A Novel Tool Using Marrow Natural Killer (NK) Cell Percentages to Predict Collection Requirements of Peripheral Blood Mononuclear Cell (MNC), Apheresis Products Used in NK Cell Adoptive Immunotherapy

Our institution is involved in two clinical trials that require a single infusion of non-expanded Natural Killer (NK) cells in recipients of haploidentical marrow transplants. NK cells are selected from non-mobilized Mononuclear cells (MNC), apheresis products collected 6 days after HPC, Marrow harvest then processed and infused the following day. The two-step processing procedure requires initial depletion of CD3⁺ T cells followed by enrichment of CD56⁺ NK cells. Results from 27 patients treated thus far show an average CD3^-CD56⁺ NK cell recovery of 54.5 ± 8.3% with 5.46 ± 0.74 log depletion of CD3⁺CD56^- T cells. The NK cell dose level target is now 5.0 x 10⁶ NK cells/kg and is limited to ≤1.0 x 10⁴ T cells/kg in these mixed pediatric and adult patients. For 21 of 26 patients (80.7%) whose adult donors underwent a standard 2-blood volume (BV) apheresis procedure, a dose level >5.0 x 10⁶ NK cells/kg was achievable. Processing a larger BV would likely improve the percentage achieving the desired target dose but would also increase the likelihood of adverse reactions, particularly citrate toxicity. It is believed that mature lymphocytes and monocytes in marrow harvested for clinical transplant come primarily from contaminating peripheral blood. Given this, we hypothesized that NK cell content in marrow MNCs would reflect NK cell content within the apheresis product and allow us to predict when a larger BV collection is needed. Immunophenotype data from 18 donors of both marrow and MNC, Apheresis products was evaluated to determine the correlation of the percentage of NK cells within the Marrow MNC population with that within the apheresis products. The % of NK cells within the MNC population was quite variable for marrow (6.1 ±3.7%, range 2.4 to 17.7%) and apheresis products (7.3± 4.6%, range 2.4 to 26.1%) but were highly correlated for individual donors (R²=0.85, p <0.0001). Based on this data, we developed a prediction workform tool requiring 3 data points from the marrow immunophenotype: 1) % of marrow lymphocytes (defined by flow cytometry as CD45^bright, CD14-), 2) the % of marrow monocytes (defined as CD45^bright, CD14+), and 3) the % of CD3-CD56+ NK cells within the marrow lymphocyte gate. These data were used to calculate NK cell content in the marrow MNC fraction. That data together with the desired NK cell dose/kg, patient weight, and average post-processing NK recovery (54.5%) was used to calculate the number of MNC cells required in the apheresis product to meet target NK cell doses. The BV to be processed was then estimated using the % MNC in the peripheral blood at marrow harvest and the known MNC collection efficiency of the device used (a conservative estimate of ~35% MNC collection efficiency was used). The prediction workform indicated that additional BV needed to be processed for one of the 3 donors for whom it was used indicated. This resulted in sufficient NK cells to meet the target infusion dose. Calcium infusion was used prophylactically to prevent citrate toxicity. Similar information could be obtained using a peripheral blood sample to determine NK cell content the day of collection. However, having information in advance from the marrow harvest allows for planning on the part of the donor and the collection team when additional collection time is needed. In conclusion, when available data from a marrow harvest performed prior to MNC, Apheresis is a useful tool to estimate the BVs needed to be processed meet infusion cell target dose.