Mark W. Lowdell1 & Emma Morris2
1Royal Free and University College Medical School, Royal Free Hospital, London, UK
2Department of Immunology and Molecular Pathology, Royal Free and University College Medical School and Department of Haematology, University College London Hospitals NHS Trust, London, UK
Immunotherapy in the form of vaccination has been part of medical practice since Jenner in the 18th century. This so-called ‘active’ immunization requires that the recipient has the capacity to mount an immune response against the antigens within the vaccine. In contrast, the infusion of antibodies or immune cells raised in other animals or individuals, in response to deliberate vaccination or prior antigen exposure, into patients at risk of infection – ‘passive’ immunization – allows treatment of immunodeficient or immunocompromised patients.
Until recently, infusion of pathogen-specific antisera was the only routine form of passive immunotherapy, equine antitetanus antisera being a well-known example. Successful passive cellular immunotherapy requires precise matching of donor : recipient histocompatibility antigens and thus advances in HLA-typing over the past 40 years has allowed this form of immunotherapy to move closer to routine treatment.
Cellular immunotherapy in haemopoietic progenitor cell transplantation
The antileukaemic activity of allogeneic bone marrow transplantation was first described, in murine experiments, more than 40 years ago, but was appreciated in the clinic only in the late 1970s when attempts at preventing graft-versus-host disease (GVHD) by T-cell depletion were sometimes frustrated by an increase in the risk of leukaemia recurrence. The clinical antileukaemic effect of GVHD was first reported in 1979 and confirmed later by registry data from the International Bone Marrow Transplant Registry (IBMTR) . The observed benefit of GVHD was particularly evident in patients transplanted for chronic myeloid leukaemia and led to the trial of posttransplant infusions of donor leucocyte infusions (DLI). The first peer-reviewed report of DLI therapy included a single patient who achieved molecular remission with no evidence of clinical GVHD, supporting the hypothesis that graft-versus-leukaemia (GVL) could be directed at leukaemia-specific or leukaemia-restricted target antigens . GVHD after DLI remained a significant clinical problem, which has been somewhat reduced by the use of incremental doses of DLI, but the search for the ‘holy grail’ of leukaemia-specific GVL in the complete absence of GVHD continues to be an active research theme.
Nonspecific T-cell immunotherapy
A pragmatic approach to the dissection of GVHD from GVL was the concept of removal of alloreactive T cells from donor grafts whilst retaining nonalloreactive cells that could mediate GVL and antiviral responses. These approaches were all based upon ex vivo stimulation of allogeneic donor T cells with normal haemopoietic cells from the recipient to provoke a clinical-scale mixed lymphocyte response. Reacting T cells were then identified by the expression of one or more activation antigens (e.g. CD25, CD69) and depleted by immunotoxin or immunomagnetic selection. Whilst possibly successful in the reduction of GVHD, the clinical trials of this approach showed no evidence of a GVL effect, although antiviral immune responses have been enhanced in some cases. The principal criticism of these studies was that too few allo-depleted T cells were infused to definitively test the hypothesis that GVHD was prevented. Subsequently an extremely thorough study of the nature of alloreactive T cell activation in vitro in a mixed lymphocyte reaction concluded that the oligoclonal T-cell response is random and hence unpredictable from day to day. These data demonstrated that not all alloreactive T-cell clones will activate in a single mixed cell reaction and thus clinically relevant minor alloreactive T cells are likely to remain and induce GVHD upon infusion.
Another nonspecific approach has been the selective depletion of CD8 T cells from DLI . Based upon the fact that the target cells of GVHD mostly lack expression of HLA-class II it has been considered that infusions of allogeneic CD4 T cells induce less GVHD. Many haemopoietic malignancies express HLA-class II antigens and are potential targets for CD4 T cells. Evidence of GVL, resolution of mixed T-cell chimerism and improved antiviral immunity in the absence of GVHD have all been reported in clinical trials of CD8-depleted DLI. Trials of this form of immunotherapy are continuing and are reporting encouraging results with respect to reversal of mixed chimerism and GVL .
Tumour-specific or tumour-restricted T-cell immunotherapy
Unselected DLI currently remains the mainstay of antitumour cellular immunotherapy following haemopoietic progenitor cell (HPC) transplantation; however, tumour antigen-specific T-cell responses can be generated by vaccination or by the generation of tumour antigen-specific T cells for adoptive transfer. T-cell-recognized tumour antigens can be divided into two main categories:
· The first are known as tumour-specific antigens (TSA), and the genes encoding TSA are only present in tumour cells and not in normal tissues.
