Transgenic CD8αβ co-receptor rescues endogenous TCR function in TCR-transgenic virus-specific T cells

Background Genetically engineered virus-specific T cells (VSTs) are a platform for adoptive cell therapy after allogeneic hematopoietic stem cell transplantation. However, redirection to a tumor-associated antigen by the introduction of a transgenic T-cell receptor (TCR) reduces anti-viral activity, thereby impeding the possibility of preventing or treating two distinct complications—malignant relapse and viral infection—with a single cell therapy product. Availability of CD8αβ co-receptor molecules can significantly impact class I restricted T-cell activation, and thus, we interrogated whether transgenic CD8αβ improves anti-viral activity mediated by native VSTs with or without a co-expressed transgenic TCR (TCR8). Methods Our existing clinical VST manufacturing platform was adapted and validated to engineer TCR+ or TCR8+ VSTs targeting cytomegalovirus and Epstein-Barr virus. Simultaneous anti-viral and anti-tumor function of engineered VSTs was assessed in vitro and in vivo. We used pentamer staining, interferon (IFN)-γ enzyme-linked immunospot (ELISpot), intracellular cytokine staining (ICS), cytotoxicity assays, co-cultures, and cytokine secretion assays for the in vitro characterization. The in vivo anti-tumor function was assessed in a leukemia xenograft mouse model. Results Both transgenic CD8αβ alone and TCR8 had significant impact on the anti-viral function of engineered VSTs, and TCR8+ VSTs had comparable anti-viral activity as non-engineered VSTs as determined by IFN-γ ELISpot, ICS and cytotoxicity assays. TCR8-engineered VSTs had improved anti-tumor function and greater effector cytokine production in vitro, as well as enhanced anti-tumor function against leukemia xenografts in mice. Conclusion Incorporation of transgenic CD8αβ into vectors for TCR-targetable antigens preserves anti-viral activity of TCR transgenic VSTs while simultaneously supporting tumor-directed activity mediated by a transgenic TCR. Our approach may provide clinical benefit in preventing and treating viral infections and malignant relapse post-transplant.


INTRODUCTION
Malignant relapse and viral infections are the two major causes for treatment failure and morbidity in patients after allogeneic hematopoietic stem cell transplantation (HSCT). 1 An ideal cellular therapy after stem cell transplant would therefore target both problems simultaneously. Virus-specific T cells (VSTs) are already a clinically validated immune effector cell therapy platform amenable to genetic redirection of antigenspecificity to tumor-associated antigens, as demonstrated with chimeric antigen receptor (CAR)-modified VST cell therapies. 2-6 CAR+ VSTs can significantly re-expand in vivo upon viral reactivation and stimulation through the endogenous T-cell receptor (TCR) and persist long-term. 6 Efforts to redirect VSTs to tumor by introduction of a transgenic TCR, [7][8][9][10][11] however, have been more problematic. Forced expression of a transgenic TCR leads to downregulation of endogenous TCRs 12 and consequent reduction in anti-viral reactivity, although anti-tumor activity can be sustained. [7][8][9][10][11] The reduction in anti-viral activity was consistent across several studies by independent groups, using a variety of different VST platforms, TCR specificities and vectors. Anti-tumor function predominated consistently, 10 11 or reactivities shifted between compartments depending on the type of antigen encountered (viral versus tumor antigen). 11 These effects are most likely explained by competition for TCR/CD3/CD8 complex signaling components used by both the endogenous anti-viral and introduced transgenic TCRs, as well as TCR mis-pairing between introduced and endogenous TCR chains, [11][12][13] and imply two important points: (i) the clinical benefit from controlling viral reactivation post transplant may be jeopardized when using TCR+ VSTs, and (ii) the capacity of TCR+ VSTs to re-expand in vivo upon viral reactivation or vaccination may be limited compared to CAR+ VSTs.
Incorporation of CD8αβ into the transgenic TCR vector enhances the function of polyclonal TCR+ T cells through multiple Open access pathways, 14 and here we investigated if this strategy rescues endogenous class I restricted anti-viral TCR function. We used a CD8-dependent TCR targeting the tumor-associated antigen survivin in the context of HLA-A*02:01 and expressed the TCR alone (TCR) 15 or in combination with CD8αβ (TCR8) 14 in VSTs targeting cytomegalovirus (CMV) and Epstein-Barr virus (EBV). We consistently generated TCR+ and TCR8+ VSTs with a predominant central memory phenotype and showed that anti-viral reactivities were restored in TCR8+ VSTs while anti-tumor function was retained.

