Impact of T cell selection methods in the success of clinical adoptive immunotherapy

Abstract

Chemotherapy and/or radiotherapy regular regimens used for conditioning of recipients of hematopoietic stem cell transplantation (SCT) induce a period of transient profound immunosuppression. The onset of a competent immunological response, such as the appearance of viral-specific T cells, is associated with a lower incidence of viral infections after haematopoietic transplantation. The rapid development of immunodominant peptide virus screening together with advances in the design of genetic and non-genetic viral- and tumoural-specific cellular selection strategies have opened new strategies for cellular immunotherapy in oncologic recipients who are highly sensitive to viral infections. However, the rapid development of cellular immunotherapy in SCT has disclosed the role of the T cell selection method in the modulation of functional cell activity and of in vivo secondary effects triggered following immunotherapy.

Electronic supplementary material

The online version of this article (doi:10.1007/s00018-013-1463-5) contains supplementary material, which is available to authorized users.

Introduction

Susceptibility to viral infections in allogeneic hematopoietic stem cell transplantation (HSCT) is the result of profoundly reduced innate and adaptive immunity caused by the immunoablative effect of the bone marrow conditioning regimens, host-versus-graft and graft-versus-host interactions which take place in the immediate post-transplant period [1, 2]. Administration of anti-viral drugs is nowadays the standard first line therapeutic treatment and it is even used as a prophylactic method in patients with high susceptibility to suffer viral primo-infection or reactivation. However, this treatment shows high regimen-related toxicity in different tissues, variability in efficacy and generation of viral-resistant variants after treatment. Therefore, many clinical trials alternate or combine anti-virals with prophylactic therapies based on infusion of virus-specific primed T cells [3]. Despite their success, there are many doubts about their application and efficacy after the reappearance of virus-associated clinical symptoms in some patients, correlating with the in vivo longevity of this therapeutic subpopulation and retention of virus reactivity. Interestingly, only a few researchers have analysed in depth if the T cell cytolytic effector function is in fact dependent on their selection method, which would alter the persistence of the adoptively transferred cells.

This review summarises the impact of classical and alternative protocols for purification of antigen-specific T cells on the establishment of immunological tolerance, the dynamics of in vivo T cell expansion and its clinical daily application. Our observations highlight the urgent need for standard technical approaches in immunotherapy facilities which would induce minor alterations in the T cells, guaranteeing a long-lived and optimum repopulating potential.

Adoptive immunotherapy precedents

The classical concept of adoptive immunotherapy is based on infusion of donor unmanipulated bulk lymphocytes (DLI) at variable intervals following bone marrow transplantation, enhancing, for example, the anti-leukaemic effect of the graft (GVL) [4]. This idea is supported from the cytogenetic remission that was observed in three patients receiving buffy coat infusion of their original marrow donors after chronic myelogenous leukemia diagnosis and conventional allograft transplant [5]. Moreover, five patients receiving nonirradiated DLI developed Epstein–Barr virus (EBV) lymphoma after transplantation. Clinical remissions were achieved within 30 days after cellular treatment [6]. Despite these promising clinical results, also corroborated by Heslop and colleagues [7], this type of cellular therapy presents a high risk for the patient’s health. Indeed, there is evidence that administration of total lymphocytes is associated with high morbidity and mortality rates, mainly due to severe graft-versus-host disease (GvHD) [8]. Alloreactive CD8+ T cells (CTLs) contained within the transferred bulk leukocyte population are directly responsible for the aggressive GvH syndrome. Consequently, different protocols have been developed to limit the presence of these self-reactive subpopulations to obtain a safer cellular product.

