Tumour cell expression of B7 costimulatory molecules and ... - Nature

3 downloads 0 Views 146KB Size Report
ate the effects of B7–1 on systemic immunity, interleukin- expression of GM-CSF did not ... CD4+ T helper cells which in turn have been activated by professional ...
Gene Therapy (1998) 5, 223–232  1998 Stockton Press All rights reserved 0969-7128/98 $12.00

Tumour cell expression of B7 costimulatory molecules and interleukin-12 or granulocyte–macrophage colonystimulating factor induces a local antitumour response and may generate systemic protective immunity H Chong1, S Todryk2, G Hutchinson3, IR Hart3 and RG Vile2 1

Division of Histopathology, United Medical and Dental Schools of Guy’s and St Thomas’ Hospitals, St Thomas’ Campus; Laboratory of Molecular Therapy, Imperial Cancer Research Fund Molecular Oncology Unit, Hammersmith Hospital; and 3Richard Dimbleby Department of Cancer Research/ICRF Laboratory, The Rayne Institute, St Thomas’ Hospital, London, UK

2

Previously, we showed that expression of B7–1 in CMT93 murine colorectal tumour cells inhibited their growth in immunocompetent animals. However, this did not result in any significant increase in systemic protective immunity, relative to that elicited by the parental tumour. To potentiate the effects of B7–1 on systemic immunity, interleukin12 (IL-12) or granulocyte–macrophage colony-stimulating factor (GM-CSF) was co-expressed with this molecule. These combinations of immunostimulatory molecules were effective in eliciting systemic immunity. We also show that expression of B7–2 led to a local antitumour response as well as significantly raised systemic immunity. In another

tumour model, K1735 murine melanoma, which is moderately immunogenic, tumours secreting GM-CSF alone were as effective as the parental tumours in generating protective immunity. Previously, we described the deleterious effect of B7–1 expression on protective immunity. Coexpression of GM-CSF did not counteract this consequence of B7–1 expression. Expression of IL-12 was extremely effective in causing rejection of inoculated tumour cells, but evoked only minimal protective systemic immunity. These results suggest that combining costimulatory molecules and cytokines may be a useful therapeutic approach in some, but not all, tumours.

Keywords: costimulation; IL-12; GM-CSF; colorectal tumour; melanoma; gene therapy

Introduction Strategies aimed at inducing therapeutic immune responses against tumour cells have focused mainly on T cells, since these represent the immune cell population which exhibit antigen specificity and memory.1,2 An effective antitumour lytic response requires activation of pre-cytotoxic CD8+ T cells by cytokines secreted from CD4+ T helper cells which in turn have been activated by professional antigen presenting cells (APCs), such as dendritic cells and macrophages, that have taken up tumour antigens.3 Professional APCs may also activate CD8+ cytotoxic T cells directly.4 Activation of T cells requires an antigen-specific signal received by the T cell receptor/CD3 complex, but for efficient and full activation to occur antigen-nonspecific signals are also needed. A crucial costimulatory signal is provided by the CD28 receptor on T cells,5 the ligands for which belong to the B7 family, including B7–1 (CD80)6 and B7–2 (B70/CD86).7,8 Professional APCs express these B7 costimulatory molecules and thus are capable of presenting antigens to T cells effectively. Costimulatory signals lead to the production of various cytokines which Correspondence: Dr RG Vile, Laboratory of Molecular Therapy, ICRF Molecular Oncology Unit, Hammersmith Hospital, Du Cane Road, London W12 0NN, UK Received 19 September 1997; accepted 27 October 1997

have autocrine and paracrine effects on the proliferation, activation and maturation of T cells.9 Many human and experimental tumours express specific antigens which may be recognised by T cells and which may act as targets for an immune rejection response.10 However, as most tumour cells do not express costimulatory molecules, tumour-specific antigens are not presented to T cells efficiently. Indeed, this may represent one mechanism by which tumour cells elude recognition by the immune system.11 Therefore, costimulatory molecules have been expressed on the surface of tumour cells to enable them to present tumourassociated antigens, together with the costimulatory signal, directly to T cells thus obviating the need for helper T cells and APCs. In support of this model, at least in primary responses against tumours in vivo, some B7–1expressing tumours have been found to elicit an effective response which is mediated by CD8+ cells independent of CD4+ cells.12,13 Immunotherapy of cancer aims to generate an effective systemic immune response capable of controlling the growth of metastatic tumours. Thus, in addition to the local response against B7-expressing tumours, it is important to determine whether such engineered tumours evoke systemic protective immunity against unmodified tumour cells. In some cases, expression of B7 in poorly immunogenic tumours has been shown to generate systemic immunity,14–17 though not in other tumour

