CD40-activated B cells (CD40-B cells) have been demonstrated to expand memory and prime naïve CD8+ T cells in healthy individuals and cancer patients (1, 2, 3, 4, 5, 6) . These findings are of particular interest to the cancer vaccine field because CD40-B cells are reliably expandable from small amounts of peripheral blood without the technical limitations encountered using dendritic cells, especially in pediatric patients. A growing number of investigators have started to use CD40-B cells instead of, or in combination with, dendritic cells, and first clinical trials are under way (2 , 3 , 7, 8, 9, 10) . Nevertheless, it remains unclear whether B cells used as antigen-presenting cells might have significant biological limitations when used as cellular adjuvant. B cells take up and present antigen primarily and most efficiently via highly epitope-specific B-cell receptors (11 , 12) . Moreover, naïve B cells have been shown to induce CD4+ T-cell tolerance (13) . It, therefore, remained unclear whether ex vivo-generated polyclonal CD40-B cells, especially under long-term culture conditions, have the capacity to take up, process, and present a broad range of antigens to CD4+ T-cells. This is of particular importance because CD4+ T-cell responses are crucial for successful tumor rejection (14, 15, 16) .
Lapointe et al. (17) addressed this important question in this journal and demonstrated that CD40-B cells generate tumor antigen-specific CD4+ T cells in cancer patients when the B cells are pulsed with lysate from tumor cells (17) . The authors conclude that these findings emphasize the role of CD40-B cells as antigen-presenting cells for immunotherapy. Melanoma antigen-specific T cells were generated from the peripheral blood mononuclear cells of two melanoma patients. Because it is most likely that these melanoma antigen-specific T cells derived from melanoma patients were expanded from a preexisting population of memory T cells, it remains to be shown whether CD40-B cells also prime naïve CD4+ T cells against neoantigens as demonstrated for CD8+ T-cells (2) . This would also extend their potential use to neoantigens such as viral antigens or previously ignored potential tumor antigens. Most likely because of low antigen concentration in tumor lysates, Lapointe et al. reported a rather low efficacy of expansion of CD4+ T cells after stimulation with CD40-B cells, (17) , whereas antigen-specific CD8+ T cell expansion was successful in more than 80% of individuals tested (2 , 6) . Here, we have addressed the induction of naïve CD4+ T-cell responses and the efficacy of ex vivo CD4+ T-cell expansion by stimulation with CD40-B cells.
The system used for CD8+ T cells (1 , 2) was modified to allow for optimal induction of antigen-specific CD4+ T-cell responses to whole proteins (tetanus toxoid and keyhole limpet hemocyanin), and to the artificial promiscuous MHC class II binding peptides PADRE-AKF and PADRE-AKX, serving as peptide-neoantigens (18) . T-cell expansion and cytokine production were assessed after 2–5 stimulations in vitro with protein- or peptide-loaded CD40-B cells in at least four T-cell lines against each of the antigens. T-cell responses against tetanus toxoid were induced in four of four T-cell lines. Similarly, two of four and three of four T-cell lines were antigen-specific to PADRE AKX and PADRE-AKF (Fig. 1, A and B) ⇓ . Analysis for IFN-γ or interleukin-4 production did not reveal any skewing toward a TH1 or TH2 profile. Successful generation of antigen-specific CD4+ T cells in this optimized system was possible in 9 (75%) of 12 of T cell lines.
Antigen uptake by CD40-B cells was mainly due to fluid-phase uptake as demonstrated by Lucifer yellow incorporation and comparable with immature dendritic cells. As expected, mannose receptor-mediated antigen uptake (Dextran-FITC loading) was low compared with immature dendritic cells. (Fig. 1C) ⇓ . Even for molecules larger than Lucifer yellow, the MHC class II pathway of antigen uptake, processing and presentation is fully functional in CD40-B cells, as demonstrated by reduced T-cell activation by keyhole limpet hemocyanin and tetanus toxoid when using agents interfering with endocytosis (2-deoxyglucose, NaN3), early and late lyso-/endosomal transport (chloroquine; bafilomycin A, data not shown), recycling (primaquine), trafficking of newly synthesized molecules (brefeldin A), loading (cathepsin inhibitor), and presentation by MHC class II molecules (MHC class II monoclonal antibody; Ref. 19 ; (Fig. 1D) ⇓ .
