HUMAN COLORECTAL TUMOURS L.G. DURRANT’’~, G.W.L. DENTON’,E. JACOBS’, M. MEE’, R. MOSS’, E.B. AUSTIN’,R.W. BALDWIN’,J.D. HARDCASTLE* and R.A. ROBINS’ ’CRC Laboratories, University of Nottingham, Nottingham, N G 7 2RD; and ‘Department of Surgery, University of Nottingham, UK A monoclonal anti-idiotypic antibody (anti-Id MAb) 708 (I&), which inhibitedthe binding of NCRC23 (IgG,) MAb to CEA and prevented radiolabelled CEA from binding to MAb NCRC23, was produced. No recognition of 3 other anti-CEA antibodies, 3 other IgG, or 2 IgM MAbs was observed with this anti-idiotypic antibody. When an immunoblotting technique was used, 708 anti-Id MAb failed to bind to isolated heavy or light chains of MAb NCRC23, whereas binding was observed with intact antibody. Mouse, rat and human lymphocytes (in vitro) were immunized with 708 anti-Id MAb and the resultant Ab3 antibodies all inhibited binding of labelled 708 anti-Id MAb to MAb NCRC23 and also reacted with CEA, showing that 708 anti-Id MAb induced anti-CEA antibody responses. Similarly, mice immunized with 708 anti-Id MAb could be restimulatedin vitro with either CEA or tumour cells expressing CEA which induced specific T-cell proliferative responses. Human tumour-infiltrating lymphocytesisolated from colorectal tumours or peripheral blood T cells from cancer patients were stimulated in vitro with 708 anti-Id MAb or an irrelevant IgG,, antibody. Six days later both sets of lymphocyteswere restimulatedwith CEA, and lymphocytes primed to 708 anti-Id MAb proliferated in response to CEA. These results suggest that 708 anti-Id M A b can act as an idiotypic replica of CEA and stimulate cellular and humoral anti-CEA immune responses. It is therefore of great interest as an idiotypic vaccine against colorectal cancer.
There is considerable interest in the development of antiidiotypic antibodies for use as vaccines against a number of diseases, including cancer. In the normal course of events, tumour growth and expansion may continue for want of an immune response, either because the tumour-associated antigens are recognized as self or due to repression of anti-tumour immune responses. Anti-idiotypic antibodies may induce a primary, i.e. an anti-anti-idiotypic, response against tumour antigen, or they may select or amplify a pre-existing immune repertoire by up-regulating a normally repressed response. Induction of an immune response by antibodies acting as “network antigens” is attractive, as anti-idiotypic antibodies may activate part of the immunological repertoire that is not susceptible to stimulation by antigen. This may act via activation of “silent clones” (Bona et al., 1981; Stein and Soederstroem, 1984; Schick et aL, 1987), or provide selective pressure to promote the induction of an idiotype for which the appropriate germ-line genes are not present (Moser et al., 1983). Thus, the possibility of activating responses not inducible by antigen makes idiotypic intervention very attractive for tumour immunot herapy. CEA is a complex glycoprotein (M, 180,000)found in 90% of colorectal cancers and 77% of gastric cancers and in an average of 87 ? 5% of epithelial cells per lesion (Durrant et al., 1989). It is therefore an appropriate target antigen for idiotypic immunotherapy. Various glycoproteins, immunologically related to CEA, are reported to be present on normal tissues. These include normal cross-reacting antigen (NCA) which has a tissue distribution similar to that of CEA but is also found on normal granulocytes and macrophages; NCA-2 and normal foetal antigen (NFA-2) which are found in meconium and adult faeces; and biliary glycoprotein I which is present in normal bowel (Rogers, 1983). A murine MAb, NCRC23, has been produced which binds to a CEA-specific epitope (Price et
al., 1987) and shows minimal cross-reactivity with normal tissues. It has also been used for the radioimmunodetection of tumours in colorectal cancer patients without any adverse reaction (data not shown). This antibody has been used to raise anti-idiotypic antibodies for immunotherapy of colorectal cancer. MATERIAL AND METHODS
