BACKGROUND OF THE INVENTION Dendritic cells ("DC") are professional antigen-presenting cells ("APC") that play a crucial role in the initiation of the immune response of both helper and cytotoxic T lymphocytes (Banchereau, et al., Nature, 392: 245,1998). Dendritic cells reside in many tissues in an immature state and are characterized by their ability to capture and process antigens. Following antigen uptake, dendritic cells undergo a maturation process and migrate to the T cell areas of lymph nodes. During maturation, dendritic cells decrease their antigen processing capacity, increase cell surface expression of MHC and costimulatory molecules, and acquire the ability to produce IL-12. These phenotypic and functional changes correlate with their capacity to induce primary T cell responses.

CD40 is member of the TNF receptor ("TNF-R") family that plays a pivotal role in both cell-mediated and humoral immune responses. CD40 has wide distribution in tissue and cells, including B lymphocytes, monocytes, hematopoietic progenitors, dendritic, endothelial and epithelial cells.

CD40L (also known as CD154), the ligand for CD40, is mainly expressed on activated CD4+ T lymphocytes. CD40 triggering on dendritic cells induces phenotypic and functional maturation of dendritic cells, i. e. upregulation of costimulatory molecules (CD54, CD58, CD80 and CD86), enhanced capacity to induce T cell proliferation and cytokine secretion including IL-1, IL-6, IL-8, IL-10, IL-12, TNFa and macrophage inhibitory protein ("MIP") la (Van Kooten, et al., Current Opinion in Immunology, 9 : 330, 1997). In vivo, activation of antigen-presenting cells, presumably dendritic cells, through CD40 crosslinking, can replace the requirement for CD4+ T cell help, for the induction of CTL responses (Schoenberger, et al., Nature, 393: 480,1998; Bennett, et al., Nature, 393: 478,1998).

Most human malignancies express tumor-associated or tumor-specific antigens but do not elicit an efficient immune response. Recently, the role of CD40 and its ligand, CD40L, in the modulation of tumor cell antigen-presenting cell functions has been outlined.

The CD40 stimulation of tumor cells could be of major importance in eliciting an efficient anti-tumor effect since it both improves the initial phase of the immune response (antigen recognition) and the effector phase (cytotoxicity/cytokine secretion) via adhesion/costimulatory molecule up-regulation, increased endogenous antigen presentation and cytokine secretion, or by mimicking T-lymphocyte CD40L signaling. These data could support the use of CD40-mediated tumor-cell stimulation for adoptive transfer immunotherapy.

Nonetheless, restoration of tumor immunogenicity could in part be counterbalanced by the anti-apoptotic and growth signals delivered to lymphoid neoplasms (Van den Hove L. E., et. al. Leukemia, 11: 572-580,1997; Plumas J., et. al. Eur J Immunol., 25: 3332-3341.1995; Schultze J., et. al. Blood Rev. 10: 111-127,1996; Schultze J. L., et. al. Proc Natl Acad Sci U S A. 92: 8200-8204.1995; Schultze J. L., et. al. Journal of Clinical Investigation. 100: 2757- 2765.1997; Fisher, D. C., et. al. American Society of Hematology 40th meeting, Miami.

1010,1998; Costello R. T., et. al. Immunol Today, 20 : 488-493,1999) by CD40 triggering.

Finding other ways to improve immunogenicity of lymphoid tumors is desirable.

As an alternative to CD40 stimulation, the present inventors focused on stimulating dendritic cells and lymphomas through HVEM, a new member of the TNF-R superfamily which can mediate, under particular circumstances, tumor apoptosis (Harrop, J. A., et. al. J Biol Chem. 273: 27548-27556,1998). The HVEM/TR2 molecule (Montgomery R. I., et. al.

Cell. 87: 427-436.1996; Kwon B. S., et. al. Journal of Biological Chemistry. 272: 14272- 14276,1997; Tan K. B., et. al. Gene. 204: 35-46,1997) was first identified for its property of mediating the entry of the herpes simplex virus (HSV) and thus was called herpes virus entry mediator (HVEM). The tissue distribution of HVEM is wide, showing expression in primary cells in lung, spleen, thymus, monocytes, B-lymphocytes and T-lymphocytes (Kwon B. S., et. al. Journal of Biological Chemistry. 272: 14272-14276,1997; Tan K. B., et. al. Gene. 204: 35-46,1997; Morel Y., et. al. J Immunol. 165: 4397-4404,2000). Solid tumor cell lines do not express HVEM, while haematopoietic cell lines, in particular of myelomonocytic lineage, express the specific mRNA which is further up-regulated by phorbol-esters. To date, three ligands have been identified for HVEM. The first one is the herpes simplex virus surface envelope. glycoprotein gD 9. Two other molecules that are members of the TNF family which bind HVEM; the trimeric lymphotoxin (X (LToc3) and LIGHT.

LIGHT, a 29 kDa type II transmembrane protein expressed by activated T-cells (Mauri, et al., Immufaity, 8 : 21,1998) is expressed on activated T lymphocytes (Morel, et

al., Journal of Immunology, 167: 2479-2486. 2001), monocytes, granulocytes and immature dendritic cells. LIGHT recognizes three different members of the TNF-R family- Herpes virus entry mediator ("HVEM"), lymphtoxim ß receptor ("LTßR") and decoy receptor 3 ("DcR3", systematic name: TNFRSF6B).

HVEM was initially characterized as a mediator of HSV-1 infection (Montgomery, R. I., et al., Cell, 87: 427-36,1996) and subsequently, HSV-1 infection was shown to be inhibited by LIGHT (Mauri, D. N., et al., Immunity, 8: 21-30,1998). HVEM is broadly expressed on cells of the immune system like T, B lymphocytes, natural killer cells (Harrop, et al., Journal of Immunology, 161: 1786,1998; Kwon, et al., Journal of Biological Chemistry, 272: 14272,1997) and dendritic cells (Salio, et al., European Journal of Immunology, 29: 3245,1999) but it is also expressed on endothelial cells.

LTßR plays a key role in the development and organization of lymphoid tissue, but it is absent on mature T and B lymphocytes, primary monocytes and peripheral dendritic cells (Murphy, et al., Cell Death & Differentiation, 5: 497,1998).

DcR3, a TNF-R lacking a transmembrane region, is expressed in lung tissue and the colon carcinoma cell line SW480, and may serve to modulate LIGHT function in vivo.

Functionally, LIGHT can mediate apoptosis of some tumor cells in vitro and in vivo (Harrop, et al., Journal of Biological Chemistry, 273: 27548,1998; Zhai, et al., Journal of Clinical Investigation, 102: 1142,1998). Although this effect appeared to require co- expression at the cell surface of both HVEM and LTßR, recent studies demonstrated that LTßR expression is necessary and sufficient (Rooney, et al., Journal of Biological Chemistry, 275: 14307,2000). LIGHT mediated apoptosis activates death signals through selective recruitment of Tumor necrosis factor receptor-associated factor (TRAF) 3 by LT (3R, implicated by their co-localization. Through its interaction with HVEM, LIGHT is also an important costimulatory molecule for T cell activation. LIGHT stimulated T cell proliferation in a three way MLR (Harrop, et al., Journal of Biological Chemistry, 273: 27548,1998), which was inhibited by a neutralizing antibody to HVEM. Moreover, blockade of LIGHT inhibited dendritic cells mediated allogeneic T cell responses (Tamada, et al., Journal of Immunology, 164: 4105,2000). LIGHT stimulation of T cells activated NF-KB (Tamada, et al., Journal of Immunology, 164: 4105,2000), induced production of INFy (Tamada, et al., Journal of Immunology, 164: 4105,2000) and led to down modulation of HEVM (Morel, et al., Journal of Immunology, 167: 2479-2486, 2001). 1n vivo LIGHT is implicated in the development of the T cell immune response in tumor and

Graft-Versus-Host Disease models in the mouse, but the molecular mechanism of LIGHT function in these models remains to be elucidated (Tamada, et al., Nature Medicine, 6: 283, 2000). Since CD40-independent pathways are implicated in the T cell help for priming of CD8+ cytotoxic T lymphocytes by dendritic cells (Lu, et al., Journal of Experimental Medicine, 191: 541,2000), we examined the effect of LIGHT on dendritic cells maturation and their capacity to prime a CTL response.

