The lymphodepletive and immunomodulatory effects of chemotherapy and radiation standard of care can be leveraged to enhance the antitumor efficacy of T cell immunotherapy. We outline a method for generating EGFRvIII-specific chimeric antigen receptor (CAR) T cells and administering them in the context of glioblastoma standard of care.
Adoptive T cell immunotherapy offers a promising strategy for specifically targeting and eliminating malignant gliomas. T cells can be engineered ex vivo to express chimeric antigen receptors specific for glioma antigens (CAR T cells). The expansion and function of adoptively transferred CAR T cells can be potentiated by the lymphodepletive and tumoricidal effects of standard of care chemotherapy and radiotherapy. We describe a method for generating CAR T cells targeting EGFRvIII, a glioma-specific antigen, and evaluating their efficacy when combined with a murine model of glioblastoma standard of care. T cells are engineered by transduction with a retroviral vector containing the anti-EGFRvIII CAR gene. Tumor-bearing animals are subjected to host conditioning by a course of temozolomide and whole brain irradiation at dose regimens designed to model clinical standard of care. CAR T cells are then delivered intravenously to primed hosts. This method can be used to evaluate the antitumor efficacy of CAR T cells in the context of standard of care.
Glioblastoma (GBM) is the most common primary malignant brain tumor and is invariably fatal. Surgical resection coupled with non-specific standard of care chemotherapy and radiotherapy fails to completely eliminate malignant cells, resulting in a dismal prognosis of less than 15 months in patients with this disease1. In contrast, immunotherapy offers a precise approach for specifically targeting tumor cells, and thus has the potential to serve as a highly effective treatment platform with reduced risk of collateral toxicity2-4. T cells engineered ex vivo to express chimeric antigen receptors (CARs) offer a versatile strategy for tumor immunotherapy. CARs are generated by fusing the extracellular variable region of an antibody with one or more intracellular T cell signaling molecule(s), in lieu of a full-length major histocompatibility complex (MHC)-restricted T cell receptor5. This mode of antibody-like antigen recognition allows for reactive antigen-specific T cells to recognize and respond to tumor antigens in the absence of MHC and can be adapted for a virtually infinite antigen repertoire.
CAR T cells engineered against a variety of tumor antigens have shown preclinical efficacy and outstanding promise in the clinic6-9. Specifically, in the context of GBM, a CAR T cell platform targeting epidermal growth factor receptor variant III (EGFRvIII), a tumor-specific mutation expressed on the cell surface10, was shown to prolong survival in glioma-bearing mice11. Despite their versatility, however, the clinical benefit of CAR adoptive therapy has not been fully realized, due in part to tumor-associated immunosuppression and immune evasion12-16 as well as challenges in establishing and maintaining antigen-specific T cells in vivo. Leveraging standard of care (SOC) with immunotherapy can potentially overcome several of these limitations, resulting in enhanced efficacy in both the preclinical and clinical setting.
SOC for post-resection GBM consists of high-dose temozolomide (TMZ), a DNA alkylating agent17, and whole brain irradiation (WBI)1. These treatments are presumed to synergize with tumor vaccines via upregulation of tumor MHC expression18-20 and the shedding of antigens by dead tumor cells17,19,21,22. Indeed, the addition of TMZ20,23 or WBI18,24 leads to enhanced antitumor efficacy of immune-based treatments in the preclinical setting. Furthermore, like many non-specific cytotoxic chemotherapeutics, TMZ is known to cause systemic lymphopenia25,26, which can be leveraged as a means of host-conditioning for adoptive therapy platforms27-29. TMZ-mediated lymphodepletion has been shown to enhance the frequency and function of antigen-specific T cells, leading to increased efficacy of an adoptive therapy platform against intracranial tumors30. In the context of CAR therapy, lymphodepletion serves as a means of host-conditioning by both reducing the number of endogenous suppressor T cells31, and inducing homeostatic proliferation32 via reduced competition for cytokines33, thus enhancing antitumor activity11,34. Given the synergistic relationship between GBM SOC and immunotherapy platforms, evaluating novel adoptive therapies and vaccine platforms in the context of SOC is critical for drawing meaningful conclusions regarding efficacy.