In immunocompetent subjects, the exposure to viral antigens and their recognition by T cells trigger T cell receptor (TCR)-signal dependent activation, which drives their expansion and differentiation and regulates the magnitude of the T cell response [9]. Virus-primed T cells are generated during this process [10]. This physiological process can be replicated ex vivo in lymphocyte cultures exposed to viral antigens. Thus, cell proliferation and differentiation into different T cell subsets can be achieved [11]. This is the basis of the incipient adoptive immunotherapy strategy suggested by Riddell and colleagues [12], and nowadays adopted by many researchers, proving that autologous CD3+ CD8+ CD4− cytomegalovirus (CMV)-specific clones can be expanded by co-culturing donor-derived peripheral-blood mononuclear cells with autologous virus-pulsed fibroblasts for 5–12 weeks. Successful adoptive immunotherapy was later performed with these virus-specific T cells. None of patients developed CMV-associated clinical disease or side effects usually associated with the treatment. The analyses of the rearrangements of the TCR β-chain variable region demonstrated in vivo clonal expansion from the transferred clones. These expanded T cells exhibited the same cytotoxicity by the parental cells from the immunocompetent donors for up to 4 weeks [13]. However, as this ex vivo expansion is undertaken in the presence of live virions, this methodology can jeopardise the patient’s safety, and, therefore, it has been limited to a restricted number of cases [14]. Nevertheless, the therapeutic potential of these findings, which have been corroborated in subsequent studies [15–49], represented a starting point for the designing of new cellular immunotherapy protocols [50, 51] (Table S1).

Improvement of virus-specific T cell production and cell enrichment methods

In the meantime, other similar approaches have also been used to generate anti-viral T cell responses. For example, dendritic cells (DCs), one of the most potent professional antigen-presenting cell (APCs) types, have been extensively used in immunotherapy protocols (Table S2). They may be used as raw biological material for in vitro T cell production of both protagonist cells of the cellular therapy products. The main advantage of using DCs is their high capacity for antigen processing and presentation of multiple epitopes in major histocompatibility complex (MHC) class I and II and their applicability to patients of all HLA types. It has also allowed the obtaining of polyclonal T cell subpopulations with the advantage that this characteristic provides synergism between different immunological subpopulations [14, 52]. Otherwise, the immunodominant T cell epitopes present in a particular given antigen have to be known to expand the patient’s T cells. In addition, it is possible to modulate the expression of their co-stimulatory molecules allowing the manipulation of the immunological synapsis that would enhance the activation or inhibition of the T cell response [53]. Similarly, EBV-transformed B-lymphoblastoid cell lines (BLCL), which were transduced with a retroviral vector encoding the immunodominant CMV^pp65, have also been used as APCs to simultaneously expand EBV- and CMV-specific CTLs. These CTLs showed class I and II viral-bispecific restriction [54]. Furthermore, BLCL can be grown in large numbers, which would enhance the therapeutic protocol. Thus, Leen et al. [55] has perfected the system: BLCL were transformed with a recombinant adenovirus expressing CMV^pp65, with the purpose of obtaining T cell preparations with trivirus-specific activity [EBV-, Adenovirus (AdV)-, and CMV-specific CTLs]. CD4 and CD8 T cell expansion was observed, leading to in vivo resolution of virus-associated clinical symptoms within the first month of the therapy.

In fact, several clinical trials are being carried out to validate a new genetic engineering methodology consisting on the redirection of the T cell antigenic specificity by the introduction of TCR genes (variable α- and β-chains) or chimeric antigen receptors (CARs) (Fig. 1) followed by infusion in the patient (Table S3).

Fig. 1.

Genetic manipulation of the TCR, showing two different methods of genetic manipulation of the therapeutic cell: recombinant T-cell receptor (rTCR) transference and chimeric antigen receptors (CAR) cloning (a). This new genetic engineering methodology consists on the redirection of the T cell antigenic specificity by the introduction of CARs using viral vectors (b) or TCR genes (variable α- and β-chains) (c)