Monoclonal antibodies NCRC23 (IgG,) and NCRC24 (IgG,,) are MAbs which recognize CEA-specific epitopes (Price et al., 1987). NCRC16 (IgG,) and NCRC17 (IgG,) are MAbs which recognize both CEA and NCA (Price et al., 1987). H65 and NCRC34 are IgG, MAbs which recognize CD5 and cytokeratin 18, respectively. C14 and NCRC30 are IgM antibodies and NCRC37 is an IgG, antibody; all of these recognize C14gp200 colorectal-tumour-associated antigen (Price et al., 1986). 381, an IgG, MAb used as the control antibody, does not bind to CEA. 730 is an anti-idiotypic mouse MAb recognizing the binding site of 791T/36 anti-tumour MAb (Doran et al., 1990). Immunization protocols and MAb production BAL,B/c mice were immunized by S.C. injection of 50 pg of MAb NCRC23 coupled to solubilized keyhole-limpet haemocyanin by glutaraldehyde fixation, and emulsified in Freund’s complete adjuvant. Animals were boosted 15 and 28 days after the primary immunization with 50 pg of the same immunogen in Freund’s incomplete adjuvant. On day 59, 4 days prior to fusion, animals were boosted with 50 pg of the same immunogen by i.v. injection. Splenocytes were fused with the mouse myeloma NSO. Culture supernatants were screened by ELISA for the presence of anti-idiotypes and positive hybridomas were cloned 3 times by limiting dilution. Isotypes were determined by means of a kit from Binding Site, Birmingham, UK. MAbs were purified from culture supernatants by affinity chromatography on protein A/Sepharose (Pharmacia, Uppsala, Sweden). Sandwich idiotype ELISA This assay was used to detect anti-idiotype antibodies in culture supernatant and to determine the specificity of the purified anti-idiotypic MAb. MAb NCRC23 or control antibodies were diluted to 5 pg/mI in PBS and added (50 pl/well) to polyvinyl chloride flat-bottomed microtitre plates. After overnight incubation at 4”C, wells were blocked with 50 p1 of BSA (1% in PBS) for 1 hr at room temperature. Culture supernatants or purified antibodies were added (50 pl/well) and incubated at room temperature for 1 hr. After washing, the biotinylated antibodies (identical to those used to coat the plate) were added at a concentration of 25 ng/well. After further washing, avidin peroxidase (BRL, Gaithersburg, MD)
’TOwhom correspondence and reprint requests should be sent. Received: October 12,1991
812
DURRANT ET AL.
was added and the plates were left for 1 hr at room temperature. After 5 washings in PBS/O.O5% Tween 20, ABTS substrate was added in H,O, and the optical density was read after 10 min at 405 nm.
Direct idiotype ELISA Microtitre plates were coated with MAb NCRC23 or other anti-CEA or IgG, MAbs at 5 p,g/ml overnight and blocked with BSA. The biotinylated 708 anti-Id MAb was added at 500 ng/ml in the presence of 0-5,000 ng/ml of unlabelled 708 anti-Id MAb for 1 hr at room temperature. After washing, avidin peroxidase was added and the assay developed. Inhibition of binding of 708 anti-Id MAb to MAb NCRC23 by anti-708 anti-id MAb antibodies Microtitre plates were coated with MAb NCRC23 and blocked with BSA. Biotinylated 708 anti-Id MAb (50 ng/well) was pre-incubated for 1 hr at room temperature with dilutions of a rat serum pool (1/3-1/30,000), a mouse serum pool (1/3-1/300,000) or human tissue culture supernatant (neat to 1/loo) before addition to the MAb NCRC23-coated plates for a further 1 hr at room temperature. After washing, avidin peroxidase was added and the assay developed as above. Direct CEA-bindingELISA CEA (10 pg/ml) was dried on to microtitre plates at 37°C overnight and blocked with BSA. Rat and mouse sera and supernatant from in vitro culture of human B cells were added, and after 1 hr at room temperature, bound antibody was detected with either rabbit anti-mouse peroxidase (Dakopatts, Copenhagen, Denmark), goat anti-rat peroxidase (Caltag, Leicester, UK) or goat anti-human peroxidase (Sigma, Poole, UK). The assays were then developed. To determine whether the anti-708 anti-Id MAbs recognized the same epitope as MAb NCRC23, rat and mouse sera and human supernatants were added to CEA-coated plates and then biotinylated MAb NCRC23 was added. After 1 hr at room temperature, the assay was developed with avidin peroxidase.
Flow cytometnc competition assays A modification of the assay of Robins et al. (1986) was used to measure inhibition of MAb NCRC23 or other anti-CEA MAbs binding to CEA-coated polystyrene beads by 708 anti-Id MAb. Briefly, MAb NCRC23 or other anti-CEA antibodies labelled with fluorescein isothiocyanate (FITC) were preincubated with 708 anti-Id MAb for 30 min at 4°C. The antibody mixture was then added to lo5CEA-coated beads and incubated for a further 30 min at 4°C. The level of FITClabelled antibody binding to CEA beads was determined by flow cytometric measurements of fluorescence intensity on a FACS IV.
