Cytotoxic T lymphocytes as immune-therapy in haematological practice

Authors

  • Ann M. Leen,

    1. Center for Cell and Gene Therapy, Baylor College of Medicine, The Methodist Hospital, and Texas Children’s Hospital, Houston, TX, USA
    Search for more papers by this author
  • Helen E. Heslop

    1. Center for Cell and Gene Therapy, Baylor College of Medicine, The Methodist Hospital, and Texas Children’s Hospital, Houston, TX, USA
    Search for more papers by this author

Ann M. Leen, Center for Cell and Gene Therapy, Baylor College of Medicine, 1102 Bates Street, 760.01, Houston, TX 77030, USA. E-mail: amleen@txccc.org

Summary

Viral infections are a significant cause of morbidity and mortality, particularly in pediatric allogeneic haematopoietic stem cell transplant recipients. Effective therapies are limited and often associated with significant side effects. Adoptive transfer of virus-reactive T cells offers a means of reconstituting antiviral immunity and this approach has been successfully used to prevent and treat cytomegalovirus, Epstein-Barr virus, and adenovirus infections in vivo. This review outlines the clinical trials that have been performed to date, and will describe future initiatives to (a) develop strategies that can increase the breadth of the viruses that can be targeted, and (b) simplify the process to extend this technology to more centers so that cellular therapy to reconstitute immunity can be more widely applied.

Hematopoietic stem cell transplantation (HSCT) has evolved into an effective strategy for the treatment of a number of haematological malignancies and many non-malignant disorders. However, successful transplantation of allogeneic hematopoietic stem cells results in immunosuppression until the donor-derived immune system reconstitutes and the duration and severity depends on the type of transplant. As a result, recipients of hemopoietic stem cells are susceptible to a wide array of serious and often lethal opportunistic infections, many of which are not amenable to conventional small-molecule therapeutics.

Increasing numbers of viral pathogens have been implicated in infectious complications after HSCT, partly due to improved diagnostic techniques, increased surveillance, and the discovery of new viruses, and partly as the result of the extension of SCT to higher risk patients who either receive more extensively manipulated products or who require more intensive and prolonged post-transplant immunosuppression. Infections caused by Epstein-Barr virus (EBV), cytomegalovirus (CMV), and herpes simplex virus (HSV), as well as by respiratory viruses, such as Respiratory Syncitial Virus (RSV), parainfluenza, and influenza are well known, while the importance of infections caused by adenovirus (Adv), BK virus, human herpesvirus (HHV)-6, and metapneumoviruses has been more recently appreciated (Flomenberg et al, 1994; Boeckh et al, 2003, 2005; Leen & Rooney, 2005; Martino et al, 2005; Myers et al, 2005; Zerr et al, 2005; Egli et al, 2007; Ison, 2007; Peck et al, 2007; Giraud et al, 2008). Antiviral drugs, which are standard therapy for some infections, are costly short-term control measures that are frequently ineffective and toxic and can induce drug-resistance. In contrast, treatment of the underlying problem, namely lack of antigen-specific T cells to control viral reactivations/infections, can offer effective and long-term protective treatment. This consideration has led to development of therapies designed to reconstitute SCT recipients with protective donor-derived virus-specific T cells.

One approach to prevent and treat these opportunistic infections in the SCT setting involves the use of donor leucocyte infusions (DLI), which consist of unmanipulated T cells isolated from the stem cell donor. But although DLI contain virus-specific T-cell precursors with the potential to protect recipients against infections (Hromas et al, 1994; Papadopoulos et al, 1994), their efficacy is limited by the low frequency of specific T cells to many common “acute” viruses (such as RSV, Adv, and parainfluenza), which are increasingly recognized as a substantial cause of morbidity and mortality after SCT compared to the frequency of alloreactive T cells. Such therapy is therefore limited by an unacceptably high risk of graft-versus-host disease (GvHD) (Heslop et al, 1994; MacKinnon et al, 1995).

Consequently, a number of groups have developed strategies to reconstitute virus-specific T cells post-transplant without inducing alloreactivity and these can be divided into two categories; (i) depletion of alloreactive T cells from DLI or (ii) adoptive transfer of selected populations of virus-specific T cells.

