Clinical-grade generation of peptide-stimulated CMV/EBV-specific T cells from G-CSF mobilized stem cell grafts

Reactivation of cytomegalovirus (CMV) and Epstein–Barr virus (EBV) worsens outcomes of allogeneic stem cell transplantation (aSCT) and remains a major obstacle to its success [1]. Within the first 100 days after aSCT, 40–50% of patients reactivate CMV, and up to 40% of all patients reactivate EBV after aSCT as determined by virus-specific PCR of cells of the peripheral blood (PB). Approximately 95% of donors and patients are seropositive for EBV, and 40–70% for CMV [2]. Both CMV and EBV reactivation after aSCT are associated with increased mortality. Reactivation of EBV bears the risk of EBV-associated post-transplantation lymphoproliferative disease [3]. Reactivation of CMV can cause pneumonia with high mortality. Therefore both viruses require preemptive treatment upon reactivation in patients after aSCT [4].

Specific antiviral therapy is only available for the treatment of CMV. However, all drugs available (Ganciclovir, Foscarnet, Cidofovir, and others) display strong side effects including bone marrow and kidney failure. Furthermore, they frequently require inpatient treatment thereby compromising quality of life and most importantly do not solve the underlying problem of missing immunological control. For EBV, no approved specific therapeutic option exists. Off-label use of Rituximab, a B-cell depleting antibody, is increasing and seems to be effective [5‐7]. However, Rituximab induces long lasting B-cell depletion resulting in frequent and obligatory transfusion of immunoglobulins. Similarly to the treatment of CMV, the fundamental problem of the lack of immunological control is not addressed with this therapy. As all antiviral therapies fail to boost the immune system, relapse of reactivation is frequent and repeated treatments are required, strongly contributing to the high costs of aSCT.

The rationale of strengthening specific T-cell immunity for both prevention and therapy of CMV and EBV reactivation therefore represents an intriguing therapeutic option. Several groups have shown that CMV- or EBV-specific T cells can be isolated or enriched from seropositive donors, and mediate viral control in aSCT patients after adoptive transfer [8‐14]. Depending on the method of isolation, virus-specific T cells are only available in a minority of donor-patient pairs, their specificity is limited to single viral antigens or epitopes, or their preparation may be inconveniently long and laborious. Here, we describe a clinical grade protocol for manufacturing multi-epitope CMV/EBV-specific T cells suitable for application after aSCT. We use a generic set of peptides representing dominant CMV and EBV CD8+ and CD4+ T-cell epitopes from different viral antigens of each virus, presented by different HLA allotypes. Thus, this protocol is applicable in more than 80% of European donors, and has a high likelihood to enrich their dominant virus-specific T-cell populations. We applied this procedure to G-CSF mobilized stem cell grafts and non-mobilized apheresis products and show that it is equally effective in the relative expansion of CMV/EBV-specific T cells. As a result, CMV/EBV-specific T cells are available shortly after transplantation within 14 days (plus the time required for microbial safety monitoring) if G-CSF mobilized stem cells are used as a T-cell source, avoiding a second apheresis of the donor. The protocol is easily applicable within clean room facilities and can be modified according to preferences of the manufacturer. This manufacturing protocol is currently used in an ongoing phase I/IIa clinical trial for prevention of CMV/EBV-reactivation after aSCT (EudraCT 2012-004240-30).

We established a GMP-compliant process for manufacturing of CMV/EBV-specific T cells by peptide stimulation of leukapheresis products. The workflow is outlined in Fig. 1. Details of manufacturing are provided in Materials and Methods. We investigated this process with a particular focus on stimulation of T cells from leukapheresis products from G-CSF mobilized donors to ensure availability of the product as early as possible after transplant and, for donor safety, to avoid a second apheresis procedure.

In this study we demonstrate that CMV- and EBV-specific T cells can be successfully enriched from CMV- and EBV-seropositive G-CSF mobilized donors by peptide stimulation in a GMP-compliant process, and specific T cells can be quickly obtained to allow for preventive application in patients after aSCT. Donor lymphocytes are stimulated with pools of defined viral peptides that represent dominant T-cell epitopes from various viral antigens with a range of HLA restrictions. A small portion of the G-CSF mobilized stem cell graft is sufficient as a source of T cells, eliminating the need for a second apheresis of the donor. Peptide stimulation resulted in expansion of T cells and increase of specific T cells, while reducing the proportion of naive T cells.