Inhibition of binding of CEA by anti-idiotype Plates were coated with MAb NCRC23 and blocked with BSA. 708 anti-Id MAb (10 p,g/ml) was added to half the plate and then 30-1,000 ng of radiolabelled lZ5ICEA were added to 708 anti-Id MAb blocked and unblocked wells. Radioactivity was counted on a gamma counter. Mouse T-cell assays Mice were immunized with 708 anti-Id MAb which had been coupled to KLH by glutaraldehyde fixation and emulsified in Freund’s complete adjuvant. Two weeks later, they were re-injected with the same immunogen in Freund’s incomplete adjuvant; 2 weeks after this, splenocytes were obtained. They were plated into round-bottomed microtitre plates in RPMI 1640 containing 5 x M mercaptoethanol and 10% FCS at 5 x lo4 cells/well. To quadruplicate sets of wells were added previously irradiated (5,000 rad) CEA-positive (MKN45) and CEA-negative (791T) cells at a density of lo4 cells/well, or
CEA antigen (10 pg/ml), or epithelial mucin antigen (10 pgiml) as a negative control. To separate quadruplicate wells were added 708 anti-Id MAb or irrelevant IgG,, MAb at 1-10,000 ng/well. Cells were incubated for 5 days at 37°C and then pulsed for 24 hr with 1 pCi of [3H] thymidine. Radioactivity was counted on a liquid scintillation counter.
Human-tumour-infiltrating lymphocyte assay Colorectal turnour was finally minced and disaggregated (Durrant et al., 1986). A single-cell suspension was sorted for infiltrating (lymphocytes, monocytes and fibroblasts: TLA) and tumour cells were sorted by the MAb magnetic-bead technique (Durrant, 1991). Alternatively, peripheral blood lymphocytes (PBL) were separated on lymphocyte separation medium (Durrant et al., 1984). Either the lymphocytes were stimulated directly, or T cells were purified on magnetic beads (Dynal, Oslo, Norway) coated with an anti-CD2 MAb, then the remaining cells were irradiated and used as antigen-presenting cells. Cells were then stimulated with either 708 anti-Id MAb or IgG,, myeloma antibody at 10 pg/ml, challenged after 6 days with CEA (10 pg/ml) or a control antigen (human epithelial mucin, 10 pg/ml), left for a further 5 days, pulsed with [3H] thymidine (1 pCi/well) and counted on a liquid scintillation counter. Immunoblotting of M b NCRC23 and 708 anti-Id MA6 MAb NCRC23 and 708 anti-Id MAb were reduced in 5% mercaptoethanol, and separated heavy and light chains were resolved in a 7.5% polyacrylamide, Phastgel (Pharmacia, Milton Keynes, UK). Non-reduced antibody was also run in this gel. Protein was transferred to nitrocellulose by Western blotting using a Phast system. Blots were blocked with 1%BSA in PBS containing 0.05% Tween-20. Biotinylated 708 anti-Id MAb was added to the MAb NCRC23 blot and biotinylated MAb NCRC23 to the 708 anti-Id MAb blot. The blots were then developed with avidin peroxidase and 0.4% AEC and H,O,(Austin et al., 1989). Statistics All results were analysed for significance by Student’s t-test. RESULTS
Generation of anti-idiotype Mitbs Out of 183 hybridoma supernatants screened, 8 were positive according to MAb NCRC23 sandwich idiotype ELISA. After cloning, an IgG,, MAb, 708, was obtained. To assess the specificity of this MAb, it was tested in sandwich idiotype ELISA against 3 other IgG, MAbs and 2 IgM antibodies. 708 anti-Id MAb could only bridge MAb NCRC23 antibody (Table I). An irrelevant anti-idiotypic MAb, 730, failed to bridge MAb NCRC23 (Table I). Similarly, specificity was also tested by direct idiotype ELISA (Fig. 1). Biotinylated 708 anti-Id MAb TABLE I - ELISA TO DETECT WHETHER THE ANTI-IDIOTYPIC MAb 708 CAN BRIDGE MAbs
MAb
NCRC23 H65 NCRC37 NCRC34 C14 NCRC30
708 MAb
Absorbance % SE (405 nm) 730 anti-Id MAb
0.812 +- 0.17 0.161 +- 0.026 0.217 & 0.01 0.105 & 0.01 0.175 ? 0.013 0.129 0.014
0.114 2 0.01 NDZ 0.101 2 0.02 ND ND 0.141 k 0.01
Antigen’
1.24
k 0.08 ND 1.115 2 0.06 0.650 2 0.02 2.247 2 0.154 2.667 t 0.119
‘Antigen was included as a positive control to show that the MAbs can bridge. The antigen for NCRC23 was CEA; that for NCRC37, C14 and NCRC30 was C14gp200; and that for NCRC34 was cytokeratin 18.-’ND, not determined.
813
INDUCTION OF ANTI-CEA IMMUNE RESPONSES
TABLE 111- INHIBITION OF BINDING OF CEA TO MAb NCRC23 BY 708 ANTI-IDIOTYPIC MAb
0.8 1
Concentration of "'1 CEA (ng)
E 0.4 3
{
\
I !
