Manufacture of Autologous CD34+ Selected Grafts in the NIAID-Sponsored HALT-MS and SCOT Multicenter Clinical Trials for Autoimmune Diseases

Abstract

To ensure comparable grafts for autologous hematopoietic cell transplantation (HCT) in the National Institute of Allergy and Infectious Diseases-sponsored Investigational New Drug protocols for multiple sclerosis (HALT-MS) and systemic sclerosis (SCOT), a Drug Master File approach to control manufacture was implemented, including a common Master Production Batch Record and site-specific standard operating procedures with "Critical Elements." We assessed comparability of flow cytometry and controlled rate cryopreservation among sites and stability of cryopreserved grafts using hematopoietic progenitor cells (HPCs) from healthy donors. Hematopoietic Progenitor Cells, Apheresis-CD34+ Enriched, for Autologous Use (Auto-CD34⁺HPC) graft specifications included ≥70% viable CD34⁺ cells before cryopreservation. For the 2 protocols, 110 apheresis collections were performed; 121 lots of Auto-CD34⁺HPC were cryopreserved, and 107 of these (88.4%) met release criteria. Grafts were infused at a median of 25 days (range, 17 to 68) post-apheresis for HALT-MS (n = 24), and 25 days (range, 14 to 78) for SCOT (n = 33). Subjects received precryopreservation doses of a median 5.1 × 10⁶ viable CD34⁺ cells/kg (range, 3.9 to 12.8) for HALT-MS and 5.6 × 10⁶ viable CD34⁺ cells/kg (range, 2.6 to 10.2) for SCOT. Recovery of granulocytes occurred at a median of 11 days (range, 9 to 15) post-HCT for HALT-MS and 10 days (range, 8 to 12) for SCOT, independent of CD34⁺ cell dose. Subjects received their last platelet transfusion at a median of 9 days (range, 6 to 16) for HALT-MS and 8 days (range, 6 to 23) for SCOT; higher CD34⁺/kg doses were associated with faster platelet recovery. Stability testing of cryopreserved healthy donor CD34⁺ HPCs over 6 months of vapor phase liquid nitrogen storage demonstrated consistent 69% to 73% recovery of viable CD34⁺ cells. Manufacturing of Auto-CD34⁺HPC for the HALT-MS and SCOT protocols was comparable across all sites and supportive for timely recovery of granulocytes and platelets.

Keywords: Autologous hematopoietic cell transplantation; CD34 selection; Clinical trial; Drug master file; Graft processing; Multiple sclerosis; Systemic sclerosis (scleroderma).

Figure 1

Comparability of flow cytometry techniques at sites. The CD34⁺ cell depleted flow-through after Isolex, from a PBSC collection from a normal volunteer, was distributed into multiple aliquots in 15-mL centrifuge tubes (approximately 2 billion cells per sample, at a cell concentration of 1 to 2 × 108/mL). Site 01, where the samples originated, shipped tubes overnight for early morning receipt to 7 other sites and shipped a tube to itself to serve as an intralaboratory control (listed as 01). In addition to the 8 shipments, 1 study tube was retained in a refrigerator at the site of origin to serve as a control to assess effect of transport on product flow cytometry characteristics (listed as 01Ref). TNC measurement and flow cytometry analysis were performed at sites. (A) Percentage TNC Recovery was calculated by dividing the TNC result at each individual site by the TNC result for the reference tube stored at the site of origin (01Ref). (B–D) Percentages were calculated based on TNC (viable + nonviable) determined at the given site. In C, note that CD3⁺ data at site 07 was not available. Site 01, FHCRC; Site 02, City of Hope; Site 03, Duke University; Site 04, Medical College of Wisconsin; Site 05, MD Anderson Cancer Center; Site 06, Ohio State University; Site 07, Dana Farber Cancer Institute; Site 08, University of Michigan.

Figure 2

Comparability of controlled rate cryopreservation at sites. The CD34-depleted flow-through after Isolex, from a PBSC collection from a normal volunteer, was distributed into multiple aliquots in 15-mL centrifuge tubes and shipped from Site 01 to the other sites and to Site 01 as described in Figure 1. Sites then processed the samples for cryopreservation in Cryocyte bags. The Cryocyte bags were subjected to controlled rate freezing at the sites, stored for 1 week, and then shipped back to Site 01 where they were thawed and measured for TNC and analyzed by flow cytometry. (A) Percent TNC Recovery was calculated by dividing the TNC result from each individual site by the TNC result from the thawed product at Site 01. (B–D) Percentages were calculated using TNC determined at the given site. Sites are as in Figure 1.

