Background
Cellular immunotherapy for the treatment of cancer has undergone remarkable growth over the past few years. Much of the recent attention that has led to this increased development has been due to encouraging phase I/II clinical trials using CAR T-cells directed against CD19 and CD22 [
1‐
3] and TCR-modified T-cells directed against NY-ESO-1, MAGE, and GP100 [
4‐
7]. Both of these approaches have their own particular nuances. While TCR-transduced T-cells are able to recognize intracellular antigens processed by major histocompatibility (MHC) proteins, CAR T-cells only target surface antigens, but these can be recognized in a MHC-independent fashion, allowing more patients to be treated with a specific viral vector.
The human papillomavirus (HPV) E6 and E7 oncoproteins are required for the induction and maintenance of the malignant phenotype of HPV-associated cancers. These oncoproteins have been postulated to be ideal targets in patients with HPV-associated epithelial cancers in that they are constitutively expressed by HPV cancers and not expressed in healthy tissue [
8]. Currently, three separate vaccines for HPV-16 are approved for clinical use and have demonstrated excellent safety profiles (Gardasil
®, Gardasil
®9, and Cevarix
®). While these vaccines are highly effective in preventing new HPV infections, they are incapable of treating established HPV infections [
9,
10].
A method was developed to expand E6/E7-specific tumor infiltrating lymphocytes (TIL) to treat patients with HPV-positive metastatic cervical cancer [
11,
12]. In a recent clinical trial (NCT01585428), three out of nine patients experienced objective tumor responses, two of which were durable complete responses [
13]. Interestingly, it was later discovered that the two patients that developed a complete response (CR) not only had viral-specific T-cells present in their blood, but also, harbored T-cells with reactivity against mutated neoantigens [
14]. The positive outcome in this trial led to efforts to target HPV-positive cancers with T-cells that are genetically engineered with TCRs against HPV-16 E6 and E7 [
15]. Preclinical studies demonstrated that these engineered T-cells had high avidity, effector cytokine production and lytic ability against target cell lines, which paved the way for three separate clinical trials at the National Cancer Institute (NCI-NCT02280811, NCT03197025, NCT02858310).
A major limiting factor for the administration of TCR-transduced T-cells is the seemingly large number of cells required to obtain a clinical response. For example, the average number of T-cells infused in the 6 of the 20 patients that experienced a partial tumor response using T-cells transduced with a MART1-specific TCR was 21.5 × 10
9 cells (range of 3.8–48 × 10
9 cells) [
6]. In another study utilizing NY-ESO-1-specific TCR transduced T-cells, the average cell dose that achieved at least a partial response was 66.4 × 10
9 cells for patients with melanoma (range of 37–130 × 10
9 cells; n = 5/11) and 62 × 10
9 cells for patients with synovial cell carcinoma (range of 50–83 × 10
9 cells; n = 4/6) [
16]. Manufacturing procedures that were used to generate the large number of cells needed in these trials utilized 6-well plates for transduction and culture flasks for expansion, which puts the product at higher risk for microbial contamination. It is notable that we and other cell therapy manufacturing centers are moving more toward automated and closed system approaches to manufacture CAR T-cells in order to improve consistency and avoid potential microbial contamination [
17]. Unfortunately, the current yield of these fully automated and closed single systems ranges from ~ 1–5 × 10
9 cells, which is sufficient to treat patients enrolled on most CAR T-cell protocols for hematologic malignancies, but insufficient for many TCR T-cell protocols which can require up to 100 × 10
9 total cells. Here, we investigated the possibility of manufacturing large numbers of transduced T-cells in closed and semi-closed, modular systems with the hope that this will result in a more safe and reliable manufacturing procedure. This procedure was designed and performed using TCR-transduced T-cells, but is easily adapted for CAR T-cells and other T-cell cultures that may require viral transduction, making it an important advancement in GMP manufacturing of cellular immunotherapy.
Methods
Study participants
The patient samples studied were acquired with informed consent from patients enrolled on National Cancer Institute (NCI)(Bethesda, MD) Protocol Number 03-C-0277, which was approved by the NCI Institutional Review Board. Leukapheresis was performed by the Surgery Branch Leukapheresis Unit. Peripheral blood mononuclear cells (PBMC) were isolated from leukapheresis samples and cryopreserved by the Surgery Branch TIL Laboratory using standard operating procedures. Healthy donor samples were collected after obtaining informed consent by the Department of Transfusion Medicine, Clinical Center, NIH on protocol 99-CC-0168.