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

224

models.18,19 CMT93 murine colorectal tumour is of relatively low intrinsic immunogenicity and previously we demonstrated that expression of B7–1 prevented the growth of the primary tumour, but did not produce a significant increase in the level of systemic immunity, relative to that evoked by the parental tumour.20 Conversely, expression of B7–1 or B7–2 in another tumour, K1735 murine melanoma, a moderately immunogenic tumour, resulted in retarded growth of the primary tumour, but the systemic immunity elicited was weaker than that generated by the parental tumour.20 Recently, the ability of B7-expressing tumours to elicit systemic immunity was suggested to be dependent not only on direct antigen presentation to T cells by tumour cells, but also on crosspriming of T cells by host professional APCs.21,22 Among possible mechanisms by which the latter route of T cell activation may occur is the suggestion that B7 may activate natural killer cells that then lyse tumour cells, thereby increasing the availability of tumour antigens for uptake by professional APCs.23 Augmentation of antitumour immune responses might then be obtained by co-expressing B7 with a cytokine which would attract and stimulate other immune effector cells. In this report, we have employed such a combinatorial approach using B7–1 with one of two cytokines. IL12 is a disulphide-linked heterodimeric cytokine with a broad spectrum of activity which is produced by activated macrophages, B cells and dendritic cells. Among other activities, this cytokine stimulates proliferation and cytotoxic activity of NK cells and T cells.24 It also induces secretion of a range of cytokines, including interferon-␥, which in turn have various secondary actions. IL-12 cooperates with B7/CD28 costimulation in stimulating lymphocytes in vitro.25,26 Moreover, IL-12 has anti-angiogenic properties which may contribute to its antitumour activity.27 Both subunits of IL-12, p35 and p40 need to be expressed within one cell to obtain biological function,28 while the p40 subunit alone or as a p40 homodimer may inhibit activity.29 Indeed, expression of p40, without p35, produces an immunosuppressive effect in vivo.30,31 In order to achieve balanced expression of both subunits we have used a bicistronic vector which contains an internal ribosome entry site derived from encephalomyocarditis virus. IL-12 induces a strong local response against a number of tumours, in the context of using recombinant IL-1232 paracrine secretion at the tumour site using engineered fibroblasts or transfected tumour cells,33,34 as well as direct in vivo gene delivery.31,35–37 It also elicits protective immunity against poorly immunogenic tumours and was effective in causing regression of pre-established tumours.33,34 GM-CSF stimulates maturation of dendritic cells38 and up-regulates the surface expression of costimulatory molecules,39 thus enhancing the antigen-presentation efficiency of APCs. This might be especially useful for eliciting responses against tumour cells which express low levels of MHC class I, since these are unable to present tumour antigens adequately by themselves.40 Indeed, GM-CSF is particularly effective in generating systemic immunity against a number of poorly immunogenic tumours.41–43 It is expected that the combination of B7 costimulation and GM-CSF would be especially beneficial, since each promotes one of the two main routes by which activation of cytotoxic T cells occurs – direct

antigen presentation and crosspriming by professional APCs. In this report, we extend our previous findings on the use of B7 molecules for therapeutic applications. We have examined the effects of the immunostimulatory molecules on local tumour growth and the generation of systemic protective immunity. The latter was assessed by rechallenging with parental tumour cells animals which had previously been inoculated with live engineered tumour cells. Live cells were used for the primary inoculation because of their superior efficacy over nonreplicating cells in evoking systemic responses.44–46 We examined the effects of B7–2 expression in CMT93 tumours and in an attempt to potentiate the effects of B7– 1 on systemic immunity, B7–1 was co-expressed with IL12 or GM-CSF. The consequence of co-expressing B7–1 with these cytokines was also investigated in the K1735 tumour model. The results presented here suggest that combining expression of B7 with cytokines may be a potentially useful therapeutic approach in certain tumours, but may be of limited benefit in others.

Results Expression of B7–2 decreased the tumourigenicity of CMT93 colorectal tumour cells in immunocompetent animals Murine B7–2 (B70/CD86) was expressed in CMT93 colorectal tumour cells using the plasmid BCMGSNeo-mB70 and expression was detected by staining with mCTLA4H␥1 (Figure 1) or a B7–2-specific monoclonal antibody (data not shown). When injected into immunocompetent syngeneic animals, CMT93 B7–2 cells formed tumours less readily compared with CMT93 cells bearing the ‘empty’ expression vector. Only half the number of mice given the B7–2-expressing cells developed tumours and these tumours appeared relatively late (Figure 2) (P ⬍ 0.001, stratified log rank test). In athymic T cell-immunodeficient nude mice, CMT93 B7–2 cells grew at the same rate as the parental cells (data not shown), suggesting that T cells mediated the antitumour response in immunocompetent mice. Expression of B7–2 in CMT93 cells elicited systemic protective immunity Systemic immunity to parental tumour cell rechallenge was examined in mice which had rejected the primary inoculation of CMT93 B7–2 cells, as well as those which had developed progressive tumours that needed surgical excision. A majority of the animals were found to be protected against the rechallenge and this represents a significant increase in protection compared with animals which had been inoculated initially with control CMT93 cells bearing the empty vector (Figure 3) (P ⬍ 0.05, stratified log rank test). CMT93 cells which co-express B7–1 and IL-12, or express IL-12 only, did not form tumours in immunocompetent mice IL-12 was expressed in parental CMT93 cells and CMT93 B7–1 cells by transfection of plasmid pCR3-IL12. The CMT93 IL-12 clone used in this study produced 8.3 ng/106 cells per 48 h of IL-12, while the CMT93 B7–1/IL12 clone produced 9.7 ng/106 cells per 48 h. When these

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

225

Figure 2 Growth of CMT93 cells expressing B7-2 in immunocompetent mice. C57BL/6 mice were injected s.c. with 5 × 106 cells, as indicated in the key, and monitored for the development of a tumour (eight mice in each group). The results are representative of three independent experiments with similar results.

Figure 1 Expression of B7-1 or B7-2 in parental and engineered CMT93 cells. (a) Parental CMT93 cells, (b) CMT93 B7-1 cells, (c) CMT93 B72 cells. Cells were incubated with mCTLA4-Hg1, indicated by bold lines, followed by FITC-conjugated rabbit anti-human IgG as described in the Materials and methods. The faint lines represent control samples where cells were incubated with normal medium instead of the fusion protein. The x axis shows fluorescence on a log10 scale and the y axis represents the relative cell number.

cells were injected into immunocompetent animals, tumours failed to develop in all cases (Figure 4a). In contrast, tumours were established readily when the IL-12secreting cells were injected into nude mice, indicating that T cells were necessary for the antitumour response to be effective (Figure 4b).

