A multifaceted program to encourage medical students' research.

Clinician-scientists are important members of a research community that has more opportunities than ever before to solve problems important to patients. Nevertheless, the number of physicians applying for and receiving grants from the National Institutes of Health (NIH) has dropped. Introducing medical students to research and relevant support mechanisms early in their education may help to reverse this trend. In 1995, the Mount Sinai School of Medicine created its Office of Student Research Opportunities (OSRO) to stimulate students to engage in research. It also appointed a new dean to direct the OSRO; the person who filled this new position was a senior faculty member involved in patient-oriented research. The OSRO advises students, identifies faculty who want to mentor students, sponsors the Distinction in Research program, organizes an annual research day, helps fund summer and full-time research, and has created an endowment to support student travel to national meetings. Between 1997 and 2000 the number of students who participated in the research day increased from 18 to 74, and the number of publications by the graduating classes increased from 34 to 58 between 1997 and 1999. Participants have presented both basic and clinical projects. The authors' experience has shown that medical students can be motivated to carry out research with appropriate encouragement from the administration and the faculty, something that may help to reverse a troubling national trend. Based upon these early successes, Mount Sinai is developing a novel five-year program to provide medical students with research training.

Surrogate markers of antitumor responses: in vitro activation of T cells by autologous tumor peptides.

The increasing ability to augment antitumor immunity in model systems has led to increased numbers of clinical trials. However, progress in detecting immune responses by patients against autologous tumors has been slow. Although a considerable number of tumor antigens, as well as peptides derived from them, and the MHC determinants together with which they are presented have been identified for melanoma, this is not so for the majority of solid tumors. Furthermore, tumor cells themselves are poor stimulators of immunity. Thus, approaches that do not depend upon defined antigens or using tumor cells as stimulators would be desirable. To attempt to measure immune responses in these situations, we tested whether total peptides, prepared from autologous tumor tissue, stimulated cytokine release by T cells. Peripheral blood mononuclear cells (PBMCs) were mixed with antigen-presenting cells (APCs), pulsed with tumor peptides, and tested in the ELISPOT assay for IFN-gamma secretion. Few spots were obtained when PBMCs were cultured with unpulsed APCs or in wells with peptide-pulsed APC alone. In contrast, a strong response was seen when PBMCs were cultured with APCs that had been pulsed with autologous total tumor peptides. This system should help to identify those immunotherapeutic approaches that induce responses against tumor cells in vivo. Because different cytokine profiles are associated with distinct arms of the immune response, testing in the ELISPOT assay may also help us understand the mechanisms responsible.

IFNgamma secretion following stimulation with total tumor peptides from autologous human tumors.

Several issues remain to be resolved before the efficacy of various approaches to elicit anti-tumor immunity in patients can be evaluated. First, in vitro assays able to detect responses by T cells primed in vivo are needed. Second, a source of tumor antigen to stimulate patients' lymphocytes in vitro is required. The ELISPOT assay is attractive, because it can be performed with a small numbers of cells and requires only short-term culture in vitro. A source of tumor antigen is more problematic, since for most tumors, tumor-associated antigens (TAA) have not been identified and/or cloned. In this report we demonstrate that autologous antigen-presenting cells (APC) pulsed with total tumor peptides from autologous tumor tissue can stimulate IFNgamma release by patients' lymphocytes in the ELISPOT assay. Thus, this approach should be considered for monitoring immune responses in clinical immunotherapy trials.

Resection of solid tumors reverses T cell defects and restores protective immunity.

We have previously reported that CTL were demonstrable early after inoculation of CMS5 fibrosarcoma cells, but that they disappeared within 3 wk. These mice were unable to reject a challenge with CMS5 tumor cells. Other studies demonstrated cell surface phenotype and signaling abnormalities of cells within the spleen. Since we assumed that such an environment would make it more difficult to elicit antitumor immune responses via immunotherapy, we asked whether resection of the tumor could reverse these abnormalities. Although early after tumor cell inoculation tumor resection leads to the development of immunity, the effect at late time points has not been studied critically. To test this, mice were inoculated s.c. with CMS5 cells and after 28 days the tumors were resected. We observed a gradual normalization of the cellular phenotype of the spleen. In particular, there was a decrease in the number of Mac1+/Gr1(high) cells and an increase in the number of CD3+ cells in the spleen within 24-48 h of tumor resection. By day 10, these values were normal. Levels of p56lck increased as well. The functional implications of these changes were illustrated by the reduced growth rate or the complete rejection of a challenge of tumor cells in the resected mice. Both CD4+ and CD8+ cells were involved in the restoration of tumor immunity. Our results suggested that tumor resection not only led to the reversal of immune suppression, but also unmasked a population of primed T cells able to mediate protective immunity.

