Lymphodepletion Followed by Donor Lymphocyte Infusions (DLI) Causes Significantly More Acute Graft Versus Host Disease Than DLI Alone.

Donor lymphocyte infusions (DLI) can produce lasting remissions in patients with relapsed CML after allogeneic HCT, but are less effective in other non-CML diseases. We hypothesized that in vivo expansion of DLI may enhance their anti-leukemia effects. Mouse adoptive transfer experiments show that up-front lymphodepletion improves in vivo lymphocyte expansion by providing lymphoid space, eliminating host anti-donor immune reactivity and by decreasing competition for growth factors that promote lymphocyte expansion. In this clinical trial, lymphodepletion was achieved with IV cyclophosphamide (Cy) 50 mg/kg once on day −6 and fludarabine (Flu) 25 mg/m² for five consecutive days (−6 to −2), a regimen we have shown previously to result in an in vivo surge of IL-15. DLI was given 48 hours after the last Flu dose and consisted of mononuclear cells (adjusted to a T-cell dose of 1 x 10⁸/kg). Fifteen patients with relapsed non-CML disease received Cy/Flu/DLI as their first treatment of relapse between 2004 and 2006. CML patients who received the same cell dose (n=28) or non-CML patients (n=35) who received a higher dose consisting of 3 daily lymphapheresis products treated from 1993–2003 were used as controls. Since there was no difference in GVHD rates between CML and non-CML patients, they were grouped together as one control group. While most control patients did not have significant leukocytopenia and were treated as outpatients, the patients receiving Cy/Flu/DLI all became lymphopenic and neutropenic by the time DLI were infused and were treated in the hospital. Patients who received Cy/Flu/DLI developed significantly more overall (60% vs. 24%, P=0.01) and grade III-IV acute GVHD (47% vs. 14%, P= 0.01) compared to controls. The interval from DLI to grade III-IV GVHD was modestly shorter in Cy/Flu/DLI patients compared to controls (mean = 17 days vs. 34 days; P=.19). In Cy/Flu/DLI patients, blood lymphocytes were collected before and at 14 days, 28 days, and 2 months after DLI for immunophenotyping. Proliferating T-cells, as measured by expression of the Ki67 marker, were significantly increased 14 days after DLI (10.3±2.8%) compared to baseline pre-DLI (1.6±0.6%, P=0.012), and were already returning to baseline levels 28 days after DLI (3.5±1.2%, P=0.19). Despite aggressive therapy, 4 of the 7 patients who developed grade III-IV GVHD died from complications of GVHD, suggesting that Cy/Flu/DLI induces immune activation that is sufficiently potent to enhance toxicity. Therefore, a decreased DLI dose (half) is currently being used in subsequent patients. These data show “proof of concept” that DLI following lymphodepleting chemotherapy leads to in vivo lymphocyte expansion and results in more severe GVHD than when giving patients DLI alone. The ability of lymphodepletion to enhance the immune effects of DLI is promising if alloreactivity correlates with remission induction. This strategy warrants further study to understand if the increased severity of acute GVHD leads to more complete or prolonged disease control and whether the immune activation can be controlled with lower cell doses.