Cell therapies offer potential treatments for currently untreatable malignancies. Approaches such as chimeric antigen receptor T (CAR T) cell therapies are being evaluated in numerous clinical trials with a number of them emerging as approved therapies, particularly in hematological tumors. Due to the growing numbers of these therapies and the complex workflow required to produce them, additional tools are needed to produce a consistent product in a closed, automated process. The Sefia Select™ system is a functionally closed instrument that was developed to enable automation of different processing steps within the CAR T cell workflow. MagnetSelect is an expanded offering for the Sefia Select™ system to enable magnetic T cell isolation to provide greater automation potential in the cell therapy process.
Introduction
CAR T cell-based therapies are acquiring an expanded role in the treatment of patients with hematological malignancies. The use of CAR T cells is approved by the US FDA for relapsed and/or refractory B-cell lymphoma, mantle cell lymphoma, B acute lymphoblastic leukemia and for relapsed and/or refractory multiple myeloma. Furthermore, a continuously increasing number of CAR T cell clinical trials are reported every year (1). As the number of CAR T cell therapies increase, the need for automation of the different phases of CAR T cell manufacturing remains one of the main objectives to decrease time and cost of production of these kinds of therapies and to improve their safety by reducing the risk of contaminations, errors, and batch failures.
The autologous CAR T cell manufacturing process is composed of different steps starting with a leukapheresis unit collected from patients (Fig 1). Frozen leukapheresis units are thawed, diluted, and washed before moving to the next step. The cell isolation step can correspond to peripheral blood mononuclear cells (PBMC) enrichment or to T cell selection via magnetic beads. After cell isolation, selective activation, viral-based transduction, and large-scale bioreactor expansion are performed to reach the appropriate cell amount. The cells are harvested by concentrating, washing, and diluting the cells with different solutions including cryoprotectant and finally split into several doses. CAR T doses are cryopreserved to be shipped and injected into the patient.
Fig 1. An example of an autologous CAR T workflow.
In this study, we focus on the first part of the CAR T cell workflow, the initial washing of the leukapheresis followed by the T cell isolation step, achieved using the MagnetSelect application. Here we show the results obtained using the MagnetSelect application in combination with the Sefia™ S-2000 cell processing instrument and the Sefia Select™ module as an automated solution for the processing of different types of input products.
Results and discussion
Donor data
Table 1 summarizes information from all donors used for generation of the performance data. A sample was taken from each donor leukopak to evaluate white blood cell (WBC) and T cell content before further processing.
Table 1. Summary of donor information
|
Fresh healthy donor leukapheresis |
Frozen healthy donor leukapheresis |
Patient model input |
|
| Biological replicate (number of donors) | 20 | 16 | 12 |
| Technical replicate (number of replicates) | 1 | 4 | 7 |
| Male vs Female | 20 Males 0 Female |
6 Males 10 Female |
8 Males 4 Female |
| Range of age (years) | 18 to 46 | 19 to 54 | 29 to 55 |
| Minimum WBC content | 8.03 × 109 | 2.32 × 109 | 20.1 × 109 |
| Maximum WBC content | 19.2 × 109 | 7.67 × 109 | 37.6 × 109 |
| Average volume | 147.7 mL | 65.1 mL | 113.7 mL |
| Minimum T cell | 24.2% | 40.7% | 1.8% |
| Maximum T cell | 69.4% | 74.3% | 2.9% |
| Average cell viability | 99.5% | 98.1% | 89.3% |
Fig 2. Phenotypic composition of fresh, frozen, and patient model input material.
Evaluating MagnetSelect performance
T cell purity, recovery, and viability drop
MagnetSelect is an automated solution to isolate T cells from either fresh or frozen input blood products. The efficiency of the process was assessed first by evaluating the T cell purity, and the T cell recovery of the final product.