In this protocol, we outline a method for the generation and intravenous administration of murine EGFRvIII-specific CAR T cells alongside TMZ and WBI in mice bearing EGFRvIII-positive intracranial tumors (see Figure 1 for treatment timeline). Briefly, CAR T cells are made ex vivo by retroviral transduction. Human embryonic kidney (HEK) 293T cells are transfected using a DNA/lipid complex (containing the CAR vector and pCL-Eco plasmids) to produce virus, which is then used to transduce activated murine splenocytes that are harvested and cultured in parallel. During the course of CAR generation, murine hosts bearing EGFRvIII-positive intracranial tumors are administered fractionated whole-brain X-ray irradiation and systemic TMZ treatment at doses comparable to clinical SOC. CAR T cells are then delivered intravenously to lymphodepleted hosts.
The following procedure is described in seven separate phases: (1) Administration of Temozolomide to Tumor-bearing Mice, (2) Whole Brain Irradiation of Tumor-bearing Mice, (3) Transfection, (4) Splenectomy and T cell Preparation, (5) Transduction, (6) CAR T cell Culture and Harvest, and (7) CAR T cell administration to Tumor-bearing Mice. These phases consist of several steps that span 6-7 days and are performed concurrently.
This protocol is based on an experimental design where 10 mice are treated with 107 CAR T cells each. This means that 108 CAR T cells will be needed; the yield should be overestimated by 5 x 107-1 x 108 to account for loss in viability. The following protocol is scaled to generate approximately 200 x 106 cells. The cells are then administered intravenously to female C57BL/6 mice with 9 day established syngeneic EGFRvIII-positive intracranial tumors, developed from the existing KR158B astrocytoma or B16 melanoma cell lines. Concomitantly with CAR T cell generation, tumor-bearing mice are administered clinically relevant doses of lymphodepletive TMZ (60 mg/kg) and WBI (16.5 Gy).
Mice were maintained and bred under pathogen-free conditions at Duke University Medical Center (DUMC). All animal experiments were performed according to protocols approved by the Duke University Institutional Animal Care and Use Committee (IACUC).
1. Administration of Temozolomide to Tumor-bearing Mice
Days 0 through 4:
2. Whole Brain Irradiation of Tumor-bearing Mice
Days 2 through 4:
3. Transfection
Day -1:
Day 0:
4. Splenectomy and T Cell Preparation
Day 0:
5. Transduction
Day 1:
Day 2:
6. CAR T cell Culture and Harvest
Days 3 and 4:
Day 5:
7. CAR T cell Administration to Tumor-bearing Mice
Day 5:
CAR T cells are generated by transduction with the EGFRvIII CAR retroviral vector11. This vector, MSGV1, was developed from the SFGtcLuc_ITE4 vector35, which contains the murine stem cell virus (MSCV) long terminal repeats, the extended gag region and envelope splice site (splice donor, sd, and splice acceptor, sa), and viral packaging signal (ψ). The EGFRvIII CAR containing the human anti-EGFRvIII single-chain variable fragment (scFv) 139, in tandem with murine CD8TM, CD28, 4-1BB, and CD3ζ intracellular regions, was cloned into the retroviral vector downstream the NcoI site (Figure 3).
Following transduction, CAR T cells can be quantified and phenotyped by flow cytometry. EGFRvIII CAR T cells can be visualized with a 2-color panel comprised of a streptavidin-phycoerythrin (SA/PE)-conjugated biotynlated EGFRvIII-derived multimer and anti-CD3 FITC. Using the culture and transduction protocols described here, we routinely observe CAR expression among 55-70% of murine splenocytes (Figure 4A)11. Alternatively, CARs can also be stained for expression using goat-anti-human F(ab’)2-biotin primary and SA/PE secondary antibodies which has been previously described36. CAR T cells can be further phenotyped by the addition of other fluorescently labeled antibodies to the two-color CAR panel. For example, staining for CD8 and CD4 shows that 70% of transduced CARs are CD8+ T cells, while 20% are CD4+ T cells (Figure 4B). CAR T cells with this expression profile have been shown to readily traffic to the brain and treat intracranial tumors.