With this in mind, we extended the initial findings by Lapointe et al. (17) and clearly demonstrated that CD40-B cells efficiently process and present antigen not only to memory but also to naïve CD4+ T cells and that it is possible to reliably expand antigen-specific T cells with CD40-B cells as antigen-presenting cells. These findings significantly widen the potential of CD40-B cells for immunotherapy.
- ©2004 American Association for Cancer Research.
Antigen-specific B lymphocytes have the capacity to uptake antigens with surface antibody and present epitopes to specific CD4+ T cells. The process involves cross-communication between B and T lymphocytes involving CD40L expression and cytokine secretion from T lymphocytes, which will then modify B-cell biology (1) . This is central and critical in the development of humoral responses. Recently, CD40-activated B cells were exploited as a source of antigen-presenting cells in the in vitro generation of CD8+ T lymphocytes specific to viral (2) and tumor antigens (3 4 5) . Several strategies were exploited to load antigens in CD40-activated B cells, including peptide pulsing (3) , RNA transfection (6) , and retroviral transduction (2) . Interestingly, CD40-activated B cells were found to be comparable with dendritic cells in their capacity to raise antigen-specific CD8+ T cells (3 , 6) . In addition, it has been demonstrated that the culture of CD40-activated B cells is technically easier compared with that of dendritic cells, and requires only small numbers of peripheral blood mononuclear cells (3) .
We have reported recently in this journal that CD40-activated B cells also have the capacity to present exogenous tumor antigens to CD4+ T cells in an immunostimulatory fashion (7) . We appreciate the positive comments and additional data provided by von Bergwelt-Baildon et al. (8) supporting our claims. Indeed, they previously succeeded in generating antigen-specific naïve and memory CD8+ T cells against different antigens (5 , 8) . On several occasions, von Bergwelt-Baildon and collaborators generated specific T cells in a high proportion of individuals, using purified peptides or proteins as a source of antigen. This contrasts with our strategy of preparing a crude lysate from tumor cells. Although the use of lysates is advantageous in that multiple antigens, including undescribed antigens, may be represented, the concentration of any particular antigen is likely to be low, perhaps limiting the efficiency of T cell generation against a specific antigen compared with the use of purified proteins or peptides. In their Letter to the Editor, von Bergwelt-Baildon et al. (8) convincingly present data demonstrating the efficiency of CD40-activated B cells in the generation of naïve, specific CD4+ T cells against several foreign antigens with purified proteins.
Many tumor antigens that have been described are “self” antigens and may be more challenging to use, compared with foreign antigens, to generate specific immune responses because of tolerance mechanisms. To explore the ability of CD40-activated B cells pulsed with purified proteins to generate immune responses to self antigens, we performed studies with recombinant gp100 protein, an antigen expressed on melanoma cells as well as normal melanocytes. With peripheral blood mononuclear cells and CD40-activated B cells prepared from five different HLA-DRβ1*0701+ melanoma patients, several gp100-reactive T-cell lines were successfully generated from three patients (Table 1) . Interestingly, T-cell lines generated from two of the patients were specific to an HLA-DRβ1*0701 epitope (gp100170–190) that we had identified previously (9) . Also, T-cell lines potentially specific to new gp100 epitopes were raised ( Table 1 ; patient 1 T-cell line B, and patient 2).
Interestingly, von Bergwelt-Baildon et al. demonstrated the ability of these activated B cells to take-up and process exogenous antigens. We suggested that the observed down-regulation of HLA-DO is involved in the efficient antigen presentation by activated B cells (7) . This is in line with recent studies by different groups elegantly showing that germinal center B lymphocytes express very little HLA-DO as compared with naïve or memory populations (10 11 12) . The assumption that DO down-regulation plays a role in our in vitro systems will need to be addressed directly and for each model antigen (13) .
Considering all of the data we obtained thus far, and the additional evidence presented in the Letter to the Editor from von Bergwelt-Baildon et al. (8) , we believe CD40-activated B cells have significant potential as a source of antigen-presenting cells for the generation of CD4+ T cells specific to foreign or tumor antigens.