Expanding multiviral reactive T cells from cord blood

Unrelated umbilical cord blood transplantation (CBT) has been used with increased frequency and success in both children and adults with malignant and nonmalignant hematologic diseases.²  Cord blood (CB) grafts contain relatively low numbers of T cells that have a naive phenotype, potentially accounting for their low incidence of acute and chronic graft-versus-host disease despite HLA mismatching. Unfortunately, these factors also delay cellular immune reconstitution, increasing the risk of morbidity and mortality from viral pathogens.^3,4 

Epstein-Barr virus (EBV), cytomegalovirus (CMV) and adenoviral infection contribute significantly to the morbidity and mortality associated with CBTs.^3,5  Medical management of these infections is challenging, because antiviral drug therapy is often ineffective and can be associated with significant toxicities. Adenoviral infections are particularly problematic and occur more commonly following CBT compared with marrow or peripheral blood stem cell transplants.⁶  Currently, there is no effective drug therapy for disseminated adenoviral disease, a frequently fatal complication of CBT that resolves only after an effective antiviral T-cell response has developed. Therefore, the development of methods to rapidly bolster T-cell immunity against viral reactivation and infection are urgently needed to reduce the mortality associated with CBT.

Leen and colleagues previously developed an in vitro method to simultaneously expand viral reactive CD4⁺ and CD8⁺ T cells against EBV, CMV, and adenovirus using antigen- presenting cells infected with an adenoviral vector encoding the immunodominant CMV-pp65 matrix protein.⁶  Using this technique, “triviral” reactive cytotoxic T lymphocytes (CTLs) were expanded from memory T cells of donors previously infected with these viruses. Importantly, when prophylactically transferred into patients undergoing T cell–depleted HCT, these CTLs had functional antiviral activity in vivo, decreasing the risk of viral reactivation and the need for antiviral therapy.

Although the expansion of viral reactive CTLs in vitro from memory T cells previously primed in vivo to viral antigens is relatively straightforward, the in vitro generation of antigen-reactive CTLs from naive T cells is more challenging.⁷  By modifying the antigen-presenting cells used to stimulate T cells as well as the cytokines contained within the culture media, in this issue of Blood, Hanley et al have developed a technique to rapidly expand from naive CB T cells, polyclonal CD4⁺ and CD8⁺ CTLs with antigen specificity for CMV, EBV, and adenovirus, the 3 most common viral infections complicating CBT.

To accomplish this feat, naive CB T cells were initially stimulated with mature Ad5f35pp65-infected dendritic cells before undergoing repeated stimulation with autologous Ad5f35pp65-infected EBV lymphoblastic cell lines (EBV-LCLs). Notably, viral reactive CTL could be generated from naive T cells only when culture media was supplemented with IL-7, IL-12, and IL-15, cytokines that are not typically required for the in vitro expansion of viral reactive CTLs from memory T cells. Using cell sorting, the authors show that viral reactive CTLs expanded from dominant naive rather than minority memory T cells contained within the CB unit.

Starting with 40 million CB nucleated cells, a median of 85.5 × 10⁶ T cells were harvested by day 21 of cell culture, far more cells than the minimum 10⁷/m² of triviral reactive CTLs previously shown necessary to restore viral immunity to CMV, EBV, and adenovirus following adoptive transfer after T cell–depleted HCT.⁶  Notably, viral reactive CTLs derived from CB had specificity for CMV and EBV antigen epitopes that were distinct from peripheral blood CTLs, lacking T cells that recognized the immunodominant epitopes targeted by peripheral blood–derived antiviral CTLs. In contrast, CB-expanded CTLs maintained the same epitope specificity as peripheral blood–derived CTLs for adenoviral-derived antigens. Despite these differences, in vitro studies showed CB CTLs had significant cytolytic function and IFN-γ secretion against PHA blasts pulsed with CMV peptide and adenoviral hexon peptide pools as well as against autologous EBV-LCLs. These data suggest that these CTLs have the capacity to recognize and kill viral-infected cells expressing CMV, EBV, and adenoviral antigens. Furthermore, although they ultimately acquired an effector memory phenotype after in vitro expansion, the use of CTLs expanded from naive T cells rather than from memory cells could potentially mediate more effective antigen-specific immunity in vivo (C. Hinrichs, verbal communication, June 22, 2009).

Although these results are promising, the expansion technique used here would require customized CTLs be generated for each cord unit and would also require 3 to 5 weeks to establish an EBV-LCL line from CB cells to stimulate viral reactive T cells following Ad5f35pp65 infection. Therefore, 35 to 50 days would be required to expand viral reactive CTLs following thawing and presumably transplantation of the primary cord unit, which is problematic considering that a significant portion of patients reactivate or develop viral infections early after CBT (ie, before day 35). Expanding viral reactive CTLs from a smaller portion of the cord unit in advance of thawing the main body of the cord unit used for the primary transplantation could potentially overcome this limitation. Notwithstanding these obstacles, the development of culture techniques used by Hanley et al to expand viral reactive CTLs from naive cord blood T cells represents a significant technological advance for the field of cellular therapy that could be used to reduce the morbidity and mortality associated with CBT.