A Syngeneic Mouse B-Cell Lymphoma Model for Pre-Clinical Evaluation of CD19 CAR T Cells

The astonishing clinical success of CD19 chimeric antigen receptor (CAR) T-cell therapy has led to the approval of two second generation chimeric antigen receptors (CARs) for acute lymphoblastic leukemia (ALL) andnon-Hodgkin lymphoma (NHL). The focus of the field is now on emulating these successes in other hematological malignancies where less impressive complete response rates are observed. Further engineering of CAR T cells or co-administration of other treatment modalities may successfully overcome obstacles to successful therapy in other cancer settings. We therefore present a model in which others can conduct pre-clinical testing of CD19 CAR T cells. Results in this well tested B-cell lymphoma model are likely to be informative CAR T-cell therapy in general. This protocol allows the reproducible production of mouse CAR T cells through calcium phosphate transfection of Plat-E producer cells with MP71 retroviral constructs and pCL-Eco packaging plasmid followed by collection of secreted retroviral particles and transduction using recombinant human fibronectin fragment and centrifugation. Validation of retroviral transduction, and confirmation of the ability of CAR T cells to kill target lymphoma cells ex vivo, through the use of flow cytometry, luminometry and enzyme-linked immunosorbent assay (ELISA), is also described. Protocols for testing CAR T cells in vivo in lymphoreplete and lymphodepleted syngeneic mice, bearing established, systemic lymphoma are described. Anti-cancer activity is monitored by in vivo bioluminescence and disease progression. We show typical results of eradication of established B-cell lymphoma when utilizing 1st or 2nd generation CARs in combination with lymphodepleting pre-conditioning and a minority of mice achieving long term remissions when utilizing CAR T cells expressing IL-12 in lymphoreplete mice. These protocols can be used to evaluate CD19 CAR T cells with different additional modification, combinations of CAR T cells and other therapeutic agents or adapted for the use of CAR T cells against different target antigens.


Introduction
Chimeric antigen receptor (CAR) T-cell therapy has shown astonishing clinical success in the treatment of CD19 + malignancies leading to the approval of tisagenlecleucel for relapsed acute lymphoblastic leukaemia 1 and axicabtagene ciloleucel for progressive large B-cell non-Hodgkin lymphoma 2 in 2017.
The importance of Interactions between cancer and the immune system in both disease progression and therapeutic mechanisms is becoming increasingly recognized 3,4,5 . For example, it is well documented that the tumor microenvironment (TME) is awash with factors that can suppress the effector functions of immune cells 6,7,8 . Alternatively priming of endogenous immune cells and epitope spreading can be key in tumor eradication and long term resistance to tumor challenge 9,10 . Both of these phenomena cannot be evaluated in xenogeneic models that lack an immune system. Likewise, systems utilizing transgenic proteins do not accurately reflect the challenge of breaking immune tolerance which is required for epitope spreading 4. Top up to 20 mL with PBS. Layer the 20 mL cell suspension gently onto 20 mL of density gradient media (Table of Materials) in a 50 mL tube. Centrifuge the resultant overlaid suspension at 800 x g for 20 min with no brake applied. 5. Harvest cells at interface layer using a sterile Pasteur pipette and transfer to a 50 mL tube. Top up to 50 mL with PBS and centrifuge at 800 x g for 10 min to wash. Discard the supernatant and re-suspend cells in complete TCM. 6. Count the number of cells using a hemocytometer. 7. Culture cells at a density of 5 x 10 6 cells/mL in complete TCM with 30 ng/mL anti-CD3ε antibody (Clone 145-2C11), 30 ng/mL anti-CD28 antibody (Clone 37.51), 100 U/mL recombinant human IL-2 and 2 ng/mL recombinant murine IL-7. Use an appropriately sized tissue culture flask for the volume of cells harvested. NOTE: Antigen-presenting cells are required for T-cell activation by CD3 and CD28 antibodies, if working with purified T cells it is necessary to coat plates with antibodies, or use magnetic beads (  (Figure 1). 4. Filter the retrovirus-containing supernatant through 0.45 μm filter to remove cell debris. Remove TCM from recombinant human fibronectin fragment-coated 6-well plates and add 2.5 mL of filtered retrovirus-containing supernatant or to each well (use complete TCM for mock transfection). Label each well as to the addition of retrovirus or mock media. 5. Centrifuge the plates at 1200 x g for 30 min at room temperature. 6. Whilst plates are spinning, collect activated T cells and count using a hemocytometer.
1. Transduction is carried out with 5 x 10 6 activated splenocytes in a total of 5 mL/well. Pellet the required number of splenocytes for mock/transduction in separate tubes by centrifugation at 500 x g for 5 min. 2. Re-suspend splenocytes at a density of 5 x 10 6 cells per 2.5 mL of filtered retrovirus-containing supernatant from step 2.6.4 or TCM as a negative control. Add recombinant human IL-2 (hIL-2) and recombinant mouse IL-7 (mIL-7) to a final concentration of 200 IU/mL and 4 ng/mL respectively.