SUMMARY OF THE INVENTION This invention is directed to novel methods of treating cancers comprising improving tumor immunospecificity by stimulating dendritic cells and lymphomas with certain TNF molecules, such as LIGHT.

Further, the present invention is directed to a method for treating dendritic cell mediated cancers comprising improving tumor immunogenicity by stimulating the dendritic cells with LIGHT in vitro to reach dendritic cell maturation and administering the mature dendritic cells to a patient in need thereof.

Further, the present invention is directed to a method for treating dendritic cell mediated cancers comprising improving tumor immunogenicity by administering to a patient in need thereof an amount of LIGHT capable of stimulating the maturation of dendritic cells in the patient.

Further, the present invention is directed to a method for treating non-Hodgkin's lymphomas comprising improving tumor immunogenicity and increasing Fas-induced apoptosis by stimulating lymphomas with LIGHT ex vivo and administering the stimulated lymphomas to a patient in need thereof.

Further, the present invention is directed to a method for treating non-Hodgkin's lymphomas comprising improving tumor immunogenicity and increasing Fas-induced apoptosis by administering to a patient in need thereof an amount of LIGHT capable of stimulating the lymphomas in the patient.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1A shows that HVEM is constitutively expressed on immature dendritic cells and that LIGHT is expressed on activated T-cell. Human immature dendritic cells (iDCs) were prepared from T-cell depleted adherent PBMCs by culture with 100 ng/ml GM-CSF and 10 ng/ml IL4 for five days. These iDC are CDla positive (>95%) and CD14 negative (<5%) (data not shown). Purified T lymphocytes were stimulated by PMA plus

ionomycin for 24h. The expression of CD40, CD40L, HVEM and LIGHT was assessed by flow cytometry using mAbs followed by FITC conjugated goat anti-mouse IgG. Filled histogram depicts specific mAb staining. Open histogram correspond to the negative control (isotypic matched antibody). These data correspond to one representative experiment out of four performed with different healthy blood donors.

Figure 1B shows that LIGHT and CD40L transfected L cells induce dendritic cells cluster formation. iDC generated as described above were incubated with 75 Gray irradiated CD32 (negative control), CD40L or LIGHT transfected L cells for 72 hours at a ratio 1 stimulator cell for 10 dendritic cells. Photomicrographs were taken on day 8 of culture with a resolution X10 and are representative of four experiments.

Figure 2 shows that LIGHT induces partial maturation of dendritic cells assessed as changes in surface phenotype. Final maturation of dendritic cells was induced by CD32 (negative control), CD40L, LIGHT or LIGHT+CD40L transfected L cells for 72h as described in Figure 1B. Cells were analysed by flow cytometry using PE conjugated mAbs and anti-HVEM mAb followed by FITC conjugated goat anti mouse IgG. Values represent % of positive cells or mean fluorescence intensity substrated of the value of matched isotype control mAb (open histogram). These data correspond to one representative experiment out of four performed with different healthy blood donors.

Figure 3A shows that LIGHT reduces macropinocytic activity of dendritic cells.

On day 8, immature dendritic cells (co-cultured with CD32 L cells) or CD40L, LIGHT and LIGHT+CD40L matured dendritic cells were incubated at 37°C or 4°C (negative control) for lh in the presence of FITC-dextran (0.5 mg/ml). Results are expressed as % positive cells and represent the mean of three independent experiments performed with different healthy blood donors.

Figure 3B shows that CD83 positive LIGHT matured dendritic cells have reduced macropinocytic activity. On day 8, LIGHT matured dendritic cells were incubated at 37°C or 4°C (negative control) for lh in the presence of FITC-dextran (0.5 mg/ml) and were then stained with CD83-PE mAb. The histograms show FITC-dextran uptake for the negative control (dotted line), CD83 negative LIGHT treated dendritic cells (solid line) and CD83 positive LIGHT treated dendritic cells (bold line).

Figure 3C shows that LIGHT cooperates with CD40L to enhance the capacity of dendritic cells to activate allogenic naive T-cells. 105 purified peripheral naive T-cells (CD4+, CD45RA+, CD45RO-) were stimulated by serial dilutions (3.103 to 14 cells/well) of irradiated (25 Gray) immature dendritic cells, CD40L-, LIGHT-or LIGHT+CD40L-

matured dendritic cells on day 8. The proliferative response was measured by [3H] Thymidine incorporation during the last 16h of a 6 day culture. Background T-cell proliferation was < 100 cpm. Results are expressed as mean cpm SD and are representative of four independent experiments.

Figure 4 shows that LIGHT synergizes with CD40L for cytokine secretion.

Immature dendritic cells, CD40L-, LIGHT-or LIGHT+CD40L-matured dendritic cells were generated as described above. On day 8, i. e. after 72h final maturation, supernatants were harvested and tested by ELISA for IL-12p75, IL6, TNFa and IL-1 (3. Results are expressed as pg/106 cells and are the mean of three experiments performed with different donor.

Figure SA shows that LIGHT and CD40L synergize for induction of specific antitumoral CTL activity. 106 CD8+ purified T-cells were cultured with 2.105 MelanA 26- 35 peptide pulsed immature dendritic cells, CD40L-, LIGHT-or LIGHT+CD40L-matured dendritic cells in the presence of IL-2 (10 UI/ml) and IL7 (Sng/ml). On day 7, CTL were harvested and restimulated at the initial ratio with the corresponding peptide pulsed dendritic cells. On day 13-14, cytotoxic activity of the resulting CTL cultures was tested at E: T ratios ranging from 100: 1 to 0.4: 1 in a standard 4h 5'Cr release assay on T2 targets in the presence or absence of MelanA 26-35 peptide (Igg/ml). Results are expressed in % specific lysis = % lysis in the presence of peptide-% lysis in the absence of peptide. One representative experiment out of three performed is shown.

Figure 5B shows that LIGHT costimulation induces enhanced IFN-secretion by CTL upon restimulation. On day 13-14, the resulting CTL were harvested and restimulated at the initial ratio by autologous MelanA 26-35 pulsed irradiated (25Gray) PBMC. After 48h, supernatants were harvested and tested by ELISA for IFN-y secretion. One representative experiment out of three performed is shown.

Figure 6 shows a model of CD4+ help for CD8+ CTL priming by dendritic cells.

CD4+ activated T cell interacts with immature dendritic cells and induce maturation through CD40L/CD40 and LIGHT/HVEM interactions. Mature dendritic cells are thus conditioned to prime CTL responses.

Figure 7A shows expression and regulation of HVEM expression in normal B lymphocytes. Resting peripheral blood mononuclear cells from a normal donor were double-stained with anti-HVEM mAb and anti-CD19 mAb (one representative experiment of 15 performed).

Figure 7B shows expression and regulation of HVEM in B cell lymphoid malignancies. White curves correspond to the negative control (isotype-matched mAb), and black curves to HVEM staining (CLL: representative of 1 of 15 experiments performed; ALL: 1 of 3 ; MCL: 1 of 3).

Figure 7C shows lymphoma cell stimulation. Lymphoma cells were incubated in medium alone (lst column), IFN-y (2nd column), CD32 transfected L-cell controls (3rd column) or CD40L-transfected L-cells (4th column) and analyzed by flow cytometry using specific antibodies. White curves correspond to the negative control (isotype-matched mAb), and black curves to the HVEM (lst row), CD80 (2nd row) or CD86 (last row).

These data correspond to 1 of 3 independent experiments with 3 different lymphomas.