Depletion of alloreactive T cells

Initial attempts to reconstitute HSCT recipients with virus-specific T cells using unmodified DLI led to GvHD in a significant number of patients. One strategy to overcome this problem is to selectively tolerize, or specifically remove, recipient-specific alloreactive T cells from donor leucocyte populations thereby enriching for virus-specific cells in the infusion product.

A number of groups have focused on inducing alloantigen-specific immune tolerance of donor T cell populations (Blazar et al, 1998; Zeller et al, 1999), taking advantage of the inhibitory/suppressive characteristics associated with CD4+ CD25+ regulatory T cells (Treg). Human Tregs isolated from the peripheral blood have been shown to suppress alloresponses in mixed lymphocyte reactions (MLR) (Baecher-Allan et al, 2001; Jonuleit et al, 2001; Levings et al, 2001), and similar results have been achieved in the murine system (Taylor et al, 2001; Hoffmann et al, 2002; Kingsley et al, 2002).

To address whether Tregs that constitutively express FOXP3 play a pivotal role in the maintenance of tolerance and the suppression of GvHD in allogeneic HSCT recipients, Rezvani et al (2006) quantitated the coexpression of FOXP3 on CD4+ T cells in 32 donor SCTs infused into human leucocyte antigen (HLA)-matched siblings and examined the incidence of GvHD in recipients. They found that patients who received a SCT with low absolute numbers of FOXP3+ CD4+ T cells were at greater risk of developing GvHD. In addition, in 21 patients with haematological malignancies who received a T cell-depleted allogeneic SCT, they found that patients who developed GvHD had significantly fewer reconstituting FOXP3+ Tregs than those who did not develop GvHD. Thus, they proposed that the selective expansion and infusion of donor Tregs at the time of transplant could reduce the risk of GvHD, without affecting virus-specific T cell activity (Rezvani et al, 2006). To this end, Hoffmann et al (2004) have developed a system for specifically expanding large numbers of functional Tregs ex vivo using cross-linked anti-CD3 and anti-CD28 antibodies together with high dose interleukin 2; thus this approach is actively being investigated as a cell-based therapy to prevent or treat GvHD.

Protocols for the physical depletion of alloreactive T cells have also been investigated as a means to reduce/prevent GvHD in vivo. Most of the published methods rely on the stimulation of donor T cells with recipient-derived APCs to activate the alloreactive T cell component, followed by specific elimination of the activated cells using antimetabolite drugs, photodepletion, or using immunotoxins or magnetic beads to target differentially-expressed cell surface molecules. The majority of groups have depleted alloactivated cells based on their expression of activation markers, including CD69 (Hartwig et al, 2006), CD25 (Andre-Schmutz et al, 2002; Amrolia et al, 2003, 2006; Solomon et al, 2005), CD134 (O × 40) (Ge et al, 2008) and CD137 (41BB)(Wehler et al, 2007), which are upregulated on donor T cells in response to stimulation with recipient cells. Barrett and colleagues have pioneered an approach to selectively deplete alloreactive cells using a TH9402-based photodepletion technique. This process overcomes fluctuations in activation-based surface marker expression as the photodepletion process targets activation-based changes in p-glycoprotein that result in an altered efflux of the photosensitizer TH9402 (Solomon et al, 2005; Mielke et al, 2008). Using this approach, pathogen-specific responses directed against CMV, Adv, Varicella Zoster Virus, HSV, Aspergillus fumigatus, Candida albicans, and Toxoplasma gondii were retained, albeit with a reduced frequency (Perruccio et al, 2008). Alternatively, methotrexate, a US Federal Drug Administration-approved drug used for the therapy of neoplastic diseases, rheumatoid arthritis and psoriasis, has also been proposed as a potential agent of selective allodepletion, and preclinical results have been promising (Sathe et al, 2007).