When starting with fresh healthy donor leukapheresis as input product, the average T cell purity obtained in the final product was 91.2% ± 4.1% with an average 76.9% ± 5.6% T cell recovery (Fig 3 and 4; n = 21). For frozen healthy donor leukapheresis, an average T cell purity of 92.1% ± 2.2% and an average T cell recovery of 63.1% ± 9.8% were obtained at the end of MagnetSelect procedure (Fig 3 and 4; n = 20). Finally, for patient model input material, final T cell purity was 89.6 ± 2.8% with a T cell recovery average of 73.6% ± 7.1% (Fig 3 and 4; n = 19).
Fig 3. CD3+ purity results for MagnetSelect.
Fig 4. CD3+ recovery results for MagnetSelect.
Fig 5. Viability drop results for MagnetSelect.
Importantly, no significant viability drop was observed between the initial and the final cellular products for any of the inputs (Fig 5). For the fresh patient model the average cell viability drop was negative (-4.49%) because the input material contained a large quantity of non-CD3+ cells (> 90%) with a lower viability to CD3+ cells that were present in a very low quantity (< 3%). Whereas, the final product contained more than 89% CD3+ cells (Fig 3).
The ability of MagnetSelect to deplete platelet and red blood cell content was also evaluated as described in Table 2. For the 21 procedures executed with fresh leukapheresis as input product, the average platelet (PLT) depletion in the final product was 99.5% ± 0.5% and the average red blood cell (RBC) depletion was 94% ± 2.3%.
Final volume accuracy
The volume accuracy of MagnetSelect final product was also evaluated. Several values were compared, including:
- The final product volume setup in the parameters (targeted) = “Expected Volume”
- The final product volume measured by the user (See Table 4 for details on method) = “Measured Volume”
- The final product volume reported on the PDF report = “Reported Volume”
Figure 6 below shows for all three scenarios (total = 60 procedures) the differences between:
- The measured and the expected volumes in blue
- The reported and the measured volumes in orange
Fig 6. Final volume accuracy obtained for all the 60 procedures (including 21 procedures with fresh healthy donor input material; 20 procedure with frozen healthy donor input material, and 19 procedures with patient model as input material) between the desired and measured volume (blue) and the measured and reported final volumes (orange). All errors have been calculated in mL.
Table 2 summarizes the results described above.
Table 2. Table summarizing the performance obtained from MagnetSelect for each input type
|
Fresh input |
Frozen input |
Patient model input |
|
| T cell purity | 90.3% ± 4.6% | 91.5% ± 2.5% | 89% ± 2.5% |
| T cell recovery | 75.9% ± 4.8% | 62.5% ± 10.4% | 72.2% ± 8% |
| Viability drop | 1.1% ± 0.9% | 2.5% ± 2.6% | -4.5% ± 4.2% |
|
Final volume accuracy: Measured vs Expected Measured vs PDF |
1.9 mL ± 1.1 mL -0.1 mL ± 1.0 mL |
1.6 mL ± 1.4 mL -0.3 mL ± 0.6 mL |
1.3 mL ± 1.4 mL -0.2 mL ± 1.2 mL |
| RBC depletion | 94% ± 2.3% | N/A | N/A |
| PLT depletion | 99.5% ± 0.5% | N/A | N/A |
| Run time (h:min) | 02:25 ± 00:03 | 03:08 ± 00:16 | 03:24 ± 00:09 |
One of the challenges of obtaining CAR T cell products of consistent quality is linked to the input product heterogenicity. Not only can they be either fresh or frozen, but their cellular composition can be drastically different as it is not only disease dependent but also patient dependent.
The flexibility of the parameter settings for MagnetSelect, allow for the application parameters to be adapted to the characteristic of the input products. This allows MagnetSelect to accommodate a variety of input products while maintaining reliable and repeatable performance.
Serial capture (Y/N) parameter
This parameter is critical to ensure a good CD3+ cell purity and is dependent on the CD3+ cell composition in the input product. It is strongly recommended to activate the serial batch capture when the input material has a low input CD3+ cells content (we have tested between 1.5% and up to 10% CD3+ cells in the input product) in order to guarantee an efficient washout of unwanted cells and a high purity of CD3+ cells (≥ 80%) in the final product (Fig 7).
