A GMP-compliant protocol to expand and transfect cancer patient T cells with mRNA encoding a tumor-specific chimeric antigen receptor

Abstract

Chimeric antigen receptors (CARs), which combine an antibody-derived binding domain (single chain fragment variable) with T-cell-activating signaling domains, have become a promising tool in the adoptive cellular therapy of cancer. Retro- and lenti-viral transductions are currently the standard methods to equip T cells with a CAR; permanent CAR expression, however, harbors several risks like uncontrolled auto-reactivity. Modification of T cells by electroporation with CAR-encoding RNA to achieve transient expression likely circumvents these difficulties. We here present a GMP-compliant protocol to activate and expand T cells for clinical application. The protocol is optimized in particular to produce CAR-modified T cells in clinically sufficient numbers under full GMP-compliance from late-stage cancer patients. This protocol allows the generation of 6.7 × 10⁸ CAR-expressing T cells from one patient leukapheresis. The CAR-engineered T cells produced pro-inflammatory cytokines after stimulation with antigen-bearing tumor cells and lysed tumor cells in an antigen-specific manner. This functional capacity was maintained after cryopreservation. Taken together, we provide a clinically applicable protocol to transiently engineer sufficient numbers of antigen-specific patient T cells for use in adoptive cell therapy of cancer.

Electronic supplementary material

The online version of this article (doi:10.1007/s00262-014-1572-5) contains supplementary material, which is available to authorized users.

Introduction

Adoptive T-cell transfer was proven to be a powerful immunotherapeutic tool during the last decades, in particular to specifically target tumors with cytolytic T cells [1]. Engineering T cells with a chimeric antigen receptor (CAR) has emerged as a strategy to redirect them to effectively recognize and lyse tumor cells in a MHC-independent fashion [2–4]. CAR-molecules are artificial immunoreceptors, which contain antigen-binding moieties, usually a single chain fragment variable region (scFv) antibody, and intracellular signaling domains which trigger T-cell activation upon antigen engagement.

Retro- and lenti-viral transductions are currently the preferred procedures to equip T cells with CARs [5–8]. Safety concerns, however, have been raised regarding permanent and high level CAR expression resulting in autoimmunity by the modified T cells [9–11]. Therefore, the transient CAR transfer is considered a safer approach, with RNA electroporation providing a robust and easy-to-perform method [12–16], which recently yielded first clinical data [17, 18].

Transient CAR expression limits the time T cells can exhibit their anti-tumor activity and brings about the need to repetitively inject CAR-modified T cells [13]. Hence, a large number of modified T cells will be required for the intended clinical application. Assumed that 10 injections of 1 × 10⁸ cells will be needed for a first round in adoptive cell therapy, we expect that approximately 1 × 10⁹ T cells in total are required for clinical application. This far exceeds the number of cells which can be obtained by leukapheresis from a tumor patient. Therefore, T-cell expansion is required before RNA transfection. Several expansion protocols have been established for the large-scale generation of T cells, developed for a variety of therapeutic approaches, i.e., receptor transfer and tumor-infiltrating lymphocytes (TIL) (reviewed by Kohn et al. [19]). The published protocols for obtaining antigen-specific T-cell clones from tumor patients are aiming at a strong proliferation over a long period of time and several rounds of stimulation, but often produced T-cell senescence which limited T-cell activity [1, 20, 21]. In contrast, the protocol described here aims at a moderate expansion of bulk T cells within a reasonably short time period (<2 weeks) with the intention to avoid extensive progression in terminal T-cell differentiation. The protocol also takes into account that month-long expansion procedures under GMP-compliant conditions are costly and detrimental for treatment of late-stage cancer patients.

Several procedures to obtain T cells from peripheral blood can be used. (1) Counterflow elutriation used for the isolation of monocytes from leukapheresis products in a closed system [22, 23] can also be used to enrich bulk T cells for subsequent manipulation. (2) Adherence of monocytes to a plastic surface leaves a non-adherent fraction (NAF), which contains mainly lymphocytes. (3) T cells from apheresis products can be expanded without further isolation procedures [13, 24].

