Rapid Generation of TCR and CD8αβ Transgenic Virus Specific T Cells for Immunotherapy of Leukemia

Background Virus-specific T cells (VSTs) are an attractive cell therapy platform for the delivery of tumor-targeted transgenic receptors. However, manufacturing with conventional methods may require several weeks and intensive handling. Here we evaluated the feasibility and timelines when combining IFN-γ cytokine capture (CC) with retroviral transduction for the generation of T cell receptor (TCR) and CD8αβ (TCR8) transgenic VSTs to simultaneously target several viral and tumor antigens in a single product. Methods Healthy donor peripheral blood mononuclear cells were stimulated with cytomegalovirus (CMV) and Epstein-Barr-Virus (EBV) peptide mixtures derived from immunogenic viral proteins, followed by CC bead selection. After 3 days in culture, cells were transduced with a retroviral vector encoding four genes (a survivin-specific αβTCR and CD8αβ). TCR8-transgenic or control VSTs were expanded and characterized for their phenotype, specificity and anti-viral and anti-tumor functions. Results CC selected cells were efficiently transduced with TCR8. Average fold expansion was 269-fold in 10 days, and cells contained a high proportion of CD8+ T central memory cells. TCR8+ VSTs simultaneously expressed native anti-viral and transgenic anti-survivin TCRs on their cell surface. Both control and TCR8+ VSTs produced cytokines to and killed viral targets, while tumor targets were only recognized and killed by TCR8+ VSTs. Conclusions IFN-γ cytokine capture selects and activates CMV and EBV-specific memory precursor CD8+ T cells that can be efficiently gene-modified by retroviral transduction and rapidly ex vivo expanded. Our multi-specific T cells are polyfunctional and recognize and kill viral and leukemic targets expressing the cognate antigens.


INTRODUCTION
Adoptive transfer of virus specific T cells (VSTs) rapidly restores antiviral immunity and prevents or treats viral infections after allogeneic hematopoietic stem cell transplantation (HSCT) (1). VSTs are both safe and effective when manufactured from the original stem cell donor or from unrelated partially HLA matched third party healthy donors (1,2), setting the stage for their use as a cellular therapy platform for the delivery of engineered receptors targeting tumor-associated antigens (recently reviewed in (3)). Indeed, leukemia targeted chimeric antigen receptors (CARs) such as CD19-CARs in B-cell acute lymphoblastic leukemia (B-ALL) or T cell receptors (TCRs) targeting Wilms Tumor 1 (WT1) or the minor histocompatibility antigen HA-1H in acute myeloid leukemia (AML) have been expressed in VSTs and infused to patients post-transplant (4)(5)(6)(7)(8)(9). Safety and some efficacy was demonstrated with CD19-CAR-modified VSTs produced from the stem cell donor (4)(5)(6), while feasibility and efficacy with TCRmodified VSTs was variable among studies (7,8).
Manufacturing of engineered VSTs is challenging and operator intensive. Certain steps are performed in open systems such as flow cytometry-based sorting (7), or require knowledge of the targeted epitope such as streptamer-selection (8). Other processes require live Epstein Barr Virus (EBV) for the generation of autologous lymphoblastoid cell lines (4)(5)(6), several types of viral vectors for the transduction of antigen presenting cells and transduction of VSTs (3,6), and prolonged ex vivo culture over several weeks (4)(5)(6)(7).
For broader applicability of such multi-antigen targeted therapies, the complexity of the production processes needs to be reduced (recently reviewed in (10)). Here, we investigated at small scale if Interferon-g (IFN-g) cytokine capture (CC) selected virus-specific memory T cells from healthy donors are sufficiently enriched and activated to directly proceed to retroviral transduction introducing an HLA-A*02:01 restricted survivin targeted TCR (11) in combination with CD8ab (TCR8) to redirect VSTs to a broad tumor-associated antigen (12,13). We have previously demonstrated that the incorporation of CD8ab as a transgene restores anti-viral activity of TCR transgenic VSTs, and redirects CD4+ T cells to the class I restricted cognate antigen (12,13). CC is attractive because it is compatible with fully closed production of VSTs independently of donor HLA and can select T cells with diverse TCR repertoires recognizing various immunogenic epitopes (14)(15)(16). Now we show that enrichment and activation of anti-viral memory T cells by CC followed by retroviral transduction reduces manufacturing time by 7-10 days, reduces the overall complexity of the process, and yields cells with simultaneous anti-viral and anti-tumor activity.

Healthy Donor Buffy Coats
Buffy coats from CMV seropositive de-identified healthy human volunteers were procured from the Gulf Coast Regional Blood Center (Houston, TX, USA). HLA-A2 status was determined by FACS analysis and HLA-A2 positive donors were selected for the experiments.

