The T-Cells’ Role in Antileukemic Reactions - Perspectives for Future Therapies’

1.1 Background (Schmid, Schmetzer) Highly specialized and sensitive defence against infections as well as tumors is provided in great part by the adaptive, T-cell-mediated part of our immune system. Those T-cell specialists arise out of a cell pool of 25-100 million distinct naive T-cell clones, after a very efficient priming phase by dendritic cells (DC), that present antigens in the context of major histocompatibility complexes (MHC), T-cell receptors (TCR) and costimulatory signals. It is well known, that T-cells are most important effectors of cellular tumor immunity and carry long lasting memory. Therefore recent tumor research has focused on the development and improvement of T-cell based immunotherapies (Barrett & Le Blanc, 2010; Smits et al., 2011). The T-lymphocyte pool includes naive (Tnaive), effector (Teff), effector memory (Tem) and central memory (Tcm), either CD8 or CD4 T-cells. After an antigen driven TCR engagement Tnaive proliferate and give rise to large numbers of Teff. Most of them die after depleting target cells. Some ‘memory T-cells’ of either Tcm or Tem phenotype remain, persist and can differentiate to Teff after a re-challenge with antigens. It has been demonstrated that memory T-cells are able to self-renewal and differentiation (Sallusto et al., 2005). However antitumor immunity is limited by regulatory T-cells (Treg) that are in general responsible for the prevention of autoimmunity, regulation of inflammatory antimicrobial or antitumor reactions and are regarded as the mediators of tolerance (Vignali et al., 2008). Treg reactions can be either mediated by inhibitory cytokines (IL-10, IL-35, TGF-s), Granzyme A/B dependent cytolysis, CD25 dependent IL-2 deprivation-mediated apoptosis, adenosine receptor mediated immunosuppression of DCs’ maturation (Vignali et al., 2008). Treg subtypes can be identified by their expression profiles and be subdivided in resting or activated Tregs of either CD4 or CD8 subtype (Miyora et al., 2009; Jordanova et al., 2008; Schick et al., 2011). Resting Tregs can convert to activated Tregs after proliferation. The functional repertoire of T-cells depends on their survival and their cooperation via cellular contact and the secretion of humoral factors. Improved knowledge of these properties should contribute to detect or select T-cells with defined markers or functional profiles for adoptive therapy.

In the past the potential of adoptively transferred T-cells to eradicate tumor cells has been intensively studied in patients with melanoma, EBV-associated tumors and especially in the context of a transplantation of allogeneic hematopoietic stem cells including immunocompetent cells in patients with lymphoid-or myeloid-derived malignant diseases (Kahl et al., 2007;Kolb et al., 2004;Schmid et al., 2011;Parmar et al., 2011;Moosmann et al., 2010).In patients with acute myeloid leukemia (AML), a malignant clonal disease of the hematopoietic stem cells, conventional chemotherapy induces remission in 60-80% of patients.However, relapse occurs in the majority of patients in the following two years (Buechner et al., 2003).Allogeneic stem cell transplantation (SCT) is considered as a curative therapy for patients with AML: Following engraftment of donor cells, the established hematopoietic chimerism persists even after discontinuation of immunsuppressive therapy, reflecting tolerance both in Graft-versus-Host and Host-versus-Graft direction (Kolb et al., 2004;Schmid et al., 2011).T-cells of the healthy donor mediate the 'graft versus leukemia' (GvL), effect which was perceived in principle as early as in the 1960ies.Clinical evidence for the efficacy of GvL reactions in AML came from the observation, that leukemia relapse after allogeneic SCT, the most frequent reason of treatment failure in AML, occurred most frequently in patients who had either been transplanted from a syngeneic twin, or had received a graft which had been depleted from donor T-cells prior to transfusion.In contrast, relapse incidence was lowest in patients developing acute or chronic Graft-versus-Host Disease (GvHD), which represents the second clinical manifestation of the allogeneic immune reaction mediated by donor T-cells (Hemmati et al., 2011;Schmid et al., 2007;Schmid et al., 2011;van den Brink et al., 2010).Based on these observations, the infusion of donor lymphocytes (DLI) was developed as a form of immunotherapy for relapsed disease after allogeneic SCT.However, up to now, not every patient, in particular those with rapidly proliferating AML, responds to or permanently benefits from those T-cell based immunotherapy.

