Idiotypic vaccination against B-cell lymphoma leads to dormant tumour

CELLULAR

IMMUNOLOGY

132, 70-83 (1991)

ldiotypic Vaccination against B-Cell Lymphoma Leads to Dormant Tumour R. J. DYKE, H. MCBRIDE, A. J. T. GEORGE, T. J. HAMBLIN, AND F. K. STEVENSON Host Immunity to Tumour Group, Lymphoma Research Unit, Tenovus Research Laboratory, General Hospital. Southampton SO9 4XY, United Kingdom Received June 26, 1990; acceptedAugust 8, 1990 Idiotypic immunoglobulin, which can be considered to bear tumour-associated antigens in the context of B-cell lymphoma, has been obtained from the splenic A3 1 tumour, purified, and used to immunise syngeneic mice. On subsequent exposure to a lethal challenge of lymphoma cells, the mice showed no overt tumour development over an observation period of 6 months, whereas mice immunised with an unrelated idiotypic immunoglobulin succumbed to lymphoma after about 20 days. Anti-idiotypic immunity persisted in protected mice, since a second exposure to a lethal tumour dose 4 months after the first challenge also failed to induce lymphoma. Antiidiotypic antibody appeared to have a major role in protection when analysed by passive transfer experiments, with no contribution from transferred cells. Protected mice were investigated for the presence of lymphoma cells 4-8 months following exposure to tumour, but the spleens, which were of normal weight and appearance, contained few or no tumour cells by phenotypic analysis. However, passage of cells dispersed from these spleens led, in 60% of cases, to tumour development in unimmunised recipients. The emergent tumours were indistinguishable from the original A3 1 lymphoma, with no evidence for variants, indicating that the cells were unable to grow in the immune mice, but that this dormant state could be disrupted by transfer. 0 1991 Academic pres\. hc.

INTRODUCTION Anti-idiotypic immunity has been shown to suppress development of idiotype-bearing lymphoma in several animal models (1, 2, 3). It has even been possible to treat incipient disease by immunising animals already bearing turnout-, thereby indicating the possibility of treating human lymphoma by this approach (4). Idiotypic immunoglobulin, therefore, provides a molecularly defined tumour-associated antigen for this disease, and by using such animal models, it has been possible to examine the nature of the immunity induced by immunisation and to follow the fate of tumour cells under anti-idiotypic attack. For lymphomas, which express immunoglobulin, but usually secrete little, the major mechanism involved in anti-idiotypic suppression of tumour appears to be antibodymediated (4) and passive transfer studies have not revealed a direct role for T cells (2, 3, 5). This contrasts with results obtained in the Ig-secreting mouse plasmacytomas where antibody would presumably be rapidly rendered ineffective; for this disease antiidiotypic T cells of the suppressor/cytotoxic phenotype appear to act against tumour (6, 7). 70 0008-8749/9 1 $3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

IDIOTYPIC

VACCINATION

AGAINST

B-CELL

LYMPHOMA

71

In spite of the efficacy of anti-idiotypic antibody in suppressing tumour, lymphoma can emerge, often months after challenge. Thus, in the mouse B-cell tumour, BCL, , where strong protection against tumour is induced by immunisation with idiotypic IgM, lymphoma can emerge after more than 100 days and on occasion even after lyear post-tumour challenge, whereas mean survival time for unimmunised mice would be about 40 days (2). However, emergence of tumour in this case does not appear to be due solely to the loss of anti-idiotypic immunity, since the developing tumours are usually variants which no longer express idiotypic IgM at the cell surface (4). The phenomenon of tumour variants which can escape antibody attack has been described frequently in animal tumours although the nature of the deleted antigen is often unknown (8). However variants with altered idiotypes have been shown to emerge in human follicular lymphoma following therapy with monoclonal anti-idiotypic antibody (9. 10). demonstrating that production of variants is not a property confined to longpassaged animal tumours. In order to investigate some of the questions surrounding immunotherapy a further model of B-cell lymphoma was studied. This is the A3 1 lymphoma, a spleen-seeking tumour. which arose in a female CBA/H mouse that had been injected with “‘Sr. The reasons for studying the A3 1 lymphoma, are as follows: first, idiotypic immunisation has only been studied in depth in two mouse lymphoma models and we wished to extend the conclusions obtained. Second, the amount of IgM expressed at the surface is closer to the range typical of human lymphoma, whereas the well-studied BCL, tumour expresses IgM at a much higher level (11). Another point is that A3 1 cells secrete very little IgM, again similar to most human follicular lymphoma ( 12). Finally, this particular lymphoma has allowed us to approach the difficult problem of the dormant state adopted by lymphoid tumours. This tantalising state has been recognised in human lymphoma for some time (13, 14) and there has been a suspicion that immune surveillance may be involved, but experimental approaches have been limited in individual patients. The model may at least tell us what to look for. MATERIALS