Generation of Retroviral Vectors and Supernatant
The design of the retroviral vector encoding the survivin-specific (s24) TCR and CD8ab has been described previously ( Figure 1A) (11)(12)(13). Retroviral supernatant was prepared by transient co-transfection of 293T cells with RD114 and Pegpam plasmids and the SFG vector containing the genes of interest (11).

Co-Culture Assay and Cytokine Detection
To determine the anti-tumor function, VSTs and BV173 cells were co-cultured at E:T ratio of 1:5 without exogenous cytokines. Supernatants from co-cultures were harvested after 24 h and were stored at -80°C for cytokine analysis. After 3 days, residual VSTs and tumor cells were enumerated using CountBright Beads (Life Technologies) and FACS analysis. Cytokines were

Chromium Release Assay
In vitro short-term cytotoxicity of VSTs was assessed using a standard 51 Cr-release assay as described previously (13). Briefly, autologous activated T cells (targets) were pulsed with the indicated peptides or pepmixes (1 mg/ml) and labeled with 51 Cr for 1 h. VSTs and target cells were incubated at various ratios for 4 h. For controls, target cells were incubated in media alone or with 1% triton-X 100 (Sigma-Aldrich) to measure the spontaneous and the maximum release, respectively. The mean percentage of specific lysis of triplicate wells was calculated as follows: [(test countsspontaneous counts)/(maximum counts − spontaneous counts)] x100%.

Statistical Analysis
Descriptive statistics was used to summarize the data. Comparison between groups was made using student's t-test or One-Way ANOVA whichever was appropriate. GraphPad prism 6 (GraphPad software, Inc., La Jolla, CA) or higher was used for statistical analysis. P values <0.05 were considered statistically significant.

IFN-g Capture TCR8+ VSTs React Against the Targeted Viral and Tumor Antigens
Next, we assessed antigen specific function of NT and TCR8+ VSTs by IFN-g ELISPOT and intracellular cytokine staining (ICS). As expected, TCR8+ but not NT VSTs produced IFN-g in response to the cognate survivin peptide (LML) or the HLA-A*02:01+survivin+ leukemia cell line BV173, targeted by the transgenic TCR (IFN-g SFCs NT vs TCR8+ VSTs; LML: 7.0±3.7 vs 577±268, p=0.003; BV173: 71.5±122 vs 925±246 p<0.0001, n=6, mean±SD) ( Figure 3A, top left). Importantly, both NT and TCR8+ VSTs showed comparable anti-viral reactivities against CMV (pp65 and IE1 pepmix) and EBV (LMP2, EBNA1 and BZLF1 pepmixes, GLC and YVL peptides) antigens, while a small but significant reduction in NLV reactivity was observed with TCR8+ VSTs ( Figure 3A). These results were corroborated by ICS where we found similar degranulation (CD107a/b), IFN-g and TNF-a levels in NT and TCR8+ VSTs in response to viral antigens ( Figure 3B). Again, the survivin derived LML peptide was only recognized by TCR8+ but not NT VSTs. Thus, IFN-g capture TCR8+ VSTs are specific for and reactive against both the targeted tumor and viral antigens, and anti-viral reactivities are not altered by the transduction and forced expression of TCR8.

IFN-g capture TCR8+ VSTs Kill Viral and Tumor Targets in vitro
We next evaluated the cytotoxicity of NT and TCR8+ VSTs in co-cultures and in a 4-hour 51 Chromium-release assay. When we co-cultured NT or TCR8+ VSTs with HLA-A*02:01+survivin+ BV173 leukemia cells, we observed significant killing of target cells by TCR8+ but not NT VSTs (residual tumor cell count NT vs TCR8+: 2.3±0.6x10 6 vs 0.04±0.07x10 6 , p=0.0004, mean±SD, n=6) ( Figure 4A, left). No difference in the VST counts at the end of the co-cultures was seen ( Figure 4A, right). We also analyzed cytokine secretion and lytic granules present in the co-   Figure 4B). We detected GZMA and perforin release into the supernatant of NT VSTs even though no cytotoxicity was observed. In a 4-hour 51 Chromium-release cytotoxicity assay, we found that activated autologous T cells pulsed with viral peptides were efficiently lysed by both, NT and TCR8+ VSTs at various E:T ratios, while un-pulsed targets were not killed ( Figure 4C, mean ±SD, n=3 donors, each plated in technical triplicates). HLA- A*02:01+survivin+ BV173 leukemia cells were only killed by TCR8+ but not NT VSTs. Donor heterogeneity with regards to CMV and/or EBV reactivity was high, as illustrated in Figure 4D. For example, anti-viral specificity from donor #9 was almost exclusively directed against EBV and not against CMV, while the other two evaluated donors showed simultaneous responses against both viruses. The anti-leukemic activity conferred by the transgenic TCR was much more consistent across donors. Thus, we demonstrate that TCR8+ VSTs generated with our approach are cytotoxic and functional against both viral and tumor targets.