CMT93 cells which co-express B7–1 and IL-12 generated systemic protective immunity Animals which had rejected CMT93 B7–1/IL-12 cells were rechallenged subsequently with parental CMT93 cells in the opposite flank. These mice showed a significantly increased level of protection compared with mice which had been given parental CMT93 cells initially and had tumours that required excision (Figure 5) (P ⬍ 0.02, stratified log rank test). In the group which had received a primary inoculation of CMT93 B7–1 cells, there was a tendency to improved protection compared with mice

Figure 3 Generation of systemic protective immunity by inoculation with live CMT93 B7-2 cells. C57BL/6 mice were given a primary inoculation of 5 × 106 CMT93 cells (six mice) or CMT93 B7-2 cells (eight mice). The key indicates the nature of this primary inoculation. Progressively growing tumours were surgically excised. All mice, including those which had rejected the primary challenge and mice which had established tumours that had undergone excision, were rechallenged in the opposite flank with 2 × 106 parental CMT93 cells. This Figure shows the growth resulting from the rechallenge. The results are representative of two independent experiments with similar results.

which had primary parental tumours but, consistent with our previous findings,20 analysis of data from independent repeated experiments failed to indicate that this increase was significant. Similarly, mice which had been injected initially with CMT93 IL-12 cells (not coexpressing B7–1) also showed a trend to enhanced protection, but this did not reach a level of significance.

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

226

Figure 5 Generation of systemic protective immunity by inoculation with live CMT93 cells expressing B7-1 and/or IL-12. C57BL/6 mice were inoculated with 5 × 106 CMT93 cells or CMT93 cells expressing B7-1 and/or IL-12 (10 mice per group). The key indicates the nature of this primary inoculation. Progressively growing tumours were surgically excised. All mice, including those which had rejected the primary challenge and mice which had established tumours that had undergone excision, were rechallenged in the opposite flank with 2 × 106 parental CMT93 cells. This Figure shows the growth resulting from the rechallenge. The results are representative of three independent experiments with similar results.

Figure 4 (a) Growth of CMT93 cells expressing B7-1 and/or IL-12 in immunocompetent mice. C57BL/6 mice were injected s.c. with 5 × 106 cells, as indicated in the key, and monitored for the development of a tumour (10 mice in each group). The results are representative of three independent experiments with similar results. (b) Growth of CMT93 cells expressing B7-1 and/or IL-12 in nude mice. Athymic nude BALB/c mice were injected s.c. with 5 × 106 cells and monitored for the development of a tumour (10 mice in each group).

CMT93 cells which co-express B7–1 and GM-CSF, or expressed GM-CSF only, were poorly tumourigenic in immunocompetent mice GM-CSF was expressed in parental CMT93 cells and CMT93 B7–1 cells using the retroviral vector, pBabeNeo GM-CSF. The clones used in these experiments produced 1.3 and 2.6 ng/106 cells per 48 h of GM-CSF, respectively. Progressively growing tumours did not develop when CMT93 B7–1/GM-CSF cells were injected into immunocompetent mice (Figure 6). Expression of GM-CSF alone also resulted in reduction of tumourigenicity. CMT93 cells which co-express B7–1 and GM-CSF elicited systemic protective immunity Animals which had received a primary inoculation of CMT93 B7–1/GM-CSF cells were rechallenged sub-

Figure 6 Growth of CMT93 cells expressing B7-1 and/or GM-CSF in immunocompetent mice. C57BL/6 mice were injected s.c. with 5 × 106 cells, as indicated in the key, and monitored for the development of a tumour (16 mice received CMT93 cells, eight mice in each of the other groups). The results are representative of two independent experiments with similar results.

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

227

sequently with parental cells. These animals showed a clear improvement in protection compared with those which had received primary parental cells (P ⬍ 0.05, stratified log rank test) (Figure 7). The protection offered by CMT93 cells expressing GM-CSF, without B7–1, was inconsistent. In the experiment depicted in Figure 7, there was a tendency towards protection but this did not reach statistical significance, while in an independent repeated experiment, these cells did elicit a significant increase in protection (data not shown).

Splenocytes from animals which had been inoculated with parental or engineered CMT93 cells secreted a Th1 pattern of cytokines Spleen cells were recovered from mice which had been inoculated 6 weeks previously with parental CMT93 cells or cells expressing B7–1, IL-12, B7–1/IL-12, GM-CSF or B7–1/GM-CSF. Mice injected with the parental cells developed tumours and these were excised. Spleen cells were also obtained from mice which had not been inoculated (naive mice). The spleen cells were cocultured with mitomycin C-treated CMT93 cells for 5 days and quantities of the following secreted cytokines were determined: IL-2, IL-4, IL-12, interferon-␥ and GM-CSF. For each experiment, spleen cells from two mice of each group were pooled. Figure 8 shows the results obtained from three independent sets of experiments (in total, three pairs of mice for each group). Neither IL-2 nor IL-4 was detected in all cases (data not shown), whereas IL-12 was secreted by cells from all groups. There was no corre-

Figure 8 Cytokine secretion by spleen cells obtained from naive mice and mice which have been inoculated with parental or engineered CMT93 cells. Spleen cells were obtained from mice 6 weeks after tumour cell inoculation. The cells were cultured in vitro with mitomycin C-treated CMT93 cells and after 5 days the quantity of cytokine secreted was determined by ELISA. Three independent experiments were performed using cells pooled from two mice per group. Each circle represents the result from a separate experiment.

lation between the amount of IL-12 produced and the in vivo tumourigenicity of the various engineered cells or their ability to elicit protective immunity. Spleen cells from all groups also produced GM-CSF and interferon␥, except cells from naive mice which consistently failed to produce detectable levels of these cytokines. Among the other groups, there was no pattern of correlation with the in vivo data.

K1735 melanoma cells which co-express B7–1 and IL-12, or express IL-12 only, did not form tumours in immunocompetent mice IL-12 was expressed in parental K1735 cells and K1735 B7–1 cells using plasmid pCR3-IL12. The clones in these experiments secreted 13 and 26 ng/106 cells per 48 h of IL-12, respectively. Co-expression of B7–1 and IL-12, or of IL-12 alone, abrogated the tumourigenicity of K1735 cells completely (Figure 9a). Even when 5 × 106 cells were inoculated, which represents 10-fold the minimal tumourigenic dose for the parental tumour line, no tumours developed. In nude mice, the IL-12-producing cells formed tumours in the majority of cases, though there was a marked delay in emergence of the tumours (Figure 9b).