Generation of anti-allogeneic cytotoxic T lymphocytes depends upon a signal, not delivered by tumor cells, that can be provided in trans in the presence of exogenous interleukin-2.

We have reported that, while immune responses were generated initially against CMS5 tumor cells, they were lost with time. In the present work, we asked whether the tumor cells contributed to this by delivering an inhibitory signal or whether a positive signal, delivered in the form of tumor-derived interleukin-2 (IL-2) or the presence of professional antigen-presenting cells (APC) would be sufficient to enable a sustained response. We observed that the presence of tumor cells did not inhibit the generation of anti-(class I) cytotoxic T lymphocytes (CTL), suggesting that the tumor cells were not directly suppressive. In addition, the lack of a response could not be attributed to insufficient levels of IL-2 alone, since even tumor cells that secreted IL-2 failed to stimulate anti-(class I) CTL. Finally, we observed that professional APC were necessary to deliver an essential signal, but only in the presence of IL-2-secreting tumor cells. Thus, CMS5 cells failed to provide essential signals necessary for CTL generation, but were not directly suppressive.

Antigen presentation by B7-nonprofessional APC does not prevent responses to solid tumor cells.

The growth of tumors in vivo often is associated with immune suppression. In this report we tested whether the expression of a known antigen by tumor cells would inhibit the development of antitumor responses when the antigen was subsequently expressed in an immunogenic form. For this, we expressed a well-characterized surrogate tumor antigen, the nucleoprotein (NP) of the PR8 virus, in solid tumor cells. Although the NP+ tumor cells were not rejected in vivo and stimulated nondetectable CTL response in vitro, T cells from these mice differentiated into CTL following i.p. inoculation of PR8 virus and their tumors regressed. The results suggested that prior presentation of tumor antigens by tumor cells does not necessary preclude a response if the peptide is subsequently presented appropriately.

Gene therapy with modified tumor cells enables T-cell activation by stimulating pathways required for signal transduction.

The expression of a variety of stimulatory molecules by tumor cells can lead to tumor rejection and the development of systemic immunity by T cells. The fact that some tumor cells naturally express such determinants leads to the hypothesis that progressive tumor growth may be a reflection of problems with the host immune system. To test this, we compared the signal-transducing ability of T cells from mice inoculated with parental tumors (PTB) with that of T cells from mice immunized with IL-2-secreting tumor cells (ITB). Our results demonstrated that following T-cell activation, higher total kinase activity was associated with the signal-transducing zeta chain in ITB mice compared with PTB mice. Western blotting following stimulation of T cells with parental or genetically engineered IL-2-secreting, B7+ tumor cells revealed increased protein tyrosine phosphorylation in lysates derived from ITB compared with PTB T cells, demonstrating that tumor-derived IL-2 could influence signaling. Taken together, the findings are consistent with the hypothesis that tumor-derived IL-2 preserves the signal-transducing ability of immunocompetent T cells, but is ineffective when they are immunosuppressed. These results suggest that IL-2-secreting tumor cell vaccines might be useful as adjuvant therapy to prevent the outgrowth of micrometastases, following tumor resection, once immune function has normalized.

Tumour cell vaccines that secrete interleukin-2 (IL-2) and interferon gamma (IFN gamma) are recognised by T cells while resisting destruction by natural killer (NK) cells.