Materials and methods

Cell culture media

R10 medium is RPMI 1640 (Cambrex, East Rutherford, NJ, USA) containing final concentrations of 10 % (v/v) heat-inactivated fetal bovine serum (FBS; PAA, GE healthcare, Piscataway, NY, USA), 2 mM l-glutamine (Cambrex, East Rutherford, NJ, USA), 100 U/ml penicillin, 100 µg/ml streptomycin (Cambrex, East Rutherford, NJ, USA), 2 mM HEPES (PAA, GE healthcare, Piscataway, NY, USA), and 20 µM beta-mercaptoethanol (Gibco, Life Technologies, Carlsbad, CA, USA).

DC-medium is RPMI 1640 containing 1 % (v/v) heat-inactivated plasma from healthy donors, 2 mM l-glutamine. For cell culture work outside the clean room facility, we added 8 ng/ml gentamicin (PAA, GE healthcare, Piscataway, NY, USA). All components were mixed under sterile conditions and filtered through a 0.22-µm sterile filter (Merck Millipore, Billerica, MA, USA). X-vivo 15 (Lonza, Basel, Switzerland) and OpTmizer (Life technologies, Carlsbad, CA, USA) are serum-free, GMP-compliant media primarily designed for the cultivation of TIL or expansion of T cells, respectively.

FACS buffer is DPBS (Cambrex, East Rutherford, NJ, USA) containing 1 % FBS (PAA, GE healthcare, Piscataway, NY, USA) and 0.02 % sodium azide (Merck, Darmstadt, Germany).

Cell lines

The CEA⁺ KATO III cell line was kindly provided by Drs S. Santegoets and T. de Gruijl, VUMC, Amsterdam. T2.A1 is a TAP-deficient T × B hybrid cell line. Cell lines were cultured in R10 medium.

T-cell isolation

All human material was obtained after informed consent and approved by the institutional review board. The elutriation of leukapheresis products was performed as previously described [22, 23] using an Elutra-cell separator (Gambro BCT, Inc., Lakewood, CO, USA). Alternatively, the peripheral blood mononuclear cells (PBMC) were isolated by standard Ficoll gradient centrifugation of apheresis products. Apheresis cells were loaded onto Lymphoprep (Axis-shield, Oslo, Norway), and cells were separated by centrifugation. PBMC were collected and washed four times with DPBS (Cambrex, East Rutherford, NJ, USA) containing 1 mM EDTA (Lonza, Basel, Switzerland). NAF cells were obtained by incubation of PBMC for 1 h in a cell factory (Nunc Doubletray; Thermo Fisher, Waltham, MA, USA) in DC-medium at standard cell culture conditions (i.e., 37 °C; 95 % relative humidity; 5 % (v/v) CO₂). The NAF cells in the supernatant were collected for T-cell expansion.

T-cell expansion

T cells were expanded immediately after obtaining the starting material (PBMC, NAF, or elutriation fraction 3). Where indicated, T cells were left resting for 24 h prior to expansion in DC-medium as they emerged from the elutriation (0.7 × 10⁶–2.6 × 10⁶ cells/ml) at standard cell culture conditions without cytokines or antibodies. T cells were cultured at 1 × 10⁶ cells/ml in the indicated medium. A final concentration of 0.1 µg/ml anti-CD3 antibody (Orthoclone OKT3; Jannsen-Cilag, Neuss, Germany) and 10³ IU/ml IL-2 (Proleukin; Novartis, Nuremberg, Germany) was added on day 0. Where indicated, anti-CD28 antibody (Miltenyi Biotech, Bergisch Galdbach, Germany) at a final concentration of 0.25 µg/ml was administered. IL-2 (10³ IU/ml) was added at day 2. On day 3, the cells were counted and diluted to 2 × 10⁵ cells/ml in cell culture medium, and 10³ IU/ml IL-2 was added; same on day 5 of culture. The total culture volume was doubled by adding fresh medium on day 7, and IL-2 (10³ UI/ml) was added. In some experiments, additional IL-7 (10 ng/ml; Peprotech, Hamburg, Germany) and IL-15 (5 ng/ml; R&D Systems, Minneapolis, MN, USA) were administered along with IL-2. Cells were harvested and counted after 10 days of expansion and subsequently transfected with RNA. The fold expansion was determined by calculating the ratio of (number of cells at the end of culture)/(number of cells at the beginning of culture). Cells were either used directly after RNA transfection or were cryopreserved as described [25].