This strategy can be used to engineer fully functional virus-specific T cells, which would otherwise be “inactivated” after chronic viral antigen exposure [56]. For example, HBV-specific T cells are deleted or exhausted in chronic hepatitis B—and HBV-related hepatocarcinoma patients. Gehring and colleagues [57] and Wood [58] have obtained fully functional TCR re-directed HBV core (18–27)—and surface (183–191; 370–379)-specific effector T cells after epitope-specific TCR cloning by retroviral transduction. These T cells exhibited cytotoxic activity towards HBV-infected cells. Kessels and colleagues [59] demonstrated the expansion and maintenance of TCR-modified transferred T cells up to 81 days after inoculation. This clearly showed their in vivo proliferative capacities. At present, different kind of tumours and viral infections are treated with this type of cell transfer therapies, in some cases demonstrating cancer regression in 30 % of treated patients [60]. A potential problem of this methodology is the association of the recombinant α/β chains with the endogenously expressed TCR chains, generating T cells with unknown and undesired new autoantigen specificities [61] and with reduced TCR surface expression. In this way, sporadic autoreactive events could potentially appear in the recipients, something to be taken into consideration. The introduction of cysteines on each recombinant TCR chain to form an extra interchain disulfide bonds between both α- and β-structures could prevent this intracellular molecular event even significantly boost the effector activity [62]. Cohen et al. [63] designed a murinisation strategy based on the substitution of the human C regions with their murine counterparts. However, these xenogenic sequences might also potentially trigger immunogenicity against the murinised TCR. Bialer et al. [64] minimally engineered human C regions with selected murine residues mediating superior chimeric TCR expression and improved activity, which would result in a more efficient pairing of the murine Cα and Cβ1, decrease the formation of mixed TCR chain dimers and minimise autoimmune manifestations.

Nowadays, human trials are in progress to evaluate the safety and feasibility of T cells transduced with CARs instead of αβTCR in adoptive cell therapy procedures [65]. Specifically, these constructs recognise tumour cell-surface molecules. They consist of the fusion of the antigen-recognition portion of a monoclonal antibody with an intracellular signalling domain capable of activating or enhancing T cell effector function by intensifying molecular signalling pathways in a MHC-dependent and -independent fashion [60, 66, 67]. Consequently, Micklethwaite et al. engineered virus-specific T cells stimulated with multispecific viral immunodominant antigens of CMV, EBV and AdV. These modified cells also exhibited anti-tumour activity that was conferred by their retroviral-mediated expression of CAR.CD19+, obtaining a bi-functional therapeutic harvest [68]. This approach allows the expansion of multivirus-specific CAR-modified CTLs (which retain their anti-viral activities), but with significantly increased ex vivo anti-tumoural activity against B-ALL blasts from patients with haematological disease. Recently, the monitoring of anti-CD19 CAR-modified T cells is possible through the use of an antibody consisting of human CD19 extracellular domains and human immunoglobulin domain. Similarly to MHC-multimer technology, the fluorescent labelling of this structure allows the direct visualisation of CAR-expressing T cells by flow cytometry, which makes this approach very attractive [69]. However, only a few tumour-specific antigens expressed exclusively by cancer cells and susceptible of being targeted have been identified. This problem could be solved by the CAR-target replacement by a fluorescein isothiocyanate (FITC) molecule. The use of cetuximab, trastuzumab, and rituximab monoclonal antibodies conjugated with FITC would expand the applicability of this tool allowing the simultaneous recognition of a variety of tumour-associated antigens by a single therapeutic product [70]. Using this molecular approach, Louis and colleagues [71] and Pule et al. [72] achieved effective anti-tumour responses in patients with advanced-stage neuroblastoma. The treated patients exhibited partial and complete tumour responses, respectively, and long-term persistence of modified T cells (beyond 96 weeks). These T cells were engineered from EBV-specific cytotoxic T cells expressing the CAR diasialoganglioside antigen-GD2. These encouraging clinical results have demonstrated the high viability of these therapeutic T cells in cancer patients. However, these results have not been reproduced by other authors. Kershaw and colleagues observed a high level of therapeutic T cells during the first few days after infusion of autologous anti-α folate receptor CAR-modified T cells in patients with metastatic ovarian cancer. These T cells were undetectable after a month of monitoring [73]. A poor choice of the pan-tumour antigens or a weak functional activity could explain these results.