MATERIALS AND METHODS
Cell lines BV173 and K562 cell lines were obtained from the German Cell Culture Collection (DSMZ) and the American Type Culture Collection (ATCC), respectively, and maintained in complete RPMI 1640 media (Hyclone; Thermo Scientific) supplemented with 10% or 20% fetal bovine serum (FBS, Hyclone), 1% penicillinstreptomycin (Gibco), and 1% glutamax (Gibco). Two hundred and ninety-three T cells (ATCC) were maintained in complete IMDM media (Hyclone) containing 10% FBS, 1% penicillin-streptomycin, and 1% glutamax. For the mouse xenograft experiments, the previously described BV173.ffLuc cell line was used. 15 Blood samples from healthy donors Buffy coats were obtained from CMV seropositive de-identified healthy human volunteers at the Gulf Coast Regional Blood Center (Houston, Texas, USA). HLA-A2 status was assessed by fluorescence-activated cell sorting (FACS) analysis and HLA-A2+ donors were selected for the experiments.
Co-culture assay and cytokine detection VSTs and BV173 cells were co-cultured at E:T ratio 1:5 in the absence of exogenous cytokines. Co-culture Open access supernatants were harvested 24 hours after plating and stored at −80°C for cytokine analysis. Residual VSTs and tumor cells were enumerated after 3 days using CountBright Beads (Life Technologies) and FACS analysis. Cytokines were quantified in supernatants using the MILLIPLEX Human CD8+ T cells magnetic beads panel (EMD Millipore) and a Luminex 200 instrument (Luminex).

51
Chromium release assay To assess the short term in vitro cytotoxic function of VSTs, a standard 51 Chromium ( 51 Cr) release assay was performed. 15 Autologous activated T cells were used as targets, pulsed with the indicated peptides or pepmixes and labeled with 51 Cr for 1 hour. Effector and target cells were incubated at various ratios. As controls, target cells were incubated in media alone or with 1% Triton-X 100 (Sigma-Aldrich) for 4 hours to determine the spontaneous and the maximum release. The mean percentage of specific lysis of triplicate wells was calculated as follows: ((test counts -spontaneous counts)/ (maximum counts − spontaneous counts))×100%.

Mouse xenograft model
Female NOD-SCID-γc −/− (NSG) mice (6 to 8 weeks old) were purchased from the Jackson Laboratory and housed at the Baylor College of Medicine Animal Facility. Mice were irradiated with 120 cGy and infused with 3×10 6 BV173.ffLuc cells/mouse through the tail vein 4 to 6 hours later. Leukemia burden was monitored by bioluminescent imaging (BLI) (photons/s/cm 2 /sr) using the Xenogen in vivo imaging system (Caliper Life Sciences). Two VST injections (2×10 6 /mouse, 3 days apart) were administered through the tail vein or the retro-orbital vein plexus beginning 24 hours after tumor injection. Leukemia growth was monitored weekly by BLI and survival recorded.

Statistical analysis
Data were summarized using descriptive statistics. Comparisons between groups were made using student's t-test, one-way or two-way analysis of variance, and Friedman or Wilcoxon test, whichever was appropriate (figure legends). Area under the curves of bioluminescent signal intensity were calculated in GraphPad Prism from day 0 to day 28 and differences between groups were analyzed with the Mann-Whitney test. Survival of mice was analyzed by Kaplan-Meier graphs and differences were analyzed with the log-rank test. GraphPad prism 6 (GraphPad software, La Jolla, California) or higher was used for statistical analysis. P values <0.05 were considered statistically significant.

Study approval
Animal studies were reviewed and approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine.