Figure 7 Generation of systemic protective immunity by inoculation with live CMT93 cells expressing B7-1 and/or GM-CSF. C57BL/6 mice were given a primary inoculation of 5 × 106 CMT93 cells or CMT93 cells expressing B7-1 and/or GM-CSF (eight mice in each group). The key indicates the nature of this primary inoculation. Progressively growing tumours were surgically excised. All mice, including those which had rejected the primary challenge and mice which had established tumours that had undergone excision, were rechallenged in the opposite flank with 2 × 106 parental CMT93 cells. This Figure shows the growth resulting from the rechallenge. The results are representative of two independent experiments.

Mice which had rejected K1735 B7–1/IL-12 cells or K1735 IL-12 cells were not protected against parental cell rechallenge Mice which had rejected the initial inoculation of K1735 B7–1/IL-12 cells or K1735 IL-12 cells were rechallenged with parental cells. Tumours resulted in most of these animals and, although there was a delay in the appearance of the tumours relative to naive control mice, this represented less resistance compared with animals which had borne established primary parental tumours (Figure 10). Expression of GM-CSF in K1735 cells, or co-expression of B7–1 and GM-CSF, resulted in delayed growth of primary tumours Expression of GM-CSF in parental K1735 cells and K1735 B7–1 cells was achieved using the pBabeNeo GM-CSF retroviral vector. The clones used in this study secreted

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

228

Figure 10 Generation of systemic protective immunity by inoculation with live K1735 cells expressing Il-12, or co-expressing B7-1/IL-12. C3H/He mice were given a primary inoculation of 5 × 105 parental or engineered cells (10 mice per group). The key indicates the nature of this primary inoculation. Progressively growing tumours were surgically excised. All mice, including those which had rejected the primary challenge and mice which had established tumours that had undergone excision, were rechallenged in the opposite flank with 5 × 105 parental K1735 cells. This Figure shows the growth resulting from the rechallenge. The experiment is representative of two independent experiments with similar results.

Figure 9 (a) Growth of K1735 cells expressing IL-12, or co-expressing B7-1/IL-12 in immunocompetent mice. C3H/He mice were injected s.c. with 5 × 105 cells, as indicated in the key (10 mice per group), and monitored for the development of a tumour. This experiment is representative of two independent experiments with similar results. (b) Growth of K1735 cells expressing IL-12, or co-expressing B7-1/IL-12, in nude mice. Athymic nude BALB/c mice were injected s.c. with 5 × 105 cells, as indicated in the key (10 mice per group), and monitored for the development of a tumour.

9.3 and 8.3 ng/106 cells per 48 h of GM-CSF, respectively. Tumours developed in most cases when these cells were injected into immunocompetent mice, although there was a delay in the appearance of these tumours relative to those arising from parental cells (Figure 11).

K1735 cells which express GM-CSF alone elicited protective immunity against parental cell rechallenge, while K1735 cells which co-express B7–1 and GM-CSF generated less immunity Mice which had been inoculated initially with K1735 cells which express GM-CSF, without B7–1, were protected against rechallenge, to a degree similar to that elicited by the parental tumours themselves (Figure 12). Previously, we reported that K1735 B7–1 generated a lower level of protection compared with that produced by the parental tumours.20 Co-expression of GM-CSF with B7–1 failed to improve the level of protective immunity.

Figure 11 Growth of K1735 cells expressing GM-CSF or co-expressing B7/1-GM-CSF in immunocompetent mice. C3H/He mice were injected s.c. with 5 × 105 cells, as indicated in the key (10 mice per group), and monitored for the development of a tumour. This experiment is representative of three independent experiments with similar results.

Discussion Some controversy exists as to the relative functional activities of B7–1 and B7–2.47 In a number of in vivo and in vitro studies, B7–1 and B7–2 were found to favour secretion of different profiles of cytokines,48–50 though other studies have not found any differences.51–53 Expression of B7–2 in CMT93 colorectal tumour cells resulted in a decrease in their tumourigenicity, similar to

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

Figure 12 Generation of systemic protective immunity by inoculation with live K1735 cells expressing B7-1 and/or GM-CSF. C3H/He mice were given a primary inoculation of 5 × 105 parental or engineered cells (eight or nine mice per group). The key indicates the nature of this primary inoculation. Progressively growing tumours were surgically excised. All, mice, including those which had rejected the primary challenge and mice which had established tumours that had undergone excision, were rechallenged in the opposite flank with 5 × 105 parental K1735 cells. This Figure shows the growth resulting from the rechallenge. This experiment is representative of two independent experiments with similar results.

the effects of B7–1 expression we had described previously.20 However, unlike B7–1, the B7–2-expressing cells induced a significant degree of protective immunity relative to that elicited by the parental tumour. The clones of B7–1 and B7–2 cells used in these studies expressed similar levels of the costimulatory molecule. Although B7–2 appeared to be superior in generating protective immunity in this tumour, this is not true for all tumours. Where these two costimulatory molecules have been compared, their effects have been found to be equal in some cases,54 while in others, either B7–117,55 or B7–216 was more potent. Although expression of B7–1 in CMT93 tumours was effective in inducing a local rejection response against the primary tumour, this did not generate a significantly increased level of systemic immunity relative to that evoked by the parental tumour.20 In an attempt to augment this response, B7–1 was co-expressed with either IL-12 or GM-CSF. IL-12 was selected for its pleiotropic effects on a variety of immune cell types and also because B7–1 and IL-12 co-operate in stimulating lymphocyte proliferation and activation in vitro.25,26 GM-CSF was used since it is involved crucially in the maturation of professional APCs, and crosspriming by these cells was expected to complement the ability of B7-expressing tumour cells to present antigens directly to T cells. Indeed, CMT93 cells which co-expressed B7–1 and one of these cytokines were found consistently to lead to a significant improvement in systemic protection. The cytokine secretion profiles of spleen cells were examined 6 weeks after inoculation with tumour cells. IL-

