1.HLA mismatched alloresponses: Exploiting the GVL potential of the haploidentical donor: We previously showed that tissue specific T cell responses are conserved across HLA disparities between patient and donor. We have now further characterized the mismatched alloresponse in man. T-cell responses to allogeneic targets arise predominantly from the nave pool of post-thymic T cells. However, in man the risk of GVHD is increased if the donor has been exposed to multiple DNA viruses such as CMV, EBV and varicella zoster virus, suggesting that memory T cells also contribute to the alloresponse. We studied the origin of the effector cell in HLA mismatched alloresponses by sorting T cell populations into memory (CD45RA-) and nave (CD45RA+) subsets before challenge with allogeneic PBMC or cell lines. We showed that both nave and memory CD4 and CD8 T cell subsets were alloreactive. We hypothesized that a significant component of the memory alloresponse is derived from cross reactivity against foreign MHC by T cells recognizing DNA viral antigens (notably CMV, EBV) which form a large component of the peripheral T cell repertoire. By sorting viral-activated T cells and screening them for IFNγrelease against a random panel of 40 T-APC from healthy individuals, we showed that there was considerable overlap between T cells with viral specificity and random mismatched individuals. These findings suggest that some mismatched alloresponses are derived from virus-specific T cells (Figure 1). The implication is that GVHD reactions could be caused by viral specific T cell lines. However in collaboration with Dr Catherine Bollards group (Baylor College of Medicine) who used HLA mismatched viral specific CTL lines to treat viral reactivation post SCT, we found that, despite the opportunity for cross-reactive (promiscuous) T cell reactions, infusions of third party virus-specific T cells did not result in GVHD. The off-target reactivity of virus-specific T cell lines raises the question of whether such reactivity might also include functional GVL reactions. This possibility was suggested when a patient with relapsed acute lymphoblastic leukemia (ALL) receiving trivirus specific CTL to treat viral reactivation after SCT entered a complete remission following T cell infusion. We will screen T cell lines used in therapy of viral disease against panels of leukemia cells to identify off-target reactivity of potential therapeutic utility. 2. Generation of high affinity leukemia-specific T cells using HLA mismatched TSA presenting cells. We next addressed the question of whether defined tumor antigens can elicit tumor specific T-cell responses when presented by HLA-mismatched APC. We hypothesized that such individual clones of such cross-reactive T cells would show a spectrum of cytotoxicities, some more against the tumor peptide and some more against the mismatched HLA molecule. We generated WT1-specific CD8+ T cell clones directed against the HLA-A2-binding WT1126-134 peptide from HLA-A2 positive (allo-HLA restricted) and negative (self-HLA restricted) individuals. Self-HLA restricted WT1-specific clones only recognized WT1 with low avidity. In contrast, allo-HLA-restricted clones showed high reactivity against HLA-A2 positive targets. Furthermore, most allo-HLA-restricted WT1 tetramer-binding T-cell clones displayed reactivity against unpulsed HLA-A2+ target cells, indicating promiscuous antigen reactivity. To characterize this promiscuous recognition, reactivity of the T cell clones against 400 randomly selected HLA-A2-binding peptides was investigated. The self-HLA restricted WT1-specific T cell clones only recognized the WT1 peptide. In contrast, the allo-HLA-restricted T cell clones recognized various HLA-A2-binding peptides. These results showed that allogeneic HLA-A2-restricted WT1-specific T cells isolated from mismatched donors may be more tumor-reactive than their autologous counterparts, but show promiscuity. While these results suggest that allo-HLA restricted purely tumor antigen specific T cells are rare, further studies are needed to determine whether such T cell lines have both antileukemia efficacy without off-target alloreactivity causing GVHD as was seen with the virus specific T cells. 3. Treatment of relapse: We evaluated the safety and efficacy of DLI from a haplo-identical donor to treat relapsed disease following HLA matched sibling SCT. Two patients relapsing with AML day 70 and 83, and one with ALL relapsing day 91 after SCT were enrolled. They were treated with fludarabine 25mg/m2 x 5 days and cytoxan 60mg/kg x 2 days, followed by a haploidentical DLI of 1 x 108 CD3+ T cells/kg. Within 12 hr of DLI all patients experienced a cytokine storm with a diffuse macular rash, mild transaminitis, and fever (>40C) without infection, resolving with 1-2 mg/Kg methylprednisolone. All patients and achieved a disease response (absence of blasts in blood and marrow), and developed marrow aplasia requiring transfusion support and a stem cell transplant from the haploidentical donor. Engraftment of all lineages occurred within 14 days of stem cell infusion. One patient died on day 18 post DLI from GVHD, one day 103 post DLI from a further relapse, and one from fungal infection day 64 after DLI. While this protocol demonstrated a potent GVL effect, there was a high mortality associated with marrow failure following DLI (graft-versus-marrow effect). The protocol has been amended to include a haploidentical stem cell boost 7 days after the DLI. 4. Leukemia vaccines: Using a combined WT1 and PR1 vaccine to treat patients with low disease burden, we demonstrated that a single dose of vaccine was safe and increased TSA specific T cells in 7/8 individuals. In a second trial, 8 patients received vaccine for a total of 6 doses in 16 weeks. Disappointingly, the initial response to vaccine was not sustained and the avidity of leukemia-specific T cells progressively fell, suggesting that the schedule using montanide adjuvant caused tolerance induction rather than boosting leukemia-specific CTL. Subsequent work has suggests that montanide may be a poor adjuvant. For this reason we plan to use imiquimod in conjunction with peptide vaccine as a more reliable adjuvant in future trials. In another study we evaluated the effect of WT1 vaccine given after cytoreductive chemotherapy to determine whether the homeostatic drive for lymphocyte recovery favored expansion of WT1 leukemia-specific T cells. While the vaccine did increase WT1 leukemia-specific T cells we also observed a rise in PR1 specific T cells (without use of PR1 vaccine) during the post chemotherapy period. We conclude that the post chemotherapy/transplant period favors expansion of many leukemia-specific T cells independently of vaccine boosting. A protocol involving vaccination of donors with PR1 and WT1 is under consideration by the FDA. SCT donors will receive a single dose of vaccine followed by a leukapheresis 10 days later to collect vaccine-boosted leukemia-specific T cells, before receiving G-CSF to mobilize CD34 cells for a T-cell depleted transplant. The boosted DLI will be transfused on day 90 post SCT and also used to treat relapsed leukemia post SCT. In a future protocol, recipients will receive vaccine boosted lymphocytes as a DLI. 5.Generating clinical grade Leukemia-specific T cells: We are studing expansion techniques for leukemia-specific T cells with peptide mixes of WT1, proteinase 3, PRAME, MAGE3, and aurora-A kinase.
The aim i s to develop a clinical grade system to generate off the shelf donor or third party tumor-specific T cells for infusion in SCT recipients to treat or prevent relapse. Preliminary data indicates that individual responses to the tumor peptides are variable, some responders recognizing a broad range of tumor antigens and other showing more restricted responses.
Showing the most recent 10 out of 37 publications