Transgenic expression of CD8αβ in VSTs enhances recognition of viral epitopes
To assess whether transgenic expression of CD8αβ enhances endogenous virus-specific TCR detection and function, we analyzed NLV pentamer staining and NLVspecific IFN-γ production of NT VSTs and CD8αβ+ VSTs. Expression of transgenic CD8αβ enhanced the detection of NLV pentamer+ VSTs at the end of the first stimulation, but was not statistically significant (%NLV+ VSTs in NT versus CD8αβ+: 3.1±4.4% versus 12.9±15%, p=ns, mean±SD, n=5) (figure 3A, left and middle panel). We also detected a trend to increased CD8α MFI in four out of five donors tested (figure 3A, right panel) but this did not reach conventional levels of significance. However, the MFI of the NLV pentamer staining was significantly higher in CD8αβ+ VSTs compared with NT VSTs (NLV MFI NT versus CD8αβ+: 1018±438 vs 2520±948, **p=0.005, mean±SD, n=5) (figure 3B). These findings are consistent with previous reports assessing the influence of CD8 co-receptor availability on major histocompatibilty complex (MHC) class I multimer stainings. 21 With regards to NLV-specific T-cell responses determined by IFN-γ ELISpot, we found a significant increase of IFN-γ spot forming cells in CD8αβ+ VSTs compared with NT VSTs (653±246 vs 835±248, *p=0.04, mean±SD, n=3, plated in duplicates or triplicates) in response to the NLV peptide ( figure 3C). Hence, forced expression of CD8αβ in VSTs enhances endogenous TCR function.

TCR+ and TCR8+ VSTs kill viral targets in vitro
We next assessed the cytolytic activity of NT, TCR+, and TCR8+ VSTs against viral antigen-presenting target cells. In a 4-hour 51 Cr-release assay against peptide/pepmixpulsed activated autologous T cells, we detected specific lysis of target cells pulsed either with NLV peptide or the pooled CMV pepmixes/peptides or the pooled EBV pepmixes/peptides (solid lines), while unpulsed target cells (dashed lines) were not killed (figure 4A, mean±SD, n=4, plated in duplicates or triplicates). Survivin LML peptide pulsed targets were killed by TCR+ and TCR8+ VSTs but not NT VSTs. Overall, we observed a trend for lower cytolytic activity against viral antigen-presenting targets by TCR+ VSTs compared with TCR8+ VSTs or NT VSTs, but the overall differences were not statistically significant due to high donor variability. Donor variability is illustrated in figure 4B where we provide a summary of specific target cell lysis obtained at the E:T ratio of 20:1, depicting individual dots per donor and average bars. Some donors lysed targets to a similar extent with TCR+ or TCR8+ VSTs (eg, donor #3, cyan), while others showed a significant drop when comparing NT versus TCR+ VSTs, and a partial restoration of the activity with TCR8+ VSTs (eg, donor #14, green). Overall, we found a significant enhancement of CMV-specific target cell lysis (both NLV peptide and CMV pool) with TCR8+ VSTs compared with TCR+ VSTs, while the analysis for the EBV pool did not reach significance. One donor had a weak baseline CMV and a strong EBV response (donor #9, purple), while the others had strong baseline responses against both viruses. Killing of LML peptide-pulsed target cells was consistent with both TCR+ VSTs and TCR8+ VSTs, and unpulsed targets were not killed.