DISCUSSION
Here we present an approach for the rapid generation of engineered human T cells with simultaneous anti-viral and anti-tumor activity ( Figure 5). With the IFN-g cytokine capture system we efficiently enriched and activated anti-viral memory CD8+ T cells that were directly amenable for retroviral transduction, significantly reducing the complexity of manufacturing compared to previously established processes. Transgenic co-expression of the CD8ab coreceptor with the tumor targeted TCR ensured sufficient co-receptor availability for both endogenous anti-viral and transgenic antitumor TCRs. T cells generated with our approach efficiently recognized and killed both viral and tumor targets. Our approach has significant advantages over other established processes for the production of engineered VSTs and allows considering moving to a semi-automated closed process. The most important advantages are that (1) no live virus is necessary for the production of lymphoblastoid cell lines, (2) viral antigen presentation and T cell activation is achieved with peptide pulsing of peripheral blood mononuclear cells and no additional generation of antigen presenting cells is necessary, (3) a one-step procedure is sufficient for T cell selection and activation that allows to directly proceed to gene-modification after a short culture period, (4) selection is performed with magnetic columns and does not require flow cytometry-based sorting, and importantly (5) the approach significantly reduces manufacturing time as well as the number of manipulations needed for product generation. IFN-g captured VSTs can be potentially modified using non-viral gene delivery systems such as transposons or CRISPR/Cas9, that are more versatile and cost-effective (18). In addition, we plan to upscale the approach and reduce the need for open manufacturing steps. For example, cytokine capture as well as gene modification and cell expansion could be adapted to the capabilities of an automated closed manufacturing platform such as for example the CliniMACS Prodigy system (16,19).
Nevertheless, we also identified some disadvantages, which include in our hands (1) the almost exclusive enrichment for CD8+ T cells, and (2) a high donor variability in T cell yield after the CC selection procedure. In fact, virus-specific CD4+ T cells play a key role in the development of long-lasting antiviral immunity by potentiating cytotoxic CD8+ T cell responses, by providing help to B cells for efficient and long-lasting antibody responses, and by direct cytotoxic effects (20)(21)(22). Upon adoptive transfer of VSTs to immunocompromised patients after HSCT, the CD4+ T cell compartment was instrumental for the development of longlasting viral control (23). The lack of CD4+ T cell enrichment in our study may be due to the fact that our stimulation with the viral pepmixes was performed over 6 hours only, compared to previous literature where the stimulation lasted 16 hours (14,24), a factor that needs to be evaluated in the future. The high variability in viral antigen-specific cell frequency is consistent with previous observations (14,16,24) and confirms the fact that circulating anti-viral memory T cell frequency varies over a broad range in different individuals.
VSTs are an interesting cell therapy platform for the development of allogeneic off-the shelf engineered T cell therapies. Several academic clinical trials have demonstrated safety and efficacy in controlling viral infections in immunocompromised patients after solid organ transplant or allogeneic HSCT with the infusion of third-party donor derived banked VSTs in partially HLA matched settings (25)(26)(27)(28)(29)(30). Third party VST cell therapy is now on the way to commercialization. Because VSTs express a viral antigen restricted TCR repertoire, they did not produce significant graftversus host disease in infused patients across studies. However, in vivo persistence was shorter when compared to VSTs derived from HLA matched donors [recently reviewed in (3)] indicating significant rejection by host T or NK cells. Recently, additional engineering strategies have been developed to confer resistance to rejection to the gene-modified VSTs (31,32) which further enhances potential future applicability as a more general cell therapy platform. In summary, we show that manufacturing of gene-engineered VSTs can be simplified and shortened by combining IFN-g cytokine capture and retroviral transduction. Our process is scalable, amenable to the use of non-viral gene delivery systems, and yields highly multifunctional T cells with both anti-viral and anti-tumor activity. Clinical translation of our approach can be envisioned in a clinical trial with the goal to prevent or treat viral infection and malignant relapse in patients after allogeneic stem cell transplant.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors upon reasonable request.

ETHICS STATEMENT
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
GB designed research, performed experiments, analyzed, and interpreted results and wrote the manuscript. CA designed research, supervised the entire study, analyzed and interpreted results and wrote the manuscript. Both authors contributed to the article and approved the submitted version.