T-cell isolation and expansion using magnetic beads

T-cell isolation and expansion were performed according to manufacturer’s protocol using Dynabeads^® ClinExVivo™ CD3/CD28 (Invitrogen, Paisley, UK). In short, donor- or patient-derived PBMC were cryopreserved and the percentage of CD3⁺ cells was determined by flow cytometry. Subsequently, the PBMC were thawed, and the CD3⁺ cells were isolated with magnetic beads at a 3:1 bead to cell ratio using a DynaMag-15 magnet (Invitrogen Paisley, UK). The cells at 1 × 10⁶ cells/ml were cultured in X-vivo 15 or OpTmizer medium with 200 IU/ml IL-2. From day three on, cells were counted daily. If the cell-concentration exceeded 1 × 10⁶ cells/ml, the cells were diluted to 5 × 10⁵ cells/ml by adding fresh cell culture medium with 200 IU/ml IL-2.

RNA transfection

The composition of the CEA-specific CAR was previously described in detail [5, 26]. The control CAR features the same signaling domains, but contains a scFv with an irrelevant specificity. The CARs were cloned into the pGEM4Z-5′UTR-sig-husurvivin-DC.LAMP-3′UTR RNA-production vector [27] (kindly provided by Kris Thielemans), and RNA was transcribed in a cell-free system using the mMESSAGE mMACHINE T7 Ultra kit (Life technologies, Carlsbad, CA, USA) and purified with an RNeasy Kit (Qiagen, Hilden, Germany) according to manufacturers’ protocols. The RNA electroporation was performed as described previously [12, 25]. T cells were electroporated using a Gene Pulser Xcell (Bio-Rad, Hercules, CA, USA) and 4-mm gap electroporation cuvettes (Peqlab, Erlangen, Germany). The cells were electroporated at 1.5 × 10⁸ cells/ml for 5 ms and 500 V (square-wave pulse) in OptiMem (Life technologies, Carlsbad, CA, USA) with a RNA concentration of 150 µg/ml.

Phenotype FACS of expanded cells and CAR surface staining after RNA electroporation

The phenotype of the expanded cells was analyzed by flow cytometry with the following antibodies: αCD14-PE, αCD25-APC (both eBioscience, San Diego, CA, USA), αCD8-FITC, αCD19-V500, αCD3-APC-H7, and αCD4-V450 (all BD Bioscience, Franklin Lakes, NJ, USA). Cells were incubated with antibody for 30 min at 4 °C in FACS buffer and subsequently washed with FACS buffer and analyzed using a FACSCanto II with FACSDiva software (both BD Bioscience, Franklin Lakes, NJ, USA). Results were evaluated using FCS Express 4 flow research edition software (DeNovo Software).

For CAR surface staining, T cells were harvested 6 h after electroporation and washed in FACS buffer. Cells were incubated with a goat-F(ab′)2 anti-human IgG-RPE antibody (Southern Biotech, Birmingham, AL, USA) for 30 min at 4 °C. The murine anti-CD3 antibody bound to the T cells was detected by staining with a goat anti-mouse Ig antibody (BD Bioscience, Franklin Lakes, NJ, USA). Samples were analyzed with a FACScan flow cytometer (BD Bioscience, Franklin Lakes, NJ, USA). Results were evaluated using CellQuest Software (BD Bioscience, Franklin Lakes, NJ, USA).

Cytokine secretion assay

T cells were used 4 h after electroporation or 2 h after thawing. T cells (5 × 10⁴) were co-cultured with the same number of tumor cells in 100 µl X-vivo 15 medium for 16 h. The cytokine concentrations in the supernatants were determined using the human Th₁/Th₂ Cytokine Kit II (BD Bioscience, Franklin Lakes, NJ, USA) according to manufacturer’s instructions.