The initial allodepletion-based clinical studies tested this strategy using a CD25 immunotoxin (IT) to deplete alloreactive T cells ex vivo (Andre-Schmutz et al, 2002; Solomon et al, 2002, 2005; Amrolia et al, 2003, 2006). In two trials, allodepleted cells were administered after haploidentical CD34-selected transplants and in the third after HLA identical sibling transplant (Andre-Schmutz et al, 2002; Solomon et al, 2005; Amrolia et al, 2006). Either recipient peripheral blood mononuclear cells or EBV-transformed lymphoblastoid cell lines were used as a source of alloantigen for stimulation. In all three studies administration of allodepleted T cells produced reconstitution of antiviral immunity without inducing GvHD. These studies indicate that the concept is feasible, and further studies to optimize the process and increase antitumor immunity are underway.

The limitations of current strategies include the time taken (4–6 weeks) to produce autologous EBV-transformed lymphoblastoid cell lines (EBV-LCL) when these cells are used as APC and the limited availability and instability of the clinical grade IT. In addition, as demonstrated in the dose escalation study by Amrolia and colleagues, a threshold number of T cells was required to produce reconstitution, so that accelerated reconstitution of EBV and CMV-specific T cells was seen only after infusion of 105 T cells/kg or more (Amrolia et al, 2006). Since the recovery of donor cells after allodepletion was approximately 10%, achieving sufficient cell numbers for infusion and for quality control assessments of potency, purity, and sterility often necessitated donor leukapheresis, which is not feasible for unrelated stem cell donors. Finally, T cells specific for a majority of pathogens circulate with much lower frequency than those specific for persistent viruses, such as EBV and CMV (Tan et al, 1999; Gillespie et al, 2000); therefore, even higher allodepleted T cell doses will probably be required to provide full spectrum protection.

Infusion of virus-specific cytotoxic T lymphocytes (CTL)

In vitro reactivation and expansion of virus-specific CTL

As viral complications in transplant recipients are clearly associated with the lack of virus-specific cellular immune responses (Cwynarski et al, 2001; Gottschalk et al, 2005), reconstitution of the host with in vitro expanded virus-specific CTLs is an attractive option to prevent and treat these diseases. In addition, infusion of in vitro expanded virus-specific T cells should be safe, because the prolonged expansion protocol needed to selectively increase the small numbers of virus-specific T cells present in peripheral blood mononuclear cells (PBMC) should also serve to eliminate residual alloreactive T cells from the final infusion product. However, the design of successful immunological strategies to treat human virus-associated diseases and malignancies requires an understanding of the effector processes that control viral infection and the mechanisms viruses use to evade such responses. To date, only a limited number of viruses have been sufficiently well characterized to allow targeting by CTL therapy (Table I).

Table I.   Clinical trials using virus-specific cytotoxic T cells post-HSCT.
Virus-specificityExpansion protocolInfused cellsPatients treatedClinical resultsReferences
  1. PTLD, post-transplant lymphoproliferative disorder; CR, complete response; PR, partial response.

