Predicting yields in autologous MNC apheresis for ATMP manufacture: applying a kinetic model for HPC apheresis

Abstract

Background & Aim In cancer immunotherapy autologous MNC apheresis provides the source material for various advanced therapy medicinal products (ATMPs). Predicting cell yields would facilitate complex therapeutic regimens. Therefore an algorithm for CD34+ HPC yield in stem cell collection was applied but failed because of considerable cell recruitment during leukapheresis. Methods, Results & Conclusion Methods 41 patients were subjected to 46 autologous MNC leukapheresis procedures with a continous flow cell separator (Spectra Optia, V11.0, CMNC programme, TerumoBCT). Blood counts and quantification of lymphocyte subpopulations were done before and after leukapheresis as well as in the apheresis product on a hematology analyzer (Sysmex K21) and a flow cytometer (FACS Verse, BD). From the results and the apheresis key data cell yields, collection efficiency (CE2) and recruitment were calculated. Furthermore, relevant data were entered into a kinetic model validated for prediction of CD34+ HPC yields in HPC apheresis. Results The yields in 46 procedures averaged to vital TNC s 1.3*10E10, lymphocytes 6.8*10E9, CD3+ T cells 3.8 *10E9, NK cells (NKs) 7.2*10E8, platelets (PLT) 16.7*10E10, respectively. Though in 29/46 separations peripheral blood CD34+ HPCs were below 1/µl, they were detectable in the corresponding apheresis products containing on average 3.8*10E7 CD34+ HPCs. The collection efficiency (CE2) amounted to 0.64 for lymphocytes, 0.61 for CD3+ T cells, 0.74 for NKs. CE2 for CD3+ and NK cells were independent from their blood count. In the model validated for the prediction of CD34+ HPC yields described by Humpe et al. the calculated amounts overestimated PLT yields (3.0fold), but underscored the harvests of NK, CD3+ T cells and lymphocytes 0.5fold each. This is probably related to the recruitment which averaged 2.7 for lymphocytes, NK and CD3+T cells each and 1.1 for PLT. As described earlier in healthy donors, cell recruitment and loss were significantly correlated. Conclusion Yields in MNC leukapheresis cannot be predicted using a kinetic model validated for the prediction of CD34+ HPC collection. This is caused by a considerable, but highly variable recruitment of cells into the apheresis product representing a novel kind of liquid biopsy. In cancer immunotherapy autologous MNC apheresis provides the source material for various advanced therapy medicinal products (ATMPs). Predicting cell yields would facilitate complex therapeutic regimens. Therefore an algorithm for CD34+ HPC yield in stem cell collection was applied but failed because of considerable cell recruitment during leukapheresis. 41 patients were subjected to 46 autologous MNC leukapheresis procedures with a continous flow cell separator (Spectra Optia, V11.0, CMNC programme, TerumoBCT). Blood counts and quantification of lymphocyte subpopulations were done before and after leukapheresis as well as in the apheresis product on a hematology analyzer (Sysmex K21) and a flow cytometer (FACS Verse, BD). From the results and the apheresis key data cell yields, collection efficiency (CE2) and recruitment were calculated. Furthermore, relevant data were entered into a kinetic model validated for prediction of CD34+ HPC yields in HPC apheresis.