When the input product contains higher CD3+ cell content (we have tested between 20% and up to 70%) it is recommended to use a single batch capture to reach a CD3+ cell purity ≥ 80% in the final product without impacting the CD3+ cell recovery (Fig 7).
Fig 7. The CD3+ cell composition in input product, post single capture, and post serial capture for two different scenarios based on observational data. The first scenario (yellow) represents an input product with an average of 40% CD3+ cells, while the second (blue) the input product contains an average of 5% CD3+ cells. The dots represent the purity (that is, repartition) of CD3+ cells in the product and the stars represent the recovery of CD3+ cells at each step of the workflow.
Other parameters
Several other parameters can also influence performance (such as, target cell purity and target cells viability) and must be chosen based on input product cell composition as well as performance expectations. Table 3 gives an idea of the influence that particular parameters can have on the final product performance.
Table 3. Summary of the MagnetSelect parameters that can influence the final product performance
|
CD3+ cells recovery |
CD3+ cells purity |
Cell viability |
|
| Intermediate volume (and associated parameters, such, as g-force and time) |
Yes | No | Yes |
| Isolation capture stage | Yes | Yes | Yes |
| Isolation capture flow rate | Yes | Yes | Yes |
| Isolation rinse volume | Yes | Yes | No |
| Isolation rinse flow rate | Yes | Yes | Yes |
Conclusion
In conclusion, we have shown:
- MagnetSelect application combined with the Sefia Select™ system is capable of performing a standard T cell isolation step of a CAR T manufacturing workflow.
- In addition, the MagnetSelect application combined with the Sefia Select™ system offers an automated and closed system capable of isolating T cells with high purity and recovery from fresh and frozen input products without being influenced by patient heterogeneity.
- De Marco RC, Monzo HJ, Ojala PM. CAR T Cell Therapy: A versatile Living Drug. Int J Mol Sci. 2023 Mar 27;24(7):6300. doi: 10.3390/ijms24076300
Input material preparation
Three different types of input product were used: either fresh or frozen healthy donor leukapheresis and a cellular model (named “Patient Model”) created in the laboratory to mimic the cellular composition of B cell lymphoma patient material.
Fresh healthy donor leukapheresis
A total of 20 healthy donor leukapheresis samples were collected in an authorized collection center in Austria (Cytocare). The leukapheresis collection was performed using the Spectra Optia device and the Continuous Mononuclear Cell (CMNC) collection program (software version 11.3). Acid citrate dextrose solution A (ACD-A) was used as anticoagulant in the following proportions: 1 volume of ACD-A solution for 12 volumes of blood. Post-collection, leukapheresis units were maintained at 4°C and shipped on the same day in refrigerated conditions. Units were processed within 48 hours (h) from collection.
Frozen healthy donor leukapheresis
A total of 10× leukapheresis were collected in two authorized collection centers in the United States (StemCell Technologies and Charles River Laboratories Cell Solutions, Inc.). Seven leukapheresis collections were performed using the Spectra Optia device and ACD-A solution as anticoagulant. Three leukapheresis collections were performed using continuous flow centrifugal technology also using ACD-A solution as anticoagulant. In both cases, following collection, leukapheresis was cryopreserved in CryoStor CS10 using a controlled-rate freezer. The collected leukapheresis units were maintained at -135°C or colder. Leukopaks were shipped with dry ice and stored in liquid nitrogen until use. The units were used within 1-year post-collection. On the day of the experiment, frozen leukopaks were thawed with the VIA Thaw™. Volume of the leukopak listed on the CoA plus an additional 10 mL was inputted into the VIA Thaw™ and selected prior to loading the leukopak.
Patient model
The third type of input product was prepared in the laboratory on the day of the procedure. Lymphocytes (T and B), as well as, NK cells and monocytes were mixed at controlled percentages to obtain a cellular product containing at least 23 × 109 total viable cells (TVC), within which ~2.5% were CD3+ cells.