The inoculation into mice of genetically engineered tumour cells that secrete IL-2 or IFN gamma results in rejection, while unmodified parental tumour cells grow progressively. In vivo studies demonstrated synergy between IL-2 and IFN gamma leading to the rejection of the transduced tumour cells. IL-2 is required for T cell proliferation and differentiation. IFN gamma induced the upregulation of MHC class I molecules that present peptides to CD8+ T cells. Furthermore, IFN gamma can correct defects in antigen processing. Thus, for T cells, IL-2/IFN gamma-secreting double cytokine tumour cell vaccines might serve as class I+ peptide/antigen presenting depots for developing effector cells. In contrast to T cells, NK cells exert spontaneous killing and kill class I+ targets less well than those that are class I-. For this reason, they may actually have a detrimental effect by destroying a class I+ tumour cell vaccine before adequate T cell stimulation occurs. Based upon this rationale, we tested the hypothesis that an unrecognised benefit of increased class I expression by tumour cells in response to IFN gamma secretion would be to enable cytokine-secreting vaccine cells to resist destruction by NK cells. Our results demonstrated that T cells recognised tumour cells secreting IFN gamma better than those secreting IL-2. NK cells, in contrast, were inhibited by tumour cells that secreted IFN gamma, but not by those that secreted IL-2. The findings suggest that, in addition to upregulating adhesion molecules, MHC molecules, and correcting defects in antigen presentation pathways, IFN gamma secretion may protect tumour cell vaccines from early NK-mediated destruction, keeping them available for T cell priming.

Molecular basis of T cell dysfunction in cancer is influenced by the paracrine secretion of tumor-derived IL-2.

Previously we reported that T cells from mice bearing parental tumors have a variety of functional and biochemical defects when compared with T cells from mice bearing IL-2-secreting tumors. The biochemical defects included reduced levels of several cytoplasmic proteins such as p56lck, p59fyn, and zeta, all of which are known to be critical for signal transduction through the TCR. Based upon these results, we determined the consequences of these alterations on downstream signaling events. First, we showed that T cells of parental tumor-bearing mice have a reduction of total in vitro kinase activity associated with the TCR/CD3 compared with T cells from mice with IL-2-secreting tumors. Second, we observed that following activation, only T cells from IL-2-secreting tumor-bearing mice had completely phosphorylated CD3-associated zeta-chains and recruited ZAP-70 to cell surface-associated TCR. In contrast, T cells from mice with parental tumors contained incompletely phosphorylated CD3-associated zeta-chains with little or no TCR-associated ZAP-70. Third, the recruitment of ZAP-70 to the TCR/CD3 complex was seen only in animals with an increase in in vitro p56lck kinase activity after T cell activation. Finally, we report that these defects in T cell signaling were associated with generalized anergy in vivo. Our findings provide a molecular basis to explain T cell suppression of mice with parental tumors and offer a hypothesis to explain the protective effect of tumor-derived IL-2.

IL-2 gene therapy of solid tumors: an approach for the prevention of signal transduction defects in T cells.

The identification of tumor-associated antigens has focused attention on the mechanisms that underlie the failure of T cells to destroy tumor cells. A deeper understanding of the process of signal transduction following the binding of ligand by the T cell receptor can help to identify underlying defects that may be involved. Gene therapy using tumor cells genetically modified to express cytokines or surface determinants is a promising technique for stimulating antitumor responses. A potential pitfall in its application to cancer, however, is that some patients' T cells are immune suppressed and may resist stimulation by such genetically engineered vaccines. Recent studies have demonstrated that T cells from tumor-bearing patients exhibit abnormalities in signal transduction events, possibly rendering them unable to respond to activation signals. Gene therapy with interleukin 2 secreting tumor cells in an animal model has been shown effective in preventing the onset of signaling defects. A more precise definition of the molecular mechanisms that enable cytokine-secreting tumor cells to stimulate specific antitumor responses may make it feasible to optimize immunotherapeutic approaches resulting in better clinical results.

The impact of gene therapy on T cell function in cancer.

Studies of signal transduction by T cells are slowly identifying the intracellular messengers that must be generated for full T cell activation to take place. The recent, convincing identification of several tumor-associated antigens (TAA) has transformed our task into trying to define the mechanisms that underlie the failure of T cells to destroy antigenic tumor cells. Although there are a variety of hypotheses that explain why tumors grow progressively, even if they are antigenic, recent evidence suggests that T cells from tumor-bearing patients exhibit abnormalities in signal transduction that render them unable to respond to appropriate activation signals, even following proper stimulation. Gene therapy with interleukin-2 (IL-2)-secreting tumor cells in an animal model has been effective in preventing the onset of these signaling defects. Discovery of the molecular mechanisms by which such cytokine-secreting tumor cells induce immune responses and how they may best be applied clinically may provide clearer indications of the directions to pursue to alter the balance between the T cell and the tumor cell in the patient's favor.

B7-1 amplifies the response to interleukin-2-secreting tumor vaccines in vivo, but fails to induce a response by naive cells in vitro.