1000 300 100 30
0.2
n
4:
Binding of '''1 CEA to NCRC23 (cpm in the presenceof:
2
SE)
~
Nothing
kG*t' (500 ne)
708 (500 ng)
55,883 f 4073 56,821 k 3871 2,795 t 191 20.222 t 361 29.416 f 581 1.243 +- 58 '488 f 5 81621 & 208 71636 f 174 2.027 f 19 1.981 2 40 214 t 13 186 t 46 194 2 16 209 f 73
0.0 1
1
I708
10 antibody]
100
1000
(pg/ml)
FIGURE 1 - Binding of 708 anti-Id MAb to 4 anti-CEA MAbs, NCRC23 (W),NCRC24 (A),NCRC16 ( 0 )and NCRC17 (O), and to another IgG, MAb NCRC37 (A). Binding of 730 anti-Id MAb to NCRC23 (0).Error bars are covered by symbols. TABLE I1 - INHIBlTION OF BINDING OF ANTI-CEA MAbs TO CEA BY 708 ANTI-IDIOTYPIC MAb
Concentration of 708 anti-Id MAb (ILgW
166 55 18 6
2 Concentration of 730 anti-Id MAb (uelml)
166
%binding ? SE NCRC23 (20 uelml)
NCRC24 (20 uelml)
NCRC17 (20 ueirnl)
2 t 0.05 3 k 0.01 86 t 0.07 105 f 1.0 101 f 1.1
56 k 0.1 86 f 3.0 96 f 1.6 99 k 2.0 94 +- 2.0
101 f 4.0 100 t 2.0 99 f 1.0 103 t 2.0 101 t 2.0
NCRC23 (20ILgiml)
18 6
96 f 1 101 f 2 104 t 5 97 t 1
2
99 f 1
55
could bind to plates coated with MAb NCRC23 but not to another CEA-specific MAb, NCRC24, or to 2 CEA/NCAreactive MAbs, NCRC16 and NCRC17, or to another IgG, antibody. 730 anti-Id MAb failed to bind to MAb NCRC23. Binding of biotinylated 708 anti-Id MAb to MAb NCRC23 was inhibited by unlabelled 708 anti-Id MAb. If an anti-idiotype MAb (Ab2) is capable of mimicking antigen, it must recognize idiotypic determinants within the antigen-combining site of Abl and therefore inhibit binding of Abl to antigen and binding of antigen to Abl. 708 anti-Id MAb inhibited the binding of MAb NCRC23 to CEA (Table 11) but had no effect on the binding of MAb NCRC17, and only at very high concentrations did it inhibit binding of NCRC24 to CEA. Equimolar concentrations of 708 anti-Id MAb and MAb NCRC23 resulted in 86% inhibition of binding. 730 anti-Id MAb failed to prevent MAb NCRC23 from binding to CEA. 708 anti-Id MAb (500 ng) also inhibited binding of radiolabelled CEA to MAb NCRC23-coated microtitre plates (Table 111) whereas 730 anti-Id had no effect.
Immunoblottingof 708 anti-Id MAb and MAb NCRC23 Idiotypes can be present on the isolated heavy and light chains of immunoglobulins or result from the association of these 2 chains. After reduction, the heavy and light chains of 708 anti-Id MAb and MAb NCRC23 antibodies were separated electrophoretically and transferred onto nitrocellulose. 708 anti-Id MAb bound to intact, but not to isolated, heavy and light chains of MAb NCRC23, and MAb NCRC23 bound to
FIGURE 2 - Immunoblots of (A-E) 708 anti-Id MAb and (E-J) MAb NCRC23. Blots (A) and (D) are of non-reduced 708 anti-Id MAb developed with anti-mouse antiserum and biotinylated NCRC23, respectively. Immunoblots (B,C,E) are of reduced 708 anti-Id MAb developed with anti-mouse antiserum, anti-kappa light chain and biotinylated NCRC23, respectively. Immunoblots (F) and (I) are of non-reduced MAb NCRC23 developed with anti-mouse antiserum and biotinylated 708 anti-Id MAb, respectively. Immunoblots (G,H,J) are of reduced NCRC23 developed with anti-mouse antiserum, anti-kappa light chain and biotinylated 708 anti-Id MAb, respectively. Molecular weights are shown to the left. intact, but not to isolated, heavy and light chains of 708 anti-Id MAb (Fig. 2).
Characterization of antisera generated by immunization with anti-708 anti-Id MAb Mice and rats were immunized with 708 anti-Id MAb or an IgG,, MAb linked to KLH in the presence of either Freund's adjuvant or the saporin adjuvant Quil A (Superfos Biosector als Vedbaek Denmark). Human peripheral blood lymphocytes were stimulated in vitro with 708 anti-Id MAb or IgG,, MAb. The resultant antisera or culture supernatants were screened for inhibition of binding of 708 anti-Id MAb to MAb NCRC23 and for direct binding to CEA. Antiserum from animals immunized with anti-Id MAb 708 KLH in the presence of either Freund's or Quil A adjuvant gave, in each case, significantly greater inhibition of binding of biotinylated 708 anti-Id MAb to MAb NCRC23 (Fig 3a,b) and higher binding to CEA (Fig. 4a,b) than did antiserum from animals immunized with an irrelevant IgG,, MAb. The antiserum from animals immunized with anti-Id MAb 708/KLH in Freund's adjuvant was superior, with end-point titres for binding to CEA of 1/1,000 to 1/3,000 for the mouse serum and 1/100 to 1/300 for the rat serum. The rat serum failed to bind to normal colon antigen (NCA), whereas weak binding was detected in the 708 anti-Id MAb Freund's adjuvant antiserum (titre 1/100), the IgG, MAb Freund's adjuvant antiserum (titre 1/10) and the 708 anti-Id MAb Quil A antiserum (titre
814
DURRANT ET AL. 1207
a
r'"
V
".la
.