Figure 8A shows the results of flow cytometry analysis of lymphoma cell following stimulation via CD40 or via HVEM. Lymphoma cells were incubated with CD32 negative control transfected L cells (CD32 L-cells; black), or LIGHT-expressing transfectants (LIGHT L-cells; white) and analyzed after 48 hours by flow cytometry using specific antibodies. Results are expressed here as % of positive cells (after subtraction of the background corresponding to the isotype-matched control). Data correspond to one representative experiments out of 5 performed with 3 different lymphomas.

Figure 8B shows the results of flow cytometry analysis of lymphoma cells following stimulation via CD40 or via HVEM. Lymphoma cells were incubated with CD32 L-cells ("unstimulated", black histograms) or LIGHT transfectants (red histograms) in absence or presence of an anti-Fas mAb with pro-apoptotic effect. Apoptosis was evaluated using flow cytometry analysis with the Apo2.7 mAb. Data correspond to one representative experiments out of 3 performed.

Figures 9A and B show the proliferation and IL-2 secretion of CD4+ T- lymphocytes in allogeneic mixed lymphocyte reaction against lymphoma cells. Allogeneic purified CD4+ lymphocyte were incubated at a 1/1 ration with irradiated lymphoma cells pre-incubated for 48 hours in different conditions. In Figure 9A, the proliferation of responding T-lymphocytes was assessed after 6 days of culture by a [3H] thymidine pulse in the last 18 hours of culture. Data represent the mean standard deviation of triplicates, and the experiments were performed with two different normal lymphocyte donors with two different lymphomas. In Figure 9B, the secretion of IL-2 by responding T-lymphocytes was assessed after 4 days of culture by an ELISA assay. Data represent the mean standard deviation of duplicates, and the experiments were performed with two different normal lymphocyte donors with two different lymphomas.

Figures 10A and B show lymphoma proliferation and apoptosis. In Figure 10A, lymphoma cells (50,000/well) were incubated with different numbers (6250 to 0/well) irradiated control transfectants (CD32 L-cells, black circles), CD40L-expressing transfectants (CD40L L-cells, black squares) or LIGHT-expressing transfectants (white triangles). After 4 days of culture, lymphoma cells were pulsed for the last 18 hours with [3H] thymidine. Data represent the mean standard deviation of triplicates, and have been performed three times. In Figure 10B, apoptosis of lymphoma cells was assessed at different times using the AP02. 7 mAb regarding different culture conditions; control transfectants (CD32 L-cells, black), CD40L-expressing transfectants (CD40L L-cells, striped) or LIGHT-expressing transfectants (white). Data represent the mean standard deviation of duplicates, and have been performed with three different lymphomas.

Figure 11 shows schematically the pleiotropic immune anti-tumor effects of LIGHT. The various effects of stimulation by LIGHT shown in this figure summarize the development of an immune reaction from dendritic cell antigen loading, maturation and migration to lymph nodes to cytotoxic T-lymphocyte (CTL) stimulation and target cell killing.

DETAILED DESCRIPTION OF THE INVENTION Applicants have discovered that the immune response could be increased by increasing the presentation of exogenously processed tumour antigens by professional antigen-presenting cells such as dendritic cells. Alternatively, the antigen- presenting cell functions of cancer cells, such as lymphomas, can be directly enhanced, thus improving both the presentation of endogenous peptides and the co-stimulation signals they provide to T-cells.

The present invention is directed to novel methods of treating cancers comprising increasing tumor immunospecificity by stimulating dendritic cells or lymphomas with TNF molecules, such as LIGHT. In one embodiment of the invention, soluble recombinant LIGHT, or LIGHT expressed on the surface of cells, is used to stimulate the maturation of dendritic cells in conjunction with stimulation via the DC40 pathway. During this period, the dendritic cells can be pulsed with tumor antigens, either as recombinant proteins, peptides or specially treated tumor cells (e. g. apoptotic bodies generated from tumor cells).

Such dendritic cells can then induce generation of tumor specific CTL responses which would eventually result in robust anti-tumor activity. In another approach, antibodies to

HVEM can be used to stimulate maturation of dendritic cells, in conjunction with stimulation via the CD40/CD154 pathway.

LIGHT, a recently identified member of the TNF superfamily, synergizes with CD40L to induce maturation of myeloid dendritic cells and to condition dendritic cells for CTL priming. Dendritic cells are potent antigen-presenting cells that control the development of T cell mediated immune responses (Banchereau, et al., Nature, 392: 245, 1998). In their immature state, dendritic cells capture antigen or apoptotic cells from sites of infection, inflammatory lesions or tissue damage and process the antigens for subsequent antigenic peptide loading to MHC Class I and Class II molecules and presentation to CD8 and CD4 T cells respectively. Following the stage of antigen capture, they begin to mature and migrate to the T cell areas of lymphoid organs to initiate the adoptive immune responses. The maturation process is a complex but sequentially highly ordered process.

Among the factors that contribute to this process are bacterial and viral products such as LPS and double-stranded (ds) RNA (Cella, et al., Journal of Experimental Medicine, 189: 821,1999) which activate dendritic cells resulting in upregulation of adhesion and costimulatory molecules and down-regulation of endocytic activities. Inflammatory cytokines TNFR and IL-lß (Cella, et al., Nature, 388: 782, 1997) or the activated T cell molecules CD40L and TRANCE (Josien, et al., Journal of Immmzology, 162: 2562, 1999; Anderson, et al., Nature, 390: 175,1997) represent endogenous stimuli that enhance the stimulatory capacity of dendritic cells.

LIGHT, like CD40L and TRANCE, is a member of the TNF family that is induced on T cells following activation (Morel, et al., Journal of Immunology, 167: 2479-2486, 2001). Upon LIGHT stimulation, only a fraction of monocyte derived immature dendritic cells acquire the fully mature phenotype characterized by expression of CD83 and high levels of HLA-DR. This is not due to insufficient level of LIGHT since stimulation at increased ratios of LIGHT-L cell: dendritic cells or with L cells expressing 3 fold higher levels of LIGHT on their surface does not improve the dendritic cell response (data not shown). Moreover, since LIGHT interaction with HVEM appears to down-modulate this receptor (Morel, et al., Journal of Immunology, 167: 2479-2486,2001), the absence of HVEM on the entire population of dendritic cells stimulated with LIGHT (Figure 2), indicates that all of the cells responded. Thus, for a substantial fraction of the iDCs, stimulation through HVEM alone is not sufficient to induce maturation.

Since LTR is not on T and B cells or dendritic cells (Murphy, et al., Cell Death & Differentiation, 5: 497,1998), LIGHT presumably stimulates dendritic cells through its

interaction with HVEM. The signaling pathways involved in dendritic cell maturation are not fully elucidated. Signal transduction pathways downstream of TNFR superfamily members have been extensively characterized in vitro. The cytoplasmic region of HVEM interacts strongly with TRAF2 but weakly with TRAF5, TRAF3 and TRAF1 (Marsters, et al., Journal of Biological Chemistry, 272 : 14029,1997). CD40, which shows significant cytoplasmic region homology to HVEM (12 identities), also interacts with TRAF2, TRAF3 and TRAF5 but, unlike HVEM, it recognizes TRAF6. In HEK-293 cells, transfection of HVEM induces activation of the transcription factors NF-kB and AP-1 (Hsu, et al., Journal of Biological Chemistry, 272: 13471,1997; Marsters, et al., Journal of Biological Chemistry, 272: 14029,1997). This activation is likely to be mediated by TRAF5 rather than TRAF2. In contrast, CD40 activates NF-kB through a TRAF2 mediated pathway (Rothe, et al., Science, 269: 1424,1995; Lee, et al., Proceedings of the National Academy of Sciences, USA, 96: 1421,1999). This is of particular interest since NF-kB activation is implicated in LPS-induced dendritic cell maturation (Rescigno, et al., Journal of Experimental Medicine, 188: 2175,1998). Using an IkB-a degradation inhibitor, it has been shown that nuclear translocation of NF-kB is necessary for upregulation of MHC class II and costimulatory molecules. In this context, the maturation of only a subpopulation of iDCs in response to stimulation by LIGHT suggests that not all of the iDC express the required complement of TRAFs and other adaptors or that the expression level of HVEM on the non-responding cells is too low to recruit a sufficient level of second messengers.