The cell mixture was resuspended in isolation washing solution (CliniMACS PBS/EDTA + 0.5% human serum albumin) at a cell density between 250 and 300 × 106 cells/mL and loaded in the processing bag PB-100.1 (PB-100.1 is the input bag for that scenario – see table 4 for detailed parameters used).
For all input materials, a sample was drawn from the starting material to determine cell counts and exact cellular composition in terms of WBCs and T cells. Prior to further processing, the exact volume of the starting material was measured using a calibrated and tared scale.
T cell isolation using MagnetSelect application
Input products were processed using MagnetSelect application on the Sefia™ S-2000 cell processing instrument combined with the Sefia Select™ system.
MagnetSelect is a modular application composed of different optional phases, schematically represented in Figure 8.
Fig 8. Schematic representation listing MagnetSelect phases and the different steps within the magnetic isolation phase.
The initial dilution step is recommended when starting with frozen input material in order to quickly reduce the concentration of cryoprotectant present in the input bag. The pre-incubation steps allow the concentration and washing of the input product and are used to reduce the platelet content in the input product. During the incubation phase, the concentrated input product is incubated with magnetic nanobeads for a user defined time, temperature, and mixing speed. The post-incubation steps allow the elimination of excess magnetic beads that did not bind to any cell and to resuspend the cellular product in a specific volume. During the magnetic isolation phase, T cells bound to beads are first captured within a column inserted in a magnetic field and then eluted as a separate fraction once the magnetic field is removed. The cell isolation can be a single or a serial isolation (Fig 8). The final formulation phase allows the concentration of the eluted product into a user defined final volume.
MagnetSelect is designed to work in combination with the CT-400.1 disposable kit and the single use process bag accessory PB-100.1, both schematically represented in Figure 9.
Fig 9. A) Schematic of the CT-400.1 disposable kit. B) Schematic of the PB-100.1 process bag. C) Table listing the different reagents connected to the disposable kit on the day of the procedure.
On the day of the experiment, the CT-400.1 disposable kit was installed on the Sefia™ S-2000 cell processing system and on the Sefia Select™ isolation module following the application guidelines. Previously prepared buffers and media were welded as listed in Figure 9C: pre-incubation washing solution (CliniMACS PBS + 1mM EDTA), isolation washing solution (CliniMACS PBS/EDTA + 0.5% human serum albumin), and final resuspension solution (either isolation washing solution or media used for further cell culture). Parameters were set up within MagnetSelect application as listed in Table 4.
Table 4. MagnetSelect parameters for each type of input product
| Parameters | Fresh healthy donor leukapheresis |
Frozen healthy donor leukapheresis |
Patient model |
| Sampling prompts | Disable | Disable | Enable |
| Beads mixing | Enable | Enable | Enable |
| Beads thermal mixer temp. control | Disable | Enable | Disable |
| Beads thermal mixer temperature (°C) | N/A | 4 | N/A |
| Initial dilution | Disable | Enable | Disable |
| Dilution volume (mL) | N/A | Volume of input product (1:1 dilution) |
N/A |
| Dilution thermal mixer temp. control | N/A | Disable | N/A |
| Dilution thermal mixer temperature | N/A | N/A | N/A |
| Post-dilution mixing prompt | N/A | Enable | N/A |
| Dilution mixing time (min) | N/A | 0:00 | N/A |
| Dilution rate | N/A | 17 | N/A |
| Pre-incubation | Enable | Enable | Disable |
| Initial line priming with product | Enable | Disable | N/A |
| Initial bag rinsing | Enable | Enable | N/A |
| Initial bag rinsing mixing prompt | Disable | Enable | N/A |
