Manipulation of costimulatory signals to enhance antitumor T-cell responses

Manipulation of costimulatory signals to enhance antitumor T-cell responses

Manipulation of costimulatory signals to enhance antitumor T-cell responses James P Allison, Arthur A Hurwitz and Dana R Leach University of Californi...

523KB Sizes 0 Downloads 30 Views

Manipulation of costimulatory signals to enhance antitumor T-cell responses James P Allison, Arthur A Hurwitz and Dana R Leach University of California Berkeley, Berkeley, USA One of the major goals of tumor immunotherapy is the induction of tumor-specific T-cell responses that will be effective in eradicating disseminated tumors. Emerging information on the role of costimulatory molecules in T-cell activation offers several new strategies for enhancing antitumor responses, including the induction of expression of costimulatory molecules on tumor cells, enhancement of the presentation of transferred tumor antigen by host antigen-presenting cells, and ex vivo antigen priming of autologous antigen-presenting cells. Current Opinion in Immunology 1995, 7:682-686 Introduction It is generally accepted that stimulation by antigenM H C via the antigen receptor (TCR) alone is not sufficient for T-cell activation. Full induction of T-cell proliferation and IL-2 production requires additional costimulatory signals provided only by 'professional' antigen-presenting cells (APCs), including dendritic cells, macrophages, and activated B lymphocytes. T C R stimulation in the absence of costimulatory signals may even result in the induction of clonal anergy. Work from a number of laboratories has implicated CD28 as the major costimulatory receptor on T cells, and B7-1 (CD80) and B7-2 (CD86) its major ligands on APCs (for reviews, see [1°-3°]). Expression of B7-1 and B7-2 is in fact largely limited to professional APCs. The poor immunogenicity of many tumors may be, in large part, a consequence of failure to express costimulatory ligands. A tumor might remain invisible to the immune system in the early stages of its growth, despite the expression of tumor antigen-MHC complexes. Later in its growth, cellular debris produced by tumor necrosis, or by local inflammatory reactions, could be processed and effectively presented by host APCs that do express costimulatory ligands. By this time, however, the tumor might have reached a sufficient mass or degree of metastasis that the belated immune response is insufficient to eradicate it. In addition, potentially reactive T cells might be rendered anergic upon stimulation by tumor antigen in the absence of costimulation during early stages of tumor growth. Several groups [4-6,7 °°] have demonstrated that tumor immunity can be enhanced by the provision of costimulatory signals. F o r example, B7-1 transfected tumor cells have been

shown to be rejected by syngeneic hosts, and this rejection results in immunity to rechallenge with the parental, B7-negative, tumor cells [4-6] (for review, see [7°°]). This review will describe recent progress in the development of strategies for enhancing costimulation in tumor immunotherapy, concentrating on the period from June 1994 to June 1995.

Costimulation by tumor cells: the tumor as an antigen-presenting cell Immunogenicity of B7-1 transfected tumor cells The first reports to document the effectiveness of B7-1 transfected tumor cells as immunogens described studies using the murine melanoma K1735 [4,5] or SA1 sarcoma cells [6]. Subsequent work demonstrated the effectiveness of B7-1 in other murine tumor models, including lymphomas (EL-4 and RMA), a murine colon carcinoma (51BLiml0) and a mastocytoma (P815) [8,9°°]. Additionally, B7-1 transfected EL4 lymphoma cells and P815 mastocytoma cells were considerably more immunogenic than vaccines comprised of untransfected cells admixed with Corynebacterium parvum [10]. Thus, induction of rejection and immunity by B7 + tumor cells does not seem to be limited to a particular cell type, but can be achieved against a variety of weakly immunogenic tumors. Despite these successes, it is clear that the expression of B7-1 is not sufficient to induce regression in all tumor models. Other B7-1 transfected melanoma cell lines (B16 and CM5153), as well as a variety of methylcholanthrene-induced fibrosarcomas, were shown

Abbreviations APC---antigen-presentingcell; CTL--cytotoxic T lymphocyte; GM-CSF--granulocyte-macrophage colony-stimulatingfactor; TCR--T-cell receptor. 682