The mixed phenotype of dendritic cells matured in response to LIGHT highlights the inverse correlation between CD83 expression and the capacity to internalize FITC-dextran by macropinocytosis. This correlation suggests that closely related signaling pathways are involved in these two steps during dendritic cell differentiation.

LIGHT stimulated dendritic cells are unable to secrete the cytokines IL-12, IL-6, TNFoc and IL-1 (3. The absence of cytokine secretion can not be attributed to insufficient stimulation (see above) or to a different kinetics than the CD40L stimulation (data not shown). It remains to be determined whether the failure to induce cytokine secretion occurs at the level of mRNA expression, protein synthesis or secretion. Although signaling via HVEM alone is inefficient at inducing cytokine production, it strongly synergizes with CD40-mediated IL-12, IL-6 and TNFa but not IL-1 (3 secretion. IL-12 is a heterodimeric molecule produced by antigen-presenting cells that plays a pivotal role during immune responses by driving T cell differentiation towards the Thl cytokine pattern. Activation of p38 MAPK is critical for CD40 induced IL-12p40 production by human monocyte derived

dendritic cells (Aicher, et al., Journal of Immunology, 163: 5786,1999). The molecular basis of HVEM and CD40 synergy in cytokine regulation remains to be elucidated but activation of the p38 MAPK pathway represent a potential candidate. Consistent with this possibility, LIGHT and CD40L synergizes in IL-12 production, a p38 MAPK dependant cytokine but not in IL-1 (3 production, a MEK/ERK dependant cytokine.

Generation of specific CTL responses by the immune system provides a therapeutic approach to cancer. Cross-presentation is a mechanism that allows exogenous antigens like tumor, viral or transplantation antigens to be presented to class I-restricted cytotoxic T lymphocyte by the antigen-presenting cells (Sigal, et al., Nature, 398: 77,1999; Huang, et al., Science, 264: 961,1994). It has been demonstrated that induction of specific CD8+ CTL responses by such a cross-priming mechanism requires cognate CD4+ T cell help (Bennett, et al., Journal of Experitneiztal Medicine, 186: 65,1997). This CD4+ T cell help is particularly necessary under non-inflammatory conditions, as occurs in most cancers, including non-inflammatory persistent tumor viruses (human papillomavirus or Epstein- Barr virus) (Toes, et al., Journal of Experimental Medicine, 189 : 753,1999). Recent publications (Schoenberger, et al., Nature, 393: 480,1998; Bennett, et al., Nature, 393: 478, 1998; Ridge, et al., Nature, 393: 474,1998) have demonstrated that the CD4+ T cell help for the cytotoxic T cell response can be bypassed by activation of antigen-presenting cells through CD40. Antigen-presenting cells also function as a"temporal bridge"in that once activated, antigen-presenting cells are conditioned to deliver, a"license to kill"to CD8+ CTLs. Ridge et al (Ridge, et al., Nature, 393: 474,1998) have implicated dendritic cells in this model, suggesting that the activation state of dendritic cells is more important than the CD4+ T cell help itself (Toes, et al., Journal of Experimental Medicine, 189: 753,1999).

The present inventors used phenotypically-defined dendritic cell populations in in vitro CTL induction experiments, as underscored by Schuurhuis et al (Schuurhuis, et al., Journal of Experimental Medicine, 192: 145,2000). Human monocyte derived iDCs, called immature dendritic cells, express low levels of HLA class II and costimulatory molecules.

Upon co-culture with CD40-L, these dendritic cells exhibit the characteristic phenotype of mature dendritic cells (Banchereau, et al., Nature, 392: 245,1998), i. e. expression of CD83 and high level of HLA class lI and costimulatory molecules. The LIGHT plus CD40L costimulated dendritic cells show a new phenotype of mature dendritic cells with enhanced functional capacity. The present inventors developed an in vitro model of cross-priming against the human melanoma antigen MelanA/Mart-1 in order to test the specific CTL priming capacity of these differentially matured dendritic cell. CD40 matured dendritic

cells elicit moderate anti-MelanA cytotoxic activity, whereas LIGHT+CD40L costimulated dendritic cells are able to prime an enhanced anti-MelanA cytotoxic response. This shows that cytotoxic T cell response mediated by CD40 signaling can be strongly modulated by LIGHT co-stimulation. These results are consistent with recent studies examining the role of LIGHT in tumor and Graft-versus-Host Disease models in mice (Tamada, et al., Nature Medicine, 6: 283,2000). Transfection of LIGHT cDNA into P815 tumor cells enhanced their immunogenicity. This effect was not due to LIGHT-induced apoptosis, as has been observed for the human colon carcinoma line HT29 (Harrop, et al., Journal of Biological Chemistry, 273: 27548,1998) and the human breast carcinoma MDA-MB-231 (Zhai, et al., Joumal of Clinical Investigation, 102: 1142,1998). Accordingly, in addition to its co- stimulatory effect on T cell activation, LIGHT co-stimulates the dendritic cells by modulating CD40 signals. The dendritic cells then achieve an activation state in which they are able to elicit an enhanced anti-tumoral CTL response.

CD40-independent pathways of T cell help for CTL priming have been reported (Lu, et al., Journal of Experimental Medicine, 191: 541,2000). In addition to LIGHT, these pathways may involve other TNF/TNFR family members. TRANCE/RANKL is reported to enhances the adjuvant properties of dendritic cells in a in vivo model of Delayed-Type Hypersensitivity (Josien, et al., Journal of Experimental Medicine, 191: 495,2000).

OX40Lis expressed on matured dendritic cells (Oshima, et al., Journal of Immunology, 159: 3838,1997) and mice lacking this genes how an impaired antigen-presenting cell function in both priming and effector phases of T cell activation (Murata, et al., Journal of Experimental Medicine, 191: 365,2000). Thus, the present invention contemplates that, in addition to using the LIGHT/HVEM interaction to prime CTL by helper cell matured dendritic cells (Figure 6), other TNF/TNFR family members may also be used. In initial studies, immunization of human subjects with autologous ex vivo modified dendritic cells efficiently primed and boosted CD4+ and CD8+ T cells (Dhodapkar, et al., Journal of Clinical Investigation, 104: 173,1999). These results show that LIGHT/HVEM enhance this response and allow its application in dendritic cell-mediated cancer immunotherapy.

This invention is further directed at a method of treating non-Hodgkin's lymphoma by using an alternate and safe way to improve tumor immunogenicity and increase Fas- induced apoptosis. B cell lymphoid malignancies are inefficient at stimulating immune responses. One way to improve their immunogenicity is stimulation via CD40. However, such stimulation can be deleterious since in"low-grade"malignancies it induces a protective effect against apoptosis and/or a proliferative signal. As shown by the present

invention, this pro-tumorigenic effects can be avoided and tumor immunogenicity can be improved by stimulation of the constitutively expressed HVEM using its ligand LIGHT.

One example performed by the inventors tested the invention in a subtype of B-cell malignancy ("mantle cell lymphoma") which is a good candidate for alternate therapeutic approaches since it is not cured by conventional treatment and has a rapid evolution.

HVEM expression was stable in time and did not decrease following lymphoma cell- stimulation, in contrast to T lymphocytes whose activation induces HVEM disappearance.

Lymphoma triggering by the HVEM ligand LIGHT increased the expression of CD86 (B7- 2), a molecule that is critical for the development of an efficient immune response, and of Fas. In addition to increasing the expression of CD86, the tumor cells become more sensitive to Fas-induced apoptosis. LIGHT triggering of lymphoma cells increased their recognition in a mixed lymphocyte response (Figure 9A). In sharp contrast with CD40 stimulation, triggering by LIGHT did not induce lymphoma proliferation (Figure 9B).