Chromium release assay

T2.A1 and KATO III cells were labeled with 100 µCi of Na₂ ⁵¹CrO₄/10⁶ cells for 1 h at 37 °C and 5 % CO₂. After washing, target cells were cultured in 96-well plates (Thermo Fisher, Waltham, MA, USA) at 1,000 cells/well. The CAR-transfected T cells were added at indicated target/effector ratios, and cells were co-incubated in triplets for 4 h. Radioactivity was determined in the supernatant, and lysis was calculated as follows: [(measured release − background release)/(maximum release − background release) × 100 %].

Results

T cells isolated by counterflow elutriation cannot be expanded to sufficient numbers for a clinical application

To obtain clinically relevant T-cell numbers, we used the T-cell-enriched fraction from counterflow elutriations of leukapheresis products from cancer patients, using melanoma patients as a source readily available to us, as starting material [22]. This allows for the generation of both a T cell and a dendritic cell (DC) medicinal product from the same apheresis in a GMP-compliant system. Elutriation fraction 3 contained on average 80.4 % T cells (n = 11; SD = 10.2 %; data not shown). Cells were stimulated with the GMP-compliant, agonistic anti-CD3 antibody OKT3, and a pharmaceutical grade IL-2 (Proleukine, Aldesleukine) for 10 days in different cell culture media (Fig. 1a), i.e., the serum-free GMP-compliant cell culture media X-vivo 15 and OpTmizer, and the in-house DC-medium used for GMP-conform DC generation [22]. The X-vivo 15 medium sustained a 1.7-fold expansion on average (Fig. 1a, left), while T-cell numbers decreased in OpTmizer medium to 14 % on average (Fig. 1a, middle). T cells expanded 8.8-fold on average in DC-medium, cell numbers of one patient, however, decreased to 30 % of the starting material (Fig. 1a), indicating that the expansion of T cells from an elutriated cell preparation in DC-medium may be of high variability, which would hamper clinical application. X-vivo 15 is available ready-to-use in GMP quality, whereas DC-medium has to be individually manufactured for each patient. Hence, X-vivo 15 was used in all further experiments.

Fig. 1.

Expansion of patient-derived T cells after elutriation. Leukaphereses from melanoma patients were elutriated, and cells from fraction 3 were expanded in presence of the anti-CD3 antibody OKT3 and IL-2 for 10 days. a Cells were expanded directly after elutriation in the following cell culture media: X-vivo 15 (n = 4), OpTmizer (n = 4), DC-medium (n = 3). b Cells were left resting for 1 day after elutriation and expanded in X-vivo 15 medium supplemented with either OKT3 antibody and IL-2 (n = 3) or additional IL-7 and IL-15 (n = 3). The different symbols indicate individual patients, and the mean fold expansion is depicted as a bar

Since T-cell numbers decreased during the first 24 h after elutriation and stimulation (data not shown), we rested the T cells without further stimulation during that period of time. This enhanced the survival of the cells during the first 24 h of culture thereby increasing the overall yield of T-cell proliferation (data not shown). Furthermore, we added IL-7 and IL-15 during subsequent T-cell activation, as these cytokines sustain a modest T-cell expansion after antigen-specific activation in vitro [28]. Cells that received OKT3 and IL-2 only expanded on average 4.26-fold, while cells that were additionally treated with IL-7 and IL-15 expanded by 5.6-fold (Fig. 1b). The addition of IL-7 and IL-15 did not produce a relevant increase in T-cell expansion under these conditions; the resulting T-cell numbers were not high enough for clinical application. While the decrease of T-cell number during the first 24 h was less severe when activation took place after resting the cells (data not shown), the resulting T-cell expansion was still insufficient.