© Current Biology Ltd ISSN 0952-7915

Costimulatory signals in antitumor T-cell responses Allison, Hurwitz and Leach 683 to grow progressively in syngeneic hosts [8,9°°]. B7-1 appears to be effective only in those tumors that are somewhat immunogenic in more conventional immunization protocols, such as repeated injections o f inactivated cells [8]. Given that B7 endows the tumor cells only with costimulatory activity, and has no effect on the expression o f M H C or tumor antigens, it is not surprising that its effectiveness is bruited to immunogenic tumors. The cellular basis for the rejection of B7-1 ÷ tumor cells, and the induction of immunity to the parental tumor, appears to be largely a result o f activation o f CD8 + T cells. Although it has been shown that B7-1 + K1735 melanoma cells or EL-4 lymphoma cells grow progressively in mice depleted o f CD8 + T cells, depletion o f CD4 + T cells has little effect on the subsequent rejection o f the tumors [4,5,9°°,10]. Similarly, depletion of CD8 + cells after rejection o f B7-1 + K1735 cells was shown to eliminate the protection against rechallenge with the parental K1735 cells, whereas depletion o f CD4 ÷ T cells had no effect [9°°]. These results are consistent with the in vitro observation that B7-1 transfected cells expressing only class I M H C antigens could induce potent, class I specific alloresponses in the absence of class II + APCs or CD4 + T cells [11], and suggest that the B7-1 transfected tumor cells themselves are serving as the APCs. As many o f the tumor cell lines in which B7-1 has proven effective express only class I M H C antigens [9°'], induction o f C D 4 + T cells would not be expected unless transfer o f class II restricted tumor antigens to host APCs had occurred (the issue of host APCs and tumor-cell presentation will be addressed below). In any event, the existing data clearly indicate that for many tumor models there is no absolute requirement for CD4 ÷ cells in the antitumor response facilitated by B7-1 expression. In other systems, CD4 ÷ T cells were shown to play an important role. In the SA1 tumor model, cells transfected with both class II M H C and B7-1 induced potent antitumor responses requiring both CD4 + and CD8 + T cells [12°°]. A similar result was obtained in studies o f a subline o f K1735 melanoma cells, that expressed significant levels o f class II M H C antigens [13°°]. In this system, cells transfected with B7-1 and the human melanoma associated antigen p97 were found to induce both CD4 + and CD8 + T-cell responses, and both CD4 + and CD8 + T cells were necessary for tumor regression. There may be an advantage in having CD4 + T cells involved in the antitumor response. Although it is clear that B7-1 costimulation o f CD8 ÷ T cells can allow IL-2 production, proliferation, and acquisition o f effector function, subsequent encounters with B7tumor cells will result in tumor-cell lysis, but not IL-2 production or proliferation [11]. With time, or as a result o f encounter with weak stimulation by antigen-MHC I complexes in the absence o f costimulation, cytotoxic T lymphocytes (CTLs) elicited by the initial immunization with B7-1 ÷ tumor cells may become exhausted and their protective effect wane. This exhaustion may account for the relatively short period o f protection that was

documented in the K1735 model [9°°]. The presence o f tumor-specific CD4 + T cells, elicited by antigen-MHC class II complexes either on the tumor cell or by cross-primed host APCs, might provide a source of exogenous help for the CTLs which could invigorate and sustain their response. Recent work has established that B7-1 + tumor cells cannot only induce protective immunity to subsequent challenge with the parental tumor, but can also eliminate pre-existing tumor. B7-1 + K1735 cells administered up to eight days after implantation o f the parental tumor led to rejection of that tumor [13°°]. Similarly, vaccination with B7-1 + EL-4 lymphoma cells beginning six days after intraperitoneal injection of parental EL-4 cells protected 80%) o f the mice from death [9°°]. Finally, treatment with B7-1+, M H C class II + SA1 sarcoma cells 9-13 days after implantation of the parental tumor, when the tumor had reached 2-5 mm in diameter, resulted in rejection of the established tumor in the majority o f mice [12°°]. These results are very encouraging, and suggest that this approach to tumor immunotherapy may find its place in the clinical treatment o f cancer. One interesting finding that has come from recent studies is the demonstration o f cross-protection against additional tumors following rejection of B7-1 + tumor cells [9°°]. Mice that had rejected B7-1 + K1735 melanoma cells were found to be resistant not only to rechallenge with the parental tumor, but also another melanoma, CM519, as well as a syngeneic fibrosarcoma. Similarly, mice that rejected B7-1 + EL-4 were resistant to rechallenge by a second syngeneic lymphoma C6VL. Although the mechanism for this cross-protection is not established, the results suggest the existence o f shared tumor antigens that can be effectively presented by B7 + tumor cells and subsequently recognized on different B7- tumor cells. There are precedents for shared tumor antigens, such as the MAGE and tyrosinase antigens of human melanomas (see review by Van den Enden and Brichard, this issue, pp 674-681). Shared tumor antigens could be exploited in a clinical setting by immunizing patients with HLA-matched, B7 + tumor cells expressing the appropriate antigens, thereby eliminating the need to produce cell lines from autologous tumor cells o f each patient. Given the wide distribution o f MAGE antigens, for example, it might be possible to generate a panel o f B7-transfected melanoma cell hnes, typed for tumor antigen and M H C expression, that could be used as vaccines for the great majority of melanoma patients. One factor that may limit the clinical application o f B7-transfected cells is the nature of the vaccine. All published studies to date reporting successful immunization with B7 + tumors have used viable cells. Inactivation o f the cells by irradiation has been reported to severely diminish immunogenicity o f B7-1 transfectants [9"]. This contrasts many reports using lymphokines, especially granulocyte-macrophage colony-stimulating factor (GM-CSF), where irradiated cells have been found to be highly immunogenic [14], and may reflect