Lymphoma cell stimulation by LIGHT therefore increases tumor cell immunogenicity, allowing its use in immunotherapy protocols with greater safety than CD40 since there are no proliferative or anti-apoptotic effects were detected. The use of LIGHT triggering in immunotherapy is further supported by its positive effects on the function of other pivotal cells of the immune network (T lymphocytes, dendritic cells), that may indirectly contribute to a better tumor control.

Primary antigen-specific T cell response requires not only antigen presentation but also costimulatory signals delivered by antigen-presenting cells, in particular via the CD28- CD80/CD86 pathway, in order to obtain effective T cell immunity instead of alloantigen- specific anergy (Guinan E. C., et. al. Blood. 1994 ; 84: 3261-3282; Boussiotis V. A., et. al., Res Immunol. 146: 140-149.; 1995; Boussiotis V. A., et. al., J Exp Med. 178: 1753-1763 1993; Gribben J. G., et. al., Blood. 87: 4887-4893,1996). One proposed model to improve immunogenicity of B-cell lymphomas 4 or CLL 1 relies on tumor cell triggering via CD40 which repairs the defect in costimulation by up-regulating CD58/LFA-3, CD54/ICAM-1, and CD80/CD86 molecules. Our data demonstrate that HVEM is expressed on all normal B-cells and in"low grade"lymphoma cells or chronic lymphocytic leukemia. This expression is long-lasting, even when lymphoma cells are stimulated. The stimulation of lymphoma cells by one of the HVEM ligands, i. e. LIGHT, induced an increased expression of CD86, but not of CD80. Experimental data support a greater role for CD86 than CD80 in conventional allogeneic MLR (Azuma M., et. al., Nature. 366: 76-79,1993), one-way MLR against dendritic cells (Caux C., et. al., J. Exp Med. 180: 1841-1847,1994), lymphoma (Plumas J., et. al. Eur J Immunol., 25: 3332-3341.1995) or acute myeloid leukemia (Costello R. T.., et. al., Eur J Immunol. 28 : 90-103,1998). This predominant effect of CD86 expression can explain the improved in vitro allogeneic immune recognition we observed following stimulation of lymphoma cells by LIGHT.

In most low-grade lymphoid malignancies and Hodgkin's disease, in vitro stimulation via the CD40 pathway improves tumor immunogenicity but also induces tumour cell proliferation and/or protection against apoptosis (Costello R. T., et. al. Immunol Today, 20: 488-493,1999). These effects partially mimic those observed in normal B-cells (Schattner E. J., et. al., J Exp Med. 182: 1557-1565,1995; Garrone P., et. al., J Exp Med.

182: 1265-1273,1995). As a consequence, the CD40 system appears to be an interesting approach for an ex vivo immunotherapy, i. e. in vitro treatment of irradiated tumor cells which are subsequently re-infused to the patient, but it raises safety problems for a direct in vivo use. The intracellular signaling delivered via HVEM contemplated by the present invention, which does not contain intracellular death domain, involves TRAF 1,2,3 and 5 and is able to activate NF-KB (Green J. M., et. al., Cell Immunol. 171: 126-131,1996) and avoids the problems associated with CD40 immunotherapy. In the present invention, the stimulation of HVEM by LIGHT failed to induce any lymphoma proliferation, even in the presence of other cytokines such as IL-4 (data not shown), in sharp contrast to the CD40/CD40L system. With respect to apoptosis, some data in solid tumors suggest that,

despite the absence of a death domain, stimulation via HVEM in conjunction with IFN-y induces programmed cell death. There was no induction of apoptosis by LIGHT treatment, even in association with IFN-y (not shown). However, LIGHT triggering was not protective against spontaneous apoptosis of lymphoma cells. Furthermore, LIGHT stimulation increased Fas expression. A parallelism between Fas expression and sensitivity to Fas- induced apoptosis is not always observed in lymphoid malignancies (Plumas J., et. al., Blood. 91: 2875-2885,1998; Iijima N., et. al., Blood. 90: 4901-4909,1997). Nonetheless, we observed that the up-regulation of Fas expression correlated, to some extent, with Fas- induced apoptosis sensitivity of lymphoma cells. This shows that LIGHT stimulation partially reversed the resistance to Fas killing. Since both immune killing and chemotherapy sensitivity partially rely on apoptosis and more specifically on Fas-driven killing (Iijima N., et. al., Blood. 90: 4901-4909,1997; Findley H. W., et. al., Leukemia. 13: 147-149,1999; Komada Y., et. al., Leuk Lymphoma. 25: 9-21.1997; McGahon A. J., et. al., Brit J Haematol. 101: 539-547,1998), LIGHT stimulation of lymphoma cells could be a way to improve tumor cell eradication, in contrast with CD40 stimulation which can protect lymphoma cells from both spontaneous and chemotherapy-induced apoptosis (Costello R. T., et. al. Immunol Today, 20: 488-493, 1999).

According to the present invention, therefore, stimulation of lymphoma cells by LIGHT retains immunogenicity restorative capacities without proliferation-inducing or anti- apoptotic effects. Moreover, LIGHT triggering may participate in the restoration of sensitivity to Fas-induced apoptosis. Finally, in vivo LIGHT stimulation may also amplify other pivotal pathways of the immune network (Figure 11) since LIGHT co-stimulates CD3-induced T-cell proliferation (Harrop J. A., et. al. J Biol Chem. 273: 27548-27556, 1998), participates in dendritic cell maturation 31 and amplifies IL-12 secretion (Morel Y., et. al., Journal of Immunol ; 167: 2479-2486, 2001). Accordingly, LIGHT stimulation is a good candidate for a"total immune therapy"approach that may simultaneously 1) enhance the intrinsic tumor immunogenicity, 2) restore tumor sensitivity to apoptosis, 3) participate in dendritic cell functional maturation, 4) directly stimulate T-lymphocyte activation, and 5) indirectly enhance T lymphocyte cytotoxicity potential via amplification of dendritic cell IL-12 secretion.

EXAMPLES I. Treatment of Dendritic Cell-Mediated Cancers A. Blood samples and cell separation Peripheral blood mononuclear cells (PBMCs) from healthy donors were isolated on Ficoll-Hypaque gradients (Costello, et al., European Journal of Immunology, 23 : 608, 1993). T-lymphocytes were isolated as the CD2+ PBMC population, corresponding to cells which adhere to sheep erythrocytes (Kamoun, et al., Journal of Experimental Medicine, 153: 207,1981) in the E-rosetting technique but fail to adhere to plastic dishes after overnight incubation in medium plus 30% FCS.

B. CD40L, LIGHTand LIGHT+CD40L transfected cell lines Full length cDNA of human CD40L and human LIGHT were cloned in pcDNA3.1/Neo and pcDNA3.1/Hygro (Invitrogen, Groningen, The Netherlands) respectively and transfected alone or sequentially by electroporation (960F, 220V) into LTK-murine fibroblasts. Stably transfected cells, selected by resistance to Geneticin, hygromycin B, or both were then selected for ligand expression by three round of FACS sorting. CD32 transfected fibroblasts are a kind gift from Schering-Plough (Dardilly, France).

C. Culture conditions and dendritic cell (DC) generation Culture experiments were performed in RPMI 1640 (Bioproducts, MA, USA) with 10% fetal bovine serum (Bioproducts). For dendritic cell generation, PBMCs were depleted of non-adherent cells by 4-hour adhesion on plastic dishes. Adherent cells were extensively washed and then cultured in RPMI 1640 (Bioproducts) 10% FCS with GM-CSF (Sandoz, Copenhagen, Denmark) at 100 ng/ml and IL-4 (Genzyme Corp., Cambridge, MA) at 10 ng/mL for 5 days. The medium was replenished with cytokines every 2-3 days. At day 5, final maturation was induced by the addition of irradiated (75Gy) L cells at a ratio of 1: 10 for an additional 72 hours.