Anti-CD3/anti-CD28 antibody-coated magnetic beads are designed to separate and expand T cells ex vivo from elutriation products [29]. Although the beads induced a strong T-cell proliferation (68.4-fold; n = 4; SD = 26.4; data not shown), the resulting T cells were not well suitable for mRNA electroporation. Sometimes these cells did not survive this rather harsh method resulting in a survival of only 6 %. The main problem, however, was that these cells produced cytokines in an antigen-unspecific manner. When they were incubated with Ag-negative targets, between 1.5 and 3 ng/ml of IL-2 and 5–16 ng/ml of IFNγ were produced. Most importantly, these concentrations were not increased, when the T cells were incubated with antigen-positive targets (data not shown).

Taken together, in our hands, none of the tested protocols and reagents led to a robust and reproducible T-cell expansion from elutriation products which would provide enough T cells for clinical application. As other groups have shown that in principle, T-cell expansion is possible with the expansion protocols and materials (i.e., beads, antibodies, cytokines) we tested [14, 28], we came to the conclusion that not the kind of stimulus, used for T cells expansion, is causing the problem, but that the T cells themselves already were adversely affected by the process of counter-flow elutriation.

T cells from late-stage melanoma patients can be expanded from PBMC

While counterflow-elutriated T cells did not result in sufficient cells after expansion, we alternatively explored activating T cells directly from the NAF or PBMC derived from apheresis products for stimulation and transfection. The idea behind this is that (1) the non-T cells in the NAF or PBMC may act in a feeder-cell-like manner, while avoiding the GMP-compliance-related problems, like the use of allogeneic material or immortalized cells and (2) the step of T-cell purification is eliminated. As patient material is scarce, we established the protocol with cells from healthy donors. Furthermore, we investigated the additional use of a GMP-compliant anti-CD28 antibody during the T-cell expansion, which would provide a co-stimulatory signal to the T cells. The stimulation of healthy donor NAF cells with the anti-CD3 antibody OKT3 and IL-2 led to a 26-fold T-cell expansion, while PBMC were expanded 49.5-fold (Fig. 2a). In the latter case, additional anti-CD28 antibody did not further improve T-cell amplification. The protocol did not only work well for healthy donor-derived cells, but also with PBMC from late-stage melanoma patients. We obtained an average of 31-fold expansion of patient-derived T cells when using the anti-CD3 antibody and IL-2 only (minimum 7.5-fold, maximum 60.5-fold); again the additional anti-CD28 antibody did not further increase T-cell amplification (29.7-fold) (Fig. 2a). These expansion rates were well sufficient for the intended clinical applications producing more than 1.5 × 10¹⁰ T cells in 10 days when starting with 2 × 10⁹ PBMC. After the expansion, we monitored the cells for residual anti-CD3 antibody. No murine antibody could be detected on the T-cell surface (Supplementary figure 1).

Fig. 2.

Expansion of patient-derived and healthy donor-derived T cells using PBMC. a The non-adherent fraction (NAF) was collected from blood of healthy donors and expanded for 10 days with OKT3 and IL-2 (left; n = 4). PBMC directly isolated from leukapheresis products of healthy donors (middle; n = 5), or melanoma patients (right; n = 4) were expanded in the presence of OKT3 and IL-2 for 10 days either in the absence (closed symbols) or presence (open symbols) of the agonistic anti-CD28 antibody. Cells were counted at the beginning and end of culture, and fold expansion was determined. The diamonds indicate individual samples, and the mean fold expansion is depicted as a bar. b PBMC from melanoma patients were expanded for 10 days using the agonistic anti-CD3 antibody OKT3 and IL-2. The amplified T cells were electroporated without RNA (mock) or with mRNA encoding a CEA-specific chimeric antigen receptor (CEA-CAR). T cells were stained for receptor expression 6 h after electroporation. An overlay histogram of one representative of three individual experiments is shown

The specificity of expanded T cells from late-stage melanoma patients is reprogrammed by electroporation of RNA encoding a chimeric antigen receptor

To demonstrate the functional activity of the expanded cells, melanoma-patient-derived T cells were equipped with an anti-CEA-CAR (as a model) or a control CAR by RNA electroporation. T cells electroporated without RNA served as controls. Furthermore, we examined whether the presence of anti-CD28 antibody during the expansion of the T cells had any positive effect on the functionality of the transfected T cells.