CMVIn vitro expandedCD8+ T cell clones14 infused prophylacticallyNo CMV infections(Walter et al, 1995)
CMVIn vitro expandedPolyclonal CTL8 with infection5 cleared after 1 dose of CTL; 1 patient cleared after a 2nd dose; 1 did not clear(Einsele et al, 2002b)
CMVIn vitro expandedPolyclonal CTL16 with viremia8 cleared w/out antivirals; 8 cleared with antivirals; 2 reactivations(Peggs et al, 2003)
CMVTetramer selectionSelected CD8+ T cells9 with viremia or postreactivation9 patients had a decrease in viral load; 8 patients cleared the infection(Cobbold et al, 2005)
CMVIn vitro expandedPeptide-specific CD8+ T cells9 infused prophylactically6/9 showed a transient increase in tetramer-reactive T cells postinfusion; 2/9 cleared reactivation w/out antivirals(Micklethwaite et al, 2007)
EBVIn vitro expandedPolyclonal CTL10 (3 as treatment and 7 as prophylaxis)3/3 controlled infection(Rooney et al, 1995; Heslop et al, 1996)
EBVIn vitro expandedPolyclonal CTL39 prophylactically treatedNo patients developed PTLD in comparison with 11·5% in control group.(Rooney et al, 1998)
EBVIn vitro expandedPolyclonal CTL6 patients treated5 patients achieved complete remission, 1 patient died (tumor mutant was resistant to CTL)(Rooney et al, 1998; Gottschalk et al, 2001)
EBVIn vitro expandedPolyclonal CTL6 with high EBV viral load5/6 showed a transient/sustained reduction in viral load post-CTL. 1 patient subsequently died of PTLD.(Gustafsson et al, 2000)
EBVIn vitro expandedPolyclonal CTL1 patient treated for EBV-LPDPatient failed to respond to therapy(Imashuku et al, 1997)
EBVIn vitro expandedPolyclonal CTL3 patients with PTLD were treated with CTL3/3 CRs(Comoli et al, 2007)
EBVIn vitro expandedPartly-matched polyclonal CTL8 patients with progressive PTLD unresponsive to conventional treatment were treated3 patients achieved CR
2 patients had no response
3 patients did not complete treatment
(Haque et al, 2002)
EBVIn vitro expandedPartly-matched polyclonal CTL33 PTLD patients who had failed conventional therapy were enrolled14 patients had a CR
3 patients had a PR
16 patients had no response
(Haque et al, 2007)
AdenovirusIFN-γ selectionSelected polyclonal T cells9 with systemic infection5/6 evaluable patients had a decrease or complete clearance of infection(Feuchtinger et al, 2006)
Multivirus (EBV, CMV, Adv)In vitro expandedPolyclonal CTL11 infused (10 prophylactically; 1 treated for Adv infection)3/3 controlled CMV reactivation w/out antivirals; 3/3 cleared EBV infection/PTLD w/out antivirals; 5 patients with infection and 1 with disease cleared Adv post-CTL(Leen et al, 2006)

EBV-specific CTL

We and others have prepared donor-derived EBV CTL, whose infusion prevented and treated EBV-driven B cell lymphoproliferative diseases (EBV-LPD) after allogeneic HSCT (Rooney et al, 1995, 1998; Heslop et al, 1996; Gustafsson et al, 2000). Malignant B cells usually express the complete panel of EBV latent viral antigens, as well as abundant co-stimulatory molecules. Thus, they are highly immunogenic, and are eliminated by circulating EBV-specific CTL in healthy individuals. However, prior to the development of effective pharmacological agents, such as Rituximab which targets the EBV-infected B cells (Kuehnle et al, 2000), the incidence of EBV lymphoproliferation in patients receiving a T cell-depleted graft from an unrelated or HLA-mismatched, related donor, was high (ranging from 1% to 25%). This led our group and others to develop protocols for the in vitro generation of EBV-specific CTLs that could be used as prophylaxis and/or treatment in these high-risk groups.

Since 1993, our group has infused over 65 stem cell recipients with donor-derived EBV-specific T cell lines and established that a dose of 2 × 107 CTL/m2 was safe and efficacious for both prophylaxis and treatment. The methodology we developed, using EBV-LCL to repeatedly stimulate T cells, produced polyclonal CTL lines with CD4:CD8 ratios ranging from 2:98 to 98:2. The first 26 patients enrolled in this study received CTL that were genetically marked with a retroviral vector containing the neomycin resistance gene (neo), allowing the collection of data about the persistence and localization of infused cells in vivo (Rooney et al, 1995; Heslop et al, 1996; Gottschalk et al, 2001; Bollard et al, 2006).

None of the patients treated with EBV-specific CTL as prophylaxis developed PTLD, in contrast with an incidence of 11·5% in a historical untreated control group, and nine patients with elevated EBV-DNA levels at the time of infusion had significantly reduced viral load levels (1–4 log reduction) within 1–3 weeks of the first T cell infusion. In patients who received neo-marked cells, specific CTL could be detected for up to 9 years post CTL. Thus, adoptive immunotherapy was established as an attractive therapy for the prevention and treatment of EBV-LPD in high-risk patients post-transplant. These results have been confirmed by several other groups (Imashuku et al, 1997; Gustafsson et al, 2000; Comoli et al, 2007; O’Reilly et al, 2007).