Following this approach, three different CAR generations have been developed, which incorporate T cell co-stimulatory signalling molecules [CD28, CD3-zeta(ξ), OX40, 4-1BB (CD137)] in their structure, improving their signalling capacities in modified T cells [74, 75]. These molecular modifications have enhanced the in vivo expansion (>1,000-fold), have increased the long-term maintenance of the engineered CAR+ T cells into the patient (for at least 6 months) and have induced trafficking to the bone marrow or cerebrospinal fluid. The infusion of CAR+ T cells that targeted CD19 or ERBB2 (HER-2/neu) and contained costimulatory domains from CD28 or CD137 and the TCR ξ chain signalling element in patients with relapse or refractory chronic lymphocytic leukemia [76–78], acute lymphoblastic leukemia [79] and metastatic colon cancer [80] generated tumour regression by a potent anti-tumour effect in all patients treated (in some cases associated with morphologic and molecular remission). However, unexpected clinical adverse events were noted in these patients including the occurrence of delayed tumour lysis syndrome accompanied by a hemophagocytic syndrome, capillary leak syndrome, non-infectious fevers, hypotension, respiratory distress syndrome, grade 3/4 lymphopenia and loss of normal B cells. In most cases, the administration of glucocorticoids or anti-cytokine therapy resolved these reversible systemic effects, although hospitalisation in intensive care units was necessary depending on the case. A cytokine-release syndrome or “cytokine storm” has also been observed in a limited number of patients, either after intensive lymphodepletion and before immunotherapy treatment or after CAR-transduced T cells infusion itself [76, 77, 79]. Most patients responded well to anti-IL-6 (tocilizumab) and/or anti-TNF (etanercept), although, in a few medically fragile patients, cytokine blockade with drugs was not effective, resulting in multiple organ dysfunction and death [78, 80]. However, the precise aetiology of these patients’ deaths remains uncertain.

Although the administration of corticosteroids in doses used to treat GvHD or antibodies specific to T cells such as alemtuzumab (CAMPATH-1H) would deplete the majority of circulating transduced cells [81], alternative approaches are necessary. In this regard, the introduction of the herpes simplex thymidine-kinase (HSV-TK) suicide transgene in the viral construct has allowed the inducible apoptosis of transduced-cells through interference with DNA synthesis on exposure to ganciclovir administered in the event of GvHD after stem cell transplantation in several phase I–II clinical trials [82–84]. Acute and chronic GvHD were successfully controlled with this suicide HSV-TK approach in the context of allo-HSCT and haplo-HSCT. However, a number of drawbacks need to be improved, such as the immunogenicity of the TK protein, the restriction of killing to dividing cells and the elimination of transduced-cells when Ganciclovir is used for the treatment of CMV infection.

For this reason, other authors have investigated in a phase I–II clinical trial the suitability of an alternative suicide gene, inducible caspase 9 (iCasp9), showing >90 % apoptosis of the modified T cells within 30 min of the administration of a specific chemical dimeriser drug [85]. This iCasp9 cell-suicide system did not change the antigen-specific functionality [86, 87] even when combined with transgenic expression of IL-2 or IL-15 [88].

The development of multimer technology has provided an invaluable method for monitoring and purification of T cells with known antigenic specificities. The basis for multimer technology resides in the recognition of antigen-specific TCRs by a recombinant class I or II molecule complex to a certain immunodominant peptide. Identification of antigen-specific cytotoxic T cells regardless of their biological activity allows the preparation of an heterogeneous T cell population which circumvents the previous phenotypic characterisations required for the identification of primed subpopulations with long-term survival capacities. Consequently, this staining technology allows the isolation of T cells with a given antigen specificity from seropositive donors without any further manipulation [89]. Nowadays, there is a wide variety of available MHC multimer molecules such as dimers, tetramers, pentamers, streptamers, dextramers and octamers which are key for studies of adoptive immunotherapy. These multimer molecules have been extensively used to identify and select CMV-, EBV- and AdV-specific T cells from healthy donors, and their transfer to immunocompromised hosts has shown excellent results [90–92] (Table S4).

Other authors have demonstrated the value of IFNγ-secreted antigen-specific CD8+ T cells for the successful reconstitution of virus-specific immunity in allogeneic bone marrow transplant recipients [93]. Manz et al. [95] has developed a high-affinity physic matrix of cytokine-secreting cells that prevents cytokine spreading [94, 96]. A number of groups have successfully used this technique for the treatment of viral infections in immunocompromised patients, increasing viral clearance and avoiding the associated disease [97, 98]. However, several authors have questioned the homogeneity of the cellular product obtained with this capture method, following the identification of non-specific NK cells, B cells and monocytes in the harvest product [99–101]. The presence of some of these cell types can have important biosafety implications in immunocompromised patients. In fact, Mutis and colleagues [102] observed association between a high frequency of histocompatibility-minor antigen-specific mismatch and the development of grade II–IV GvHD. This pathological event after adoptive immunotherapy was explained by the presence of allo-reactive clones in the infused product [103].