· The second group, called tumour-associated antigens (TAA), are expressed at elevated levels in tumour cells but are also present in normal cells.
The majority of T-cell-recognized tumour antigens in humans are TAAs. The significance of this is that a low level of gene expression in normal cells can lead to the inactivation of high avidity T cells by immunological tolerance mechanisms. As a consequence, low avidity T-cell responses in patients are often inadequate in providing tumour protection. Therefore, TSA are theoretically the most desirable target antigens for cellular immunotherapy (vaccination or adoptive transfer), as there is no pre-existing immunological self-tolerance and TSA-specific immune responses are unlikely to damage normal tissues. Unfortunately TSAs with specific mutations are often invisible to cytotoxic T-lymphocytes (CTL) as a result of impaired antigen presentation due to competition with normal cellular antigens for proteasomal degradation, transportation by TAP molecules and binding to MHC. To date, most tumour antigens indentified as CTL targets are TAAs.
The majority of antitumour vaccination trials in humans has been against melanoma antigens and not in the context of stem cell transplantation. In these situations, vaccination can lead to TAA or TSA-reactive CTL responses, but there has rarely been a corresponding clinical benefit.
Recently, vaccination against the Wilms' tumour antigen 1 (WT1, a leukaemia-associated antigen) has been shown to induce WT1-specific T-cell responses in patients with myeloid malignancies. In the next 5 years it is anticipated that phase I/II clinical trials will test whether vaccination against WT1 epitopes early posttransplant can augment the reconstitution of WT1-specific CTL and act as maintenance immunotherapy.
TCR gene transfer
The inability to generate antigen-specific T cells is a serious limitation of adoptive cellular therapy for cancer. As discussed above, tumour antigens are often poorly immunogenic and patients are frequently immunocompromised as a consequence of the tumour burden or as a result of previous therapies. TcCR gene transfer offers a strategy to produce T cells with a TCR specific for a tumour antigen (antigen-specific T cells) independent of precursor frequency . Over the last 5 years, TCR gene transfer has been demonstrated to redirect reliably the antigen specificity of a given population of T cells via the introduction of a cloned TCR using retroviral transduction. This allows for the rapid generation and expansion of tumour antigen-specific T cells. Both the specificity and avidity of the TCR-transduced T cells are similar to the parental CTL clone from which the TCR has been isolated. Tumour antigen-specific TCR-transduced T cells have been shown to provide tumour protection in murine models and result in re-call responses up to 3 months post adoptive transfer. Recently, the first clinical trial using TCR-transduced T cells in melanoma patients has been published, demonstrating that TCR-transduced autologous T cells can have an antitumour effect. Patient T cells were transduced with a retroviral construct encoding the alpha and beta chains of the MART-1-specific TCR to target melanoma tumour cells. This milestone study demonstrated the feasibility and potential of TCR gene therapy. However, further modifications are required to the approach in order to maximize the clinical benefit. These include:
· modifications of the TCR construct to enhance cell surface expression of the introduced TCR and reduce the incidence of mispairing with endogenous alpha and beta chains (Figure 43.1);
· optimization of the conditioning regimen used prior to adoptive transfer;
· generation of functional TCR-transduced helper T cells.
Fig 43.1 Schematic illustrating mis-pairing with endogenous TCR chains by the introduced TCR chains following retroviral TCR gene transfer.
Chimeric antigen receptor (CAR) modified T cells
The inability to generate antigen-specific T cells (see above) can be overcome by inserting a B-cell receptor with an intracellular TCR domain through gene transfer – a chimeric receptor. Murine monoclonal antibodies are readily generated to human proteins, some of which may be tumour restricted or at least tumour associated. CD19 is a protein expressed on B cells from the pre-B-cell stage to the mature B cell but is absent from plasma cells. T cells transfected with a sequence encoding the F(ab) domain of an anti-CD19 and the intracellular domain of CD3ζ can bind to CD19 expressing leukaemia cells and be triggered functionally as if they received signalling through the CD3–TCR complex. This is not a new concept; CAR-T cells have been tested experimentally for many years but the clinical effects have been transient due to failure of the cells to survive in vivo.
Recently the first successful clinical trial of CAR-transduced T cells used in adoptive immunotherapy was reported from Pennsylvania State University . This group combined the anti-CD19 CAR with the sequence encoding human CD137, the costimulatory receptor 4-1BB, which is involved in T-cell survival. These CAR-T cells targeted CD19 expressing chronic lymphocytic leukaemia cells in vivo and successfully matured into memory T cells, which were detectable in recipient blood samples for prolonged periods.