2 and IL-4 were not detected in these assays, whereas IL12 was present in all cases. The presence of IL-12 may reflect constitutive production by macrophages within the spleen cell population. Interferon-␥ and GM-CSF were not released by splenocytes from naive mice, but were produced by splenocytes from all the other groups, including mice which had been inoculated with parental tumour cells. This correlates with the observation that under the experimental conditions used here even the parental cells showed a tendency to elicit resistance against rechallenge (Figures 3 and 5, and data not shown). However, amongst the various groups which had been inoculated, there was no correlation of cytokine production with levels of systemic immunity. Previously, we reported that even though expression of B7–1 in K1735 melanoma caused the primary tumours to grow more slowly, the protective immunity induced by these tumours was inferior to that elicited by parental tumours.20 Expression of GM-CSF alone generated a level of protective immunity similar to that produced by the parental tumours but when co-expressed with B7–1, GMCSF failed to reverse the inhibitory effect of B7–1. The basis of this B7–1-mediated reduction in protective immunity is unclear and thus it is difficult to explain the failure of GM-CSF to improve immunity in this situation. However, one difference between the K1735 and CMT93 tumours which may be of significance and which may contribute to the different effects of B7–1 in these tumours, is that CMT93 cells express surface MHC class I molecules, whereas these are undetectable on K1735 cells. As such, K1735 cells which express B7–1 may still be inefficient at presenting antigens to T cells directly. IL-12, with or without B7–1, was powerful in abrogating the growth of K1735 cells completely. No tumours developed even when the cell inoculum was increased by 10-fold that of the parental minimal tumourigenic dose. However, mice which had rejected IL-12-secreting K1735 cells showed little protection against parental cell rechallenge although, compared with naive mice, there was a clear delay in the emergence of the tumours (Figure 10). Why the IL-12-secreting cells were weak at generating protective immunity, in spite of their potent effects against the primary inoculation of tumour cells, is uncertain. Since these cells were rejected, the quantity of tumour antigens to which the host was exposed would be very low compared with mice injected with parental K1735 cells, in which progressively growing tumour masses were established. Possibly, IL-12-secreting K1735 cells in the primary challenge induced a particularly strong local response, including ‘nonspecific’ effector cells such as NK cells, which may result in such early elimination of tumour cells as to prevent sufficient time for necessary interaction with T cells. In support of this possibility is the observation that when IL-12-secreting K1735 cells were injected into nude mice, there was a marked delay in the emergence of tumours compared with mice given parental cells (Figure 9b), indicating that non-T effector cells, such as NK cells, must have participated in the initial reaction against the IL-12-secreting cells. This contrasts with the situation in the CMT93 tumour model, where IL-12-producing cells grew readily in nude mice (Figure 4b) suggesting that in this instance T cells were involved in the early reaction against these cells in immunocompetent mice. Interestingly, others have recently reported that co-expression of B7–1 and IL-

229

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

230

2 in another tumour resulted in brisk rejection of the cells but that the systemic immunity elicited was relatively weak.56 This situation bears some similarities with our results with IL-12-secreting K1735 cells. The combination of B7–1 and IL-12 was effective in generating systemic immunity in other models. Thus, B7– 1-expressing tumour cells co-injected with an equal number of IL-12-secreting tumour cells or fibroblasts,57 or B7– 1-expressing tumour cells together with recombinant IL12,58 both elicited systemic immunity. These approaches contrasted with ours where we employed tumour cells which were engineered to co-express B7–1 and IL-12. More comparable with our experiments, the advantages of co-expressing B7–1 with IL-12,59 or B7–2 with GMCSF,60 have also been reported, but in these cases the effects on systemic immunity of expressing B7 or the cytokine alone were not determined. Therefore, it is difficult to ascribe these findings to the combined effect of these molecules. Our study differed in that we also investigated the effects of expressing B7, IL-12 and GM-CSF separately as well as in combination. The effects of co-expressing B7–1 with other cytokines have also been examined. The combination of B7–1 and IL-2 was effective in reducing tumourigenicity,61 but an unexpected effect of this combination on systemic immunity was noted, as discussed above.56 B7–1 has also been combined with IL-462 or IL-745 and in these cases better protective immunity was generated compared with any of these molecules alone. We show here that it was possible to improve the ability of B7–1-expressing CMT93 tumour cells to generate systemic immunity by co-expression of cytokines. The choice of cytokines was based on properties which were predicted to complement those of the B7–1 cells. Expression of B7–1 was expected to make tumour cells competent at direct antigen presentation to T cells, while co-expression of the cytokine would enhance activation of T cells and other effector cells and promote crosspriming by professional APCs. Therefore, combinations of costimulatory molecules and cytokines represent a strategy which could be exploited for therapeutic applications, although our results with the K1735 tumour suggest that such an approach may not always be useful nor does efficacy in stimulating local responses necessarily correlate with ability in generating effective protective immunity. Experiments involving direct delivery of such combinations of immunostimulatory genes to established tumours in vivo are currently underway to evaluate their usefulness for inducing rejection of pre-established tumours and their efficacy in eliciting systemic immunity in this context.