AND

METHODS

A31 Twnouv CBA/H mice were bred and maintained under conventional conditions in the animal unit of the laboratory. Mice were 8-12 weeks of age when used. A31 tumour cells from passages 105-l 15 were kindly provided by Dr. L. M. Cobb (M.R.C. Radiobiology unit, Harwell). The tumour was maintained in vivo by ip injection of CBA/H mice with 5 X lo5 spleen cells prepared from animals in the terminal stage of disease. This terminal stage is reached within 3-4 weeks of tumour cell passage and is identified by generalised piloerection and abdominal distension, the latter due to a grossly enlarged spleen, often reaching 20X normal weight (15). Tumour cells were prepared from the spleen by teasing out the tissue and filtering through a coarse screen. Disaggregated lymphocytes were isolated by centrifugation on Ficoll-Hypaque (Lymphoprep) (Nycorned, Oslo, Norway) and washed in medium (MEM) (GIBCO, Paisley) containing 10% fetal calf serum (Tutt et al.. 1985). Cells were only passaged on in this way eight times before reverting to the frozen tumour cell stocks to restart the in vivo passaging programme.

72

DYKE

ET AL.

Production of Idiotypic IgM Idiotypic IgM from the A3 1 tumor was obtained following a “rescue” hybridisation of A3 1 cells with the nonsecreting BALB/c mouse myeloma line P3/NSO/l-Ag4-1 (NS-0). In order to establish a stable IgM-secreting hybridoma it was necessary to monitor large numbers of hybrids and carry out strict cloning by limiting dilution. Measurement of immunoglobulin secretion was by enzyme-linked immunosorbent assay (ELISA) using goat anti-mouse p chain (Nordic Immunologicals, Maidenhead) or rat anti-idiotypic antibody as coating antibody and goat anti-mouse p chain linked to HRP (Nordic) for detection (11). The selected clone was injected into pristaneprimed (CBA X BALB/c) Fl mice and the resulting ascitic fluid collected. Idiotypic IgM was purified by precipitation with 50% saturated ammonium sulphate followed by size separation on a column of Ultragel AcA22 (LKB, Bromma, Sweden). Assessment of purity was by SDS-PAGE, under reducing conditions. Staining of the SDS gel for protein with Coomassie blue dye revealed the presence of only two bands, one at approximately 76 kDa which corresponds to the immunoglobulin heavy chain, and the other in the light chain position at approximately 26 kDa. A control monoclonal IgM was similarly prepared from a parallel ascitic fluid containing secreted IgMK from the hybridoma TIB-200 (HNKl) obtained from the American Type Culture Collection (ATCC) (Rockville, MD). The IgMs were coupled to keyhole limpet haemocyanin (KLH) (Calbiochem Bioscience, Cambridge) using glutaraldehyde. Briefly, the IgM (0.5 mg/ml in PBS) was mixed with an equal volume of KLH (0.5 mg/ml in PBS). Aqueous glutaraldehyde (Sigma Chemicals, Poole) was added to give a final concentration of O.l%, and the solution rocked at room temperature for 4 hr before thorough dialysis against PBS.

Monoclonal Antibodies Hybridoma cell lines producing antibodies with defined specificity for I-Ab,d,q and I-Edsk(M5/114.15.2), H-2 antigen (all haplotypes) (M1/42.3.9.8) and a Thy- 1.2 antigen (30-H 12) were obtained from the ATCC. Rat monoclonal antibodies against idiotypic determinants or heavy chain constant regions of the A3 1IgM were produced following the protocol described for the BCl, lymphoma (2). In brief IgM from the “rescue” hybridoma culture fluid (see above) was purified and complexed to sheep anti-mouse Fabp following a defined method (16). The complexes were emulsified in complete Freund’s adjuvant (CFA) (Difco, East Molesley, UK) and used to immunise LOU rats. A total of 0.6 mg of complexes was divided between three subcutaneous injections. A final injection of aqueous complexes was given ip, and 3 days later the rat spleen cells were fused with the mouse myeloma cell line P3/NSl/l-Ag4-1 (NS-1). Resulting hybridoma colonies were screened by ELISA for antibody against A3 1IgM or control IgM (TIB200). IgM was coated onto the plates and any anti-Id or anti-p antibodies were detected with goat anti-rat IgG coupled to HRP (Calbiochem Bioscience, Cambridge). By this procedure it was possible to select an anti-Id and an anti-p constant region clone for expansion. Selected clone supernatants were checked for reactivity against A3 1 tumour cells or normal splenic lymphocytes by immunofluorescence. The specificity of the rat anti-Id antibody for the tumour cells was confirmed by positive staining of the A3 1 cells and the lack of reactivity of this antibody toward normal splenocytes.