DISCUSSION
We show that transgenic CD8αβ alone or in combination with a transgenic TCR promotes endogenous TCR function in VSTs expressing a second transgenic TCR. TCR8 expression in VSTs rescues endogenous anti-viral activity to levels comparable to non-transduced VSTs while maintaining anti-tumor function mediated by the transgenic TCR. These benefits are observed both in vitro and in vivo. Thus, our approach overcomes the competition for essential signaling components that leads to The concept of providing simultaneous anti-viral and anti-tumor function in one cell therapy product was clinically validated when autologous CAR modified VSTs were shown to be able to target both EBV and GD2 in patients with neuroblastoma. 2 After adoptive transfer, GD2-CAR+ EBVSTs demonstrated safety, anti-tumor function, and long-term persistence in patients. 2 3 Since then, the approach was extended to CD19-CARs expressed in allogeneic stem cell donor-derived VSTs and infused to children with high risk B-cell acute lymphoblastic leukemia in remission after allogeneic HSCT, [4][5][6] in whom in vivo expansion of CD19-CAR+ VSTs during viral reactivation was observed. 6 Similarly, vaccination of subjects with varicella zoster virus may be another approach that augments expansion of VSTs expressing a GD2 CAR in patients with sarcoma or neuroblastoma (NCT01953900). 6 Thus, in principle, stimulation through the endogenous virus-specific TCR and co-stimulation on antigenpresenting cells in vivo can produce significant re-expansion and function of infused CAR+ VSTs.
It is more challenging to extend the VST platform to target tumor antigens by means of transgenic TCRs rather than CARs. Endogenous and transgenic TCRs compete for the same TCR signaling complex components (including CD3 chains and the CD8αβ co-receptor), and introduced TCR chains may mis-pair with endogenous TCR chains, leading to an overall reduced anti-viral activity. [7][8][9][10][11][12] The use of murine-constant regions in the transgenic TCR can significantly reduce mispairing. 13 Transfer of CD8αβ can redirect CD4+ T cells to class I-restricted antigens, and the transgenic CD4+ T cells become cytotoxic hybrid cells. 14 22-27 We have recently analyzed the transcriptional consequences of TCR8 expression in polyclonal CD4+ and CD8+ T cells by single cell RNA sequencing and experimental validation and showed that transgenic TCR8 has multiple advantages in both CD4+ and CD8+ lineages, promoting  Open access an overall enhanced anti-tumor function. 14 We now investigated the impact of transgenic CD8αβ on endogenous class I-restricted TCR function in CD8+ T cells as we hypothesized that combining CD8αβ co-receptor as a transgene with a tumor-targeted TCR is one potential means of overcoming the limitation of reduced anti-viral activity in TCR+ VSTs. We first confirmed that anti-viral specificity and activity was reduced when our survivinspecific TCR is expressed in VSTs. Next, we demonstrated that TCR8 rescued endogenous anti-viral TCR specificity and activity, using pentamer staining, IFN-γ ELISpot, degranulation assay and ICS, as well as cytotoxicity assays. TCR8+ VSTs retained specificity and anti-viral activity that was comparable to NT VSTs. We also showed that the transgenic CD8αβ effect on endogenous TCR function was preserved in vitro even in the absence of a transgenic TCR. These findings indicate that (i) TCR8+ VSTs respond to viral infections to a comparable level as NT VSTs, and (ii) TCR8+ VSTs are ready to re-expand on viral challenge. Importantly, the in vivo anti-tumor function of TCR8+ VSTs was retained. However, the in vivo anti-viral activity of TCR8+ VSTs remains to be validated, an analysis that will be a component of ongoing and proposed human studies. 2-4 6 Several features of the CD8αβ co-receptor most likely explain our findings, including its role in TCR-pMHC recognition and modulation of antigensensitivity, 28 recruitment of Lck to the immune synapse, and activation of signaling components that are crucial during early T-cell activation, 29 and, as recently demonstrated by single cell RNA sequencing, its impact on a variety of transcriptional pathways upon tumor challenge that support enhanced anti-tumor function. 14 An additional advantage of VSTs as a platform for the delivery of tumor-antigen-specific T-cell function to treat cancer patients lies in the fact that transgenic VSTs can be generated from well-characterized healthy donors for allogeneic cell banks. Indeed, third party donor allogeneic 'off-the-shelf' banked VSTs have been clinically validated for the treatment of viral infections or EBVassociated lymphoproliferation in immunocompromised patients after HSCT or solid organ transplant. 18 30-33 Third party donor VSTs could be engineered with TCRs or CARs, banked, and thereby provide a readily available 'off-the shelf' product that does not require genome editing for safe infusion in immunocompromised patients. 34 Furthermore, additional engineering can be incorporated to protect VSTs from host NK and/or T-cell mediated allograft rejection, furthering their potential to become a safe universal 'off-the shelf' T-cell product with long in vivo persistence. 35 36 Stimulation of T cells with viral peptide-pulsed autologous DCs followed by retroviral transduction yielded TCR+ or TCR8+ VST lines consistently composed of CD4+ and CD8+ VSTs with a predominant central memory phenotype and high transduction efficiencies. TCR8 expression was associated with a better preservation of less differentiated VSTs in the CD4+ compartment. 14 The only HLA restriction of the product is defined by the transgenic TCR used to redirect VSTs to the cancer. Feasibility and safety of our manufacturing have already been demonstrated in CAR+ VST clinical trials and are GMP compliant. 2-4 6 In summary, we show that transgenic CD8αβ rescues endogenous anti-viral TCR function in VSTs when combined with a tumor-targeted transgenic TCR. Our TCR8+ VST product provides significant anti-tumor function in vivo. We propose that the use of a combined TCR8 vector may better maintain in vivo long-term benefits of VSTs than cells that lack this modification. Our clinically validated VST platform is amenable to genetic engineering with TCR8 and ready to clinically assess the safety and efficacy to prevent and treat viral infection and malignant relapse after allogeneic HSCT.