The establishment of good manufacturing practices (GMP)-based facilities and methodologies to achieve homogenous functional T cell preparations would favour the development of these techniques to a clinical scale improving the standardisation of the process of cell selection methods for immunotherapy applications. Thus, the above-mentioned results would justify an increased investment of funds and resources to develop a safe approach to implementing these promising genetic and cellular therapies in medical practice. To identify funding sources to support the viral constructs and cellular products manufactured under GMP conditions, and to overcome the complex legal barriers, should be a priority for academia and industry, which we need to further develop these advanced technologies [104].

The therapeutic virus-primed T cell status determines its long-term persistence in the recipient

So far, it has been possible to reactivate the memory CTL pool through ex vivo antigen presentation by B cells-, fibroblasts-, DCs- and BLCL-based systems. This is possible because of the high prevalence of CMV, EBV and AdV infections in the human immunocompetent population. Theoretically, this implies that memory rather than naïve T cells would be the best candidate to be selectively expanded ex vivo. However, there is no doubt about the notable influence of the in vitro culture period to obtain differentiation status-specific phenotype T cell clones and their permanence ability in vivo [105]. A few years later, this natural phenomenon was demonstrated by Berger and colleagues [106], who have shown, in an experimental model of macaque with persistent CMV infection, that for its treatment the use of this ex vivo expansion system from effector memory T cell clones can hamper both correct homing to lymph nodes and bone marrow (BM) and their survival for a limited time in vivo. On the other hand, if T cells are derived from central memory clones, they retain their ability to respond to CMV, expand in vivo and undergo phenotypic conversion to both central- and effector-memory T cells. Therefore, selection and infusion of more incipient primed cells would ensure that, after the therapeutic procedure, virus-specific cytotoxic T cells would permanently re-establish the immune memory response [106–108]. The combination of this enrichment system with molecular engineering would have a high therapeutic potential. Thus, Wang et al. [109] isolated polyclonal T cells with a central memory-like phenotype (CD8+ CD45RA− CD62L+) which exhibited anti-tumour activity after CD19-specific CAR expression. Recently, the enrichment of CD8+ CD62L+ T cells using magnetic microbeads technology and priming with peptide-pulsed APCs before transduction with a lentivector encoding CD19-CAR has allowed the engineering of CMV- and EBV-specific modified T cells. These cells were further selected with reversible streptamers and showed equivalent T cell responses to tumours and endogenous viral-bispecific TCRs [110]. These results have shown that the previous selection of specifically primed T cells significantly increases the efficiency of the harvest. The studies conducted by Hinrichs and colleagues [111, 112] went further by showing, in a transgenic murine model of adoptive immunotherapy, that T cells from the naïve compartment resist terminal differentiation and possess the highest expansion potential and anti-tumour activities. This finding implies that the pre-priming of lymphocyte subpopulations improves gene modification, leading to a high transduction efficiency and transgene expression. The use of these T cells would require in vitro priming and extensive expansion from both seropositive and seronegative donors indistinctly. Following this line, Hanley et al. has achieved extensive ex vivo expansion of naïve T cells isolated from cord blood utilising EBV-infected B cells as APCs, after their modification with adenoviral vectors. In this way, they have generated large numbers of CMV-, EBV- and AdV-specific T cells [113]. Interestingly, expanded cytotoxic T cells from naïve precursors exhibited anti-leukemia activity after transplantation [114]. In addition, this methodology would bypass terminal differentiation, as demonstrated by Gattinoni and colleagues. Otherwise, terminally differentiated T cells would be less effective at triggering disease regression in vivo [105, 115, 116]. However, data published by several research groups indicate that naïve T cell populations contain alloreactive precursors. Thus, these authors have found high frequencies of GvHD in mice infused with naïve T cells in comparison with memory T cells [117, 118]. In agreement with this observation, Distler and colleagues [119] have demonstrated alloreactivity of sorted naïve T cells against single class I or II mismatched MHC alleles, questioning the validity of naive T cells as a substrate for adoptive immunotherapy in BM transplant recipients. Recently, Gattinoni and colleagues [120, 121] have identified in humans a lytic T cell memory subset with phenotypic (CD45RA+ CD62L+ CD95+ CCR7+ IL-7Rα+ IL-2Rβ+ CXCR3+) and functional (IFNγ+ IL-2+ TNFα+) characteristics shared by naïve T and stem cells [122]. Consequently, the use of T cells derived from naïve precursors requires an exhaustive analysis before its routine application in human therapy.