CMV-specific CTL

Adoptive transfer of in vitro activated and expanded CMV-specific CTL has also been used as prophylaxis and treatment of CMV infections post-transplant. In a pioneering study (Walter et al, 1995), CD8+ CMV-specific T cell clones, reactive against CMV virion proteins, were isolated and expanded from the blood of bone marrow donors and administered to 14 patients prophylactically at weekly intervals in doses escalating from 3·3 × 106/kg to 1 × 109/kg, beginning 30–40 d post-transplant. While the majority of patients lacked any evidence of CMV-specific anti-viral activity preinfusion, after the first infusion responses were detected in all recipients. However, there was no evidence of CD8+ T cell persistence in patients who did not have a concurrent recovery of CD4+ T cells, highlighting the importance of helper T cells in the maintenance of anti-viral activity in vivo. In a number of infused recipients the authors could directly correlate CMV T cell immunity and the persistence (for up to 12 weeks) of transferred T cells by following rearranged TRAV and TRBV genes as molecular markers. Neither CMV viremia nor disease developed in any of the treated patients.

Peggs et al (2003) used a slightly different approach to prevent and treat CMV infections post allogeneic HSCT by producing and characterizing polyclonal CMV-specific CTL using DCs pulsed with CMV antigens derived from a CMV-infected human lung fibroblast cell line to stimulate reactive T cells. Sixteen patients with CMV infection were infused with CMV-specific CTL (1 × 105 CTL/kg) at a median of 36 d post-transplant. The authors were able to reconstitute immunity without inducing GvHD after infusion of relatively small doses of cells because they obtained impressive in vivo expansion of the adoptively-transferred cells, and in eight cases further antiviral drugs were not required (Peggs et al, 2003).

Einsele and colleagues have used polyclonal CMV-specific CTL lines, generated using CMV lysate to activate both CD4+ and CD8+ T cells, as treatment for HSCT patients with persistent or recurrent CMV infections despite the prolonged use of anti-viral medications (Einsele et al, 2002a,b). CTL infusions were effective in reducing the viral load in seven of eight treated individuals and the results were sustained in five, and transient in two patients (Einsele et al, 2002b). Thus, the adoptive transfer of donor-derived CMV-specific T cells is capable of reconstituting immune responses against this virus, and protecting patients against the development of CMV disease or late recurrences.

To avoid the use of lysate or CMV antigen for T cell stimulation, another group adoptively transferred donor-derived CMV peptide-specific T cells to adult and pediatric transplant recipients who had mostly undergone non-myeloablative HSCT without in vitro or in vivo T cell depletion (Foster et al, 2004; Micklethwaite et al, 2007). The adoptively-transferred T cells were activated and expanded using dendritic cells pulsed with the immunodominant CD8+ HLA-A2 restricted epitope NLVPMVATV derived from pp65. In six of nine patients who received T cells there was evidence of an increase in the frequency of specific T cells postinfusion, but the rises were modest and did not persist beyond several days to weeks, which may again be related to lack of CD4+ T cell help. Two of nine patients reactivated CMV, one while receiving high-dose corticosteroids. In the other patient, the reactivation was short and self-limiting. Three patients developed GvHD within 14 d of receiving CMV-specific T cells and one patient died of GvHD. Since all had had GvHD prior to infusion, and the incidences of GvHD coincided with a reduction in corticosteroid dosage, it is unclear whether these recurrences were related to the T cells but the possibility that the T cells may have caused or exacerbated GvHD is a concern.

The residual alloreactive potential of the cells will be more carefully assessed in a follow-up study from the same group using dendritic cells transduced with an adenoviral vector expressing the full length pp65 antigen to reactivate polyclonal CD4+ and CD8+ T cells against multiple different epitopes for infusion purposes. This will overcome the limitations of the previous study which included (a) restriction of treatment to HLA-A2 positive patients, (b) the administration of CTL lines with single peptide specificity, and (c) lack of CD4+ T cell help in the infusion product.

Trivirus-specific CTL

More recently, we have produced bivirus- and trivirus-specific CTL lines containing polyclonal antigen-specific CTL targeting EBV and Adv, or EBV, CMV, and Adv, respectively. The CTL lines were generated by genetically modifying monocytes and EBV-LCL with a chimeric adenoviral vector that was either not expressing a transgene (for the generation of bivirus-specific CTL), or an adenoviral vector expressing CMV-pp65 as a transgene. The EBV-LCL served as the source of EBV antigens, while the adenoviral vector stimulated Adv-specific T cells, and pp65 was used to activate CMV-pp65-specific T cells. Using this protocol we were able to consistently activate and expand CTL lines with the appropriate specificities (Leen et al, 2006).