0001
,001
.01
.1
/**
I
0
1
,0001
,001
.01
1
1
DILUTION OF MOUSE SERUM (1TTITRE)
DILUTION OF MOUSE SERUM (1TTITRE)
b
0.0 ,0001
,001
.01
1
1
,0001
0
DILUTION OF RAT SERUM (1ITITRE)
120 110
1
c 0.71
T
T-T
100
CI
,001
.01
1
1
DILUTION OF RAT SERUM (1TTITRE)
o,61
c
0.5
90
0
f 80 s
70 0.3 60
0.2
50
1
3 9 27 TISSUE CULTURE SUPERNATANT (1TTITRE)
81
1
DILUTION
2
6 F TISSUE
4
8
16
32
CULTURE SUPERNATANT FROM B CELLS
FIGURE 3 - Binding of 708 anti-Id MAb to MAb NCRC23 in the presence of (a) 708 anti-Id MAb immune mouse serum, (b) 708 anti-Id MAb immune rat serum and (c) 708 anti-Id-MAbstimulated human-B-cell supernatant. Animals were immunized with 708 anti-Id MAb in Freund's adjuvant (W), 708 anti-Id MAb in Quil A adjuvant (O), IgG, in Freund's adjuvant ( 0 )or IgG,, in Quil A (0).In (c), the human-B-cell supernatants were stimulated with 0 (a)or 10 (0)ng/ml of 708 anti-Id MAb or 10 ng/ml irrelevant IgG,, MAb (H). In a and b the error bars are covered by the symbols.
FIGURE 4 - Binding of (a) 708 anti-Id MAb immune mouse serum, (b) 708 anti-Id MAb immune rat serum and (c) 708 anti-Id-MAb-stimulated human B-cell tissue culture supernatant to CEA. Animals were immunized with 708 anti-Id MAb in Freund's adjuvant (W), 708 anti-Id MAb in Quil A adjuvant (O), IgG, in Freund's adjuvant ( 0 )or IgG,, in Quil A (0).The human B cells were stimulated in vitro with either 0 (a), 1(A), 10 (M)or 100 ( 0 )ng/ml708 anti-Id MAb or 10 ng/ml irrelevant IgG MAb (A). The error bars are covered by the symbols. *p > O.OOl,zb**p > 0.01.
1/ 10). Similarly, culture supernatants from human peripheralblood mononuclear cells which had been stimulated in vitro with 708 anti-Id MAb contained antibodies which inhibited 708 anti-Id MAb binding to MAb NCRC23 (Fig. 3c) and bound directly to CEA (Fig. 4c). Unstimulated B-cell cultures of B cells stimulated with an irrelevant IgG,, MAb produced no such antibodies (Fig. 3c, 4c).