In vitro Validation of CAR T cell Activity
1. Seed syngeneic target CD19 + tumor cells with or without luciferase expression at a density of 1 x 10 4 cells in 100 μL TCM/well in a 96-well Ubottom tissue culture plate. 2. Add 1 x 10 4 CD19 CAR T cells/well in a volume of 100 μL/well to achieve an effector to target (E:T) ratio of 1:1.
NOTE: E:T ratios should be established for each CAR construct and target cell line. 3. Use T cells alone and tumor cells alone as negative controls and T cells stimulated by phorbol-myristate-acetate (PMA) (50 ng/mL) and ionomycin (1 μg/mL) as positive control for Interferon gamma (IFNγ) release. Co-culture cells at 37 °C, 5% CO 2 for 16-24 h. 4. Following co-culture, centrifuge the plates at 500 x g for 5 min and collect the supernatant for further IFNγ and IL-12p70 ELISA analysis.
NOTE: This can be stored at -80 °C. 5. Re-suspend cell pellets in 100 μL of PBS containing luciferin (final concentration of 1.5 mg/mL). Incubate the plates for 10 min at 37 °C. Then measure the luminescence from each well with a suitable luminometer. NOTE: Exposure times must be optimized for cell lines and density. Representative results are shown in Figure 3a. Ex-vivo cytotoxicity of CAR T cells can be modified to express luciferin by co-culture with cell lines expressing target antigen. As CAR T cells kill target cells, luciferin is released, therefore a reduction in luminometry signal is correlated with cell kill. Non-transduced cells can often have an effect on target cell viability, particularly over long incubation periods. Measure the concentration of murine IFNγ and IL-12p70 in the supernatant according to the manufacturer's ELISA protocols. Representative results are shown in (Figure 3b and 3c). Ex-vivo activation of CAR T cells by co-culture with cell lines expressing target antigen can be assayed by analyzing supernatant contents using ELISA. The ratio of CAR T cell to target cells and length of co-culture period must be optimized for each CAR construct, target cell line and analyte. PMA and ionomycin treatment can be used as a positive control to confirm quality of T cells and their ability to respond.  (Figure 4).

NOTE:
Establishing A20 lymphoma can take over 2 months with a suboptimal take rate. This can be improved by the use of cyclophosphamide 1 day prior to the delivery of lymphoma cells. In order to study lymphoreplete mice, we identified a dose of cyclophosphamide that could increase efficiency of lymphoma without causing lymphodepletion.  (Figure 5a and b). 5. Inject a single dose of 1 x 10 6 CAR T cells by IV injection into lymphoreplete mice bearing established lymphoma.
NOTE: (Important) Dosing levels must be established for each CAR construct using a dose escalation schedule to ensure that any possible toxicities arising from CAR T cells are characterized and can be addressed. Though anti-mouse CD19 CAR T cells do not display toxicities, CAR T cells can give rise to unexpected toxicities. Where multiple CAR constructs and transduction efficiencies are not identical, the total number of T cells administered should be kept equal by the addition of non-transduced T cells into cell preparations. 6. Monitor disease progression weekly through IP injection of 100 μL of 30 mg/mL luciferin and imaging using an in vivo bioluminescence imaging system (Figure 5c). 7. Closely monitor mice for signs of toxicity and euthanize any mice that show early signs of hind limb paralysis (HLP) or pathological tumor burden before any suffering can arise. NOTE: Toxicities from A20 lymphoma can include hind limb paralysis through tumor invasion of the meninges. Check regularly for early signs of altered gait. Likewise, large IP tumors can arise which can lead to discomfort shown by altered behavior. 8. Monitor survival of mice for 60 -100 days (Figure 5d). Perform euthanasia by a schedule-1 method upon conclusion of the experiment. NOTE: Patients undergoing CAR T-cell treatments undergo a range of regimens to achieve lymphodepletion before the administration of CAR T cells which significantly increases the engraftment of adoptively transferred CAR T cells. This can be replicated in mice with total body irradiation (TBI) (Figure 6). 6. On the next day, inject 1 x 10 6 CAR T cells in 100 μL of PBS via tail vein injection into mice bearing established tumors.
7. Collect blood samples via tail vein bleeds after 7 days. 8. Add red cell lysis buffer to each blood sample, then prepare for flow cytometry as described in section 3. Analyze CAR T cell persistence in the circulation by flow cytometry (Figure 2).
Copyright © 2018 Creative Commons Attribution 3.0 License October 2018 | 140 | e58492 | Page 5 of 11 NOTE: Addition of counting beads immediately prior to cytometry allows determination of the number of CAR T cells per milliliter of blood. 9. Monitor disease progress as described in steps 5.1.5 -5.1.8 (Figure 7).