IDIOTYPIC

VACCINATION

AGAINST

B-CELL

LYMPHOMA

73

Immunisation with Idiotypic IgM Immunisation with IgM or IgM-KLH conjugate was carried out as described (George et ul., 1988b) using aqueous antigen (50 pg) emulsified with an equal volume of CFA. Injections were SCgiven on Days 1, 22, and 36. Animals were then used on or after Day 50. Blood for serum analyses was taken from the tail vein.

Immunisation hvith Tzunour Cells Syngeneic mice were immunised with freshly prepared irradiated (2000 rad) A31 tumour cells in either aqueous suspension or emulsified with CFA. The time sequence was as for idiotypic protein immunisation and cells were injected SCat doses of 2 X lo6 for the aqueous suspension or 1 X lo7 for cells in CFA. Control mice were injected with either PBS or CFA alone.

Measurement of’Syngencic Antihod?. Levels Syngeneic anti-idiotypic antibody was measured by ELISA, using purified idiotypic IgM coated to the plate at a concentration of 500 rig/ml. After incubation with dilutions of immune or control mouse sera, bound mouse IgG was detected by HRP-sheep anti-mouse Fey antibody (Serotec, Blackthorn, Bicester). The detecting antibody is specific for y-chain, thereby avoiding recognition of the coated IgM antigen. This system would recognise only the IgG class of antibody and would fail to detect antiidiotypic antibody of the IgM class, although little would be expected at the stage of immunisation under investigation. An arbitrary standard was established using a pool of immune sera, which was assigned a value of 100 U/ml, and all levels were measured in comparison with this.

The ability of antibodies to mediate complement-dependent cytotoxicity was tested using 5’Cr labelled A3 1 cells. A3 1 cells (3 X 10’) were labelled by suspending the cells in 1 ml MEM (GIBCO, Paisley) with 100 &I of NazS’Cr04 (Amersham International, Aylesbury) for 30 min at 37°C before washing four times in MEM. Labelled cells (0.1 ml) at 1 X lo6 cells/ml were exposed to 0.1 ml antibody for 15 min at 4°C. Fresh rat serum (300 ~1 of a 1:2 dilution) was then added as a source of complement, the temperature raised to 37”C, and incubation continued for 30 min. The cells were then centrifuged and 300 ~1 of supernatant was drawn off for radioactive counting. The % cytotoxicity was calculated as follows: specific cytotoxicity (%) = [A - C]/[B - C] X 100, where A = cpm released in the presence of antibody, B = cpm released by detergent lysis ( 1% NP-40) and C = cpm released in the presence of medium alone.

For the analysis of protective antibody, pooled immune or control serum at different doses was transferred ip into recipient mice, and 5 X lo4 tumour cells were then injected into an intramuscular (im) site immediately afterward. For analysis of protective spleen cells a modified Winn assay was used. Spleen cells from immune or control mice. each pooled separately, were mixed with tumour cells at a ratio of 100: 1 (spleen cells:tumour), and incubated together for 30 min at 37°C

74

DYKE

ET AL.