Body weight | Volume |
25 g | 0.50 ml |
24 g | 0.48 ml |
23 g | 0.46 ml |
22 g | 0.44 ml |
21 g | 0.42 ml |
20 g | 0.40 ml |
19 g | 0.38 ml |
18 g | 0.36 ml |
17 g | 0.34 ml |
16 g | 0.32 ml |
15 g | 0.30 ml |
14 g | 0.28 ml |
13 g | 0.26 ml |
12 g | 0.24 ml |
11 g | 0.22 ml |
Table 1: Temozolomide dose based on animal weight. Temozolomide doses were calculated based on a total dose of 60 mg/kg.
D | d | # Fractions | BED | SOC |
16.5 | 5.5 | 3 | 62 | GBM |
14 | 7 | 2 | 63 | GBM |
20 | 4 | 5 | 60 | GBM |
12 | 6 | 2 | 48 | LGA |
16 | 4 | 4 | 48 | LGA |
21 | 3 | 7 | 52.5 | LGA |
12 | 3 | 4 | 30 | Met |
16 | 2 | 8 | 32 | Met |
Table 2: Clinically relevant biologically equivalent radiation doses. Radiation doses were calculated according to BED = D[1+d/(α/β)], where D = total dose, d = fractionated dose, and α/β = 2). The total dose administered as WBI to the murine host is shown in terms of the fractional doses delivered and the human dose that is modeled by those dose fractions. For model purposes, the malignancy for which the BED is standard of care is also shown.
Figure 1: Timeline for treatment of tumor-bearing mice and concurrent ex vivo CAR generation. Standard of care chemotherapy and whole brain irradiation begins 4 days after tumor implantation (A). During this time interval, CAR T cells are generated and cultured ex vivo (B) and delivered after a full course of host conditioning.
Figure 2: Layout and settings for radiation delivery. X-rays are delivered according to a dosimetry of 320 kV and 10 mA, with a variable duration chosen according to the desired dose (A). The focal field of X-ray radiation is a narrow 2.5 cm area. Sedated animals are arranged according to the grid (B) such that their heads lie in the area of highest X-ray intensity; (C) shows the view from one end of the X-ray beam. Approximately 4 animals can be under the irradiator during one course of X-ray administration according to this grid layout (B, D). Lead tubing is used to shield the body, leaving only their heads exposed for WBI (D).
Figure 3: The modified SFGtcLuc_ITE4 retroviral vector, MSGV1, was used for transduction and generation of CAR T cells. The EGFRvIII CAR insert containing the human anti-EGFRvIII single chain variable fragment (scFv) 139, in tandem with murine CD8TM, CD28, 4-1BB, and CD3ζ intracellular regions, is flanked by murine stem cell virus (MSCV) long terminal repeats (LTR). Upstream of the CAR insert are the extended gag region and envelope splice site (splice donor, sd, and splice acceptor, sa), and the viral packaging signal (ψ).
Figure 4: EGFRvIII CARs are expressed on the surface of T cells. 48 hr following transduction, T cells were stained for surface expression of EGFRvIII CARs using a multimer composed of LEEKKGNYVVTDHC-K(biotin)-NH2/streptavidin-PE. Untransduced T cells from the same donor were also stained as a negative control. Data shows the number of CAR T cells (y-axis) positive for staining by the PE-conjugated EGFRvIII multimer (x-axis), gated on CD3+ cells as a T cell marker, where expression was found to be 60.9% (A). CAR T cells stained for CD4-PerCP-Cy5.5 and CD8-APC show an expression profile of 21.7% and 70.9%, respectively (B).