Under standard T-cell-electroporation conditions [25], we were able to electroporate 90 × 10⁶ cells per cuvette. From our experience with GMP-compliant DC electroporations, we know that a maximum of 30 cuvettes can be handled in the production process. This allows for the transfection of 2.7 × 10⁹ T cells.

After electroporation, the T cells were checked for CAR expression. The transfection efficiency was 96 ± 0.88 % (Fig. 2b). Freshly modified CAR T cells secreted IL-2, TNF, and IFNγ upon CEA-mediated, specific CAR stimulation, whereas those cytokines were secreted at background levels by control CAR-transfected or mock-electroporated T cells (Fig. 3a). The CAR-mediated activation was antigen-specific, since incubation with CEA-negative target cells did not induce CAR-T-cell activation (Fig. 3a). The addition of the agonistic anti-CD28 antibody during T-cell expansion did not yield any significant improvement of the antigen-specific cytokine release (Fig. 3a).

Fig. 3.

CAR-RNA-transfected T cells of melanoma patients secrete cytokines after antigen-specific stimulation. PBMC from melanoma patients were expanded for 10 days using the agonistic anti-CD3 antibody OKT3 and IL-2, in the presence or absence of the agonistic anti-CD28 antibody. The amplified T cells were electroporated without RNA (neg.) or with mRNA encoding a CEA-specific chimeric antigen receptor (αCEA) or a CAR of irrelevant specificity (con.). Cells were either stimulated 4 h after electroporation or were cryoconserved 2 h after electroporation. Fresh cells (a), or thawed cells, which were left resting for 2 h after thawing (b), were co-incubated with T2.A1 (CEA⁻) or KATO III (CEA⁺) cells over night. Concentrations of IL-2, TNF, and IFNγ in the supernatant were determined. All depicted data are the mean of three independent experiments. Error bars indicate standard deviation

Cryopreservation of modified T cells is a prerequisite for staggered injections and required for safety testing of the individual cell batch prior to clinical application. The overall yield after electroporation and cryoconservation was 25.7 ± 2.9 % of the T cells (data not shown). We assayed whether modified T cells retained their ability to secrete cytokines in an antigen-specific manner after cryopreservation. Although T-cell-secreted cytokine levels were lower compared to the levels of freshly manipulated cells (Fig. 3a), the response was still antigen-specific (Fig. 3b). Again the additional use of the anti-CD28 antibody during the expansion did not substantially improve cytokine secretion (Fig. 3b).

Experiments with cells from healthy donors revealed that our expanded T cells were able to lyse tumor cells antigen-specifically after CAR-RNA transfection in a standard chromium release assay, but also that cryoconservation reduced their cytolytic capacity (Supplementary figure 2). Hence, we investigated the cytolytic capacity of T cells from melanoma patients after cryopreservation. T cells engineered with the anti-CEA-CAR lysed CEA⁺ KATO III cells, whereas the control cells did not (Fig. 4b). Additional anti-CD28 antibody during T-cell expansion did not increase their cytolytic activity (Fig. 4b). T cells, which were transfected with an irrelevant control CAR, did not lyse KATO III cells (Fig. 4b). No relevant degree of antigen-independent cytolysis or lysis of antigen-negative T2.A1 cells was obtained (Fig. 4a). We conclude that patient-derived, ex vivo-expanded and CAR-modified T cells recognize tumor cells in an antigen-specific fashion and that this property is still preserved after cryopreservation of the cells.

Fig. 4.