in medium. After centrifugation, cells were taken up in PBS for im injection into the back leg. Each mouse received 0.1 ml containing 5 X lo6 spleen cells and 5 X lo4 tumour cells, and was then monitored for survival. Immunojluorescence Immunofluorescent studies were carried out with the FACS III (Becton-Dickinson Electronics, Mountain View, CA). Direct immunofluorescent studies were carried out using FITC-sheep anti-mouse IgD (Nordic Immunologicals, Maidenhead) and FITCsheep anti-mouse IgM (Binding Site Ltd., Birmingham). Idiotypic, p chain constant region, and K chain determinants were detected by indirect immunofluorescent techniques using rat antibodies (rat anti-Id and rat anti-p as described earlier and rat antiK (HB58) obtained from the ATCC) and detected by FITC-mouse anti-rat IgG (Jackson Immunoresearch Labs., Avondale). In order to detect syngeneic mouse IgG antibody bound to A31 tumour cells, FITC-sheep anti-mouse Fey (Serotec Ltd., Oxford) was used. Immunoperoxidase Staining This technique generally detects both cell surface and cytoplasmic antigen. A suspension of tumour cells was cytocentrifuged on microscope slides and air-dried for 1 hr. The slides were fixed in cold, dry acetone and washed in PBS, and then exposed to rat monoclonal antibodies for 1 hr at room temperature. They were then washed in PBS and incubated for 30 min with HRP-mouse anti-rat IgG (Jackson Immunoresearch Labs., Avondale) at 1:80 dilution. After further washes, the slides were developed using 3-amino-9-ethylcarbazole (Sigma Chemicals, Poole) substrate for 10 min (George et al., 1987). Slides were then washed in water and counter-stained with haematoxylin for 8 min, and washed again before mounting. RESULTS Preparation of Idiotypic IgM The procedure of cell hybridisation with a mouse myeloma cell line was used to increase the secretion of IgM by the A3 1 cells, since the secretion of IgM by untreated A3 1 tumour cells in swirling culture was found to be very low (20 ng/hr/ 1O7cells). Following hybridisation it was found that out of 250 hybrids, 9 were able to secrete IgM, and unless these were rigorously cloned they were rapidly outgrown by nonsecreting colonies. However, once cloned, the secreting hybridoma proved stable and secreted 12 pg IgM/ml when cultured in small flasks. Ascitic fluid from mice passaged with the hybridoma contained IgM at about 0.5 mg/ml and sizing column separation showed this to be largely pentameric. Phenotype of A31 Tumour The A31 tumour cells express IgM and lower levels of IgD at the cell surface, as shown in Fig. 1a, The indirect immunofluorescence analysesof the A3 1 cells revealed the presence of idiotype, K chains, and p chains at the surface (Fig. 1b).

IDIOTYPIC

VACCINATION

AGAINST

B-CELL

LYMPHOMA

75

a

Log Fluorescence

b

Anti-id

Anti- j.4

Log Fluorescence

Intensity

Anti-K

Control

Intensity

FIG. I. Expression of immunoglobulin by tumour cells. Spleen cells from tumour-bearing animals were examined by immunofluorescence techniques. using the FACS Ill. Data were collected using a log amplifier. (a) Cells were treated with directly conjugated sheep polyclonal antibodies against mouse n chain (-), or 6 chain (---). or normal sheep Ig control (. -. -). (b) Cells were examined by indirect fluorescence techniques using rat monoclonal antibodies directed against the A3 1 tumour idiotype, mouse p chain. mouse K chain, or an irrelevant human tumour idiotype.

Protective Immunisation

against Turnour

The protective effect of immunisation with idiotypic IgM coupled to KLH on tumour development is shown in Fig. 2. No mouse immunised with A31IgM-KLH prior to tumour challenge developed lymphoma. This protection was specific since the group immunised with control IgM conjugated to KLH showed no increase in survival when compared with mice immunised with adjuvant alone or with unimmunised mice (Fig. 2). Pooled results from this and other experiments revealed that from a total of 19 animals immunised with A3 1IgM-KLH, all survived > 100 days after tumour passage, whereas 24 control animals immunised with a control IgM coupled to KLH died on Day 34 & 6.5. The effect of increasing the number of tumour cells from 5 X IO3 to 5 X 1O5in the challenging dose is also shown in Fig. 2, and it is clear that full protection was still obtained. Immunisation with idiotypic IgM alone, i.e., not linked to KLH, was tested (Fig. 2) and complete protection was recorded for the group immunised with A3 1 IgM and challenged with 5 X lo3 tumour cells, whereas the animals in the group immunised with control IgM had died by Day 34 following tumour challenge.

76

DYKE ET AL.

A

0 0

50

100

Days post-tumour

150

200

250

challenge

FIG. 2. Effect of immunisation with idiotypic IgM on the survival of mice following tumour challenge. CBA/H mice were immunised with A3llgM-KLH (-&), A3lIgM (+), control IgM-KLH (+) or adjuvant only (t). The animals were then challenged with either 5 X IO’ (-) or 5 x IO5(---) A31 cells ip on Day 0.

One group of five immune mice has been observed for 8 months following tumour challenge and no tumour has emerged, indicating that immunisation with idiotypic IgM confers a high degree of protection against tumour challenge. In order to assessthe persistence of protection following immunisation, mice which had survived a challenge of 5 X lo3 tumour cells were rested for 4 months and then rechallenged with a second similar dose of tumour. Age-matched controls died within 33-6 1 days (mean of 44 days), whereas all five immune mice are still alive and tumourfree at Day 120.

Immunisation with Tumour Cells The effect of immunising mice with A31 tumour cells was investigated, using irradiated cells either in suspension in PBS or emulsified in CFA. Results of challenging the immunised animals with 5 X lo3 live tumour cells (Fig. 3) indicate that no protection had been obtained by either procedure. Analysis of the serum of the immunised mice by ELISA or FACS revealed no detectable anti-idiotypic antibody or anti-tumour cell antibody, respectively.