negative human tumour-cell lines (Table IV). The splenocytes gave a significantly higher proliferative response to the CEApositive than to the CEA-negative cell lines and to CEA antigen compared with breast tumour mucin. The splenocytes also proliferated in response to re-challenge with 708 anti-Id MAb and also to IgG,, (Table IV). Human tumour-infiltrating lymphocytes (TILs) were separated on magnetic beads coated with an anti-CD2 MAb from an enzymatically disaggregated single-cell suspension of colorectal tumour cells. These lymphocytes were stimulated in vitro in the presence of irradiated antigen-presenting cells with
Cellular immune responses induced by 708 anti-Id MAb Mice primed in viva with anti-Id MAb 708/KLH were screened for proliferation against CEA-positive and CEA-
INDUCTION OF ANTI-CI?A IMMUNE RESPONSES
815
leads to the induction of anti-tumour effects, resulting in suppression of tumour growth. In a syngeneic rat system, tumour suppression correlates with induction of a cellular rather than an antibody response (Dunn et al., 1987). Incorporation of ['HI thymidine (cpm ? SD) Three approaches to anti-idiotypic therapy of human cancer In vitro challenge 708 anti-Id IgG, have been reported. Colorectal patients have been treated primed primed (Herlyn et al., 1987) with heterologous (goat) anti-idiotypic 1,191 t 188 923 2 196 791T Cells (lo4) antibodies raised against the MAb 17-1A. Melanoma patients 961 t 227 MKN45 Cells (lo4) 2,274 2 182 have been treated with a mouse anti-idiotypic MAb which mimics the high-molecular-weight human melanoma antigen 708 anti-Id MAb 10 pg/ml 4,847 t 436 4,833 2 48 (Mittleman et al., 1990). Both of these trials revealed some 1 pg/ml 6,993 2 649 ND favourable responses although they were aimed at inducing IgG,, MAb humoral rather than cellular immune responses, and both 7,659 2 54 5,516 t 528 10 pg/ml induced anti-species antibodies. Our own trial in colorectal 6,567 2 118 ND 1 pg/ml cancer patients with a human anti-idiotypic MAb, which 2,745 2 149 274 t 64 CEA recognizes the binding site of the anti-tumour MAb 791T/36, 240 2 5 EMA 657 2 92 showed cellular anti-tumour immune responses in 5 out of 6 patients and high levels of plasma IL-2 in 4 out of 6 patients ND. not determined. (Robins et al., 1991). No antibody responses were observed, and there was no associated toxicity. Two of the patients had stable disease, for a period of 15 months in one case, and were alive and well at the time of writing. The aim of this study was to investigate other tumourx 6 associated antigens as potential targets for idiotypic immunoW n therapy. CEA is a promising target as it is a homotypic z 5 adhesion molecule widely expressed on colorectal tumours. Cancer patients are almost always tolerant to this antigen but 5 4 idiotypic immunization has been shown to break self-tolerance i= in other systems. In this communication we describe the production and characterization of an anti-idiotypic MAb 2 3 W which acts as a network antigen (Kohler et al., 1989). LL 708 anti-Id MAb reacts specifically with NCRC23 MAb and A 2 can inhibit binding of this antibody to CEA and of CEA to this 0 U antibody. The binding of 708 anti-Id MAb to MAb NCRC23 n 1 requires an association of the heavy and light chains, as immunoblotting of separated chains destroys recognition. n " Analysis of the Ab3 response in animals immunized with 708 TLA-1 TLA-2 PEL-2 PEL-3 anti-Id MAb clearly shows that Ab3 antibodies which prevent 708 anti-Id MAb from binding to MAb NCRC23 and which FIGURE 5 - Proliferative response of human-tumour-infiltrating lymphocytes to 708 anti-Id MAb and CEA antigen. Lymphocytes bind directly to CEA are produced. Similarly, immunization of were treated with 708 anti-Id MAb (m) or IgG,, (0) and after 6 human lymphocytes in vitro with 708 anti-Id MAb resulted in days they were rechallenged with CEA for a further 7 days. human antibodies which recognized both 708 anti-Id MAb and Alternatively,lymphocytes were primed with 708 anti-Id MAb (m) CEA antigen. These results provide strong evidence that the or IgG,, (B) and challenged 6 days later with human epithelial former can act as a replica of a B-cell epitope on CEA. antigen. Several other research groups have also identified CEA as a potential target for idiotypic immunotherapy. Monestier et al. either 708 anti-Id MAb or an irrelevant IgG, antibody and (1989) produced 2 murine MAbs which recognized determithen 6 days later they were challenged with CEA (Fig. 5). Only nants within the antibody-combining site of a CEA-specific TILs which had been primed with 708 anti-Id MAb prolifer- MAb. Kohler's group have recently produced a mouse MAb to ated in response to challenge with CEA. Peripheral blood the combining site of a CEA-specific mouse MAb which lymphocytes were also isolated from 2 cancer patients. TILs induces anti-CEA antibodies in mice (Bhattacharya-chatterwere also isolated from the tumour of 1 of these patients for jee et aL, 1990). This anti-anti-idiotypic mouse anti-serum has direct comparison. Either lymphocytes or T cells, primed with a fine specificity which is similar, but not identical, to that of 708 anti-Id MAb in the presence of antigen-presenting cells, the original anti-CEA mouse MAb. responded with enhanced proliferation to a second challenge 708 anti-Id MAb is, however, unique in that it is the first with CEA antigen. The T cells failed to respond if they had anti-idiotypic replica of CEA which can not only prime mouse been primed with IgG, or if they had been primed with 708 splenocytes to proliferate in response to CEA and CEAanti-Id MAb and then challenged with an irrelevant antigen expressing human tumour cells but can also prime humansuch as epithelial much antigen. tumour-infiltrating lymphocytes and peripheral-blood T cells from cancer patients to proliferate in vitro specifically to CEA antigen. The epitopes recognized by T cells are small linear DISCUSSION peptides which are easily presented by major histocompatibilAnti-tumour responses induced by anti-idiotypic manipula- ity molecules. Even though 708 anti-Id MAb only binds to tion have been demonstrated against both murine- (Forstrom intact MAb NCRC23, the T-cell epitope which stimulates cell et aL, 1983; Lee et aL, 1986) and rat-tumour-associated proliferation may be a linear determinant expressed on the antigens. It has been reported (Kennedyet al., 1985; Raychaud- heavy or light chain. This has been shown for an anti-idiotypic huri et al., 1987) that stimulation of anti-idiotypic responses antibody which mimics mycobacterial protein. Ab2 only bound TABLE N - PROLIFEKATION OFlN WVO-PKIMED 708 ANI'I-ID .MAb AND IgG,, MAh MOUSE SPLENOCYTES IN RESPONSE TO CEA-NEGAIIVE (791TJ, CEA-POSITIVE (MKN45J TUhlOUR CELL I.ISES, CEA, EPITHEI.IAL MUCIV. 7U8 AKTI-ID MAh AND IeG,, MAh
71 T
816
DURRANT ET AL.