Materials and methods Cell lines K1735 murine melanoma cells and CMT93 murine colorectal tumour cells engineered to express B7–1 have been described previously.20 All cell lines were grown in Dulbecco’s modified Eagle’s minimal essential medium supplemented with 10% (vol/vol) foetal calf serum. Expression plasmids and retroviral vectors Subcloning was performed using standard recombinant DNA techniques.63 The expression vector BCMGSNeo-

mB70 was a kind gift of Dr M Azuma.8 Plasmid pCR3IL-12 is a bicistronic expression vector based on the pCR3 vector (Invitrogen NV, Leek, The Netherlands) in which the p35 and p40 subunits of IL-12 are linked by the IRES sequence of the encephalomyocarditis virus, excised from the pCITE-1 vector (Novagen, Madison, WI, USA). The IL-12 subunits were generously provided by HoffmannLa Roche (Nutley, NJ, USA). Transfection was performed by calcium phosphate precipitation, as described previously.20 The retroviral vector, pBabeNeo GM-CSF, was constructed from the pBabeNeo vector.64 Production of vectors and infection of cell lines have been described previously.65 Briefly, vector plasmid was transfected into GP+envAM12 packaging cells66 and after incubation in selection medium for 3 weeks, surviving colonies were pooled and viral supernatant was harvested and used to infect target cells. Selection media contained 1.25 ␮g/ml puromycin (Sigma, Poole, UK) or 1 mg/ml G418 (Gibco, Paisley, UK). In all cases, compared with the parental lines, the engineered cell lines had similar morphology and growth rates in vitro.

Flow cytometry Detection of B7–2 expression by staining with the fusion protein mCTLA4-Hg1 (a gift from Dr P Lane67) or GL1 antibody, rat IgG2a (anti-B7–2; Pharmingen, Cambridge Bioscience, Cambridge, UK) has been described previously.20 Detection of GM-CSF production from cell lines Infected cell lines (106 cells) were seeded in normal medium and 48 h later, the supernatant was harvested and the cells trypsinized and counted. GM-CSF was assayed by ELISA using antibody pairs obtained from Pharmingen. Functional assay for detection and quantification of interleukin-12 IL-12 was quantified using an assay based on the ability of this cytokine to stimulate production of interferon-␥ by lymphocytes. Fresh splenocytes (2.5 × 106 cells in 0.5ml aliquots) were placed in wells with 0.5 ml of the test supernatant. A range of diluted recombinant IL-12 (Pharmingen) was used as standards. After incubation for 48 h, the supernatant was assayed for interferon-␥ by ELISA, using antibody pairs (Pharmingen). The assay for IL-12 was quantified by comparing the amount of interferon-␥ produced in the test samples with that induced by the recombinant IL-12 standards. Detection of cytokines secreted by splenocytes in vitro Fresh spleen cells were obtained from naive mice or mice which had been inoculated with a variety of engineered CMT93 cells. Spleen cells (5 × 106 cells) were mixed with mitomycin C-treated CMT93 cells (1 × 105 cells) in 2-ml volumes and incubated for 5 days. The spleen cells were then counted and the supernatant assayed for IL-2, IL-4, IL-12, GM-CSF and interferon-␥ by ELISA, using antibody pairs (Pharmingen). In vivo injection of tumour cells C3H/HeN mice, C57BL/6 mice and BALB/c nude mice were obtained from colonies bred at the Imperial Cancer Research Fund. Mice were age- and sex-matched for individual experiments. To establish subcutaneous tumours,

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

5 × 105 K1735 cells or 5 × 106 CMT93 cells were injected s.c. (100 ␮l) into the flank region. Animals were examined daily until the tumour became palpable, whereafter the diameter in two dimensions was measured three times a week. Animals were killed or had tumours surgically excised (see below), when tumour size was approximately 1.0 × 1.0 cm.

Rechallenge of mice After tumour excision mice were rechallenged with s.c. injection of 5 × 105 parental K1735 cells or 2 × 106 parental CMT93 cells on the opposite flank. Mice which had rejected the initial inoculation of tumour cells were also rechallenged. All groups of mice in any one individual experiment were rechallenged using the same preparation of cells on the same occasion, 7–10 days following the most recent surgical excision that had been performed on any mouse in the cohort. A naive group of mice was also injected with these cells at the same time. Animals were monitored as described above. Statistical analyses of in vivo experiments Data from animal experiments were analyzed by plotting Kaplan–Meier curves using the ‘occurring event’ as the time at which a tumour appeared. A tumour was considered present when a palpable mass ⬎0.2 cm was noted. Comparisons were made using the log rank test. Data from independent experiments were combined by using the stratified log rank test to obtain a more sensitive comparison.68

Acknowledgements This work was supported by the Imperial Cancer Research Fund. HC held an Imperial Cancer Research Fund Clinical Research Fellowship. ST holds a Postdoctoral Fellowship supported by the Lewis Family Charitable Trust. We are grateful to Hoffman-La Roche (Nutley, NJ, USA) for providing us with the IL-12 subunits.

References 1 Colombo MP, Forni G. Immunotherapy. I: Cytokine gene transfer strategies. Cancer Metas Rev 1996; 15: 317–328. 2 Tepper RI, Mule JJ. Experimental and clinical studies of cytokine gene-modified tumor cells. Hum Gene Ther 1994; 5: 153–164. 3 Pardoll DM. Cancer vaccines. Immunol Today 1993; 14: 310–316. 4 Huang AYC et al. Role of bone marrow-derived cells in presenting MHC class I-restricted tumor antigens. Science 1994; 264: 961–965. 5 Linsley PS, Ledbetter JA. The role of the CD28 receptor during T cell responses to antigen. Ann Rev Immunol 1993; 11: 191–212. 6 Freeman GJ et al. Structure, expression and T cell costimulatory activity of the murine homologue of the human B lymphocyte activation antigen B7. J Exp Med 1991; 174: 625–631. 7 Freeman GJ et al. Cloning of B7–2: a CTLA-4 counter-receptor that costimulates human T cell proliferation. Science 1993; 262: 909–911. 8 Azuma M et al. B70 antigen is a second ligand for CTLA-4 and CD28. Nature 1993; 366: 76–79. 9 Gimmi CD et al. B cell surface antigen B7 provides a costimulatory signal that induces T cells to proliferate and secrete interleukin 2. Proc Natl Acad Sci USA 1991; 88: 6575–6579. 10 Boon T et al. Tumor antigens recognised by lymphocytes. Ann Rev Immunol 1994; 12: 337–365. 11 Vile RG, Chong H, Dorudi S. The immunosurveillance of cancer:

12

13

14

15

16

17

18

19

20

21 22

23 24

25

26

27 28 29

30

31

32

33

specific and non-specific mechanisms. In: Dalgleish AC, Browning MJ (eds). Tumor Immunology. Cambridge University Press: New York, 1996, pp 7–38. Chen L et al. Costimulation of antitumor immunity by the B7 counterreceptor for the T lymphocyte molecules CD28 and CTLA-4. Cell 1992; 71: 1093–1102. Townsend SE, Allison JP. Tumor rejection after direct costimulation of CD8+ T cells by B7-transfected melanoma cells. Science 1993; 259: 368–370. Cavallo F et al. Co-expression of B7–1 and ICAM-1 on tumors is required for rejection and the establishment of a memory response. Eur J Immunol 1995; 25: 1154–1162. Dunussi-Joannopoulos K et al. Irradiated B7–1 transduced primary acute myelogenous leukemia (AML) cells can be used as therapeutic vaccines in murine AML. Blood 1996; 87: 2938–2946. Martin-Fontecha A et al. Heterogenous effects of B7–1 and B7– 2 in the induction of both protective and therapeutic anti-tumor immunity against different mouse tumors. Eur J Immunol 1996; 26: 1851–1859. Gajewski TF et al. Tumor rejection requires a CTLA4 ligand provided by the host or expressed on the tumor. J Immunol 1996; 156: 2909–2917. Ramarathinam L et al. T cell costimulation by B7/BB1 induces CD8 T cell-dependent tumor rejection: an important role of B7/BB1 in the induction, recruitment and effector function of antitumor T cells. J Exp Med 1994; 179: 1205–1214. Chen L et al. Tumor immunogenicity determines the effect of B7 costimulation on T cell-mediated tumor immunity. J Exp Med 1994; 179: 523–532. Chong H et al. Expression of co-stimulatory molecules by tumor cells decreases tumorigenicity but may also reduce systemic antitumor immunity. Hum Gene Ther 1996; 7: 1771–1779. Huang A et al. Does B7–1 expression confer antigen-presenting cell capacity to tumors in vivo? J Exp Med 1996; 183: 769–776. Cayeux S et al. Influence of gene-modified (IL-7, IL-4, and B7) tumor cell vaccines on tumor antigen presentation. J Immunol 1997; 158: 2834–2841. Wu T-Z et al. A reassessment of the role of B7–1 expression in tumor rejection. J Exp Med 1995; 182: 1415–1421. Trinchieri G. Interleukin-12: a proinflammatory cytokine with immunoregulatory functions that bridge innate resistance and antigen-specific adaptive immunity. Ann Rev Immunol 1995; 13: 251–276. Kubin M, Kamoun M, Trinchieri G. Interleukin-12 synergizes with B7/CD28 interaction in inducing efficient proliferation and cytokine production of human T cells. J Exp Med 1994; 180: 211–222. Murphy E et al. B7 and interleukin-12 co-operate for proliferation and interferon-␥ production by mouse T helper clones that are unresponsive to B7 costimulation. J Exp Med 1994; 180: 223–231. Voest EE et al. Inhibition of angiogenesis in vivo by interleukin 12. J Natl Cancer Inst 1995; 87: 581–586. Schoenhaut DS et al. Cloning and expression of murine IL-12. J Immunol 1992; 148: 3433–3440. Mattner F et al. The interleukin-12 subunit p40 specifically inhibits effects of the interleukin-12 heterodimer. Eur J Immunol 1993; 23: 2202–2208. Kato K et al. Local production of the p40 subunit of interleukin12 suppresses T-helper 1-mediated immune responses and prevents allogeneic myoblast rejection. Proc Natl Acad Sci USA 1996; 93: 9085–9089. Chen L et al. Eradication of murine bladder carcinoma by intratumor injection of a bicistronic adenoviral vector carrying cDNAs for the IL-12 heterodimer and its inhibition by the IL12 p40 subunit homodimer. J Immunol 1997; 159: 351–359. Brunda M et al. Antitumor and antimetastatic activity of interleukin-12 against murine tumors. J Exp Med 1993; 178: 1223–1230. Tahara H et al. Effective eradication of established murine tumors with IL-12 gene therapy using a polycistronic retroviral vector. J Immunol 1995; 154: 6466–6474.