Analysis of Immunity by Passive Tranqfer The role of antibody-mediated or cellular mechanisms in protection obtained by immunising with idiotypic IgM was investigated by transferring either immune serum or spleen cells into unimmunised recipients which were then challenged with tumour. The results (Fig. 4) show that passively transferred serum from mice immunised with A3 1 IgM conferred protection, Control serum from mice immunised with the idiotype prepared from a different IgM was not protective. No effective role for transferred cells could be demonstrated. Further investigations using lightly irradiated (500 rad) recipient mice also failed to reveal any protective role for transferred cells.

Investigation of Protective Anti-Idiotypic Antibody Sera from immune mice were tested for anti-idiotypic antibody by ELISA. In mice immunised with A3 1IgM-KLH the average anti-idiotypic antibody (anti-Id) level was

IDIOTYPIC

0

VACCINATION

20

AGAINST

40

Days post-tumour

B-CELL

60

77

LYMPHOMA

60

100

challenge

FIG. 3. Effect of immunisation with irradiated tumour cells on the survival of mice following tumour challenge. CBA/H mice were immunised with I X IO’ irradiated A3 I cells in CFA (+). 2 X IO6 irradiated A31 cells in PBS (+). CFA only (+). or PBS only (G-). The animals were then challenged with 5 X lo3 viable A3 I cells ip on Day 0.

129 U/ml. Mice immunised with unconjugated A3 1IgM gave a mean value of 5 U/ ml for anti-Id levels, whereas mice immunised with the control IgM gave no detectable levels of anti-Id (~0.25 U/ml). Thus. immunisation with IgM coupled to KLH induced higher levels of anti-Id; however mice with low levels of antibody appeared to be still protected against tumour. In order to investigate this further, nonimmunised mice were given ip injections of graded doses of a pooled immune serum of known antibody level, and then challenged with 5 X IO3 A31 cells in an im site. The results (Table 1) show that immune serum is highly protective and that doses down to 2.8 U/mouse were effective, 2.8 U being equivalent to 48 ~1 of this immune serum. Since the serum volume of a mouse was estimated to be approximately 1 ml then 1 U of anti-idiotypic antibody/mouse is approximately equivalent to a final serum concentration of 1 U/ml. Dilutions of immune serum were tested for their ability to bind to the A31 cell surface using immunofluorescence techniques. The anti-Id concentrations chosen were

0

50

100

Days post-tumour

150

200

250

challenge

FIG. 4. Passive transfer of immunity against tumour. CBA/H mice were inoculated on Day 0 with 5 X 10“ A31 tumour cells im together with 300 ~1 immune serum ip (4) 300 ~1 PBS ip (G), 5 X lo6 immune spleen cells (A) or 5 X IO6 control spleen cells (-&).

78

DYKE

ET AL.

TABLE

1

ANTI-IDIOTYPIC ANTIBODY IN CELL BINDING AND PROTECTION

Anti-Id dose (U/ml) 22.0 17.1 11.0 5.5 2.8 1.4 0.8 0.2 0.1 0.01 0

Anti-Id binding to A3 I tumour cells (S)”

Survival time (days)

100

>150, >150, ,150, >I50 >l50, r150, >150, >150 >150, >150, >150, >I50 >150, >150,~150, >150 >150, >150. ,150, >I50 112, >l50, >150, >I50 50, 54, 54, 59 39, 39, 40, 43 39, 39, 39, 40 32, 36, 36, 39 33, 36, 39, 41

98 103 89 56 32 18 9 2 -1 0

’ % Anti-Id binding = [MFI (sample) - MFI (min)]/[MFI(,,,, > MFI,,,“J X 100. Min MFI, binding of normal mouse serum; Max MFI, binding of saturating anti-Id level (equivalent to 22.0 U/ml); MFI, mean fluorescence intensity.

equivalent to the levels used in the passive transfer experiment. The antibody doses which showed complete protection in the transfer experiment gave a relative binding percentage of >50% in the fluorescence assays. The minimal protective dose (1.4 U/ ml of antibody) showed a decreased, although still detectable, binding to A3 1 cells.

Blocking of Anti-Id in Vivo and in Vitro Protective immune serum was further investigated for its specificity in binding to target tumour cells using FACS analysis. The immune serum bound specifically to the A31 cells when compared to serum from mice immunised with a control IgM, and this binding was completely inhibitable by purified idiotypic IgM (Fig. 5).