When small total numbers of trivirus-specific CTL (ca 2 × 105/kg) were administered to allogeneic HSCT recipients receiving a graft from any donor who was EBV and CMV seropositive, the infused cells could expand in vivo and appeared able to protect against all three viruses (Leen et al, 2006). Similar results were achieved when bivirus-specific CTL targeting EBV and Adv were administered to patients receiving a graft from a CMV seronegative donor (unpublished observations). Thus, by administering CTL lines targeting multiple viruses simultaneously, broad spectrum treatment from a single infusion of cells could be offered.

The T cell immune response to Adv is less well understood than that directed against either EBV or CMV (Flomenberg et al, 1995; Smith et al, 1996; Hamel et al, 2002; Leen et al, 2004a,b; Tang et al, 2004, 2006; Veltrop-Duits et al, 2006). Therefore we sought to extensively characterize the Adv-specific T cell component present in the bivirus- and trivirus-specific CTL lines in order to identify the specificities of T cell responses directed against the adenoviral capsid hexon antigen. In total, we screened 26 CTL lines produced for clinical use (13 bivirus and 13 trivirus lines) using a peptide library spanning the entire sequence of the hexon protein (Leen et al, 2006, 2008). Of the 26 lines screened, 25 contained an Adv-specific T cell component, and we confirmed the responsiveness of these CTL lines to previously published epitopes as well as identifying 33 new CD4- and CD8-restricted hexon epitopes. However, there were significantly fewer Adv epitope-specific responses in the trivirus-specific CTL lines generated using the Ad5f35pp65 vector than in the bivirus-specific CTLs generated using the Ad5f35null vector. Thus, it appears that “antigenic competition” can limit the range of antigens/epitopes recognized, and that triviral specificity in a single line seems to be close to the limit of viral-target recognition by the immune system. Thus, extending multivirus-specific CTL therapies to include specificity for and protection from additional viruses may prove problematic, especially for infection with “acute” viruses (e.g parainfluenza and RSV) for which low frequencies of reactive memory T cells circulate in peripheral blood.

Overcoming the limitations of CTL immunotherapy

Although the clinical trials outlined above have all shown promise in the clinical setting, this type of adoptive immunotherapy has been confined to centres with specialized Good Manufacturing Practice (GMP) laboratories. There are three main reasons for this; (i) the complexity of the process and length of time required for the in vitro expansion of virus-reactive T cells, (ii) the limited breadth of viruses that can be targeted in a single CTL line, and (iii) lack of facilities for T cell preparation.

Rapid selection protocols

The time required for complicated CTL generation protocols is one issue that limits broader application of strategies to reconstitute antiviral immunity. Currently, the generation of EBV-, bivirus-, and trivirus-specific CTL lines requires 4–6 weeks to generate EBV-LCL for use as antigen presenting cells (APCs), followed by 4–6 weeks to produce sufficient CTLs for infusion, sterility testing, and functional analysis. Other protocols require the use of dendritic cells (DCs) as APCs, however their generation is also time consuming, and DCs are unable to proliferate in vitro. Consequently large blood volumes are required to produce sufficient DCs for CTL expansion, and therefore cell numbers can be limiting. In addition, each reagent needed for the generation of a CTL line must be available as a clinical grade product, necessitating extensive and expensive analysis and testing.

Although viral infections are frequently evident in patients early post-transplant (<30 d), due to the length of time required for CTL generation, most CTL lines must be prophylactically rather than therapeutically produced. However, if more rapid techniques for CTL production were available or if accurate methods for disease prediction were developed, then CTL therapy could be applied as standard of care rather than as an investigational therapy. Several groups have been developing more rapid and less cumbersome methods to produce virus-specific T cells for infusion (Feuchtinger et al, 2004, 2006; Rauser et al, 2004; Cobbold et al, 2005; Fujita et al, 2008). Two methods have been used clinically; tetramer selection and interferon-γ (IFN-γ) selection, and the studies confirmed that small numbers of antigen-activated T cells could expand substantially in vivo in the presence of antigen, and provide protection against the targeted pathogen.