684

Cancer differences in the mechanism o f antigen presentation. The data are consistent with the B7 transduced tumor cell itself functioning as the APC. In contrast, in the cytokine-transfected tumor models, T cells seem to be primed by tumor antigens transferred to host APCs (see below). Irradiation o f B7 transfected cells may result in accelerated necrosis, with the cells not persisting long enough to effectively present antigen. Another possibility is that irradiation interferes with the APC function of the tumor cells. It is clear that additional work is needed on dose regimens and alternative means o f inactivation.

and lymphocyte function-associated antigen-1 (LFA-1) and the elevated expression of M H C class I and class II [22°]. The hybrid cells were rejected by syngeneic animals, and this rejection resulted in T cell dependent immunity to the parental tumor line. Although this approach has not yet been extended to other tumor models, it may be useful in cases where dividing tumor cells are not available for transfection or transduction.

Enhancement of host costimulation with GM-CSF Immunogenicity of B7-2 transfected tumor cells The identification o f a second CD28 ligand, B7-2 [2°,3°], prompted interest in its effectiveness o f costimulating tumor cell responses. B7-1 and B7-2 appear to have similar affinities for binding CD28, although there are differences in the kinetics o f the binding [15]. There has been some controversy as to whether B7-1 and B7-2 have different effects in costimulation. It has been suggested that costimulation by B7-2 favors T-helper cell type 2 responses [16,17]. In other reports using B7-1 or B7-2 transfected cells, however, no differences were found in the outcome o f costimulation [18,19]. Although the exact outcomes orB7-1 versus B7-2 costimulation remain unclear, two reports have documented the effectiveness o f costimulation via B7-2 in eliciting antitumor immunity [20°,21]. B7-2 transfected P815 cells were shown to regress in a process that required CD8 +, but not CD4 + T cells [20°]. This regression resulted in immunity to wild-type P815. Transfection o f B7-2, by itself or in combination with B7-1, had no effect on growth o f the non-immunogenic MCA120 fibrosarcoma. In another study [21], murine MC38 carcinoma cells transduced with recombinant vaccinia virus vectors containing either B7-1 or B7-2 genes were both rapidly rejected upon inoculation into syngeneic hosts. These reports establish that B7-2 expression can enhance responses to some tumors in a manner similar to that o f B7-1. More detailed studies o f mechanisms o f rejection will reveal whether the two costimulatory ligands are in fact equivalent in eliciting tumor immunity.

Fusion of tumor cells with antigen-presenting cells The work described above demonstrates that transfection o f tumor cells with individual costimulatory molecules can correct deficits in costimulation and enhance immunogenicity o f many tumors. Immunogenicity o f transfected tumors, however, might still be impaired by insufficient expression o f other molecules required for efficient T-cell activation, such as M H C and adhesion molecules. One way to overcome these problems would be to fuse tumor cells with professional APCs, and use the resulting hybrid cells as a vaccine. It was reported that fusion o f rat hepatocarcinoma cells with syngeneic activated B cells induced expression o f B7

As mentioned above, it has been shown that irradiated tumor cells transfected with GM-CSF elicit potent antitumor responses [14]. A subsequent report demonstrated that effective priming o f CD8 + CTLs in this system resulted from antigen presentation by cells derived from host bone marrow, and not by the tumor cells themselves [23°°]. GM-CSF apparently enhances differentiation and activation o f host APCs, especially dendritic cells, which then process and present transferred tumor antigen. This is in contrast to the mechanism o f immunization with B7-transfected tumor cells, where the tumor cells themselves are capable o f antigen presentation (H Levitsky, personal communication). Additionally, although B7 transfectants usually induce CD8 + T cells, GM-CSF induced immune responses to M H C class I + tumor cells involved both CD4 + and CD8 + T cells, indicating that transferred antigen is presented in the context o f both M H C class I and class II molecules [24]. On the other hand, GM-CSF transduced cells were also capable o f eliciting protection against M H C class Itumor cells in an manner dependent on natural killer cells. Thus, by enhancing host APC function, inherent deficits in antigen presentation and costimulation may be overcome.