Parental and interleukin-2 (IL-2)-secreting CMS5 tumor cells were transfected with the B7-1 costimulatory molecule to amplify anti-tumor responses. CMS5 cells transfected with B7-1 grew more slowly in vivo than did parental CMS5 cells. Moreover, tumor cells secreting levels of IL-2 too low to cause rejection alone were rejected following transfection with B7-1. To determine whether the expression of B7-1 enabled the tumor cells to activate T cells directly, their ability to stimulate in vitro functional responses by T cells was examined. We found that neither B7-1+ nor IL-2-secreting, B7-1+ CMS5 cells stimulated naive spleen cells to proliferate or to become cytotoxic. In contrast, restimulation of primed T cells by B7-1+ CMS5 cells resulted in stronger cytotoxicity responses than seen following restimulation by parental CMS5 cells. Lysis was even higher if the B7-1+ tumor cells also secreted IL-2. Our results suggest that the expression of costimulatory molecules can augment responses generated by vaccinating with IL-2-secreting tumor cells. Furthermore, they are consistent with the hypothesis that the initiation of an anti-tumor response by naive T cells may depend upon initial antigen presentation by another unidentified cell and that the major action of IL-2-secreting and/or B7-1+ tumor cell vaccines might be to potentiate the response of already primed cells.

Abnormal signal transduction by T cells of mice with parental tumors is not seen in mice bearing IL-2-secreting tumors.

There is considerable evidence to demonstrate that immune function is abnormal in tumor-bearing mice, perhaps accounting, at least in part, for progressive tumor growth. In an attempt to generate an antitumor response, we used retroviral vectors to express IL-2 cDNA in CMS5, a murine fibrosarcoma. Mice inoculated with unmodified tumor cells suffered progressive tumor growth, whereas tumors secreting IL-2 were rejected or grew slowly. Animals bearing unmodified but not IL-2-secreting tumors also were immunosuppressed. On the basis of these observations, we were interested in how IL-2 secretion by the tumor cells prevented the onset of hyporesponsiveness. To identify biochemical differences between T cells of mice with parental vs slowly growing IL-2-secreting tumors, we examined signal transduction after activation through the CD3/TCR complex. Protein tyrosine phosphorylation was altered and calcium flux was reduced in cells of mice with parental tumors compared with animals with slowly growing IL-2-secreting tumors. In addition, levels of protein for the tyrosine kinases p56lck and p59fyn, as well as the TCR-zeta-chain, were reduced. These differences in signal transduction were observed for T cells of mice with parental and IL-2-secreting tumors of the same size, demonstrating that differences in tumor size alone could not explain our findings. Thus, IL-2 secretion by tumors seems to be able to prevent immunosuppression by maintaining normal signal transduction in T cells, facilitating the generation of antitumor responses.

Human tumor vaccines and genetic engineering of tumors with cytokine and histocompatibility genes to enhance immunogenicity.

Genetically engineered tumor cells can be used as vaccines in order to stimulate an immune response. To date, tumor cells have been modified in vitro so that they secrete cytokines or express histocompatibility molecules that they naturally fail to express. These tumor cells differ in the types of immune responses they induce and in whether the responses have local or systemic efficacy. Many questions have been raised during the past year, including whether allogeneic or autologous tumor cells should be employed and whether there may be a risk of inducing autoimmune disease along with the antitumor response. Nevertheless, because of the paucity of available therapies for patients with advanced cancer, investigators must attempt to refine the approaches used in order to minimize patient risk while maximizing tumor cell destruction.

Functional defects are associated with abnormal signal transduction in T cells of mice inoculated with parental but not IL-2 secreting tumor cells.

CMS5, a murine fibrosarcoma, was transduced with the IL-2 gene to stimulate the host antitumor response. Previously, we described that splenocytes isolated from parental tumor bearing (PTB) mice and IL-2-secreting tumor bearing (ITB) mice differed significantly in their proliferative responses when restimulated with IL-2-secreting tumor cells. In this report we extend these results by showing that the inability of PTB cells to proliferate when stimulated with parental CMS5 cells is not corrected by providing a source of costimulation. Furthermore, we demonstrate that T cells from PTB animals exhibit defective signaling through the T-cell receptor, defined by decreased protein tyrosine phosphorylation, that correlates with the impairment of functions attributed to both CD4+ and CD8+ T cells. In contrast, T cells from ITB animals are normal in this regard. The results suggest that immunosuppression underlies functional differences between PTB and ITB splenocytes and that defects in signal transduction may be responsible for the lack of normal functional responses.