Construction and production of clinical-grade E6 and E7 TCR recombinant retroviral vector
The PG13-MSVG1-E7 TCR (B3) retroviral vector and PG13-MSGV1-oDCA2-E6 TCR retroviral vector was prepared and cryopreserved following cGMP conditions in the Surgery Branch Vector Production Facility (SBVPF), NCI, NIH.
Preparation of peripheral blood mononuclear cells
Healthy donor volunteers and patients who were enrolled into clinical protocols signed informed consent that were approved by the Institutional Review Board of the National Institute of Health. Peripheral Blood Mononuclear Cells (PBMCs) were collected by leukapheresis and were enriched for lymphocytes by automated ficoll density gradient separation using COBE2991 Cell Processor (TERUMO, Lakewood, CO). Following separation, PBMCs were either immediately cultured for transduction or cryopreserved for later culture/transduction. For feeder cells used in the Rapid Expansion Protocol (REP), PBMCs from healthy donors were cryopreserved after density gradient separation.
Generation of E6 and E7 TCR transduced T-cells
Fresh or frozen and thawed PBMCs were cultured with TCR-300 CM which contains AIM V medium (Gibco, Grand Island, NY), 5% heat-inactivated human AB Serum (Valley Biomedical, Winchester, VA), 2 mM GlutaMax (Gibco), and 300 IU/mL IL-2 (Prometheus Laboratory, Inc. San Diego, CA). The cells were stimulated with 50 ng/mL soluble anti-CD3 (Miltenyi Biotec, Auburn, CA) and placed in flasks for 48 h at 37°, 5% CO2. On day 2 post activation, the cells were transduced with E6 TCR or E7 TCR retroviral vectors by spinoculation and overnight incubation in a closed bag system. PermaLife bags (OriGen Biomedical, Austin, TX) were coated with RetroNectin (Takara Bio, Mountain View CA) overnight. The bags were then incubated with a blocking solution containing PBS with 2.5% HSA, and rinsed with a wash solution (WS) containing HBSS and 25 mM HEPES. Retroviral vector (15 mL) was added into each bag, and centrifuged at 2000g for 2 h at 32 °C. Viable cells (15 × 106) were added into each bag to a final concentration of 0.5 × 106/mL, and the bags were centrifuged at 1000g for 15 min at 32 °C. The bags containing the cell and viral suspension were placed in a 37 °C incubator overnight. The procedure was repeated on day 3 for the 2nd transduction. On day 4, the transduction was stopped and cells were diluted to 0.4 × 106 cells/mL with fresh TCR-300 CM. Cell were expanded until day 7–10. The transduced cells were harvested and cryopreserved or initiated fresh in the REP.
Rapid expansion protocol (REP) for transduced cells
REP was initiated with fresh or cryopreserved transduced cells. The transduced cells were cultured with irradiated (50 Gy) allogeneic PBMCs from three healthy donors as “feeder” cells at a ratio of 1 to 100. The cultures were initiated in closed, gas-permeable G-REX500MCS vessel (Wilson Wolf Manufacturing, New Brighton, MN). For each G-REX500MCS, 10 × 106 viable cells and 1 × 109 irradiated feeders were cultured in 800 mL of REP-3000-5 CM containing AIM-V medium, 2 mM GlutaMax, 3000 IU/mL IL-2 and 5% heat-inactivated human AB Serum, and supplemented with 30 ng/mL of anti-CD3. The vessels were incubated at 37 °C in 5% CO2. Four days after culture initiation, 800 mL of REP-3000-5 CM was added to each vessel to a final volume of 1600 mL. On day 7, additional 1200 mL of REP-3000-5 CM was added to each vessel. On day 11, REP-3000-0 CM was prepared, which contains AIM-V medium, 2 mM GlutaMax, and 3000 IU/mL IL-2. Thousand seven hundred milliliter of REP-3000-0 CM was added to each flask to a final volume of 4500 mL. The cells were harvested on day 14 of culture. At harvest, the supernatant of each G-REX500MCS vessel was removed by GatherREX (Wilson Wolf Manufacturing) to reduce volume of cell suspension for concentration and wash. The cell suspension was then concentrated and washed using the LOVO device (Fresenius Kabi, Lake Zurich, IL). The wash solution is plasmalyte-A (Baxter, Deerfield, IL) supplemented with 0.5% HSA (Baxter). After the washing procedure was complete, the cell product was supplemented with 4% HSA in plasmalyte-A.