231

B7 with IL-12 or GM-CSF induces systemic immunity H Chong et al

232

34 Cavallo F et al. Antitumor efficacy of adenocarcinoma cells engineered to produce interleukin-12 (IL-12) or other cytokines compared with exogenous IL-12. J Natl Cancer Inst 1997; 89: 1049–1058. 35 Bramson JL et al. Direct intratumoral injection of an adenovirus expressing interleukin-12 induces regression and long-lasting immunity that is associated with highly localized expression of interleukin-12. Hum Gene Ther 1996; 7: 1995–2002. 36 Caruso M et al. Adenovirus-mediated interleukin-12 gene therapy for metastatic colon carcinoma. Proc Natl Acad Sci USA 1996; 93: 11302–11306. 37 Rakhmilevich AL et al. Gene gun-mediated skin transfection with interleukin-12 gene results in regression of established primary and metastatic murine tumors. Proc Natl Acad Sci USA 1996; 93: 6291–6296. 38 Caux C et al. GM-CSF and TNF-␣ co-operate in the generation of dendritic Langerhans cells. Nature 1992; 360: 258–261. 39 Larsen CP et al. Regulation of immunostimulatory function and costimulatory molecule (B7–1 and B7–2) expression on murine dendritic cells. J Immunol 1994; 152: 5208–5219. 40 Levitsky HI et al. In vivo priming of two distinct antitumor effector populations: the role of MHC class I expression. J Exp Med 1994; 179: 1215–1224. 41 Levitsky HI et al. Immunization with granulocyte–macrophage colony-stimulating factor-transduced, but not B7–1-transduced, lymphoma cells primes idiotype-specific T cells and generates potent systemic antitumor immunity. J Immunol 1996; 156: 3858–3865. 42 Dranoff G et al. Vaccination with irradiated tumor cells engineered to secrete murine granulocyte–macrophage colonystimulating factor stimulates potent, specific, and long-lasting anti-tumor immunity. Proc Natl Acad Sci USA 1993; 90: 3539– 3543. 43 Castleden SA et al. A family of bicistronic vectors to enhance both local and systemic antitumor effects of HSVtk or cytokine expression in a murine melanoma model. Hum Gene Ther (in press). 44 Allione A et al. Immunizing and curative potential of replicating and nonreplicating murine mammary adenocarcinoma cells engineered with interleukin (IL)-2, IL-4, IL-6, IL-7, IL-10, tumor necrosis factor ␣, granulocyte–macrophage colony-stimulating factor and ␥-interferon gene or admixed with conventional adjuvants. Cancer Res 1994; 54: 6022–6026. 45 Cayeux S et al. Tumor cells cotransfected with interleukin-7 and B7.1 genes induce CD25 and CD28 on tumor-infiltrating T lymphocytes and are strong vaccines. Eur J Immunol 1995; 25: 2325–2331. 46 Santodonato L et al. Cure of mice with established metastatic Friend leukemia cell tumors by a combined therapy with tumor cells expressing both interferon-␣1 and herpes simplex thymidine kinase followed by ganciclovir. Hum Gene Ther 1996; 7: 1–10. 47 Boussiotis VA et al. The role of B7–1/B7–2:CD28/CTLA-4 pathways in the prevention of anergy, induction of productive immunity and down-regulation of the immune response. Immunol Rev 1996; 153: 5–19. 48 Freeman GJ et al. B7–1 and B7–2 do not deliver identical costimulatory signals, since B7–2 but not B7–1 preferentially costimulates the initial production of IL-4. Immunity 1995; 2: 523–532. 49 Lenschow DJ et al. Differential effects of anti-B7–1 and anti-B7– 2 monoclonal antibody treatment on the development of dia-

50

51

52

53 54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

betes in the nonobese diabetic mouse. J Exp Med 1995; 181: 1145–1155. Kuchroo VK et al. B7–1 and B7–2 costimulatory molecules activate differentially the Th1/Th2 developmental pathways: application to autoimmune disease therapy. Cell 1995; 80: 707–718. Lanier LL et al. CD80(B7) and CD86(B70) provide similar costimulatory signals for T cell proliferation, cytokine production and generation of CTL. J Immunol 1995; 154: 97–105. Levine BL et al. CD28 ligands CD80(B7–1) and CD86(B7–2) induce long-term autocrine growth of CD4+ T cells and induce similar patterns of cytokine secretion in vitro. Int Immunol 1995; 7: 891–904. Schweitzer AN et al. Role of costimulators in T cell differentiation. J Immunol 1997; 158: 2713–2722. Yang G et al. Antitumor immunity elicited by tumor cells transfected with B7–2, a second ligand for CD28/CTLA-4 costimulatory molecules. J Immunol 1995; 154: 2794–2800. Matulonis U et al. B7–1 is superior to B7–2 costimulation in the induction and maintenance of T cell-mediated antileukemia immunity. J Immunol 1996; 156: 1126–1131. Cayeux S et al. Lack of correlation between rejection of tumor cells co-expressing interleukin-2 and B7.1 and vaccine efficiency. Eur J Immunol 1997; 27: 1657–1662. Zitvogel L et al. Interleukin-12 and B7.1 co-stimulation cooperate in the induction of effective antitumor immunity and therapy of established tumors. Eur J Immunol 1996; 26: 1335– 1341. Coughlin CM et al. B7–1 and interleukin-12 synergistically induce effective antitumor immunity. Cancer Res 1995; 55: 4980–4987. Chen PW et al. Tumor cells transfected with B7–1 and interleukin-12 cDNA induce protective immunity. Ann NY Acad Sci 1996; 795: 325–327. Parney IF et al. Granulocyte–macrophage colony-stimulating factor and B7–2 combination immunogene therapy in an allogeneic Hu-PBL-SCID/beige mouse–human glioblastoma multiforme model. Hum Gene Ther 1997; 8: 1073–1085. Salvadori S et al. B7–1 amplifies the response to interleukin-2secreting tumor vaccines in vivo, but fails to induce a response by naive cells in vitro. Hum Gene Ther 1995; 6: 1299–1306. Cayeux S et al. Coexpression of interleukin-4 and B7.1 in murine tumor cells leads to improved tumor rejection and vaccine effect compared with single gene transfectants and a classical adjuvant. Hum Gene Ther 1996; 7: 525–529. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press: New York, 1989. Morgenstern JP, Land H. Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 1990; 18: 3587–3596. Vile RG et al. Systemic gene therapy of murine melanoma using tissue-specific expression of the HSVtk gene involves an immune component. Cancer Res 1994; 54: 6228–6234. Markowitz D, Goff S, Bank A. Construction and use of a safe and efficient amphotropic packaging cell line. Virology 1988; 167: 400–406. Lane P et al. Expression and functional properties of mouse B7/BB1 using a fusion protein between mouse CTLA4 and human gamma 1. Immunology 1993; 80: 56–61. Altman DG. Practical Statistics for Medical Research. Chapman & Hall: London, 1991.