Enhancing Graft-vs-Tumor (GVT) Activity of Donor Leukocyte Infusions by Ex-Vivo Expansion and Activation with CD3/CD28 Co-Stimulation.

Donor leukocyte infusions (DLI) can induce a potent GVT effect for some patients who relapse after allogeneic stem cell transplantation (SCT). However, GVT effects are most pronounced for patients with early phase CML and less effective for acute leukemia or other hematologic malignancies. One hypothesis for disease resistance suggests that donor T cells are not appropriately activated in-vivo to induce an anti-tumor response. Ex-vivo co-stimulation of T cells via CD3 and CD28 might overcome disease-induced anergy and augment GVT activity. To explore the possibilities for this approach, we performed a phase I trial of DLI followed by escalating doses of ex-vivo co-stimulated donor T cells (activated, or aDLI) for patients with relapse after allogeneic SCT (excluding CP-CML). Activated donor T cells are produced by co-stimulation and expansion by exposure to Dynal-magnetic beads coated with anti-CD3 (OKT3) and anti-CD28. 18 patients (11 male) with a median age of 45 (range 12–57) were treated for relapse after matched sibling SCT for ALL (7), AML (4), NHL (2), CLL (1), CML-BC (1), myeloma (1), Hodgkin’s disease (1) and lymphoblastic lymphoma (LL, 1). The median time from SCT to relapse was 11.5 mo (2–90) and relapse to DLI was 6 wks (2–31). All patients received standard unstimulated DLI (median 1.5 x 10⁸ mononuclear cells/kg, range 0.9–3.5) followed 10 days later by ex-vivo co-stimulated donor T cells as aDLI. Patients with acute leukemia or BC-CML first received standard induction chemotherapy followed by DLI at the nadir. aDLI was dose escalated between every 3–4 patients from 1 x 10⁶ CD3+ cells/kg to 1 x 10⁸ CD3+ cells/kg in 5 dose levels. Ex-vivo co-stimulation resulted in a median 130-fold expansion of CD3+ cells (range 32–526) consisting of 56% CD4+ cells (range 30–84%) and 34% CD8+ cells (range 5–71%). The infusion of activated donor T cells was well tolerated with minor infusional toxicity at the highest dose level. Following aDLI, 7 patients developed acute GVHD (5 grade I-II skin only, and 2 grade III skin and liver). 2 patients have chronic GVHD limited to the oral mucosa and 2 have extensive chronic GVHD. No patient died of complications related to GVHD. 6/16 evaluable patients achieved CR (3/5 ALL, 1/4 AML, 1/1 CLL and 1/2 NHL) and 5 of 6 remain alive in CR a median of 25 mo after aDLI (range 8–43). 1 patient with ALL in CR relapsed 8 months after aDLI despite prior grade II aGVHD and died of disease progression. 1 patient with APML has residual disease detectable only by PCR 5 wks after aDLI and 4 other patients remain alive with disease (AML, ALL, HD, NHL) a median 4 mo after aDLI (range 3–5). 7 patients died of disease (3 ALL, CML-BC, MM, AML, LL) a median of 4 mo after aDLI (range 1–14). Functional assays show that immunity against pathogens such as EBV and CMV are retained after aDLI. These data demonstrate that adoptive transfer of co-stimulated allogeneic T cells is feasible, can induce a durable CR in a subset of patients, and does not result in excessive GVHD or other toxicity. In diseases where conventional DLI has been disappointing, response rates are impressive. Further studies to enhance GVT activity and tumor-specificity of aDLI are therefore warranted.