Cell counts and flow cytometry
Cell counts were performed using the Advia 120 automated hematology analyzer (Siemens Healthcare, Erlangen, Germany) and Cellometer Auto 2000 (Nexcelom Bioscience, Lawrence, MA). Flow cytometry was performed with a FACSCanto II (BD Biosciences, San Jose, CA) using CD3, CD4, CD8, CD14, CD15, CD19, CD45 and CD56 antibodies (BD Biosciences). The expression of E6 TCR and E7 TCR was assessed by flow cytometry using antibodies that recognize murine components within the TCR construct (anti-mouse TCRβ).
Cytotoxicity assays
Killing activity was determined using the xCELLigence RTCA MP (Acea Bioscineces Inc., San Diego, CA) instrument and was calculated by measuring electrical impedance in the culture plates caused by the adhering target cell lines. Addition of the non-adhering TCR cells at a ratio of 1:1 (E:T) results in decreasing electrical impedance measured in the culture wells due to cell death and cytolysis of the target cells. Cytolytic activity was measured in percentage against wells that contain either only target cells or effector cells. Electrical impedance was calculated every 15 min.
Cytokine secretion assays
E6 or E7 TCR transduced T-cells were co-cultured with the target cell lines at a ratio of 1:1 for 24 h in 96-well plates. The plates were centrifuged and the supernatants removed and stored in − 30 °C. ELISA kit’s were purchased from RnD Systems (Minneapolis, MN) and were used according to manufacturer’s instructions. Briefly, Assay Diluent was mixed with supernatant sample (diluted 1:10 in advance) and incubated in room temperature for 2 h. Consecutively, the microplates were washed 4× and either TNF-α or IFN-γ conjugates were added and incubated for further 2 h. The microplates were washed and Substrate solution was added to the microplates and incubated for about 20 min in room temperature protected from light. Stop solution was added to each well to stop the reaction. On the same plates and for each assay, a standard curve was created according to manufacturer’s instructions from the provided Stock solution of known concentration. The microplates were read for Optical Density using an AccuSKAN FC Microplate photometer (Thermo Fisher Waltham, MA) at 450 nm with a wavelength correction at 540 nm. The optical density results were analyzed and plotted using Prism. All optical densities were corrected against the optical densities of control wells on the microplates. The standard curves created (with R2 values of 0.9911 and 0.9964 for IFN-γ and TNF-α respectively) were used to calculate the optical densities to actual concentrations. All samples were done in triplicate.
Graphs and statistical analysis
Graphs were generated using Graphpad Prism 7 Software. For cytotoxicity assays and cytokine secretion assays, P values were calculated using Students t test for comparing like groups. P < 0.05 was considered statistically significant and is illustrated with an asterisk (*).
Discussion
The current work highlights the use of closed and semi-closed modular systems for clinical scale manufacturing of E6- and E7-specific TCR transduced T-cells. Manufacturing methods to generate TCR-transduced T-cells at some cell therapy centers generally utilize open lymphocyte enrichment, open 6-well plates for transduction, or open tissue culture flasks for expansion. Here, these procedures have been replaced with semi-automated closed lymphocyte enrichment, closed tissue culture bags, and closed G-REX containers, considerably limiting the potential for microbial contamination and increasing the consistency of the cultures. We show that using these methods, we are able to generate a large number of highly functional T-cells capable of recognizing target tumor cells. The process is reproducible and robust, and the performance of patient samples appears similar to those of healthy donor samples. This study used a single G-REX culture vessel to obtain approximately 20 × 109 T-cells, but the cell yield can easily be increased further by using multiple vessels.