Tumour-restricted natural killer cell immunotherapy
Some of the earliest trials of antitumour cellular immunotherapy were based upon infusion of NK cell activating cytokines or of ex vivo activated NK cells. Most of the early trials were in the autologous setting and, with the notable exception of a single report of acute myeloid leukaemia (AML) patients after autologous haemopoietic progenitor cell (HPC) transplant, were uniformly disappointing. However, these early trials were conducted before the complex mechanisms underlying NK cell function were understood.
Human NK cells are controlled by a variety of inhibitory and stimulatory signals though cell surface receptors, which allow them to distinguish between normal and malignant or infected cells. These receptors fall into one of four families:
· killer immunoglobulin-like receptor (KIR);
· C-type lectins;
· immunoglobulin-like transcript (ILT);
· natural cytotoxicity receptor (NCR).
The first two families include both inhibitory and activating receptors whilst the ILT and NCR families contain only activating receptors. All human NK cells express multiple receptors from each family and it is now apparent that functional subsets of NK cells exist.
In the 1980s Klaus Karre first demonstrated that murine NK cells preferentially lysed MHC class I negative tumours. This led him to construct his ‘missing self’ hypothesis, in which he proposed that NK cells are inhibited from lysis of normal cells that express MHC class I but are capable of lysing MHC class I negative tumour cells. Murine NK cells express surface receptors for MHC class I molecules and their ligation transduces inhibitory signals that prevent NK-mediated lysis. As implied above, the human NK regulatory system is more complex. Killer immunoglobulin-like receptors (KIRs) bind to HLA class I molecules and the majority transduce inhibitory signals upon ligation by their specific HLA class I ligand. KIR molecules are classified on the basis of the number of extracellular domains and the length of the intracellular domain. For example:
· KIR2DL1 is a molecule with two extracellular domains and one long intracellular domain. This KIR binds to the family of HLA-C molecules with asparginine in position 77, the so-called ‘type 2’ HLA-C. The inhibitory signal is via an ITIM in the long intracellular domain.
· KIR2DL2, in contrast, binds to the remaining ‘type 1’ HLA-C molecules, those with a serine in position 77.
· KIR3DL1 binds to HLA-Bw4 alleles.
· KIR3DL2 binds HLA-A3 or A11 alleles.
· Other KIRs have been shown to bind HLA-G (Figure 43.2).
Fig 43.2 Ligands controlling NK cell activation and triggering.
In a co-culture of NK cells, autologous normal cells and HLA-class I deficient tumour cells one can demonstrate the colocalization of the KIR : HLA interaction between the NK and normal cell whilst the tumour cells show no such signalling (see Plate 43.1 in the plate section).
Given the concept that NK cells are ‘kept in check’ by inhibitory signals initiated through binding HLA molecules on normal somatic cells one might imagine that the KIR repertoire of any given individual is determined by their HLA type. This is not, however, the case and many healthy individuals will maintain NK cell clones that lack the appropriate inhibitory KIR molecules. However, the C-type lectin, NKG2A, which forms a heterodimer with CD94, appears to be universally expressed on human NK cells and presumably can provide the requisite inhibitory signals through its ligation to HLA-E. Most healthy individuals sustain a population of NK cells that lacks both KIR and NKG2A; these cells appear to be hyporesponsive to activating stimuli.
The clinical relevance of the understanding of NK cell inhibition is evident in haploidentical HPC transplantation where certain HLA class I mismatches generate a situation of HLA : KIR incompatibility. Highly significant reductions in relapse have been reported among AML patients receiving HLA : KIR incompatible haploidentical HPC grafts compared to patients receiving grafts in which the donor NK repertoire is matched to the HLA type of the patient.
Whilst these data stand on their own merit, much remains to be done to understand the mechanisms behind their results since the KIR effect seems to be limited to excessively T-cell-depleted haploidentical grafts and there is little or no role for CD94/NKG2A or for the activating receptors that appear so important in autologous NK cell function.
The implication from the haploidentical transplant data is that the lack of KIR-mediated NK cell inhibition is sufficient to initiate NK activation and lysis. However, since the patients with known KIR : HLA incompatibility did not experience NK-mediated GVHD one must assume that normal cells failing to inhibit NK cells were spared. NK activating signals may be provided via numerous receptors, although their ligands remain largely unknown. Recent published work has shown that, like T cells, NK cells generally require more than one signal to initiate cytokine secretion or lysis and that these signals may need to be provided sequentially to the cell.
Despite the lack of a complete understanding of human NK biology, clinical trials of allogeneic NK immunotherapy are already underway and some have been reported.