| Initial bag rinsing volume (mL) | 25 | 25 | N/A |
| Centrifugation g-force pre-incubation (× g) |
85 | 85 | N/A |
| Centrifugation time pre-incubation (min) | 07:00 | 07:00 | N/A |
|
Intermediate volume |
1.3 mL for every 109 Total Viable Cells |
1.3 mL for every 109 Total Viable Cells |
N/A |
| Washing cycles pre-incubation | 1 | 1 | N/A |
| Incubation | Enable | Enable | Enable |
| Incubation bead solution volume | 1 µL of beads for every 1 × 106 CD3+ for CD4 and CD8 beads | 1 µL of beads for every 1 × 106 CD3+ for CD4 and CD8 beads | - |
| Incubation total volume | Target density of 6 ×107 CD3+ cells/mL + bead volume | Target density of 6 ×107 CD3+ cells/mL + bead volume | Input material volume + bead volume |
| Incubation time (min) | 15:00 | 15:00 | 30:00 |
| Incubation mixing | Enable | Enable | Enable |
| Incubation temperature control | Disable | Disable | Disable |
| Incubation temperature | N/A | N/A | N/A |
| Post-incubation | Enable | Enable | Enable |
| Centrifugation g-force post-incubation (× g) | 300 | 300 | 300 |
| Centrifugation time post-incubation (min) | 06:00 | 06:00 | 06:00 |
| Intermediate volume post-incubation | 1.3 mL for every 109 Total Viable Cells |
1.3 mL for every 109 Total Viable Cells |
1.3 mL for every 109 Total Viable Cells |
| Washing cycles post-incubation | 1 | 1 | 1 |
| Isolation volume (mL) | 100 | 100 | 100 |
| Isolation (1) capture stages | 1 | 1 | 2 |
| Isolation (1) capture flow rate (mL/min) | 12 | 12 | 21 |
| Isolation (1) rinse volume (mL) | 50 | 50 | 50 |
| Isolation (1) rinse flow rate (mL/min) | 12 | 12 | 40 |
| Serial capture | Disable | Disable | Enable |
| Isolation (2) capture flow rate (mL/min) | N/A | N/A | 12 |
| Isolation (2) rinse volume (mL) | N/A | N/A | 50 |
| Isolation (2) rinse flow rate (mL/min) | N/A | N/A | 12 |
| Centrifugation g-force final formulation (× g) | 300 | 300 | 300 |
| Centrifugation time final formulation (min) | 06:00 |
06:00 | 06:00 |
| Intermediate volume final formulation | 20 | 1.3 mL for every 109 Total Viable Cells |
20 |
| Final volume (mL) | 50 to 200 | 50 to 150 | 50 to 200 |
Initial dilution was only activated when starting with frozen leukapheresis as the input material, because of the presence of cryoprotectant within the product. The pre-incubation phase, which allows the reduction of the platelet content in the starting material, was disabled when using the patient model as input product, as the patient model did not contain any platelets or red blood cells to washout.
Parameters related to the incubation phase were also adapted depending on the type of input product used. In the case of healthy donor leukapheresis, the appropriate quantity of beads to be used for each procedure was calculated based on the leukapheresis cell content using the following proportion: 1 µL of each bead type for 1 × 106 CD3+ cells present in the leukapheresis, up to a maximum of 7.5 mL of each bead type (equivalent to an entire product vial). The incubation volume was calculated to obtain a CD3+ cell density of 6 × 107 cells/mL during the incubation phase and added to the total bead volume. When the patient model was used as starting material, a fixed volume of magnetic beads was used for each procedure. In this case, the incubation volume corresponded to the sum of the input product volume and of the total bead volume.
Finally, isolation phase parameters were setup according to the input material cell composition. The single batch capture was used for fresh and frozen input material whereas the serial capture was selected when the patient model was used as input product.
Methods used to assess application performance
Cell samples were taken from the initial and final bags for each experiment and subjected to different types of analysis, including cell counting and flow cytometric analysis, to evaluate MagnetSelect performances. Table 5 reports the list of the evaluated performances and a description of the test used to determine them.