CAR-RNA-transfected T cells of melanoma patients lyse antigen-bearing tumor cells. PBMC from melanoma patients were expanded for 10 days using the agonistic anti-CD3 antibody OKT3 and IL-2, in the presence or absence of the agonistic anti-CD28 antibody. The amplified T cells were electroporated with mRNA encoding a CEA-specific chimeric antigen receptor (αCEA) or a CAR of irrelevant specificity (con.). Cells were cryoconserved 2 h after electroporation. Thawed T cells were assayed for cytolytic activity in a standard 4 h chromium⁵¹ release assay, with a T2.A1 (CEA⁻) and b KATO III (CEA⁺) as target cells at the indicated ratios. The release of chromium into the supernatant was determined, and lysis was calculated. Mean values from three independent experiments are depicted. Error bars indicate standard deviation

A schematic representation of the established expansion protocol is depicted in Fig. 5. The phenotype of the expanded cells was determined and showed that more than 90 % of the expanded cells were CD3⁺ T cells. The T cells were composed of 29 % CD4⁺, 61 % CD8⁺, 3 % CD4⁺ and CD8⁺, and 7 % double negative cells. Contaminating CD19⁺ cells (i.e., B cells) and CD14⁺ cells (i.e., monocytes) were <3 % and <5 %, respectively (data not shown). At large, the results demonstrate that this protocol to amplify and manipulate T cells is robust and easy-to-perform, can be applied to T cells from late-stage cancer patients, and can be performed under GMP-compliant conditions. Starting from 2 × 10⁹ PBMC, at least 1.5 × 10¹⁰ T cells can be generated. Of these, 2.7 × 10⁹ T cells can be electroporated and cryopreserved in compliance to GMP, resulting in 96 ± 0.88 % transfection efficiency, and a yield of 25.7 ± 2.9 % after electroporation and cryoconservation. In total, 6.7 × 10⁸ CAR-expressing functional T cells ready to be used for clinical application can be generated.

Fig. 5.

Scheme of the preferred protocol. PBMC are seeded at a concentration of 10⁶/ml in X-vivo 15 supplemented with 10³ IU/ml Proleukin (IL-2) and 0.1 µg/ml anti-CD3 antibody (OKT-3). Fresh IL-2 (10³ IU/ml) is added at the indicated time points (10³ IU/ml final volume). On day 3, the cells are counted and diluted to 2 × 10⁵ cells/ml in X-vivo 15. The volume is doubled with X-vivo 15 on day 7. Cells are electroporated on day 10 and cryoconserved 2 h after electroporation

To facilitate an easy implementation of the presented procedure, we included detailed GMP-compliant operational protocols (Supplementary standard operating procedure 1 and 2). These documents can easily be adjusted to the prerequisites of the local regulatory authorities and therefore provide help for the application of an intended clinical study. We also provide flowcharts of the process (Supplementary flow charts 1, 2) as it is usually requested by medical regulatory bodies to include those in Investigational Medicinal Product Dossiers (IMPD).

Discussion

As different therapies, based on the adoptive transfer of T cells, showed promising results in the immunotherapy of cancer, various protocols for the activation and expansion of T cells have emerged depending on the final clinical product.

The standard method for CAR transfer is retroviral transduction [5–8]. To this end, T cells are activated and subsequently transduced with the CAR. The successfully transduced T cells are then expanded to clinically relevant numbers for administration. This method is feasible only for stably transduced cells as they retain their chimeric receptor. In contrast, if a transient expression of the receptor is desired, a different expansion protocol is required, as T-cell proliferation after receptor transfer would result in dilution and eventual loss of the transferred receptor.

Given the severe and even lethal off-target toxicities, which were recently observed after stable receptor transfections despite careful preclinical testing [30, 31], we believe that mRNA-based transfection of TCR and CAR will emerge as a standard transfer method to ensure safety and to screen for adverse effects. In case the tested receptor shows no unwanted off-target or off-tissue reactivity, it could then be transduced retrovirally. Furthermore, in conjunction with combination therapies, such as checkpoint-blockade, adoptive therapy using transiently reprogrammed T cells may exert clinical effects based on a change of the tumor microenvironment from suppressive to inflammatory, combined with a first wave of tumor destruction, triggering the induction of endogenous tumor-specific T cells. This will require feasible, highly reproducible, and GMP-compliant methods and procedures to generate sufficient numbers of mRNA-electroporated T cells from cancer patient-derived material. Although several protocols for T-cell activation and expansion were published (reviewed in: [32, 33]), these were developed for retroviral transduction and detailed protocols to generate activated T cells for mRNA electroporation are not published, even though the feasibility and safety of adoptive therapy with CAR-mRNA-transfected T cells was recently shown [18].