A remarkable study in which minimally conditioned patients with AML received bolus infusions of partially enriched IL2-activated NK cells from HLA-mismatched donors without concomitant HPC transplant showed engraftment of donor NK cells in the presence of recipient T, B and myeloid cells . At the highest NK dose and the greatest level of preinfusion conditioning with cyclophosphamide and fludarabine, 5 of 19 patients achieved complete remission. Donor NK cells were detected in their peripheral blood and bone marrow. Despite the engraftment of haplomismatched NK cells the patients maintained normal bone marrow function and normal levels of autologous T cells, B cells and granulocytes.
The antileukaemic effect was relatively short lived in this trial but the data support the safety of such an approach and it is possible that such patients could receive multiple courses of NK cell infusions to maintain control of residual disease. The concept of repetitive passive cellular immunotherapy is novel and contrary to the design of most current approaches, which have been conceived within a mindset of ‘cure by vaccination’. However, most tumour antigens elicit relatively weak immune responses and the physiological immune response to tumours may be one of control rather than eradication.
Passive cellular immunotherapy of infectious disease
Possibly the most remarkable clinical results from cellular immunotherapies have been seen in the treatment of opportunistic viral infections in immunocompromised patients. Most of these trials have been in the posttransplant setting, particularly in recipients of allogeneic HPC grafts:
· The earliest studies involved infusion of enormous numbers of cloned CMV-reactive CD8 T cells, which caused resolution of refractory CMV disease in patients postallogeneic HPCT.
· Subsequently, others elegantly demonstrated the specific resolution of posttransplant EBV-driven lymphoma following infusion of donor-derived anti-EBV CTLs.
Ex vivo generation of very large numbers of antiviral T cells is complex and expensive. However, in 2003, a phase I trial of allogeneic donor-derived CMV-reactive T cells grown for 21–28 days ex vivo on monocyte-derived dendritic cells that were pulsed with fixed whole CMV was reported. These expanded cells were infused into patients with molecular evidence of CMV reactivation postallogeneic HPCT and 8/16 patients resolved the reactivation without recourse to antiviral chemotherapy. No patient received a dose greater than 105 T cells per kilogram body weight and the average dose of CMV-specific T cells in each dose was no greater than 200–300 per kilogram. Despite this incredibly low dose of cells, virus-specific T cells were detectable in the peripheral blood of responding recipients at levels equivalent to a 35 000-fold expansion. The small numbers of cells infused in this demonstrated that the production of donor-specific cell therapies could be cost effective.
Despite the acknowledged clinical success of these trials neither led to a wide-scale adoption of cellular therapy due to the extreme technical complexity of cell therapy production. However, with recent advances in the availability of clinical-grade reagents and disposables the translation of laboratory-grade procedures to clinical application has advanced rapidly. For several years immunologists have been able to immunomagnetically select specifically activated T cells on the basis of the secretion of gamma-interferon and its capture on the cell surface with a bispecific antibody complexed to a paramagnetic nanoparticle. This approach selects both CD4 and CD8 cells .
An alternative approach is the use of multimeric recombinant MHC class I complexes loaded with an immunodominant viral peptide antigen restricted to the specific class I antigen . These HLA-multimers can be complexed with the same sort of paramagnetic nanoparticles used in the gamma-secretion process described above and directly select antiviral-specific CD8 T cells from donor blood. These two patented technologies are now produced to clinical grade and are already central to a number of trials, including a phase III trial of allogeneic immunotherapy of CMV reactivation post-HPCT – the first multicentre randomized clinical trial of directed donation cellular immunotherapy.
There is undoubted promise in the clinical application of cellular immunity and the field has advanced very substantially in the last 5 years. However, the true potential of adoptive cellular immunotherapy remains constrained by the perceived need for directed donations (autologous or HLA-matched allogeneic) and by confusion over the regulatory framework in which the therapies fall. The first issue is the greatest barrier although some recent studies do support the feasibility of the ultimate goal of ‘off the shelf’ products . The haploidentical NK study discussed above used NK cells from HLA-mismatched donors and demonstrated transient engraftment. A group in Edinburgh, UK, recently used ‘off the shelf’ HLA-mismatched T-cell lines to treat posttransplant EBV lymphoma in recipients of renal transplants .
Technical advances facilitating translational research in cellular immunotherapy
In Europe, since the ratification of the EU Clinical Trials Directive in member states in 2004, cellular immunotherapies have becomes susceptible to regulation as investigational medicinal products (IMPs). Whether a specific cell therapy product constitutes an IMP is determined by the relevant authority in each member state but, once a product is regulated as an IMP then production must meet Good Manufacturing Practice (GMP), and this has been difficult in the field of cellular immunotherapy. However, a number of European companies now manufacturer CE-marked reagents, consumables and devices for clinical-grade cell production. Closed and semi-closed systems are available for handling large-volume cell suspensions. Gas-permeable cell culture and expansion bags allowing closed-system culture are now widely available and the availability of clinical-grade cytokines is improving.