Log Fluorescence

Intensity

FIG. 5. Blocking of the binding of syngeneic anti-idiotypic antibody to tumour cells by idiotypic IgM. Spleen cells from tumour-bearing animals were examined by indirect immunofluorescence techniques using the FACS III. The tumour cells were incubated with 140 U/ml syngeneic anti-idiotypic antibody (-), 140 U/ml anti-idiotypic antibody and 40 fig/ml A3 1IgM ( . . . ), 140 U/ml anti-idiotypic antibody and 40 rg/ ml control IgM (. - * -) or 40 pg/ml control IgM (---). Data were collected using a log amplifier.

IDIOTYPIC

VACCINATION

AGAINST

B-CELL

79

LYMPHOMA

In order to implicate specific anti-idiotypic antibody in the transferred protection, 40 pg A3 1IgM in 50 ~1 PBS was added to 140 U immune serum in a 50-~1 volume. This level of A31IgM was shown, by the above FACS analysis, to block the anti-Id activity of the immune serum. The mixture was then injected into recipient mice along with a lethal dose of tumour cells. The results in Fig. 6 show the survival curves for these animals. Animals that received 140 U/mouse of immune serum survived > 100 days post-tumour challenge, whereas mice that had received the antigen-blocked serum died at the same time as the control group.

Complement-Dependent Cytotosicity Immune sera with known levels of anti-idiotypic antibody ranging from 57 to 260 U/ml were tested for their ability to lyse target A3 1 cells in the presence of rat serum as a complement source. A rat monoclonal antibody of the IgG2a subclass with reactivity against the MHC Class I antigen was included in the assay as a positive control. A saturating concentration of this antibody (10 wg/ml) gave 49% specific “Cr release under the conditions used. The rat anti-idiotypic antibody, which is also of the IgG2a subclass, was less effective, giving 23% cytotoxicity at antibody saturation (10 pg/ml). However, no detectable cytotoxicity was mediated by the syngeneic mouse immune sera even at a saturating concentration of 65 U/ml of anti-idiotypic antibody.

In order to investigate dormancy a group of five immunised mice which had survived a challenge by 5 X 10’ tumour cells for > 100 days were assessed for the presence of dormant tumour. In these studies A3 1 cells were sought both by phenotypic analysis and also by their ability to grow in unimmunised recipient mice. An ip injection of 100 A3 1 cells is known to represent a lethal dose (Cobb et al., 1986). This was confirmed when mice given 100 cells were found to survive 55 days f 12 (data not shown). In experiments where immunised mice were challenged with 5 X lo3 A31 cells ip if as few as 2% escape destruction (i.e., 100 cells) then from the above information the mice should succumb to tumour by Day 55. Thus any mouse surviving > 100 days following

100 z ‘5 .s

80

: ‘,

60

‘E g

40

a s

20 0 0

20

40

60

Days post-tumour

80

100

120

passage

FIG. 6. Inhibition of the protective effect of anti-idiotypic antibody by free idiotypic IgM. CBA/H mice were inoculated on Day 0 with 5 X lo3 A3 I tumour cells im together with 140 U syngeneic anti-idiotypic antibody ip (55). 140 U anti-idiotypic antibody and 40 wg A3llgM ip (+3), or PBS only (+).

80

DYKE

ET AL.

a challenge of 5 X lo3 A31 cells is either cured, or a possible carrier of dormant tumour. The five immunised mice were sacrificed on Days 112, 121,237,238, and 246 posttumour challenge, and their spleens taken. These organs, which were all of normal weight and appearance, were processed for analysis. A FACS profile for the spleen from the mouse sacrificed on Day 112 indicated a minor population of idiotypepositive tumour cells among the normal lymphoid cells (Table 2). In the spleenstaken from the other four mice, no idiotype-positive cells were detectable. The absence of idiotype-positive cells is recorded in Table 2 as ~2% since mixing experiments indicated that this was the lowest percentage of tumour cells detectable by this technique. Cytocentrifuged cells were examined for idiotypic IgM by the immunoperoxidase technique in order to detect tumour cells which may be negative for surface IgM but positive for cytoplasmic IgM. No significant tumour cells could be detected. However, the passageof lymphocytes prepared from the spleens of the sacrificed mice into unimmunised recipients (5 X lo5 cells per mouse) revealed the presence of tumour in three-fifths of the long-term survivors (Table 2). In these recipients the tumour developed at a rate comparable with that expected from an ip injection of 5 X 1O3A3 1 cells, implying that the dormant tumour cells account for approximately 1% of the passagedcells. Injection of cells prepared from either lymph nodes or liver of the same sacrificed mice into unimmunised recipients failed to cause tumour, indicating that there were no tumour cells in these tissues. The expression of idiotypic IgM on the emergent tumour was indistinguishable from that of the parental A3 1 cells in all cases,with no evidence for variant tumours. In addition there was no difference in the amount of MHC Class I or II antigen expressed at the cell surface of either the parental A31 cell or the emergent tumour (Fig. 7). DISCUSSION The goal of immunising patients against their tumour has long been sought. However, some initially promising results in animal models ( 17, 18) were not reproduced in the clinic (19), so the whole approach came to be regarded with some scepticism. One of the major problems of these early experiments was that the putative tumourassociated antigens were not characterised. This is now being remedied by the use of molecular biological techniques which are revealing oncogene products as possible candidates (20), thus resurrecting the possibility of specific immunisation. The idiotypic