Ex vivo priming of host antigen-presenting cells The effectiveness o f cross-priming of T-cell responses by host APCs with transferred tumor antigen suggests another strategy for tumor immunotherapy. It has been demonstrated that isolated murine dendritic cells, after short-term culture in GM-CSF and incubation with tumor fragments, can immunize mice against subsequent challenge with tumor [25,26°,27°]. Methods employing GM-CSF, tumor necrosis factor-or, and IL-4 have recently been developed that allow the generation o f large numbers o f dendritic cells from murine or human progenitor cells, including progenitors found in the blood o f adult humans [28-30,31°,32°]. The immunotherapy strategy that is emerging would be to generate dendritic cells from a patient's own blood, pulse them with tumor fragments (or antigenic peptides, if identified), and then reintroduce them into the patient

Costimulatory signalsin antitumor T-cell responsesAllison, Hurwitz and Leach 685 [27°]. This approach would avoid the necessity for the modification of the tumor cells and their reintroduction into patients.

Enhancement of T-cell responses by blockade of negative costimulatory signals All of the above techniques involve enhancing antitumor immunity by the provision of positive costimulatory signals to T cells. In addition to their interaction with CD28, however, B7-1 and B7-2 molecules also bind efficiently to CTLA-4, an activation induced T cell surface molecule that is highly homologous to CD28 (for reviews, see [1",2°]). In fact, CTLA-4 binds these ligands with an affinity that is orders of magnitude higher than CD28 [15]. The function of CTLA-4 has remained somewhat obscure. Two recent reports [33,34] suggest that CTLA-4 delivers a signal that inhibits T-cell activation, perhaps by interfering with CD28 costimulation. Should this be the case, then it is possible that blockade of the inhibitory CTLA-4 signals in vivo could result in augmented T-cell responses, thus enhancing the protective response to B7-transfected tumors.

References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: • of special interest •of outstanding interest 1. Allison JP: CD28-B7 interactions in T-cell activation. Curr Opin • Immunol 1994, 6:414-419. See annotation [3•]. 2. June CH, Bluestone JA, Nadler LM, Thompson CB: The B7 and • CD28 receptor families. Immunol Today 1994, 15:321-331. See annotation [3•]. 3. Sharpe AH: Analysis of lymphocyte costimulalion in vivo using • Iransgenlc and 'knockout' mice. Curr Opin Immunol 1995, 7:389-395. This paper and [1•,2 .] provide recent reviews of mechanisms of costimulation in T-cell activation. 4.

5.

6.

Chen L, Ashe S, Brady WA, Hellstrom I, Hellstrom KE, Ledbelter JA, McGowan P, Linsley PS: Costimulalion of antitumor immunity by the B7 connterreceptor for the T lymphocyte molecules CD28 and CTLA-4. Cell 1992, 71:1093-1102. Townsend S, Allison JP: Tumor rejection after direct costimulation of CD8 + T cells by B7-transfected melanoma cells. Science 1993, 259:368-370. Baskar S, Ostrand-Rosenberg S, Nabavi N, Nadler LM, Freeman GJ, Glimcher LH: Constitutive expression of B7 restores immunogenicity of tumor cells expressing truncated major histocompatibillty complex class II molecules. Proc Natl Acad 5ci USA 1993, 90:5687-5690.

7. •"

Ostrand-Rosenberg S: Tumor immunotherapy: the tumor cell as an antigen-presenting cell. Curr Opin Immunol 1994, 6:722-727. This review describes earlier work on the enhancement of tumor immunogenicity by transfection with MHC class II genes and/or B7-1. 8.