D. Flow-cytometry studies For cell surface staining, cells were processed following standard procedures and analysis was performed on a FacScan flow cytometer (Becton Dickinson, Immunocytometry Systems, CA, USA). The mAbs directed against HVEM (12C5 and 20D4, both murine IgGl) and LIGHT (2C8, murine IgG2b) were generated at SmithKline Beecham by conventional hybridoma methodology from mice immunized with the respective recombinant proteins and screening the hybridomas by ELISAs. The mAbs to

CDla, CD3, CD4, CD8, CD14, CD19, CD25, CD40, CD54, CD56, CD69, CD83 and HLA- DR were purchased from Beckman Coulter (Marseille, France). The mAb to CD80 was from Beckton Dickinson, and the mAbs to CD86 and CD154 from Pharmingen. Cell surface CD40 and HVEM were quantified on immature dendritic cells by indirect immunofluorescence staining using QIFIKIT (Dako, Glostrup, Denmark).

E. Primary MLR Serial dilutions (3.103 to 14 cells/well) of irradiated (25Gy) stimulator cell were cultured in triplicate with 105 allogenic naive CD4+ T-cells in 96-well round bottom plate (Costar, Corning NY). Naive CD4+ T-cells were prepared from purified T-cells by three rounds of negative depletion using magnetic beads (Beckman Coulter) incubated with mAbs to CD8 and CD45RO (Beckman Coulter). Proliferation of T-cells was monitored by measuring methyl- [3H] thymidine (luCi/well ; Amersham, GB) incorporation during the last 16h of a 6-day culture. Thymidine uptake was counted on a gas-phase p-counter (Matrix 9600, Packard). Immature dendritic cells or dendritic cells matured by co-incubation with L cells expressing CD154, LIGHT or LIGHT plus CD154 were used as stimulators.

F. Macropinocytosis assay Mannose Receptor-mediated fluid phase macropinocytosis, measured by the cellular uptake of FITC-dextran, was used as a surrogate marker for antigen capture. On day 8,105 of the immature and matured dendritic cells were incubated in media containing FITC-dextran (0.5 mg/ml) (m. w. 40000; Sigma) for lh at 37°C or 4°C (negative control).

After four washes with cold PBS, cells were analysed by flow cytometry for uptake of the FITC-dextran.

G. Cytokine determination After 72h of final maturation, dendritic cell cultures were harvested and cell free supernatants were frozen. After thawing, cytokine concentrations were quantified by ELISA: IL-12p75, IL-lp (R&D Systems, Minneapolis, MN) and IL6, TNFa (Beckman Coulter, France).

H. Induction of specific anti MelanA CTL Immature and matured dendritic cells were pulsed for 2h at 37°C in serum free X- Vivo 15 (Biowhittaker) with MelanA 26-35 peptide (ELAGIGILTV) (101lg/ml) together with b2 microglobulin (3)-tg/ml). After two washes, 2.105 peptide pulsed dendritic cells were cultured with 106 CD8+ purified T-cells (>95% by flow cytometry), obtained by two round of negative selection (reference-or state antibody used) from purified T-cells, in 2 ml CTL medium in the presence of IL-2 (10 UI/ml) and IL7 (5ng/ml). The CTL medium is

IMDM (Biowhittaker) supplemented with L-Arg 550, uM, L-Asn 240M, L-GIn 1. 5mM, 1% Penicillin-Streptomycin and 10% pooled human serum. On day 7, CTLs were harvested and restimulated at the initial ratio with the corresponding peptide pulsed dendritic cells.

Cytotoxic activity was measured on day 13-14 in the slCr release assay described below.

L Cytotoxicity assay The T2 cell line was labeled by incubating 106 cells in IOOgCi sodium [5 1 Cr] chromate for 2h at 37°C and washing three times. Labeled target cells (103) and serial dilutions of effector cells in triplicate were incubated in RPMI 10% FCS in 96-well V- bottom plates at 37°C for 4h in the presence or absence of MelanA 26-35 (lg/ml) and a 30-fold excess of unlabeled K562 cells. Supernatants were then analysed in microplate scintillation counter (Topcount, Packard). Percentage lysis was determinated for each triplicate experiment as [(experimental 5'Cr release-spontaneous 5'Cr release)/ (maximal 5'Cr release-spontaneous 5lCr release)] x 100. Results are expressed in % specific lysis (% lysis in the presence of peptide-% lysis in the absence of peptide). One representative experiment out of three performed is shown.

J. IFN-ysecretion assay On day 13-14,106 CTL obtained as described were stimulated by 2.105 peptide pulsed autologous irradiated (25 Gy) PBMC. After 48 hrs, the levels of IFN-y in the culture supernatants were measured by ELISA (OptEIA, Pharmingen, San Diego, USA).

K. HVEM is constitutively expressed on immature dendritic cells while LIGHT is induced upon T cell activation HVEM, like CD40, is expressed on peripheral blood monocyte derived immature dendritic cells but at a lower level. Using the QIFIKIT quantification system (unlabelled mAbs followed by FITC conjugated goat anti mouse IgG), we estimate that there are 510,000 50,000 molecules of CD40 and 30,000 + 8,000 molecules of HVEM at the cell surface (data not shown). LIGHT and CD154 are not expressed on resting T-lymphocytes but both are upregulated at the cell surface following activation (Morel, et al., Journal of Immunology, 167: 2479-2486, 2001 and figure l. A). These cellular distributions suggest that, similarly to CD40L and CD40, the interaction of LIGHT with HVEM is important in T-cell communication with dendritic cells.

L. LIGHT induces phenotypically mature dendritic cells Immature dendritic cells (iDCs) were generated from T-cell-depleted adherent PBMC by a five day culture in the presence of GM-CSF and IL4. They were then incubated for 72 hrs with irradiated stable L cell transfectants expressing similar levels of CD32 (negative control), CD40L, LIGHT or LIGHT plus CD40L at a 1: 10 ratio of iDCs: stimulator. LIGHT transfected L cells induced small cell clusters (Figure 1B) whereas co-cultures with of CD40L (Figure 1B) or CD40L plus LIGHT (not shown) transfected L cells formed large homotypic cell clusters. No clusters were observed with control CD32 transfected cells and the dendritic cells remained non-adherent. The cell surface phenotype of these differentially matured dendritic cells were compared by flow cytometry. CD40L stimulation led to complete dendritic cell maturation. The CD40L-treated dendritic cells expressed CD83, were HLA-DRhigh and expressed high levels of adhesion (CD54) and costimulatory (CD80 and CD86) molecules. CD40L stimulation also enhanced the expression of its own receptor (CD40) but had little or no effect on HVEM. LIGHT stimulation induced maturation of a subpopulation of the iDCs. In different experiments, 20 to 40% of the dendritic cells acquired the mature phenotype of CD83+ and HLA-DRhigh This subpopulation also showed increased levels of the co-stimulatory molecules CD86 and CD40 but, unlike CD40L-stimulated dendritic cells, there was little effect on CD80.

Consistent with the observation of dendritic cells cluster formation, LIGHT stimulation induced up-regulation of the adhesion molecule CD54. Moreover, LIGHT down-regulated its own receptor, HVEM, as previously observed on T-lymphocytes (Morel, et al., Journal of Immunology, 167: 2479-2486, 2001. Immature dendritic cells co-cultured with L cells expressing both LIGHT and CD40L acquired a mature phenotype distinct from that obtained by CD40L or LIGHT stimulation alone. In comparison to stimulation with CD40L alone, these dendritic cells expressed higher levels of HLA-DR, CD54 and costimulatory molecules (particularly CD86) and showed down modulation of HVEM. These phenotypic changes were specific for CD40L and LIGHT stimulation since addition of mAbs to CD40L and/or LIGHT (10gg/ml) completely inhibited the elaboration of these markers. Similarly, iDCs cultured with CD32 transfected L cells maintained the immature phenotype of CD83-, HLA-DR°W, CD80'°W and CD86'°W.

