Introduction
Epstein Barr virus (EBV) is a human herpes virus with a prevalence of
over 90% in adults [1]. Acute infection is generally asymptomatic,
but can present clinically as infectious mononucleosis (glandular fever)
and infection persists latently throughout life in B lymphocytes.
Intermittent viral reactivation drives proliferation of infected cells,
which can transform to malignant lymphoblasts in the absence of immune
surveillance. In immunocompetent individuals, EBV-specific T cells
control these reactivations [2]. However, iatrogenic
immunosuppression of patients following haematopoietic stem cell
transplant (HSCT) or solid organ transplant (SOT) can result in
potentially fatal EBV-driven post-transplant lymphoproliferative
disorder (PTLD), where EBV-transformed B cells develop into an
aggressive B cell lymphoma [3].
Adoptive immunotherapy using Epstein-Barr Virus (EBV)-specific T cell
lines has proven to be an effective clinical treatment for patients with
rituximab-resistant or refractory PTLD [4]. The Scottish National
Blood Transfusion Service (SNBTS) EBV-specific T cell bank has delivered
third-party partially HLA-matched EBV-specific T cells for PTLD patients
for over 15 years, allowing a rapid therapeutic intervention following
diagnosis, with an initial 52% complete response rate in a phase 2
multicentre trial [5]. As of December 2020, over 100 SOT and HSCT
patients have been treated with the more recent second-generation SNBTS
T cell bank, issued under a Medicines and Healthcare Regulatory Agency
(MHRA) Manufacturing Specials licence for therapeutic use, and a recent
follow-up study of 64 patients with refractory disease treated between
2011 and 2017 indicates that overall survival at 3 years post treatment
was more than 40%, and higher (62%) in SOT patients [6]. This
corresponds well with other clinical studies using adoptive therapy with
EBV-specific T cells, where overall response rates were 63-68%
[7-9].
Clinical manufacture of allogeneic EBV-specific T cells for therapy
requires a generation protocol that conforms to current good
Manufacturing Process (cGMP) regulations. Manufacture of the current
SNBTS second-generation SNBTS T cell bank was undertaken using a
conventional process for generating EBV-specific T cells, using
mononuclear cells (MNC) co-cultured over 6-8 weeks with autologous
EBV-transformed lymphoblastoid cell lines (LCL). The LCL act as
antigen-presenting cells (APC) to induce proliferation of T cells
recognizing EBV viral proteins such as BZLF1 and BMLF1 involved in early
lytic cycle transactivation, and viral latency proteins such as EBNA1,
EBNA2, EBNA3, LMP1 and LMP2 [10,11]. Multiple rounds of stimulation
with irradiated autologous LCL are required to induce EBV-specific T
cell expansion to a suitable therapeutic dose [12]. The introduction
of rapid expansion protocols in development of new T cell therapies
[13] indicates a need for both robust, rapid methods for
EBV-specific T cell manufacture, and improved analytic methods to assess
the quality of the material throughout.
Replacement of standard flask culture with gas-permeable rapid expansion
devices (G-Rex flasks, Wilson Wolf) is now established for T cell
culture, and T cell lines are generally initiated in culture with a
ratio of 40 MNC : 1 LCL for the first stimulation round. Subsequent
stimulation rounds are generally at a 1 T cell : 5 LCL ratio in rapid
expansion protocols [14], though variations in this approach are
evident between studies [15-18]. As part of this study, we
determined the optimal number of stimulation rounds and optimized our
previous manufacture protocol by transfer to EU GMP-compliant culture
reagents (medium and cytokine) and G-Rex flask culture throughout. We
confirmed appropriate T cell: LCL stimulation ratio parameters for
culture and in each case utilised an extensive multi-parameter flow
cytometric analysis approach to dissect the composition of the cultures.
The critical quality attributes of the SNBTS therapeutic EBV-specific T
cells products are stringent and include viability, T cell lineage, and
absence of other lymphocytes; for this process optimization study we
introduced a broader flow cytometric analysis panel which allowed
characterization of the T cell compartment to a much higher level of
discrimination. The new panels were able to determine T cell
development, differentiation and activation status, and provide markers
of efficacy through intracellular cytokine analysis. We utilized these
panels to determine the efficacy of the modified protocol in comparison
with the current process, and assess quality from starting mononuclear
cells through to final T cell product. Visualizing this data is complex,
and we applied t-Stochastic Neighbour Embedding (t-SNE) algorithms to
incorporate multiple parameters into a single image, allowing visual
comparison of the T cell phenotype. This approach provides clear
evidence for an improved process for manufacture of virus-specific T
cells to a standard suitable for clinical use.