Vaccination with IL-2-secreting tumor cells stimulates the generation of IL-2-responsive T cells and prevents the development of unresponsiveness.

Infection of CMS5 tumor cells with retroviral constructs containing interleukin-2 (IL-2) cDNA and selection in medium supplemented with G418 resulted in the isolation of clones which secreted IL-2. Whereas injection of parental tumor cells resulted in progressive tumor growth, tumor cells secreting high levels of IL-2 were rejected. Furthermore, in animals vaccinated with IL-2-secreting cells, the immunosuppression associated with the inoculation of parental tumor cells did not develop, and these animals resisted a challenge with viable tumor cells. To better understand the functional differences in the anti-tumor responses of immune and tumor-bearing mice which are at the basis for these diverse responses, we used an in vitro model to analyze interactions between splenic lymphocytes and tumor cells. Spleen cells isolated from either tumor-bearing or immune mice proliferated vigorously when cultured alone for 6 days, but much less in the presence of parental tumor cells. This effect could not be transferred with supernatant from tumor cell lines. Spleen cells from tumor-bearing mice remained unresponsive, while those from immune mice proliferated well in response to IL-2-secreting tumor cells. Only spleen cells from immune animals were able to develop cytotoxicity against CMS5 cells following in vitro restimulation. These results are consistent with the interpretation that exposure to parental tumor cells inhibited cell-mediated anti-tumor responses by a mechanism that involved cell-to-cell contact.(ABSTRACT TRUNCATED AT 250 WORDS)

Tumor-bearing animals contain suppressed antitumor effectors the function of which can be unmasked by IL-2.

CMS5 fibrosarcoma cells were infected with retroviral constructs containing interleukin-2 (IL-2) cDNA and selected in G418. Parental tumor cells and those that produced IL-2 were injected in vivo. Whereas injection of parental tumor cells resulted in progressive tumor growth, those secreting high levels of IL-2 were rejected. Furthermore, the immunosuppression associated with inoculation of parental tumor cells was not seen. To understand the failure of mice to reject non-IL-2-secreting tumor cells, functional responses of spleen cells from immune and tumor-bearing mice were studied in vitro. As expected, immune spleen cells proliferated under a variety of conditions but were inhibited in the presence of parental tumor cells. Even spleen cells from tumor-bearing animals responded well in the absence of parental tumor cells or in the presence of parental tumor cells, if supplied with adequate levels of IL-2. These results suggest that both tumor-bearing and immune mice generate antitumor effectors but that the cells might be functionally suppressed because of their inability to secrete IL-2 after contact with parental tumor cells.

Retroviral gene transfer induced constitutive expression of interleukin-2 or interferon-gamma in irradiated human melanoma cells.

Cytokines are important modulators of host antitumor responses. Two of these cytokines, interleukin-2 (IL-2) and interferon gamma (IFN-gamma), are produced after antigen-induced activation of helper lymphocytes. The cytokines are released into the immediate vicinity where they either interact with the appropriate receptors on effector cell populations or are rapidly degraded. To mimic this physiologic release of cytokines at the effector-target site, we used retroviral vectors to transduce melanoma cells with the IL-2 or IFN-gamma cDNA. Five melanoma cell lines were transduced with IL-2- or IFN-gamma-containing vectors and secreted IL-2 at 1 to 40 U/mL/10(6) cells/24 h or IFN-gamma 1 to 8 U/mL/10(6) cells/24 h, respectively. After gamma irradiation, these cells continued to secrete cytokines for about 3 to 4 weeks. Secretion of IFN-gamma induced upregulation of major histocompatibility complex class I molecules in a subset of melanoma cell lines. IL-2 production by human melanoma xenografts induced tumor rejection in BALB/c nu/nu mice, showing the in vivo effect of this cytokine. This study shows that (1) human melanoma cells can be stably transduced with cytokine-containing retroviral vectors; (2) cytokines are secreted constitutively by the transduced tumor cells and have the expected biologic effects in vitro and in vivo; and (3) after gamma irradiation, cytokines continue to be secreted for several weeks. These data suggest that irradiated cytokine-secreting allogenic or autologous tumor cells can be used in vaccination protocols for cancer patients.