M. LIGHT down-regulates pinocytic activity and cooperates with CD40L to enhance dendritic cell-mediated allogenic T-cell responses During the maturation process, dendritic cells lose their ability to capture exogenous antigen and in turn acquire potent antigen presenting capacity. We examined the

effect of LIGHT stimulation on these functional parameters of dendritic cell maturation. On day 8, immature dendritic cells, CD40L-, LIGHT-and LIGHT +CD40L-matured dendritic cells, generated as described above, were harvested and evaluated for their macropinocytic activity. FITC-dextran uptake was measured by flow cytometry and the results expressed as % of positive cells (Figure 3. A). Under the control condition of incubation with CD32 L cells, most of the immature dendritic cells (79 + 4% of positive cells) captured FITC- dextran whereas few of the CD40L matured dendritic cells (12 + 7% positive cells) showed uptake. LIGHT stimulation resulted in the loss of the capacity to take up FITC-dextran for a subpopulation of the dendritic cells (47 4% positive cells). This subpopulation corresponds to the CD83+ dendritic cells (Figure 3B), consistent with the expected phenotype of mature dendritic cells. As expected, LIGHT plus CD40L matured dendritic cells, like CD40L matured dendritic cells, did not capture FITC-dextran (9 1% positive cells).

The antigen presenting capacity of these differentially matured dendritic cells was then investigated in a primary allogeneic MLR (Figure 3C). CD40L matured dendritic cells show a stronger capacity to active T cells compared to immature dendritic cells. LIGHT matured dendritic cells showed only modest enhancement and only at dendritic cells: T cell ratios of <1: 500., In contrast, LIGHT in combination with s CD40L markedly increased of T-cell proliferation relative to CD40L alone.

N. LIGHT synergies with CD40L for cytokine secretion The dendritic cell maturation process is associated with cytokine synthesis.

Secretion of immunoregulatory and pro-inflammatory cytokines plays a pivotal role during T-cell priming in the lymphoid organs. As expected, CD40L stimulation induced high levels of secreted IL-12, IL-6, TNFa and IL-1 (3 compared to those produced by immature dendritic cells (figure 4). Consistent with their weak effect in the MLR, LIGHT matured dendritic cells only modestly induced IL-12 and TNFa and had no effect on IL-6 and IL-1 (3, relative to control iDCs. This low response could not be attributed to a toxic effect of LIGHT on cytokine secretion pathways since IL-8 secretion was not inhibited (data not shown). Moreover, LIGHT cooperated with CD40L to induce 5 to 15 fold higher levels of IL-12, IL-6 and TNF (X than observed for CD40L alone. In contrast, this combination had little effect on L-lés secretion.

O. LIGHT and CD40L synergize in tlze primifag of specific anti-tttrnor CTL and their production of INFy

Recent publications (Schoenberger, et al., Nature, 393: 480, I998 ; Bennett, et al., Nature, 393: 478,1998; Ridge, et al., Nature, 393: 474,1998) have proposed a two step model for the induction of cytotoxic T-cell responses in which dendritic cells play the role of"temporal bridge". First helper T-cells activate the dendritic cells via the interaction of CD40L and CD40. Then, dendritic cells are conditioned to prime CTL responses. The present Inventors developed an in vitro model of CTL priming against a tumor antigen consistent with this hypothesis. Immature and matured dendritic cells were pulsed with MelanA 26-35 peptide and incubated with autologous purified CD8+ T cells. After two rounds of stimulation, CD8+ T cells were tested for their anti-MelanA CTL activity against T2 cells (Figure 5A). LIGHT matured dendritic cells, like immature dendritic cells, were unable to prime CTL activity. As predicted by the model described above, CD40L matured dendritic cells induced MelanA specific lysis ranging from 8 1 % to 18 2 %, at effector: target ratios greater than 10: 1. LIGHT plus CD40L matured dendritic cells induced a marked increase of the anti-MelanA cytotoxic activity relative to CD40L alone, showing specific lysis from 17 2 % to 34 1 % over the same range of effector: target ratios.

IFN-y levels were measured to further assess the relative priming activity of the dendritic cells matured in the presence of LIGHT and/or CD40L. On day 13-14 after the initial dendritic cell stimulation, T-cells were harvested, challenged for 48 hrs with autologous irradiated MelanA pulsed PBMC and tested for the levels of IFN-y level secretion. Consistent with the cytotoxicity assays, LIGHT matured dendritic cells, much like immature dendritic cells, failed to prime T cells for IFN-y secretion. CD8+ T cells obtained by co-culture with CD40L matured dendritic cells produced only a modest level of IFN-Y. In sharp contrast, CD8+ T cells primed by LIGHT plus CD40L dendritic cells secreted large amounts of IFN-Y.

II. Treatment of non-Hodgkin's lymphoma by stimulation with LIGHT to improve tumor immunogenicity and increase Fas-induced apoptosis.

A. Blood samples, cell separation and cell lines In order to have sufficient number of tumor cells, 3 patients with circulating lymphoma cells > 95% of mononucleated cells were selected; isolation of tumor cells was performed on Ficoll-Hypaque gradients and controlled by flow cytometry analysis.

Peripheral normal blood mononuclear cells (PBMCs) of healthy donors were isolated on

Ficoll-Hypaque gradients (Costello R., et. al. Eur J Immunol. 23: 608-613,1993). T- lymphocytes from normal donors were isolated as the CD2-positive PBMC population, corresponding to cells which adhered to sheep erythrocytes (Kamoun M., et. al. J Exp Med.

153: 207-212,1981) in E-rosetting technique but failed to adhere to plastic dishes after overnight incubation in medium + 30% FCS. The CD4+ T-cells were isolated by two rounds of negative selection using magnetic beads (Beckman Coulter, France) and anti-CD8 (8E17, D Olive) mAbs. Purity of CD4+ cells was evaluated by flow cytometry analysis and was in all cases > 95%. The U266BL and XG7 (plasma cell lines), Daudi and Raji (Burkitt's lymphoma), DEL (Hodgkin's lymphoma) cell lines were obtained from the ATCC.

B. Culture conditions and Primary Mixed Lymphocyte Reactiozt (MLR) For lymphoma proliferation, culture experiments were performed in RPMI 1640 (Bioproducts, MA, USA) with 10% fetal bovine serum (Bioproducts). Lymphocytes were cultured at 106/mL. For MHC-unmatched MLR against lymphoma cells, responding CD4+ T lymphocytes were isolated from 3 unrelated healthy blood donors as previously described. Culture experiments were performed in RPMI 1640 (Bioproducts) with 10% fetal bovine serum (Bioproducts), 1 % L-glutamine (Life Technologies, Gaithersburg, MA, U. S. A), 1% Na pyruvate (Life Technologies) and 5.10-5 (3-mercaptoethanol (Sigma Chemical Co, Saint-Louis, MO, U. S. A). Lymphoma cells had y-irradiation at 50 Gy, were then incubated (from 5 x 104 per well) with T-lymphocytes (5 x 104 per well) for 6 days and pulsed for the 10 last hours with [3H] thymidine (Amersham, Buckinghamshire, United Kingdom). Thymidine incorporation was assessed with a direct ß counter (Matrix 9600, Packard Instruments, Rungis, France). For lymphoma cell stimulation experiments, 50 Gys y-irradiated L cells stably transfected with CD40L or control CD32 (kindly provided by Dr.

J. Banchereau, Schering-Plough, Lyon, France) or LIGHT (GlaxoSmithKline) were added (1 x 105/mL) to lymphoma cells (1 x 106/mL).

To induce Fas-dependent apoptosis, we used the CH-11 anti-Fas (Immunotech) mAb at 125 ng/ml, a concentration (per the manufacturer's recommendation) corresponding to 90% death of the WR19L-12a Fas-transfected cell line. The lymphoma cells were incubated for 48 hours with the appropriate stimulus (medium alone or LIGHT stimulation) and then the anti-Fas mAb was added for a 24-hour incubation.