Figure 3

Stability of normal donor Auto-CD34⁺ HPC with storage. CD34⁺ cell enrichment using Isolex was performed on PBSCs from a G-CSF mobilized normal volunteer, and the enriched CD34⁺ cells were aliquoted into 4 Cryocyte bags and subjected to controlled rate cryopreservation. Bags were thawed at 1 week and at 1, 3, and 6 months and measured for TNC and analyzed by flow cytometry. TNC recovery was calculated by dividing TNC at the respective time point by TNC in the initial product before cryopreservation. Percentage viable CD34⁺ cell recovery was calculated by dividing number of viable CD34⁺ cells at the respective time point by the number of viable CD34⁺ cells in the initial product before cryopreservation. Percentage viable CD3⁺ recovery was calculated similarly.

Figure 4

Auto-CD34⁺HPC composition by study and device. By Device: Median percentage of viable CD34⁺ cells constituted well above the 70% specification, with lots prepared on CliniMACS having a higher percentage on average as compared with Isolex (A). Lots prepared on Isolex had a higher percentage of residual CD3⁺ cells on average, compared with CliniMACS (B). Final product viability met specifications for all lots with no more than 15% dead cells in lots processed on either device (F). The percentage of B cells, natural killer cells, and monocyte content was also analyzed (C, D, and E). B cell (range, 0% to 2.1%) and monocyte (range, .1% to 6%) content in the lots processed on CliniMACS was uniformly low. B cell (range, 0% to 38.9%) and monocyte (range, .01% to 26.6%) contamination was greater for the lots processed on Isolex. B cell content for lots processed on Isolex was >10%, with 12 lots having B cell content > 20%. Of note, lots from a given patient tended to have similar levels of B cell contamination (data not shown). Although the median proportion of monocytes was the same for both devices (.34%), 5 lots processed with Isolex had ≥4% monocytes, with 1 as high as 26.6%. Lots from both devices had very low numbers of natural killer cells (<5.5%). By Study: The composition of lots prepared from patients on the HALT-MS and SCOT studies were very similar (A–F).

Figure 5

Days from start of collection to transplant by study. The number of days Auto-CD34⁺HPC were stored after apheresis collection, CD34⁺ selection, and cryopreservation, before infusion on the day of transplant, is illustrated for HALT-MS and for SCOT. The number of days was calculated from the start of collection to the day of transplant. Collection may have occurred over a series of days to ensure an adequate dose.

Figure 6

Viable dose per kg administered by study. Graft dose was calculated as viable cells administered/kg RBW. (A,B) Doses meeting or exceeding target CD34⁺ dose for each respective study are shaded in red (A). Mean doses are denoted with the filled diamond. To achieve these doses, subjects received a mean and median of 2 lots of Auto-CD34⁺HPC per subject (Table 1C). For HALT-MS, endpoint status was event-free survival (EFS), defined as survival without death or disease activity from any one of disability progression, relapse, or new lesions on magnetic resonance imaging (C,D). Participants were evaluated through 5 years post-transplant. For SCOT, endpoint status was EFS, defined as survival without death or organ failure (C,D). Organ failure included respiratory failure defined as decrease in diffusing capacity of the lung for carbon monoxide > 30% or forced vital capacity > 20% from baseline, renal failure defined as requiring dialysis or renal treatment, or cardiac failure defined as left ventricular ejection fraction < 30% or New York Class III. Participants were evaluated through 54 months post-transplant. All HALT-MS subjects had their products processed on the Isolex, whereas SCOT subjects had their products processed on either the CliniMACs or the Isolex device. Of the 8 subjects on SCOT who met endpoint, 1 (12.5%) had their graft prepared on CliniMACs and 7 (87.5%) had their graft prepared on Isolex. Similarly, of the 25 SCOT subjects who did not meet endpoint, 5 (20.0%) had their graft prepared on CliniMACs and 20 (80.0%) had their graft prepared on Isolex.

Figure 7

Engraftment kinetics by study. Neutrophil recovery: For HALT-MS, neutrophil engraftment was defined as absolute neutrophil count > 500/μL for 2 consecutive readings. For SCOT, neutrophil engraftment was defined as absolute neutrophil count >500/μL for 3 readings on consecutive days. Platelet recovery: Days to last platelet transfusion was calculated from date of transplant. We note recovery of platelet counts > 20,000/ μL could be expected to be about 7 days after the last platelet transfusion for HALT-MS or SCOT subjects. However, these data were not documented precisely for our studies, as some individuals were discharged from the transplant center and therefore did not receive daily platelet counts in this timeframe. (A and B) The mean is denoted with the filled diamond. (C and D) Pearson correlations are presented as a measure of association.