to intact Abl (Rees et al., 1987) but a specific T-cell response could be stimulated by heavy chain alone (Rees and Lamb, 1990). Similarly a putative T-cell epitope has been identified in the CDR2 domain of an anti-idiotypic MAb 2F10 which induces protection against transfer and growth of L1210/GZL tumour (Raychaudhuri et al., 1990).
708 anti-Id MAb appears to be suitable for immunotherapy of gastrointestinal cancer. As it is a mouse antibody, it may be necessary to immunize with antibody fragments or genetically cloned Fab to avoid anti-mouse humoral responses which may interfere with repeated immunizations. Alternatively, the antibody could be humanized (Reichmann et aL, 1988).
REFERENCES
AUSTIN,E.B., ROBINS,R.A., DURRANT,L.G., PRICE, M.R. and BALDWIN, R.W., Human monoclonal anti-idiotypic antibody to the tumour-associated antibody 791T/36. Immunology, 67,525-530 (1989). BHATTACHARYA-CHATTERJEE, M., MUKERJEE, S., BIDDLE,W.. FOON, K.A. and KOHLER,H., Murine monoclonal anti-idiotype antibody as a potential network antigen for human carcinoembryonic antigen. J. Immunol., 145,2758 (1990). BONA,C.A., HERBER-KATz, E. and PAUL,W.E., Idiotype-anti-idiotype regulation. I. Immunization with a Levan binding myeloma protein leads to the appearance of auto-anti-(anti-idiotype) antibodies and to the activation of silent c1ones.J. ap. Med., 153,951-961 (1981). DORAN,M., DURRANT, L.G., AUSTIN,E.B., BALDWIN,R.W. and ROBINS,R.A., A syngeneic anti-idiotypic antibody to an anti-tumour antibody which induces cellular and humoral anti-tumour responses. Brit. J. Cancer, 62,500 (1990). DUNN,C.P.L.., JOHNSON, C.A., STYLES,J.M., PEASE,S.S. and DEAN, C.J., Vaccination with syngeneic monoclonal anti-idiotype protects against a tumour challenge. Immunology, 60,181-186 (1987). DURRANT, L.G., Rapid and efficient method for separating tumour infiltrating cells and tumour cells from colorectal tumours. J. immunol. Meth., (1991). (In press). DURRANT, L..G., PARKER,M., KENWORTHY, N. and TAYLOR,G.M., Characterisation of a human x mouse T cell hybridoma and identification of a clone secreting and binding interleukin-2. Immunology, 52, 119-124 (1984). DURRANT, L.G., ROBINS,R.A., ARMITAGE, N.C., HARDCASTLE, J.D. and BALDWIN, R.W., Association of antigen expression and DNA ploidy in human colorectal tumours. Cancer Rex, 46, 3543-3549 (1986). DURRANT, L.G., ROBINS,R.A. and BALDWIN, R.W., Flow cytometric screening of monoclonal antibodies for drug or toxin targeting to human cancer. J. nat. Cancer Inst., 81,688-691 (1989). I. and FORSTROM, J.W., NELSON,K.A., NEPOM,G.T., HELLSTROM, HELLSTROM. K.E., Immunization to a syngeneic sarcoma by a monoclonal auto-anti-idiotypic antibody. Nature (Lond.), 303, 627-629 (1983). HERLYN, D., WETTENDORF, M., SCHMOLL, E., ILIOPOULOS, D., SCHEDEL,I.. DREIKHAUSEN, V., RUB, R., ROSS,A.H., JAKSCHE, H., SCRIBA, H., Anti-idiotype immunization of cancer paM. and KOPROWSKI, tients: modulation of the immune response. Proc. nat. Acad. Sci. (Wash.), 84,8055-8059 (1987). KENNEDY, R.C., DREESMAN, G.R., BUTEL,J.S. and LANFORD,R.E., Suppression of an in vivo tumour formation induced by SV40transformed cells in mice receiving anti-idiotypic antibodies. J. exp. Med., 161,1432-1436 (1985). T., SRINIVASAN, S., KAVERI,S., MORKOHLER,H., KIEBER-EMMONS, ROW, W.J.W., MULLER, S., KANG,C-Y. and RAYCHAUDHURI, S., Short analytical review: revised immune network concepts. Clin. Immunol. Immunopatliol., 52,104-116 (1989). LEE,V.K., I~ELLSTROM,K.E. and NEPOM,G.T., Idiotypic interactions in immune responses to tumour associated antigens. Biochim. biophys. Acta, 965,127-128 (1986). T., YANG,H., YAMADA, M., MITTLEMAN, A., CHEN,Z.J., KAGESHITA, BASKIND, P.,GOLDBERG, N., Puccio, C . , AHMED,T., ARLIN,2. and S., Active specific immunotherapy in patients with melaFERRONE, noma. J. chi. Invest., 86,2136-2144 (1990).