The manufacturing process was specifically designed so that it incorporated several distinct, but highly specialized modules so that it has a high degree of versatility and would be able to be performed at most standard cell processing facilities. First, automated enrichment of lymphocytes was achieved through the use of closed density gradient separation using the cobe2991 cell processor (Terumo, BCT). Next, after T-cells are activated and monocytes are removed, cells are transduced with retroviral vector and expanded for a period 7 days in closed culture bags. Part II of the manufacturing process incorporates a REP where transduced T-cells are stimulated with allogeneic irradiated feeder cells and anti-CD3 and allowed to further expand for 14 days within a closed GREX500. Upon cell harvest of one GREX500 using an automated closed GatheRex device, cells are in the range of 40 × 10
6 cells/mL in a total of 500 mL. These cells are transferred to a LOVO cell wash/concentrate instrument (Fresenius Kabi) where they are washed with plasmalyte A containing HSA and resuspended at a final concentration of 10–330 × 10
6 cells/mL for infusion (depending on dose level). One of the strengths of this manufacturing schema lies within its high degree of flexibility. This procedure would allow for the removal/insertion of other elements to easily fit the particular needs of multiple protocols, and we predict that this method will be effective for manufacturing other T-cell therapies. We previously found that G-REX flasks can be used for REP of tumor infiltrating lymphocytes (TIL), however, that method involved the use of smaller open system G-REX vessels and the open transfer of TIL at the time of harvest from the G-REX vessels [
12]. The method described here involves the use of G-REX flasks with fivefold greater volume and when these larger vessels are used, media and cells can be added and removed without opening the system. Earlier renditions of this protocol and other protocols similar to this [
20], which yielded ~ 20–30 × 10
9 cells, required up to 40 T-175 flasks during the first 7 days of the REP. Each of these flasks required opening for media exchange at least once before they were finally transferred to cell culture bags for further expansion. The closing of the vessels and the reduction in total vessel number not only significantly reduce the volume of media and supplements required for manufacturing process, but also limit the chances of microbial contamination, and thus reduce the potential impact on patient well-being, cost and convenience. This system could also be used for manufacturing CAR T-cells, however, most CAR T-cell protocols do not require REP since a relatively small quantity of transduced T-cells are administered.
We find that this process results in a product that is very consistent. Although we observed slight differences in cell growth and transduction efficiency between E6 and E7 cultures, these same differences were not observed among the patient samples versus healthy donor within each of these groups and are most likely the result of the retroviral vector used during the manufacturing process.
Upon examining the kinetics of cell growth when using this procedure, we find that CD14+ monocytes are the first to disappear from the culture and are nearly completely eliminated by as early as day 2 of culture (Fig.
3b, d) due in large part to plastic adherence prior to transduction. This is followed by a disappearance of CD19+ B-cells, and finally CD56+ NK cells which persist in very small numbers outward to day 7 of culture, at which time point > 97% of the cells are CD3+ T-cells. Transduction efficiencies at day 7 are very high for both E6 transduced T-cells (> 68%) and E7 transduced T-cells (> 87%).
Resting T-cells exist in three general states, naïve (T
naive), central memory (T
CM), and effector memory (T
EM) [
21]. Most of the T-cells at the time of harvest using this procedure have the cellular phenotype of T
EM cells (> 90%), which is consistent with other products manufactured in a similar manner [
22]. A potential limitation to this manufacturing process is that the current general consensus in the field is that a more naïve phenotype is associated with a more robust T-cell with the greater potential for long-term persistence. Clinical trials ongoing here at the NIH (NCT01087294) and by others are exploring this topic in more detail, and our group is currently testing a fully closed flow cytometer capable of sterilely sorting naive T-cells for these purposes (MACSQuant Tyto, Miltenyi).
Functional cell killing assays and cytokine release assays showed that the manufactured cells had a high degree of specificity for target antigen (Figs.
6,
7). In general, the main variations that were observed in these assays were again between E6 and E7 transduced cells rather than between healthy donor and patient samples within these groups, indicating potential differences in TCR affinity toward the target or other vector/construct related issues rather than differences stemming from the cell manufacturing process or patient variability.
Conclusion
In summary, clinical manufacturing procedures for T-cell therapies that achieve a safe and consistent final product represents a major challenge in the field of cellular immunotherapy. Here, we addressed the issue of improving safe manufacturing procedures by introducing multiple semi-closed and closed-system modules aligned in tandem to generate E6- and E7-specific TCR modified T-cells. This process is currently being implemented here at the NIH Clinical Center for two ongoing trials (NCT03197025, NCT02858310). The modular nature of this approach allows for the highest degree of flexibility, which is of great benefit to most cell therapy manufacturing centers that must accommodate multiple phase I/II trials potentially using multiple cell types. Inherent inter-patient variability due to disease status and natural variation are among the factors that contribute the most to the challenge of obtaining a consistent final product. We are able to show that this manufacturing procedure results in a consistent final product and that there is little difference between products generated from healthy donors versus products generated from patients with HPV-associated cancers. Adaptation of these procedures at other phase I/II sites may help streamline the cell therapy manufacturing process, and allow for a quicker transition into later phase studies for successful products.
Authors’ contributions
SLH, JJ, NG designed the study and summarized the data. SLH drafted the manuscript and all authors contributed to the revision of the manuscript. All authors read and approve the final manuscript.