C. Flow-cytometry studies Cells were processed following standard procedures and analysis was performed on a FacScalibur flow cytometer (Becton Dickinson, Immunocytometry Systems, CA, USA).

The following mAbs were used: the anti-HVEM 12C5 and 20D4 (mouse IgGl, SmithKline Beecham), anti-CD3, anti-CD4, anti-CD8, anti-CD14, anti-CD19 (Beckman Coulter), anti- CD80, anti-CD54, anti-CD58 (Beckton Dickinson), anti-CD86, anti-CD40L and anti-Fas (Pharmingen). Apoptosis was detected using the AP02. 7 mAb (Immunotech), that recognizes a 38 kDa protein localized on the membrane of mitochondria whose expression appears to be restricted to cells undergoing programmed cell death.

D. Cytokine production and assay T-lymphocytes were incubated with lymphoma cells in RPMI 1640 (Bioproducts, MA, U. S. A) with 10% fetal bovine serum and supernatants were harvested after a 2-day incubation. The IL-2 was measured using an immunoenzymatic assay with a sensitivity of 5 pg/ml (Immunotech).

E. HVEM expression by Iyznphoid cells (Table I and Fig. I) HVEM baseline expression in a panel of lymphoid cells was examined. As seen in Table I below, HVEM was not expressed in the most differentiated B cells, i. e. in the two plasma cell lines we evaluated (U266BL and XG7). In contrast, HVEM was expressed in all peripheral blood B-lymphocytes from all the healthy donors we tested (Table I, and Fig. 1 panel A, one representative experiment) and in all the eight Epstein-Barr Virus cell lines derived from PBMCs (data not shown). HVEM expression on Burkitt's lymphoma depended on the cell line tested since it was positive on Raji cells but not on Daudi. We failed to detect HVEM on the Hodgkin cell line we tested. HVEM was not expressed at the surface of acute lymphoblastic leukemias (Table I, and Fig. 7B, first row, phenotype of three representative samples). In contrast, we observed a consistent expression of HVEM in all the chronic lymphocytic leukemia and mantle cell lymphomas we tested (Table I, and Fig. 7B, second and third rows, phenotype of three representative samples). We choose to focus our study on this latter type of lymphoma that represents a lymphoid malignancy with very poor prognosis in the absence of a curative therapy.

HVEM expression time-course on lymphoma cells during a ten days culture period was tested. The 48-hour data are shown here (Fig. 7C). Either under unstimulated (medium alone or CD32-transfected fibroblasts) or stimulated (IFN-y or CD40L-transfected fibroblasts) conditions, HVEM expression (first row) was still detectable with no significant differences between the various conditions, and no difference regarding the

baseline level (not shown). A decrease of HVEM expression was observed by day 6 (not shown), and may reflect impaired viability of lymphoma cells. The up-regulation of the CD80 and CD86 adhesion/costimulation molecules (second and third rows) controls the efficiency of the IFN-y and CD40L stimulation.

The data shows that 1) constitutive expression of HVEM in lymphoma cells is very similar to normal B lymphocytes, and 2) this expression is stable in contrast with T lymphocytes, whose activation induces HVEM down-regulation (Morel Y., et. al. J Immunol. 165: 4397-4404,2000).

Table 1 : Expression of HVEM on Primary Cells and Cell Lines of B Lymphoid Origin Cell Type No. of Tested % Positive MFI Samples Cells Normal Primary B Lymphocytes 15... ++ EBV Cell Lines 8 U266BL (Plasma Cell Line) 1 XG7 (Plasma Cell Line) 1 Daudi cell line (Burkitt's lymphoma) 1 Raji cell line (Burkitt's lymphoma) 1 +++ + DEL cell line (Hodgkin's lymphoma) 1-- Chronic Lymphocytic Leukemias 15 +++ ++ Acute B Lymphoblastic Leukemias 3-- Mantle Cell Lymphoma 3 Data are presented as the % positive blasts or mean fluorescence intensity after background subtraction. Samples which were found to contain less than 5% blast or had a MFI < 5 were defined as negative and quoted-. For positive samples, the mean % of positive cells was quoted as follows: 5-10%, ; 10-20% +; 20-50%, ++; > 50%, +++. The MFI was quoted as follows: 5-10, ; 10-20 +; 20-100, ++; > 100, +++. One of the 7 AML tested expressed HVEM with high intensity (+++/++).

F. Flow cytometry analysis of lyniphon ? a cellfollowitig stimulatioll by LIGHT In order to better define the effects of HVEM triggering, the surface phenotype of lymphoma cells after a 48-hours incubation with LIGHT or CD32 (control) transfected L cells (Fig. 8A) was analyzed. The stimulation of lymphoma cells by LIGHT failed to up- regulate CD80, CD54 and CD58. In contrast, there was an increase in CD86 and Fas expression compared to the CD32 stimulated cells. Comparable results were observed for the other lymphoma tested (data not shown).

We then assessed the correlation of Fas expression with Fas-induced apoptosis using the anti-Fas mAb CH-11 at a concentration (per the manufacturer's recommendation) expected to induce significant apoptosis. Detection of apoptotic cells was performed using the AP02. 7 surface expression, which is restricted to cells undergoing apoptosis. Two hours after thawing, less than 10% lymphoma cells were labeled by AP02. 7 mAb (data not shown). After 48 hours of culture in standard medium without cytokine addition (Fig. 8B, left histograms), a percentage of lymphoma cells underwent spontaneous apoptosis but this was not different in the culture condition with CD32-transfected fibroblast (black histogram, 40% apoptotic cells, MFI = 13), compared to LIGHT-stimulated lymphoma cells (red histogram, 34% apoptotic cells, MFI = 12). Following anti-Fas treatment, unstimulated lymphoma cell apoptosis increased to 51 % although the expression level remained low (MFI=16). LIGHT stimulated lymphoma cells had higher sensitivity to Fas-induced apoptosis with AP02. 7 positive cell increasing to 68% and a two-fold higher intensity of staining (MFI=32). Comparable results were obtained with the other lymphoma tested (data not shown).

G. Proliferation and IL-2 secretion of CD4+ T-lymplaocytes in allogeneic rnixed lynphocyte response against lymphonaa cells (Fig. 9) We tested the modulation of lymphoma cell immunogenicity induced by LIGHT triggering. As seen in panel A, stimulation of lymphoma cells via HVEM induced greater proliferation of responding CD4+ T-lymphocytes than incubation with medium alone, CD32-expressing control fibroblast or IFN-Y. Of note, the addition of IFN-in conjunction with LIGHT stimulation did not have any synergistic or additive effect. With respect to IL-2 secretion (panel B), pre-stimulation of lymphoma cells by LIGHT allowed a relevant secretion by responding lymphocytes (~ 250 pgml/250, 000 cells/mL), higher than the level obtained by lymphoma cell pre-incubation with medium alone, CD32-expressing control fibroblast or even IFN-y (< 60 pgml). We failed to observe a synergistic or additive effect by addition of IFN-y to LIGHT stimulation.

H. Lymphoma proliferation and apoptosis (Fig. 10) As seen in panel A, lymphoma cells stimulated by CD40L-transfected cells displayed high level proliferation. In sharp contrast, stimulation by LIGHT-transfected cells did not induce any proliferation (even with higher level of stimulating cells, i. e.

50,000/wells, which correspond to a 1/1 ration, data not shown). In addition to these data, panel B shows that CD40L stimulation (striped bar) decreases the spontaneous apoptosis of lymphoma cells cultured for 24 hours in the absence of cytokines. In contrast, no apoptosis prevention is observed with LIGHT transfectant, with a percentage of apoptotic cells comparable to control obtained by lymphoma cell culture with CD32-expressing fibroblasts.