Several strategies to generate or isolate CMV- and/or EBV- specific T cells for adoptive transfer are currently available. Direct isolation of specific T cells without any period of cultivation generally requires the use of recombinant HLA/peptide multimers and magnetic or flow-cytometric sorting [22, 23]. This method is elegant and allows for high cell purity, but its routine clinical application is challenging since it cannot be used for all HLAs and epitopes, and a continuous stock of various manufacturing-grade multimer reagents and a separation device are needed. Antigen stimulation and subsequent isolation based on T-cell reactivity, e.g. IFN-γ secretion or CD137 up-regulation [24‐28] is more versatile in terms of the T-cell specificities that can be included and can be performed within one or 2 days, but T-cell yield may be low and thus characterization of the product is often limited. Therefore, simplified approaches that take advantage of antigen-specific expansion of T cells in vitro by peptide stimulation have again attracted interest and have been explored in the clinical setting with promising results [29, 30]. One protocol [31] uses a general procedure similar to ours, but employs overlapping peptide pools covering 12 antigens from five viruses including EBV and CMV, employing approximately 2000 peptides in total, which is certainly a strong advance toward optimizing the breadth of the achievable response. In designing the present protocol, in which we used a total of 34 well-characterized peptides, 10 HLA class-I-restricted and 7 class-II-restricted peptides from each of the two viruses, we have considered it advantageous to strike a middle ground between simplicity and complexity, based on the following considerations. Our exclusive use of defined peptides with known HLA class I and II restrictions will make it possible to use HLA class I multimers (and potentially class II multimers) for a comprehensive quantitative analysis of the specific T cells in the product, and to track them in patients. Since we avoid the use of thousands of potentially irrelevant peptides, we decrease the probability that T cells with unwanted specificities that accidentally cross-react with some of these peptides will be activated and expanded. Moreover, it is very likely that our specificities of choice will be more efficiently activated if the number of potentially competing peptides in the loading reaction is limited. Still, it is important to include peptides from a range of antigens from different phases of viral replication. For CMV, it is far from clear which T-cell antigens provide the best protection from viral reactivation [28, 32, 33], and the situation for EBV is complex as well. Our choice of a limited number of epitopes and HLA restrictions will nevertheless ensure that a majority of donor/patient pairs can be covered. The idea that the use of a full-length viral antigen (or antigen-covering peptide library) is HLA-independent may be illusory, since a given viral protein will contain only epitopes presented by a finite, sometimes rather limited, number of HLA allotypes.

The presence of some remaining naïve T cells of unknown specificity in our T-cell cultures may be a safety issue in patients after aSCT, since alloreactive T cells are mainly found in the naïve compartment [34]. However, adoptive transfer of peptide-stimulated virus-specific T cells with isolation [25, 28] or without isolation [29, 30] was proven safe so far. Immunodominant peptides with known allogeneic cross-reactivity such as the EBV-derived FLR [35] were deliberately excluded from the peptide mix to minimize the risk of acute GvHD after adoptive transfer.

Furthermore, we anticipate using our T-cell product for prevention of viral reactivation at an early stage after transplantation, when most patients are still receiving calcineurin inhibitors such as cyclosporine A for acute GvHD prophylaxis. There is evidence that activated T cells become cyclosporine A resistant and therefore can exert their function in vivo despite ongoing immunosuppression [36].

A recent study described that the functionality of antiviral T cells from G-CSF mobilized stem cell grafts or after G-CSF treatment in vitro is reduced [37]. This is consistent with our observation of reduced cytokine secretion in early stages of the T-cell culture (data not shown). Furthermore, Rutella et al. described that serum of G-CSF treated healthy donors reduced lymphocyte proliferation but this effect was overcome by addition of IL-2 [38]. However, after completion of our stimulation protocol, our data show no evidence for impairment of virus-specific T cells due to the use of material from G-CSF mobilized donors. Nawa et al. [39] described a reduced stimulatory potential of monocytes from G-CSF mobilized donors in mixed lymphocyte reactions, potentially indicating reduced potential of such monocytes to stimulate T-cell responses; however, they described that T-cell function was not directly altered by G-CSF mobilization. Others described that the cell count of dendritic cells per µl PB is increased after mobilization [15, 16] bearing the advantage that these dendritic cells as APCs may contribute to effective expansion of antigen-specific T cells in protocols such as ours. During our cultivation process, however, G-CSF is absent, and we observed that 9-day cultures from G-CSF mobilized donors resulted in equal relative expansion of virus-specific T cells in comparison to cultures from non-mobilized donors, and T-cell function, as measured in ELISpot, was even better for mobilized donors. In line with our results, Samuel et al. showed that a successful expansion of CMV-reactive T cells from G-CSF mobilized donors is possible [40]. In addition, Chevallier et al. reported no significant differences in the composition of T-cell subsets of peripheral blood of G-CSF mobilized and non-mobilized donors [16]. The reduction of naïve T cells in the products of both donor types is encouraging, as the naive T-cell compartment is considered the main source of alloreactive T cells [34].

One major advantage of this manufacturing protocol is its cost efficiency. First, in comparison to current standard therapies of CMV or EBV reactivation (ganciclovir, rituximab, etc.), specific T-cell therapy in general is a competitive option in terms of costs. Second, among possible protocols to prepare specific T cells, the present one is particularly economic, since it does not depend on expensive proprietary technology for specific cell cultivation or isolation, but uses only general-purpose machinery that has multiple uses in transfusion medicine departments. Multiple doses can be produced in one manufacturing run, and can be stored for repeated application if necessary. Our protocol is versatile, and will allow for including peptides from additional viral or other microbial pathogens without any changes in the workflow.

We are currently conducting a phase I/IIa clinical trial exploring a preventive strategy using this T-cell product in patients after aSCT (Eudract 2012-004240-30). Peptide-stimulated T cells are stored over liquid nitrogen and administered at a dose of 5 × 10⁴/kg CD3+ cells/kg body weight on days 30, 60, and 90 after aSCT. Aside from safety we aim at demonstrating efficacy of this T-cell product. In a more long-term perspective, we are confident that simplified protocols to prepare CMV/EBV-specific T-cells for adoptive therapy after aSCT will help make this therapy more widely available, especially for prophylactic intervention to prevent reactivation and disease caused by these widespread viruses. We believe this may decrease virus reactivation rates, thereby reducing hospitalization time and need for medication prone to side effects, and improving quality of life for patients after aSCT.