Conclusions The results of enhancing tumor costimulation in a number of animal models is encouraging. As shown above, the techniques for enhancing antigen presentation and costimulation are varied, and each has its own potential advantages. To optimize fully the therapeutic potential, it may be necessary to combine approaches such as inducing concurrent expression of B7 and GM-CSF or blocking negative signals while enhancing costimulation. Any technique that requires extensive modification of tumor cells, however, may be too cumbersome and expensive for widespread clinical application. Additionally, any manipulation of host immune responses has inherent dangers. Enhancing presentation of tumor antigens may also lead to the presentation of self antigens normally tolerated by the host and could result in autoimmune responses. Finally, there is the possibility that long-term growth of a tumor could result in the induction of clonal anergy in potentially reactive T cells, which may not be overcome by later provision of costimulatory signals. Although it is clear that there are obstacles to be overcome, recent advances in our understanding of the role of costimulatory signals in regulating the immune response offer encouragement that antitumor immunotherapy may become an effective clinical tool in the foreseeable future.

Chen L, McGowan P, Ash• S, Johnston J, Li Y, Hellstrom I, Hellstrom KE: Tumor immunogenicity determines the effect of B7 costimulatlon on T cell-mediated tumor immunity. J Exp Med 1994, 179:523-532. 9. Townsend SE, Su FW, Atherton JM, Allison JP: Specificity •• and longevity of antitumor immune responses induced by B7-lransfected tumors. Cancer Res 1994, 54:6477-6483. This paper describes aspects of the immunity induced by several B7-1 transfected tumor cell lines. It shows that the longevity of protection differs in two different tumor models. Cross-protection is reported in two different tumor models. Elimination of a pre-existing tumor is also demonstrated. Finally, this paper reports that irradiation greatly diminishes immunogenicity of B7-transfected tumor cells. 10.

Chen L, McGowan P, Ashe S, Johnston iV, Hellstrom I, Hellstrom KE: B7-1/CD80-lransduced tumor cells elicit better systemic immunity than wild-type tumor cells admixed wilh Corynebacterium parvum. Cancer Res 1994, 54:5420-5423.

11.

Harding FA, Allison JP: CD28/B7 interactions allow the induction of CD8 ÷ CTLs in the absence of exogenous help. J Exp Med 1993, 177:1791-1796. 12. Baskar S, Glimcher L, Nabavi N, Jones RT, Ostrand-Rosenberg •* S: Major histocompatibility complex class I1+B7-1 + tumor cells are potent vaccines for stimulating tumor rejection in tumorbearing mice. J Exp Mecl 1995, 181:619~o29. See annotation [13••]. 13. "•

Li Y, McGowan P, Hellstrom I, Hellstrom KE, Chen L: Costimulation of tumor-reactive CD4 ÷ and CD8 ÷ T lymphocytes by B7, a natural ligand for CD28, can be used Io treat established mouse melanoma. J Immunol 1994, 153:421-428. This paper and [12 "•] demonstrate that B7-I transfected tumor cells can be used to induce regression of pre-existing tumors, and that both CD4 + and CD8 ÷ cells are required for the effect. 14.

15.

Dranoff G, Jaffee E, Lazenby A, Golumbek P, Levitsky H, Brose K, Jackson V, Hamada H, Pardoll D, Mulligan RC: Vaccination wilh irradiated tumor cells engineered to secrete GM-CSF slimulates potent, specific, and long lasting antitumor immunity. Proc Natl Acad Sci USA 1993, 90:3539-3543. Linsley PS, Greene JL, Brady W, Bajorath J, Ledbetter JA, Peach R: Human B7-1 (CD80) and B7-2 (CD86) bind with

686

Cancer similar avidities but distinct kinetics to CD28 and CTLA-4 receptors. Immunity 1994, 1:793-801. 16.

Kuchroo VK, Das MP, Brown JA, Ranger AM, Zamvil SS, Sobel RA, Weiner HL, Nabavi N, Glimcher LH: B7-1 and B7-2

costimulatory molecules differentially activate the TH1/TH2 developmental pathways: application to autoimmune disease therapy. Cell 1995, 80:707-718. 17.

FreemanGI, Boussiotis VA, Anumanthan A, Bernstein GM, Ke XY, Rennert PD, Gray GS, Gribben JG, Nadler LM: 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. 18.

26.

Flamand V, 5ornasse T, Thielemans K, Demanet C, Bakkus M, Bazin H, Tielemans F, Leo O, Urbain J, Moser M: Murine dendritic cells pulsed in vitro with tumor antigen induce tumor resistance in vivo. Eur J /mmunol 1994, 24:605-610. This paper demonstrates that dendritic cells can be primed ex vivo and used to induce tumor immunity. See also [25].



27.