Microenvironment of the recipient, a highlighted variable in the multivariate equation of biological immunotherapy

The host lymphoablative conditioning regimen essential for haematopoietic transplant carried out before adoptive T cell transfer-based immunotherapy may enhance anti-tumour responses by the modulation of the microenvironment through a range of mechanisms, including the inhibition of endogenous CD4+ CD25+ FOXP3+ regulatory T cells (T_regs), upregulation of MHC class I proteins, increase in the pool of peptides available for presentation, T cell trafficking, potentiation of innate immunity and increase in homeostatic cytokines (IL-2, IL-7, IL-15 and IL-21) [123–127]. This subject has been the central issue of different experimental studies for over a decade in radiated or pharmacologic-treated recipients.

Furthermore, T_regs suppress effector T cells by a number of mechanisms, including an increase of the activation threshold of effector T cells, expression of inhibitory costimulatory molecules, induction of anti-inflammatory biochemical pathways, direct or indirect killing, consumption of proinflammatory cytokines or production of immunoregulatory cytokines [128]. The downregulation of T_regs by exogenous immunostimulatory agents could therefore potentially improve the migratory properties, engraftment and cytolytic activity of the transferred T cells. Similar actions have been demonstrated through combined therapy of daclizumab (humanised anti-CD25 monoclonal antibody) with peptide vaccine, administering to breast cancer patients. In this clinical trial, T_regs eradication in situ and reprogramming induces robust augmented of physiological CTL and T helper response [129, 130]. Modulation of inducible T_regs as a consequence of the preparative regimen for transplant (immunosuppression) triggers inhibitory counterproductive cellular mechanisms that could limit the therapeutic potential of the infused cells.

Common γ-chain cytokines, including IL-7 and IL-15, have been reported to induce vital cellular activity such as proliferation of human T cells in the absence of TCR stimulation, furthermore avoiding apoptosis and maintaining cell metabolism after transfer into the lymphopenic host (homeostatic expansion). The absence of some of these homeostatic cytokines could result in the metabolic atrophy of infused T cells, which leads to delayed growth and proliferation following viral stimulation [125, 131]. Thus, it is known that expression of IL-7 and IL-15 receptors is key to the establishment of resting memory cells in cellular therapy procedures, as both cytokines synergistically drive T cells through this crucial checkpoint in their differentiation process [132]. Shen et al. demonstrated that downregulation of both IL-7R and IL-15R is likely to be a contributing factor for the poor survival of therapeutic influenza-specific memory CTLs in the respiratory tract. According to these authors, there could even exist molecular mechanisms that would condition their survival depending on the particular tissue [133]. On the other hand, this in vivo expansion is also thought to be driven by different factors such as self-peptides and other antigens. Furthermore, exposure to viral antigens during the period of profound lymphopenia results in a significant boost of cellular immunity [90, 134, 135], and it is a mandatory requirement for antigen-specific immunological recovery in the transplanted patient.