TABLE 2 DORMANCY STUDIES

Day of sacrifice Day Day Day Day Dav

112 121 237 238 246

TOTumour cells (using FACS analysis)

Survival of naive recipients (days)

6 <2 <2 <2 12

37, >100, >lOO, >I00 >lOO, >lOO, ZlOO 38,42, > 100 >lOO, >lOO, >I00 54, 56, > 100

IDIOTYPIC VACCINATION AGAINST B-CELL LYMPHOMA Control

Anti-id

Anti-Class

I

Anti-Class

81

II

t 13

E 2 5 0

FIG. 7. Expression of immunoglobulin and MHC antigens by tumour cells escaping from dormancy. Spleen cells from an immunised animal, which had survived tumour challenge for 237 days, were passaged into an unimmunised recipient. No tumour cells were detected by FACS at time of passage. Emergent tumour cells from this animal. and from an animal with normal A3 I tumour, were examined by indirect fluorescence techniques using the FACS III with rat monoclonal antibodies directed against the A3 I idiotype. MHC class I or class II antigens. or an irrelevant idiotype. Data were collected using a log amplifier.

determinants expressed by B-cell turnouts have always been an exception to the catalogue of poorly defined tumour-associated antigens in that immunoglobulin has a well-defined molecular structure. Its use in prophylactic immunisation against mouse myeloma has been clearly demonstrated (2 I), and the successful application of a similar approach to lymphoma has brought the idea closer to the clinic (2). The major problem for lymphoma is in producing sufficient idiotypic Ig as a pure product for immunisation. and this proved difficult in the A3 1 model. Only by strict cloning of the “rescue” hybridomas was it possible to obtain a stable clone which secreted idiotypic IgM at a suitable level. This has also been difficult for human lymphoma, and the use of cloned Ig genes may provide a solution to the problem (22). Once the IgM was available from A3 1 it proved to be efficient in establishing specific anti-idiotypic immunity when used either conjugated to KLH or alone. These results were very similar to those obtained in the BCL, lymphoma (2) but different from the 38C 13 tumour where it was necessary to conjugate IgM to KLH for protection (3). However the protection obtained for A3 1 was extraordinarily strong with no (19/ 19) immunised mice developing tumour after challenge, even with a large inoculum (5 X IO”) of cells. Interestingly, attempts to induce protection by immunising with A3 1 cells failed. confirming an earlier attempt (15). For the idiotypic antigen this may be dose-related since the IgM content of lysed A3 1 cells has been measured at 2 pg per 10’ cells whereas immunisation with pure protein used 50 pg per injection, i.e., a total of 150 Kg. Certainly no anti-idiotypic antibody was raised by the cellular immunisation. It was more surprising that no protective immunity was generated; other lymphoma models have shown quite strong anti-tumour immunity by this procedure although it has been difficult to delineate the antigen responsible. In some cases a viral product may be important in increasing tumour antigenicity (23). It is possible that the A31 tumour does not have virally encoded cell surface antigens, or that they are not immunogenic.

82

DYKE

ET AL.