Tetramer selection

Cobbold and colleagues attempted to decrease the complexity of in vitro CTL generation by isolating a pure population of CMV peptide-specific CD8+ T cells directly from donor peripheral blood using staining with specific tetramers followed by selection with magnetic beads (Cwynarski et al, 2001; Cobbold et al, 2005). This process facilitates the selection and direct infusion of a virus-specific population of cells without ex vivo manipulation, allowing the therapeutic rather than prophylactic use of isolated CTL.

Small numbers (median 8·6 × 103/kg) of selected cells were infused to nine patients within 4 h of selection and, despite small starting cell numbers, CMV-specific CD8+ T cells were detectable in all patients within 10 d of infusion. Further, T cell receptor clonotype analysis showed evidence of persistence in two patients studied. The infused T cells also demonstrated antiviral activity in vivo, particularly in the case of one of the patients whose CMV reactivation was refractory to antiviral drugs, but was controlled within 8 d of adoptive T cell transfer.

Although this study was extremely promising, it has not been reproduced due to the increased regulatory constraints that have been applied to translational laboratories regarding the use of clinical grade reagents. To date, multimers have not been produced to these GMP standards, limiting their use. In the future, GMP grade products may be available, but additional challenges remain including the limited number of CD4+ multimers available. Also, this approach is limited to patients who express HLA alleles for which viral peptides are available and for which the circulating frequency of reactive T cells in peripheral blood is detectable by multimer staining. Thus, infections associated with viruses where the circulating frequency of reactive T cells is much lower than for CMV, will not be amenable to this type of therapy.

IFN-γ capture

The IFN-γ Capture Assay is designed for the quantification and specific isolation of live antigen-stimulated IFN-γ secreting T cells. This technology has been used to rapidly and specifically isolate donor-derived Adv-specific T cells for infusion purposes (Feuchtinger et al, 2004, 2005, 2006). In a pilot study, virus-specific T cells were isolated and infused into nine children with systemic Adv infection after SCT following a short in vitro stimulation with adenoviral antigen. The selected cells were polyclonal, and the infusion of small numbers (1·2–50 × 103/kg) was safe, with no acute toxicities or GvHD induction. In five of six evaluable patients, there was an increase in AdV-specific T cells postinfusion which correlated with a decrease in viral load.

Thus, this approach was feasible, GMP-compatible, and could easily be applied to other infectious pathogens where a source of clinical grade antigen was available. However, although this method of T cell selection was rapid, cells with specificity to only a single virus were produced. Moreover, because of the small numbers of cells selected, it was not possible to characterize the infused product, making it difficult to correlate the phenotype and functional activity of the infused cells with subsequent clinical outcome.

Our group has tried to overcome these shortcomings by merging our trivirus-specific CTL generation technology with the IFN-γ Capture assay to rapidly and specifically isolate antigen-specific T cells directed against EBV, CMV, and Adv for immediate infusion. Further, we have developed a protocol to simultaneously expand a small portion of the selected product using autologous feeder cells and a cocktail of cytokines. This in vitro expansion of small numbers of selected cells facilitates extensive in vitro characterization of the infusion product, allowing us to correlate clinical outcome with the infused cells (Fujita et al, 2008).

These preliminary studies have shown that virus-specific T cells isolated by either tetramer selection or IFN-γ capture are safe and have activity when adoptively transferred to patients with active infection. However both methods rely on the availability of relatively large blood volumes (leukapheresis products), which are difficult to collect from unrelated donors recruited through national or international panels.

Targeting additional viruses

As discussed above, CTL lines that simultaneously target three different viruses have been used clinically with promising results. However, extending this technology to additional pathogens requires (a) extensive immunological characterization to identify immunogenic and protective antigens that can be targeted using T cell therapy, and (b) development of protocols to allow the generation of multivirus-specific CTL lines without loss of antigen specificity due to competition.