To date, the essential role of T helper cells for the maintenance of CD8s is a controversial issue, and different studies have sometimes shown inconsistent results. Cobbold and colleagues [90] observed no correlation between CMV-specific CD4+ T cells and the circulating level of CD8+ T cell. However, this is likely due to the low sample size (n = 5), so a declining trend in both subpopulations was observed in 4/5 patients 60 days from the time of infusion of the therapeutic product. Several studies confirmed this trend for a relatively long time. Riddell et al. [12] demonstrated that the adoptive transfer of donor CMV-specific CTLs in transplant patients results in a fast therapeutic activity, although their long-term maintenance is hampered by the absence of the appropriate T helper subpopulation. In line with these observations, Rosenberg and colleagues reported that HIV-1-infected subjects without acquired immunodeficiency syndrome development show a high virus-specific CD4+ proliferative response and an extremely vigorous CTL concomitant response [136]. The particular mechanisms whereby CD4+ T cells maintain effective anti-viral immunity are poorly understood, but they could be related to the orchestration of CTL precursor activity. According to this, there are already studies raising tumour-specific CTL responses in which a simultaneous activation of the T helper subpopulation is found after cellular vaccination with class I and II peptide-loaded DCs vaccines. With this strategy, bi-functional CD4+ activity is generated, resulting in increasing CTL proliferation and T_reg inhibition [137]. In order to examine the relationship between both protective immunological populations and their role during the immunological recovery process in the post-transplant period, the development of MHC class II multimer complexes is being encouraged by biotechnology companies [138]. In fact, more studies are incorporating them as a research tool. As a matter of fact, the use of T helper cells as a therapeutic product has also been put in question, and the limited experience with these cells in adoptive cell therapy is insufficient to obtain objective conclusions. Quezada and colleagues transferred tyrosinase-related protein 1-specific TCR CD4+ cells into an irradiated RAG2^−/− mouse model with advanced melanoma. Tumour eradication was mediated by cytotoxic CD4+ T cells, which appeared after acquisition of IFN-γ, TNF-α, IL-2, granzyme B and perforin expression. Other endogenous immunological subpopulations (helper- and cytotoxic-T, B and NK cells) did not play an essential role in the anti-tumour effect of these infused cells [139]. The lymphopenic microenvironment and the depletion of CD25+ FOXP3+ CTLA-4+ CD4+ T_regs apparently played a crucial role in the in vivo priming, expansion and activation of the exogenous naïve CD4+ cells and their cytotoxic-like phenotype conversion [140]. Furthermore, these conditions allowed the long-term establishment of a CD4+-memory subpopulation [141].

Finally, it is important to mention that many of these cytokines may play a role in manipulating the microenvironment in the context of allogeneic HSCT, and that there are ongoing clinical trials evaluating several modifications of the different conditioning regimens including non-chemotherapy-based conditioning regimens for transplantation [142, 143].

In vitro modulation of the functional ability of viral-specific CTLs by multimer complexes

In contrast to the other T cell selection techniques, multimer technology (Fig. 2) developed by Altman et al. [144] allows the identification and enrichment of viral-specific CTLs without altering their differentiation status. However, recent studies with soluble experimental MHC class I-tetramers have shown that continuous in vivo administration of an MHC multimer, induces unexpected outcomes in the antigen-reactive CD8+. An increase in the frequency of annexin V staining was observed, and could be attributed to the induction of cellular anergy or activation of induced cell death [145]. In the first case, anergy can be triggered by a strong “signal 1” provided by the binding of the experimental tetramer with its specific transgenic TCR, in the absence of “signals 2” (co-stimulation) and “signals 3” (cytokine priming) [53]. Instead, a limited expansion of these CD8+ T cells and their effector activities are compromised by the continuous presence of the soluble tetramer within the transgenic mice, leading to either clonal exhaustion or anergy, resulting in T cell dysfunction. In agreement with this, Neudorfer and colleagues [146] have observed in vitro that both peptide-specific activation-dependent cytotoxic activity and proliferation capacity of primed T cells were impaired and decreased following staining with conventional tetramers. Thus, gene transcription after tetramer complex–TCR interaction has previously been shown in experimental models. Other authors have confirmed these observations [147, 148], and some have even demonstrated the loss of protective capacity to Listeria monocytogenes in BALB/c mice after transference of CTLs pre-treated with MHC-tetramers [149]. However, administration of CMV-specific CTLs in stem cell transplant recipients that had been previously selected using tetramer complexes contributed efficiently to the control of virus dissemination [90]. In this study, the authors demonstrated persistence of CMV-specific CTLs at least 110 days after infusion, even in patients without CMV-primed CTLs before cell transfer, which suggested expansion of the infused cells. This would discard anergy in the infused T cells, at least for this period of study.

Fig. 2.