Grabbe S, Beissert S, Schwarz T, Granstein RD: Dendritic cells



as initiators of tumor immune responses: a possible strategy for tumor immunotherapy? Immunol Today 1995, 16:117-121.

This is an current review suggesting a promising strategy for immunotherapy. 28.

Sallusto F, Lanzavecchia A: Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregnlated by tumor necrosis factor alpha. J Exp Ivied 1994, 179:1109-1118.

29.

Xu S, Ariizumi K, Caceres-Dittmar G, Edelbaum D, Hashimoto K, Bergstresser PR, Takashima A: Successive generation of antigen-presenting, dendritic cell lines from murine epidermis.

Levine BL, Ueda Y, Craighead N, Huang ML, June CH:

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. 19.

Lanier LL, O'Fallon S, Spinoza C, Phillips IH, Linsley PS, Okumura K, Ito D, Azuma M: 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.

30.

20.

Yang G, Hellstrom



immunity elicited by tumor cells transfected with B7-2, a second ligand for CD28/CTLA-4 costimulatory molecules. J Immunol 1995, 154:2794-2800.

KE, Hellstrom

I, Chen L: Antitumor

This paper demonstrates that B7-2 expression can result in tumor rejection and in induction of immunity in a manner similar to that of B7-1. 21.

Hodge JW, Abrams S, Schlom J, Kantor JA: Induction of antitumor immunity by recombinant vaccinia viruses expressing B7-1 or B7-2 costimulatory molecules. Cancer Res 1994, 54:5552-5555.

22. •

Guo Y, Wu M, Chen H, Wang X, Liu G, Li G, Ma J, Sy MS: Effective tumor vaccine generated by fusion of hepatoma cells

with activated B cells. Science 1994, 263:518-520. This paper reports that fusion of a tumor cell with an activated APC endows the tumor cell with the capacity to induce potent immunity. 23. •°

J Immunol 1995, 154:2697-2705.

Huang AY, Golumbek P, Ahmadzadeh M, Jaffee E, Pardoll D, Levitsky H: Role of bone marrow-derived cells in

178:1067-1076. 31. •

Romani N, Gruner S, Brang D, Kampgen E, Lenz A, Trockenbacher B, Konwalinka G, Fritsch PO, Steinman RM, Schuler G: Proliferating dendritic cell progenitors in human blood. J Exp Med 1994, 180:83-93. See annotation [32•]. 32. •

24.

25.

Levitsky HI, Lazenby A, Hayashi RJ, Pardoll DM: In vivo priming of two distinct antitumor effector populations: the role of MHC class I expression. J Exp Med 1994, 179:1215-1224. Grabbe S, Bruvers S, Gallo RL, Knisely TL, Nazareno R, Granstein RD: Tumor antigen presentation by murine epidermal cells. J Immunol 1991, 146:3656-3661.

Bernhard H, Disis ML, Heimfeld S, Hand S, Gralow JR,

Cheerer MA: Generation of immunostimulatory dendritic cells from human CD34 + hematopoietic progenitor cells of the bone marrow and peripheral blood. Cancer Res 1995,

55:1099-1104. This paper and [31"] describe methods for obtaining large numbers of dendritic cells from precursors in the blood of adults. 33.

Walunas TL, Lenschow DJ, Bakker CY, Linsley PS, Freeman GI, Green JM, Thompson CB, Bluestone JA: CTLA-4 can function as a negative regulator of T cell activation. Immunity 1994, 1:405-413.

34.

Krummel MF, Allison JP: CD28 and CTLA-4 deliver opposing signals which regulate the response of T cells to stimulation. J Exp &,led 1995, 182:459--465.

presenting MHC class I.restricted tumor antigens. Science 1994, 264:961-965. This paper demonstrates that host-derived APCs are responsible for induction of T-cell responses to tumor cells transduced with GM-CSF. It is shown that tumor cells that do not express MHC class I genes can cross-prime CD8 + CTL responses to class I restricted tumor antigens.

O'Doherty U, Steinman RM, Peng M, Cameron PU, Gezelter S, Kopeloff I, Swiggard WJ, Pope M, Bhardwaj N: Dendritic cells freshly isolated from human blood express CD4 and mature into typical immunostimulatory dendritic cells after culture in monocyte-conditloned medium. J Exp Med 1993,

JP Allison, A Hurwitz and D R Leach, Cancer Research Laboratory, Department of Molecular and Cell Biology, University of California, Berkeley, California 94720, USA. JP Allison E-mail: [email protected]