One of the most significant advances in the field has been the development of CE-marked clinical-grade immunomagnetic cell sorters. These are now widely used for the specific selection of subsets of haemopoietic progenitor cells and other leukocytes and can even select antigen-reactive cells on the basis of cytokine secretion, multimeric HLA-peptide reagents or expression of activation markers.
As the regulatory position becomes clearer more trials will be conducted to good clinical practice and the regulatory authorities will gather more evidence and experience of the field. In the not too distant future hospital blood banks may become more of a ‘cell pharmacy’ than ever before.
1. Allogeneic GVL by DLI is proof-of-principal of cellular immunotherapy.
2. Cellular immunotherapy of viral infections is becoming an alternative to antiviral chemotherapy.
3. HLA-matching may not be necessary.
4. Technical and regulatory difficulties in production of cell therapies are being overcome.
1. Horowitz MM, Gale RP, Sondel PM et al. Graft-versus-leukemia reactions after bone marrow transplantation. Blood 1990; 75: 555–562.
2. Kolb HJ, Mittermueller J, Clemm C et al. Donor leukocyte transfusion for treatment of recurrent chronic myelogenous leukemia in marrow transplant patients. Blood 1990; 76: 2462–2465.
3. Shimoni A, Gajewski JA, Donato M et al. Long-term follow up of recipients of CD8-depleted DLI for the treatment of CML relapsing after allogeneic progenitor cell transplantation. Biol Blood Marrow Transplantation 2001; 7: 568–575.
4. Orti G, Lowdell M, Fielding A et al. Phase I study of high-stringency CD8 depletion of donor leukocyte infusions after allogeneic hematopoietic stem cell transplantation. Transplantation 2009; 88: 1312–1318.
5. Xue SA & Stauss HJ. Enhancing immune responses for cancer therapy. Cell Mol Immunol 2007; 4: 173– 184.
6. Kalos M, Levine BL, Porter DL et al. T cells with chimeric antigen receptors have potent antitumor effects and can establish memory in patients with advanced leukemia. Sci Translational Med 2011; 3: 95ra73.
7. Miller JS, Soignier Y, Panoskaltis-Mortari A et al. Successful adoptive transfer and in vivo expansion of human haploidentical NK cells in patients with cancer. Blood 2005; 105: 3051–3057.
8. Peggs K, Thomson K, Samuel E et al. Directly selected cytomegalovirus-reactive donor T cells confer rapid and safe systemic reconstitution of virus-specific immunity following stem cell transplantation. Clin Infect Dis 2011; 52: 49–57.
9. Cobbold M, Khan N, Pourgheysari B et al. Adoptive transfer of CMV-specific CTL to stem cell transplant patients after selection by HLA-peptide tetramers. J Expl Med 2005; 202: 379–386.
10. Leen AM, Myers GD, Sili U et al. Monoculture-derived T lymphocytes specific for multiple viruses expand and produce clinically relevant effects in immunocompromised individuals. Nat Med 2006; 12: 1160–1166.
11. Haque T, Wilkie GM, Jones MM et al. Allogeneic cytotoxic T cell therapy for EBV-positive PTLD: results of a phase II multicentre clinical trial. Blood 2007; 110: 1123–1131.
Kadowaki N & Kitawaki T. Recent advance in antigen-specific immunotherapy for acute myeloid leukemia. Clin Dev Immunol 2011; 2011: 104926. Epub 19 October 2011. Review.
Kolb HJ, Schattenberg A, Goldman JM, Hertenstein B, Jacobsen N, Arcese W, Ljungman P, Ferrant A, Verdonck L, Niederwieser D, van Rhee F, Mittermueller J, de Witte T, Holler E, Ansari H; European Group for Blood and Marrow Transplantation Working Party. Chronic leukemia. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. Blood 1995. 1 September; 86(5): 2041–2050.
Linley AJ, Ahmad M & Rees RC. Tumour-associated antigens: considerations for their use in tumour immunotherapy. Int J Hematol 2011, March; 93(3): 263–273. Epub 1 March 2011. Review.
Vincent K, Roy DC & Perreault C. Next generation leukaemia immunotherapy. Blood 2011, 15 September; 118(11): 2951–2959. Epub 6 July 2011. Review. PMID: 21734234 [PubMed – indexed for MEDLINE].