MONESTIER, M., DEBBAS,M.E. and GOLDENBERG, D.M., Syngeneic anti-idiotype monoclonal antibodies to murine anti-carcinoembryonic antigen monoclonal antibodies. Cancer Rex, 49,123-127 (1989). M., LAI, A.C.K., NORTON,G., REALE, MORAN,T.M., MONESTIER, M.A., THOMPSON, M.A., SCHULMAN, J.L., RIRLET,R. and BONA,C.A., Characterisation of variable region genes and shared cross-reactive idiotypes of antibodies specific for antigens of various influenza viruses. Viral Immunol., 1,l-12 (1987). MOSER,M., HIERNEUX, L.J. and URBAIN, J., Idiotypic manipulation in mice: BALBic mice can express the cross-reactive idiotype of A/J mice. Proc. nat. Acad. Sci. (Wash.), 80,4474-4479 (1983). PRICE,M.R., EDWARDS, S. and BALDWIN, R.W., Association of the Y hapten with glycoproteins, glycolipids and carcinoembryonic antigen in colorectal carcinoma. Cancer Lett., 33,83-89 (1986). S., JACOBS,E. and BALDWIN, R.W., Mapping PRICE,M.R., EDWARDS, of monoclonal antibody-defined epitope associated with carcinoembryonic antigen, CEA. Cancer Immunol. Immunother., 25,lO-15 (1987). RAYCHAUDHURI, S., KANG,C-Y., KAVERI,S-V., KIEBER-EMMONS, T. and KOHLER,H., Tumor idiotype vaccines. VII. Analysis and correlation of structural, idiotypic and biologic pro erties of protective and non protective Ab2. J. Immunol., 145,760-76y (1990). RAYCHAUDHURI, S., SAEKI,Y., CHEN,J-J, IRIBE,H., FUJI, H. and KOHLER,H., Tumour-specific idiotype vaccines. 11. Analysis of the tumor-related network response induced by the tumor and by internal image antigens (Ab2beta). J. Immunol., 139,271-280 (1987). REES, A.D.M. and LAMB,J.R., Human T-cell recognition of an “internal image” of mycobacterial antigen in an anti-idiotypic antibody. In: A. Osterhaus and F. Uytdehaag (eds.), Idiotypic networks in biology and medicine, pp. 195-201, Elsevier, Amsterdam (1990). REES,A.D.M., SCOGING, A,, DOBSON, N., PRAPUTPIITAYA, K., YOUNG, D., IVANYI,J. and LAME,J.R., T-cell activation by anti-idiotypic antibody: evidence for the internal image. Europ. J. Immunol., 17, 197-201 (1987). REICHMANN, L., CLARKE, M., WALDMANN, H. and WINTER,G., Reshaping human antibodies for therapy. Nature (Lond.), 332, 323327 (1988). ROBINS,R.A., DENTON,G.W.L., HARDCASTLE, J.D., AUSTIN,E.B., BALDWIN, R.W. and DURRANT, L.G., Anti-tumor immune response and interleukin-2 production induced in colorectal cancer patients by immunization with human monoclonal anti-idiotypic antibody. Cancer Res., 51,5425-5429 (1991). ROBINS,R.A., LAXTON,R.R., GARNETT,M., PRICE,M.R. and BALDWIN,R.W., Measurements of tumour-reactive antibody and antibody conjugates by competition, quantitated by flow cytometry. J. immunol. Meth., 90,165-171 (1986). ROGERS,G.T., Carcinoembryonic antigens and related glycoproteins. Molecular aspects and specificity.Biochim. biophys.Acta, 695,227-249 (1983). SCHICK,M.R., DREESMAN, G.R. and KENNEDY, R.C., Induction of an anti-hepatitis B surface antigen response in mice by non-internal image (Ab2p) anti-idiotypic antibodies. J. Immunol., 138, 3419-3424 (1987). STEIN,K. and SOEDERSTROEM, T., Neonatal administration of idiotype or anti-idiotype primes for protection against E. coli K13 infection in mice.J. exp. Med., 160,1001-1007 (1984).