Multimer technology allows the identification and enrichment of viral-specific CTLs. A diagrammatic representation of the structure of three different multimers: tetramers, pentamers and streptamers together with the IFN-gamma (γ) catch reagent and their mechanism of action. The IFNγ-capture, pentamers and streptamers but not the tetramers are currently produced under GMP conditions

Therefore, experimental and physiological microenvironments show conflicting results about the long-term cellular effect induced by the multimer on the virus-specific CTL. However, it is currently unknown whether pentameric constructs would also induce anergy in these T cells. Pentamer and tetramer technologies are similar tools that use the same molecular approach, so it is theoretically possible that a similar impairment of T cell phenotype and function could occur in vitro using pentamers. However, Uhlin et al. demonstrated the presence of functional EBV-specific CTLs that induced regression of an EBV-driven lymphoma in vivo up to 189 days after infusion. It is worth mentioning that complete regression of EBV infection-associated lymphoma occurred in a patient diagnosed with post-transplant lymphoproliferative disease after therapeutic treatment [91].

One possible explanation to this mismatch between in vivo and in vitro results is the possibility that T cell hyporesponsiveness negatively modulated by multimer engagement can be reversed by a cytokine storm generated in the lymphopenic recipient. That assessment is proposed by Brown and colleagues [150], who demonstrated that anergic anti-tumour CD8+ T cells restore their function after transfer into RAG2^−/− immunodeficient recipients promoting tumour rejection. Specifically, Teague et al. [151] has shown that exogenous addition of IL-15 rescued and expanded previously tolerised cytotoxic T cells in vitro. However, no data in the literature have been published concerning the association between cellular functional immunomodulation induced by the multimer complexes and the influence of the host microenvironment.

As discussed above, the influence of tetramer and pentamer staining on T cell functional status remains an important unresolved issue. This could substantially limit their clinical applicability. In order to address this problem, a reversible human MHC/peptide multimer, called streptamer, was constructed by Neudorfer and colleagues [146], using the molecular technology proposed by Knabel et al. [149]. “Naïve” antigen-specific T cells can be obtained after multimer complex staining. The reversible binding between both structures (streptamer–TCR) allows its easy disengagement through exposure to a competitor molecule. Following dissociation, the streptamer-treated CTLs are functionally indistinguishable from untreated T cells. Wang and colleagues [152] demonstrated that in vitro treatment of OT-I TCR-transgenic CTLs with peptide-loaded OT-I-streptamers markedly increased ³H-thymidine incorporation and upregulated early activation markers. The biochemical analysis of the signalling pathways in this assay identified several signalling molecules which were regulated after streptamer engagement. Sustained phosphorylation of Akt and ERK1/2 was observed, possibly increasing Bcl-x_L expression, thus resulting in cell survival. According to these authors, streptamer engagement is not “silent”. However, this positive effect may be favourably used in adoptive immunotherapy protocols. The performance of cell enrichment methods using streptamer complexes has significantly improved in just a few years [103, 110, 153]. Recently, this methodology has even been combined with other genetic approaches, in which in vitro analysis has shown excellent T cell-dependent anti-viral and anti-tumoural activities, thus demonstrating the bi-specific system validity and the potential of this multimeric methodology [110]. However, it is unclear which signalling pathways may be activated by the streptamer–TCR interaction, or whether this would have synergistic effects with other signal transduction cascades. Whether these molecular changes in T cells would affect their therapeutic behaviour once administrated to the cancer patient is an open question. More detailed studies will be necessary to clarify this issue.

Concluding remarks: interrelationship between in vitro and ex vivo potential of the virus-specific primed T cells

The particular physiological characteristics of the recipient’s haematopoietic progenitors, the cellular composition and purity of the therapeutic product, and the selection method used to isolate or develop antigen-specific T cells will determine the long-term persistence of transferred lymphocyte populations. In some cases, the altered function and differentiation of these cells significantly affected the performance of adoptive immunotherapy used in SCT patients, making it difficult to predict clinical results. In future, multivariate analysis will be necessary to understand the interaction between all these physiological variables and to determine the optimal method of achieving anti-viral T cell immunity.

Electronic supplementary material

Below is the link to the electronic supplementary material.