Table 5. MagnetSelect performance parameters under evaluation and reporting details on the methods used to determine them
| Criteria to evaluate |
Test execution description |
|||
| Final product volume accuracy | All volume measurements were performed using a calibrated and tared scale. Final bag volume was measured by subtracting the weight of the filled and empty final bag. The volume was calculated based on a 1 g/mL density. Errors were assessed by comparing: • target versus measured final volume • measured versus reported final volume |
|||
| Cell viability drop | Appropriately sized samples were withdrawn from the input and final bags. Cell viability was measured using an automated cell counter (NucleoCounter NC-200). Cell viability drop was calculated by subtracting cell viability measured in the input and final bags. | |||
| T cell recovery | Appropriately sized samples were withdrawn from the input and final bags. Cell counts were measured using an automated cell counter (NucleoCounter NC-200) and CD3+ cell % were assessed by flow cytometer. CD3+ cell recovery was calculated by dividing the CD3+ cell count obtained in the final bag by the CD3+ cell count measured in the input bag. | |||
| T cell purity | An appropriately sized sample was withdrawn from the final bag post MagnetSelect procedure. CD3+ cell purity was assessed by flow cytometry. | |||
| RBC and PLT depletion | Appropriately sized samples were withdrawn from input and final bags. PLT and RBC counts were measured using cell analyzer (Sysmex) and depletion was calculated by dividing the PLT and RBC values obtained in the final bags by the PLT and RBC measured in the input bag. | |||
| Washout efficiency | Samples were withdrawn from the input bag and all final bags. We performed ELISA on both samples to assess washout efficiency of a specific solute. |
|||
| Procedure duration | Procedure start and end times were measured and procedure duration was calculated by subtraction. | |||
Cell counts and viability measurements
Cell concentration and viability were measured using the NucleoCounter NC-200 automated cell counter (ChemoMetec). For each sample, the exact volume of the input product was measured using a calibrated scale (1 g = 1 mL conversion was applied) and at least 3 dilutions were prepared using a diluent that was matched to the media of the input product. Samples were diluted to ensure the cell concentration value was within the linear range of the NucleoCounter (5 × 104 to 5 × 106 cells/mL). If the input product contained a cryoprotectant agent, dilution factor was not lower than 5. Diluted sample was carefully mixed by pipetting and then loaded into a Via1-Cassette (ChemoMetec). The cassette was then inserted in the NucleoCounter NC-200 for analysis. For each sample, the dilution factor was setup in the parameters in the NucleoCounter NC-200 window to obtain the correct cell concentration post sample processing. Once all triplicates were analyzed, the average cell concentration and cell viability were calculated. The total cell number was calculated by multiplying the average cell concentration by the volume of the product.
Flow cytometry
5 × 105 cells were washed with FACS buffer (either BD Stain Buffer or DPBS containing 1% HSA and 1 mM EDTA) and centrifuged at 400 × g for 5 min. Cells were resuspended with 25 µL blocking buffer per test, containing 10 µL human FcR Blocking Reagent (Miltenyi Biotec) in BD Stain Buffer, and incubated for 10 min at 4°C (Optional step). Cells were then stained for 20 min at 4°C with the following antibody cocktail: CD45-BV421 (HI30, BD), CD3-APC (SK7, BD or Biolegend), CD56-FITC (NCAM16.2, BD), CD14-PE (MφP9, BD), CD19-PE-Cy™7 (HIB19, BD), CD8-APC-Cy™7 (RPA-T8, BD) and CD4-PerCP-Cy™5.5 (SK3, BD). Cells were washed with FACS buffer, centrifuged at 400 × g for 5 min, and resuspended in FACS Buffer. The stained cells were analyzed on the CytoFLEX Flow Cytometer (Beckman Coulter) or a BD FACS Canto II (Becton Dickinson). A minimum of 20 000 single cell events were acquired. Within the PBMC (CD45+) population, monocytes (CD14+ cells), NK cells (CD3-CD56+ cells), T cells (CD3+CD56-), NKT cells (CD3+CD56+), and B cells (CD19+) were identified. Flow data were analyzed using FlowJo version 10 software.