The primary role of anti-idiotypic antibody in protection against A3 1 tumour, clearly demonstrated by passive transfer, parallels results obtained in several other models of lymphoma (2, 3). This reinforces the suggestion that lymphoma cells, which express Ig but secrete little, may be attacked most effectively by antibody, whereas the plasmacytomas, which express Ig poorly and secrete considerable amounts of idiotypic protein, can be dealt with more effectively by cellular mechanisms (4). The ability of the transferred antibody to suppress A3 1 tumour is very strong, diluting out at 2.8 U/ mouse which is equivalent to 48 ~1 of immune serum. If the blood volume of the mouse is taken as 2 ml, this closely parallels the dilution curve of pooled immune serum tested on tumour cells by immunofluorescence, indicating that there must be sufficient antibody in the injected dose to show >50% binding to the tumour-associated idiotype in order to confer complete protection. The mechanism of protection remains obscure in this and other models (2, 3, 24). Again no involvement of complement-mediated lysis could be demonstrated, and the suggested role of antibody-dependent cellular cytotoxicity (ADCC) remains a possibility which we are investigating. However the use of mouse effecters in ADCC may prove difficult since attempts to demonstrate ADCC using unsensitized mouse lymphoid cells have not always been successful (25). The finding of dormant tumour cells in the spleens of three-fifths of mice which had been completely healthy for several months after challenge is intriguing, especially since dormant tumour can exist in patients with lymphoma or other neoplasms (26). For the A3 1 lymphoma the number of such cells was very low, but they were capable of causing tumour in recipient mice. The presence of these cells in the spleens, possibly stimulating a low-grade immunity, may account for the observation that immune mice previously challenged with tumour were able to resist a second challenge 4 months later. Since the emerging tumour cells were identical to the parental lymphoma, with no variants, we now have the opportunity to investigate the splenic environment surrounding the dormant cells in order to find out if antibody alone, with no attached bullet, is the magic ingredient in suppressing tumour. ACKNOWLEDGMENTS The authors thank Dr. E. M. Shevach (National Institute of Health, Bethesda, MD) for kindly providing us with some of the antibodies used in this study. Our thanks also go to the staff of the animal house for monitoring the survival of the experimental groups. This work was supported by Tenovus. Andrew George is a Beit Memorial Fellow.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9.

Stevenson, F. K., and Gordon, J., J. Immunol. 130, 970, 1983. George, A. J. T., Tutt, A. L., and Stevenson, F. K., J. Immunol. 138, 628, 1987. Kaminski, M. S., Kitamura, K., Maloney, D. G., and Levy, R., J. Immunol. 138, 1289, 1987. George, A. J. T., and Stevenson, F. K., Intern. Rev. Immunol. 4, 271, 1989. Campbell, M. J., Carroll, W., Kon, S., Thielemans, K., Rothbard, J. B., Levy, S., and Levy, R., J. Immunol. 139,2825, 1987. Jorgensen, T., and Hannestad, K., Eur. J. Immunol. 7, 426, 1977. Abbas, A. K., Perry, L. L., Bach, B. A., and Green, M. I., J. Immunol. 124, 1160, 1980. Weinhold, K. J., Goldstein, L. T., and Wheelock, E. F.. J. Exp. Med. 149, 745, 1979. Meeker, T. C., Lowder, C. J., Cleary, M. L., Stewart, S., Warnke, R., Sklar, J., and Levy, R., N. Engi. J. Med. 312, 1658. 1985.

IDIOTYPIC

VACCINATION

AGAINST

B-CELL

LYMPHOMA

83

IO. Carroll. W. C.. Lowder. J. N.. Streifer. R., Warnke. R.. Levy. S.. and Levy, R., J fz,xp. Med. 164, 1566. 1986. I I Tutt. A. L.. Stevenson, F. K.. Slavin. S.. and Stevenson. G. T., Immurzolo~~ 55, 59. 1985. 12. Hamblin. T. J., Abdul-Ahad, A. K.. Gordon. J., Stevenson. F. K.. and Stevenson. G. T., &if. J. Cunwr 42, 495, 1980. 13. Krikorian, J. G.. Portlock, C. S., Cooney. D. P.. and Rosenberg, S. A.. Cuncrr 46, 2093. 1980. 14. Wheelock, E. F.. Weinhold, K. J.. and Levich. J.. Adr. Cunwr Res. 34, 107. 1981. 15. Cobb. L. M.. Glennie. M. J.. McBride. H. M.. Breckon, G.. and Richardson. T. C.. hi/. J. Chnwr 54, 807. 1986. 16. Stevjenson, G. T.. Smith. J. L., and Hamblin. T. J., In “Immunological Investigations of Lymphoid Neoplasms” (R. C. Nairn. Ed.). Churchill Livingstone. Edinburgh, 1983. 17. Gross, L., C’otwr Rc\. 3, 326. 1943. 18. Kronman. B. S., Wepsic. H. T.. Churchill, W. H. Jr.. Zbar, B.. Borsos. T.. and Rapp, H. J.. Scicnc~c 168,257. 1970. 19. Morton. D. L.. &win. Onr~ol 13, 180. 1986. 20. Bernards. R.. Destree. A.. McKenzie. S.. Gordon. E.. Weinberg. R. A., and Panicali. D.. Pm ~%‘:a//. /Icad. S