Representative Results
For high efficiency transduction of T cells, it is necessary to obtain fresh retroviral particles. Transfection of the Plat-E cell line with pCL-Eco producer plasmid and pMP71 retrovirus plasmid gives rise to the secretion of retroviral particles into the cell supernatant. When a fluorescent marker gene, such as mCherry, is encoded in the retrovirus, successful transfection can be confirmed by fluorescence microscopy (Figure 1). Virus-containing supernatant from transfected Plat-E cells is used to transduce T cells via 2 rounds of spin-fection on fibronectin fragment-coated plates. The efficiency of transduction can be determined 4 days post transduction via flow cytometry. Successfully transduced cells express the marker gene encoded in the retrovirus (Figure 2). Transduction efficiencies range from ~ 50 -90% efficiency with first generation receptors to 1 0 -40% with CAR constructs close to the retroviral packaging capacity. While marker gene expression shows successful retroviral transduction, it is paramount to show functionality of CAR T cells upon engaging with cells that express target antigen on their surface. Target cell lines modified to express luciferase can be used in luciferase assays to test the degree of cell-kill by CAR T cells directly (Figure 3A). The release of effector cytokines from CAR T cells upon co-culture with target cells, determined by ELISA, can also be used as an indirect measure of CAR T cell cytotoxicity (Figure 3B and 3C).
CAR T cells produced in this protocol can be evaluated in lymphoreplete mice by establishing systemic A20 lymphoma with a 100 mg/kg dose of cyclophosphamide (injected intravenously), 1 day prior to IV injection of 5 x 10 5 A20 cells (Figure 4). IP injection with luciferin and image capture using an in vivo bioluminescence imager can be used to monitor tumor burden using a constant ROI and exposure time throughout (Figure 5A-C). CAR T cells modified to express IL-12 are capable of eradicating systemic lymphoma with lymphodepleting pre-conditioning giving disease-free survival in about 25% of mice ( Figure 5D). Lymphodepleting preconditioning, achieved by 5 Gy TBI 1 day prior to the IV administration of CAR T cells, significantly improves engraftment (Figure 6). In this model, first generation CAR T cells are capable of eradicating systemic A20 lymphoma, typically inducing disease-free survival in 100% of mice (Figure 7).

Discussion
Syngeneic mouse models allow the testing of disease progression and therapy while maintaining an intact immune system. This is paramount when it comes to therapies that interact with the immune system and in particular for immunotherapeutic agents.
The protocol described here has two critical work streams, the first one is genetically modifying mouse T cell to express CARs. This requires 7 days from initiation to the validation of the transduction. Concomitant with the production of CAR T cells is the establishment of systemic lymphoma in mice. Should CAR T cell production fail, or quality being insufficient, there is typically not enough time to produce replacement cells before mice succumb to lymphoma. It is therefore critical that researchers using these models accurately perform tumor dosing and disease progression studies in order to successfully time the production of CAR T cells for therapeutic administration.
Typical reasons for low T-cell transduction efficiency includes poor transfection efficiency of producer cells, typically caused by poor plasmid purity or inaccurate determination of the pH of transfection media. It is recommended to check the efficiency of producer cell transfection before proceeding with the full protocol as poor transfection will limit the efficiency of T-cell transduction. Recombinant human fibronectin fragments can be collected and stored at -20 °C for re-use, however, multiple freeze-thaws result in reduced transduction efficiency. Swift processing of mouse spleens after collection is also important for obtaining high yields of viable T cells.
It should be noted that the protocol described here utilizes A20 cells expressing luciferase. This is preferred as it provides the ability to measure systemic tumor burden by bioluminescence imaging. However, in the presence of a functional immune system, responses to luciferase could skew the results. We have previously tested immune reactions of surviving mice to marker transgenes 17 . It is key to replicate key experiments using A20 cells free of transgenes to validate that these do not play a significant role in tumor eradication by immune cells.
While clinical agents can only be used in vivo in immune-deficient mice, the use of mouse CAR T cells against mouse cancer cells allows us to evaluate the contributions of the immune system to therapeutic efficacy or disease progression. This protocol could be utilized for the pre-clinical evaluation of CARs targeting B-cell lymphoma or other CARs with additional modifications such as secretion of IL-12 as described here. It must be noted that although the interplay between immune cells can be evaluated in syngeneic mouse models, they may not accurately recapitulate , tissue distribution of target antigen expression may vary between humans and mice and experienced toxicities may be radically different. It is therefore essential to utilize ex vivo and xenogeneic models to corroborate results.
In summary, the syngeneic lymphodepleted and lymphoreplete model of lymphoma recapitulate patients with and without prior chemo/ radiotherapy. This provides a model system in which to mimic the clinical settings to allow the testing of a range of therapeutic strategies that will be important with the coming wave of new immune therapy agents.
With the use of pre-conditioning, it will be noted that all the mice typically clear the lymphoma. With up to 90% complete response rates in humans, this is representative. However, the challenges for CD19 CAR T-cell therapy will hinge on preventing the high frequency of relapses observed that are often CD19. Relapses have not been observed in this model up to, and often beyond 100 days. Modifications to mimic the relapses seen in the clinic could help with the future challenges of CD19 CAR T-cell therapy.

Disclosures
David Gilham works for Celyad which is involved in the production of CAR T cells. The rest of the authors have nothing to disclose.