The treatment timeline described here was designed to model clinical standard of care and leverage its effects for CAR adoptive therapy. CAR T cell doses, TMZ regimens, and radiotherapy administration can be modified to enhance in vivo T cell activity, lymphodepletion, and tumor killing. TMZ regimens can be increased to yield host myeloablation and increased expansion of adoptively transferred cells30. Furthermore, the lymphodepletive effects of TMZ can be recapitulated by low-dose (4 – 6 Gy) single-fraction total body irradiation (TBI), thus circumventing the tumoricidal properties of SOC. The radiation regimen applied here mimics a biologically equivalent dose of 60 Gy; however, the number of fractions and fractional doses can be tuned to mimic other clinical doses (Table 2). Importantly, our model of SOC utilizes biologically equivalent doses of WBI, rather than external beam radiation delivered focally to the tumor and margins, as is often utilized in the clinic, and thus may lead to a lesser dose of radiation delivered to the tumor and an increased risk of toxicities to healthy mouse brain. Despite these limitations, however, WBI is an accepted technique for modeling clinical radiotherapy in a preclinical setting.
CAR T cell doses can be increased or decreased to observe dose effects on tumor killing and overall survival11. Additionally, the timing of treatment and route of delivery can be modified to see differences in efficacy. For instance, CAR T cells can be delivered intracranially to ensure activity at the tumor site. The yield of CAR-positive T cells can vary depending on the efficiency of transduction. One way of ensuring a successful transduction is by carrying a green fluorescent protein (GFP)-control throughout the procedure. To do this, 1 HEK293T 10 cm PDL plate can be allotted for GFP transfection and transduction of a single well of T cells to screen for GFP on days 3-5. A critical factor in transduction efficiency is the extent of T cell activation and proliferation upon addition of viral supernatant. We note an important discrepancy between our method of T cell activation, which is achieved by the addition of ConA to the culture medium, and that used in the clinical setting, achieved by CD3 and CD28 agonist monoclonal antibodies. Although the latter is routinely used for generating patient retroviral transduced CAR T cells and has also been successful in transduction of murine T cells, we previously determined that ConA stimulation led to better and more consistent transductions37. We have also observed that overall CAR T cell yield can vary based on when step 5.18 is performed. If you experience poor CAR T cell yield, we recommend performing step 5.18 immediately after step 5.17 instead of waiting for the 5-6 hr incubation. Although this may impact the efficiency of transduction, we have not observed significant differences in CAR surface expression in our studies.
Following transduction, CAR T cells can be quantified and phenotyped by flow cytometry; we stain our EGFRvIII CAR T cells with a 2-color panel comprised of a streptavidin-phycoerythrin (SA/PE)-conjugated biotynlated EGFRvIII-derived multimer and anti-CD3. Alternatively, CARs can also be stained for expression using goat-anti-human F(ab’)2-biotin primary and SA/PE secondary antibodies which has been previously described36. We routinely observe CAR expression among 55-70% of murine splenocytes11. We have previously shown that delivery of T cells with this CAR expression profile can readily traffic to the brain and treat tumors. TMZ- or TBI-induced lymphopenia enhances clonal expansion of adoptively transferred cells, increasing the frequency of CAR-positive cells and their overall antitumor response. CAR T cells can be tracked in vivo by staining peripheral blood with the appropriate tetramer; this is a useful technique for understanding CAR T cell survival over time and the effects of host-conditioning on CAR T cell function and persistence. Alternatively, if no tetramer is available for the antigen receptor, a fluorescent reporter can be included in the CAR vector construct, or CAR T cells can be labeled ex vivo with carboxyfluorescein succinimidyl ester (CFSE) prior to administration for in vivo tracking.