A number of groups have begun to analyze the protective immune response directed against community viruses, such as BK virus, RSV and influenza, in order to identify immunogenic antigens that can be targeted using in vitro generated CTL (McMichael et al, 1981; Gotch et al, 1987; Co et al, 2008). The problem of sustained multivirus specificity is also being addressed, and one way to avoid in vitro antigenic competition may involve the prevention of activation-induced cell death (AICD). Vella et al (1997, 1998) reported that the addition of cytokines, such as IL-7 and IL-15, prevented AICD in vitro, and current initiatives involve modifying CTL generation protocols to incorporate these cytokines and the assessment of the specificity, phenotype, and function in the resultant lines.

Partially-matched allogeneic lines

Crawford and colleagues used a different strategy to provide a rapidly available product by generating a bank of polyclonal EBV CTL lines for the treatment of EBV-associated diseases in SCT and solid organ transplant recipients. In a pilot study, eight patients with progressive PTLD were treated with closely HLA-matched CTL generated from an unrelated third-party blood donor (Haque et al, 2002). In the original trial, three patients attained complete remission following treatment, and no patient showed evidence of GvHD postinfusion. More recently, the same group has reported a larger phase 2 multicentre trial, treating 33 patients with EBV positive PTLD that had failed conventional therapy. The CTL lines for infusion were chosen based on the best-fit HLA match with the patient, and the results were encouraging with an overall response rate of 64% at 5 weeks, and 52% at 6 months (Haque et al, 2007).

The 33 PTLD patients treated in this trial were enrolled from 19 transplantation centres, who requested the best matched “off the shelf” product. Due to the partial HLA matching of the CTL used in the trial the authors expected limited persistence of the CTL in vivo. Therefore, to counteract this, 30 of the 33 treated patients received ≥2 CTL infusions. Overall, only one patient developed detectable anti-alloantibodies directed against non-shared HLA antigens, but no patient developed GvHD post-CTL infusion, demonstrating the safety of this approach. Importantly, the authors reported a significant increase in response rate with increased number of HLA matches between CTL donor and recipient (Haque et al, 2007).

Based on these encouraging results, our group has developed a similar multicentre approach to evaluate whether trivirus-specific CTL lines, generated at our centre, can have similar activity. We will establish a bank of trivirus-specific CTL lines, which will be extensively characterized immunologically, and then frozen in aliquots. Following the identification of a patient with an EBV, CMV, and/or an Adv infection/reactivation post-SCT that is persistent despite standard therapy, we will ship an appropriate CTL line to the treatment centre for immediate infusion. The appropriate CTL line will be chosen based on best HLA match and viral specificity.

CTL generation from seronegative donors

Although the generation of donor-derived virus-specific CTL lines has proven feasible from seropositive donors, production of such lines from seronegative adult donors and from cord blood is more challenging because of the antigen-inexperienced or “naïve” status of the T cells (Savoldo et al, 2002; Comoli et al, 2006). However, Park et al (2006) have recently developed an approach for the generation of CMV-specific T cells from cord blood using CMV antigen-loaded DCs to stimulate T cells. Culture in the presence of IL-7 and IL-12 and weekly antigen re-stimulation resulted in the generation of CMV-specific IFN-γ-producing CTL, which were consistently detectable by the fourth week of stimulation. Therefore, although more complicated and time consuming than the seropositive setting, such an approach may eventually lead to effective clinical strategies to prevent or treat opportunistic viral infections in the cord blood transplantation setting (Park et al, 2006). Alternatively, infusion of allogeneic banked CTL, as described above, may also confer protection from viral infections prior to endogenous T cell recovery.

Conclusions

Viral infections post-HSCT are an increasing cause of morbidity and mortality post-transplant. Adoptive immunotherapy to reconstitute antiviral immunity is an attractive prophylaxis and/or treatment option and the focus of a number of groups is to simplify the process to make this option more broadly available for clinical use.

Acknowledgements

AML and HEH are supported by NIH Grants PO1 CA94237, U54HL081007, P50CA126752, the GCRC at Baylor College of Medicine (RR00188), and a Specialized Center of Research Award from the Leukemia and Lymphoma Society, an Infection and Immunity award from Baylor College of Medicine to AML and HEH and a Doris Duke Distinguished Clinical Scientist Award to HEH.

Ancillary