In a preclinical study, Almasbak et al. [14] performed the activation and expansion of T cells for CAR-RNA transfection with antibody-coated magnetic beads. The published protocol, however, contains steps, which are challenging to perform in a GMP-compliant fashion. The daily resuspension of the cells, the cell-counting, and the adjustment of the T-cell concentration require opening of the T-cell cultures and therefore increase the risk of contaminations. The freeze–thaw cycle of the cells before separation and cultivation [14] may negatively affect the cells and is tedious to perform under GMP conditions. Furthermore, automated culture systems are a prerequisite for any T-cell therapy to become broadly available [33]. This also precludes frequent opening and individual adjustment of T-cell cultures, which are a part of the common bead-based protocols. In addition, in our hands, the bead-based expansion did not reliably yield T cells of sufficient quality for RNA transfer.

A recent study by Maus et al. [18] revealed that the repetitive injection of activated, CAR-RNA-transfected T cells can lead to anaphylaxis. The authors conclude that this is due to the generation of human anti-mouse antibodies (HAMA) against the murine scFv fragment of the transfected CAR or other components of the cellular product. The fact that the Ab-derived parts of the CAR, especially the Fc ‘spacer’ domain can be recognized by the hosts Fc-receptors [34], probably contribute severely to the immunogenicity of the CAR-molecules. By introducing several point mutations into the Fc-spacer part of the CAR, binding of the CAR to Fc-receptors of innate cells can be prevented [34]. This would protect CAR-transfected cells from phagocytosis and hence prevent presentation of CAR-derived epitopes. Yet, immune responses to the murine scFv can further be addressed by using either a humanized scFv or a fully human CAR. Alternatively, Maus et al. [18] suggest to adjust the treatment protocol to avoid long intervals without injections, hence circumventing the isotype switch from IgG to IgE and therefore prevent an anaphylactic shock.

In our case, the use of murine antibodies as OKT3 may theoretically contribute to this issue. Hence, we controlled the expanded T cells before electroporation for residual OKT3 and detected no remaining murine antibody. This in-process-control was also integrated as release criterion into the intended GMP protocol.

A relevant observation we made during the establishment of the protocol, and which we want to share with other researchers to save them from excessive but pointless work, is the fact that T cells purified by counterflow elutriation were not suitable for expansion with any of the protocols we tried. Personal communications with scientists from other laboratories, namely Michael Aigner (Medizinische Klinik 5, Universitätsklinikum Erlangen) and Özcan Met (Center for Cancer Immune Therapy (CCIT) University Hospital Herlev), revealed that these problems with elutriation-derived T cells have also been observed in other groups. During elutriation, the cells are in a single cell suspension in medium for an extended time. The lack of nutrients, cytokines, and cell–cell contacts, as well as the physical stress, may affect T cells stronger than other cells. Therefore, even if it may be tempting to use this source, because large quantities of T cells are generated as by-product from the isolation of other cells—like monocytes—we strongly advise not to use them.

The methodology presented here allows the generation of high numbers of T cells (1.5 × 10¹⁰). With standard laboratory electroporation devices (e.g., the Gene Pulser Xcell or the ECM 830 from BTX), it is feasible to electroporate 2.7 × 10⁹ of these. Currently, electroporation requires a class A clean room, as these devices do not offer an operation in a closed system. However, recently a high throughput electroporation device from the company MaxCyte became commercially available and other companies are developing similar systems, which can be integrated into a GMP-compliant closed system. These will allow to transfect the complete batch of cells generated, which would increase the output approximately fivefold. In parallel, commercial vendors start to offer custom-made mRNA in large quantities for affordable prices. In light of these developments, we consider the protocol presented here to be a solid basis for T-cell manufacturing for clinical application in the field of adoptive T-cell transfer.

Electronic supplementary material

Below is the link to the electronic supplementary material.