Depending on the tumor model, immunotherapy platform, and SOC regimen, clinical SOC administered alongside adoptive therapy or vaccines can lead to toxicities. TMZ is a DNA alkylating agent that leads to non-specific cell killing, and intensified doses can result in weight loss and morbidity in murine hosts. High-dose and prolonged fractions of WBI can result in hydrocephalus. Furthermore, inflammatory effects from strong vaccine and/or adoptively transferred T cell responses can lead to morbidity due to toxic cytokine storms. It is also important to note the effects of systemic anesthesia on morbid animals. If TMZ and WBI lead to significant toxicities, then ketamine/xylazine doses can be decreased by 20-25% to reduce the risk of mortality, as animals only need to be sedated and immobilized for 10 min for delivery of small radiation doses (2-8 Gy).
Host conditioning is critical for all immunotherapy platforms that include adoptive transfer. As immunotherapy trials are often evaluated within the context of SOC, preclinical evaluation of immunotherapies should be done within this setting to recapitulate the clinical scenario. We present here one example of leveraging the lymphodepletive and tumoricidal effects of SOC with adoptive therapy using CAR T cells. This regimen is a powerful tool that can be applied to other immunotherapy platforms to evaluate the effects of combined SOC.
The authors have nothing to disclose.
The authors would like to acknowledge Dr. Laura Johnson and Dr. Richard Morgan for providing the CAR retroviral construct. The authors also thank Giao Ngyuen for her assistance with dosimetry for whole brain irradiation. This work was supported by an NIH NCI grant 1R01CA177476-01.
Name of Material | Company | Catalog Number | Comments/Description |
pCL-Eco Retrovirus Packaging Vector | Imgenex | 10045P | Helper vector for generating CAR retrovirus |
Concanavalin A | Sigma Aldrich | C2010 | Non-specific mitogen to induce T cell proliferation and viral transduction |
Retronectin | ClonTech/Takara | T100B | Facilitates retroviral transduction of T cells |
Lipofectamine 2000 | Life Technologies | 11668-019 | Transfection reagent |
DMEM, high glucose, pyruvate | Life technologies | 11995-065 | HEK293 culture media |
RPMI 1640 | Life Technologies | 11875-093 | T cell culture media |
Opti-MEM I Reduced Serum Medium | Life technologies | 11058-021 | Transfection media |
200 mM L-Glutamine | Life technologies | 25030-081 | T cell culture media supplement |
100 mM Sodium Pyruvate | Life technologies | 11360-070 | T cell culture media supplement |
100X MEM Non-Essential Amino Acids Solution | Life technologies | 11140-050 | T cell culture media supplement |
55 mM 2-Mercaptoethanol | Life technologies | 21985-023 | Reducing agent to remove free radicals |
Penicillin-Streptomycin (10,000 U/mL) | Life technologies | 15140-122 | T cell culture media supplement |
Gentamicin (50 mg/mL) | Life technologies | 15750-060 | T cell culture media supplement |
GemCell U.S. Origin Fetal Bovine Serum | Gemini Bio Products | 100-500 | Provides growth factors and nutrients for in vitro cell growth |
Bovine Serum Albumin (BSA), Fraction V—Standard Grade | Gemini Bio Products | 700-100P | Blocks non-specific binding of retrovirus to retronectin-coated plates |
Pharm Lyse (10X concentrate) | BD Biosciences | 555899 | Lyses red blood cells during splenocyte processing |
70 µm Sterile Cell Strainers | Corning | 352350 | Filters away large tissue particles during splenocyte processing |
100 mm BioCoat Culture Dishes with Poly-D-Lysine | Corning | 356469 | Promotes HEK293 cell adhesion to maximize proliferation after transfection |
Temozolomide | Best Pharmatech | N/A | Lyophilized powder prepared on the day of administration |
Dimethyl Sulfoxide | Sigma Life Sciences | D2650 | Necessary for complete dissolution of temozolomide |
Saline | Hospira | IM 0132 (5/04) | Solvent for temozolomide and ketamine/xylazine |
Ketathesia HCl | Henry Schein Animal Health | 11695-0701-1 | Ketamine solution |
AnaSed | Lloyd Inc | N/A | Xylazine sterile solution 100 mg/mL |