Aims (Schmetzer)
The ability of T-cells to eliminate tumor cells and even to cure tumors has been demonstrated in experimental animal models.In man the development of effective T-cell therapies to treat human tumors remains still a challenge.Tumor antigens that elicit curative responses have been identified in animal models; in man tumor associated antigens are also known, but their use is limited due to HLA-restriction and limited duration of responses.This applies also to hematological malignancies as AML.In the last years several approaches have been made to further characterize T-cells and to explore the role of soluble and cellular factors on the regulation and mediation of antitumor reactions in AML-patients.Known tumor antigens may help to identify specific T-cells and their subtypes.This could be monitored in the course of treatment and they could be prepared for further treatment.However the majority of potential tumor antigens is unknown.Their presence may be deduced from T-cell reactions initiated and mediated by leukemia-derived dendritic cells (DC), presenting the whole antigenic repertoire of the leukemic cell.In analogy to T-cell reactions against known tumor antigens DC-stimulated T-cells reacting against unknown tumor antigens may be analysed against healthy cells before further use.In addition the identification of possible immune escape phenomena in cases without successful SCT or DLI or without antileukemic functions ex vivo could contribute to develop strategies to overcome those immunological barriers.
Using these methods we could evaluate antigen expression profiles on T-cells, antigen presenting cells or blast cells as well as cytokine secretion profiles assigning these profiles to cellular subtypes and correlate them with antileukemic reaction profiles or the clinical response to immunotherapies.Moreover we could contribute and quantify reaction profiles of 'antigen stimulated' or '-unstimulated' as well as of (specifically) selected, enriched or cloned T-cells against blast targets.Statistical evaluations were performed with standard excel programmes or SPSS software.

Experimental key results
In the following chapter the most important experimental results generated by our group are summarized.Responsible co-workers and cooperation partners are given and their contributions listed below.In haploidentical SCT we and others observed that female-donors (especially mothers) show a higher GvL reactivity against male-patients (particularly sons) compared to all other haploidentical donor-recipient combinations for AML patients (Stern et al., 2008).These effects could be due to Y-chromosome-encoded male minor histocompatibility antigens (minor-H-Ags, mHAs) recognized by female alloimmune effector (memory) T-cells, immunized during pregnancy, in the context of a GvL-reaction (Ofran et al., 2010).We studied the expression profiles of Y-chromosome genes in healthy male stem cell donors compared to male AML patients in order to possibly detect new AML-typical Y-restricted expression patterns.Blasts from male patients with acute (myelo) monocytic leukemia and monocytes from healthy male donors as the healthy control counterpart were used to determine and compare the expression profiles of Y-chromosome genes.We could detect several genes being up-regulated in male AML-cells.Among those, we focused on PCDH11, VCY, TGIF2LY, known to be expressed only in male tissues and its X-chromosome-encoded homolog TGIF2LX (Blanco-Arias et al., 2002;Skaletsky et al., 2003;Gallo et al., 2011).In a next step, we studied the immunological impact of the identified Y-encoded genes.We analyzed the proteins encoded by those four genes for the presence of nonameric peptides that could potentially bind HLA-A0201 molecules.Such analysis was performed with the help of publicly available peptide-motif scoring systems (http://bimas.dcrt.nih.gov/molbio/hla_bind/ and http://www.syfpeithi.de;Rammensee et al., 1995).High-scoring peptides were not found for PCDH11, whereas HLA-A0201-binding peptides could be identified in silico for the VCY, TGIF2LY and TGIF2LX genes.Their effective binding efficacy was determined in a standard HLA-A0201-T2 binding assay: Two peptides derived from VCY as well as from TGIF2LX and one derived from TGIF2LY were able to bind to the HLA-A0201molecules of the T2 cells.Furthermore these peptides were tested for their ability to induce a CD8+ T-cell response: we selected CD3+ T-cells from female volunteers and cultured them in the presence of T2 cells loaded with the different Y-chromosome-encoded peptides for four weeks.We could show that the VCY-encoded peptide G54 stimulated female effector Tcells as shown by a specific lysis of G54-loaded T2 target cells in a chromium release assay (Figure 3.1.1-1).Further research will focus on the immunogenic G54-peptid particularly with respect to our assumption that mothers (who had given birth to a son) bear CD8+ Tcells being reactive to male-specific antigens, leading to a strong GvL effect.Moreover, peptides restricted to other HLA-types will also be investigated.

T-cells addressing known (leukemia
In conclusion we identified three new potentially antileukemic, male-specific human genes being upregulated in blasts of male AML-patients.These might be valuable targets for T-cells. As a rule Y-chromosome coded antigens are expressed in all cells of the organism.However some data have been reported, that differentially spliced forms of the Y-linked UTY (ubiquitously-transcribed tetratricopeptide-repeat-gene, Skaletsky et al., 2003) may be restricted in a tissue restricted fashion (Warren et al., 2000).UTY is highly conserved in man and animal species.We worked out in a dog model in vitro and in vivo that the Ychromosome coded mHA UTY might be a promising candidate target-structure to improve GvL immune reactions after SCT: female T-cells were stimulated with either autologous (female) DCs, loaded with three different UTY-derived (male) peptides or with allogeneic, donor-matched male cells (PBMCs) endogenously expressing UTY.We identified 3 out of 15 identified UTY-encoded peptides bearing immunological potential to stimulate 'antimale'-(i.e.anti-UTY-) directed immune reactions.Amongst those, W248 showed highest immunogenic potential in both in vitro and in vivo settings: In vitro expanded CTLs specifically recognized mainly bone-marrow (BM) from DLA-identical male littermates in an MHC-I-restricted manner (Figure 3.1.1-2;in vitro).In vivo, comparable W248-(UTY-) specific reactivity against BM was also obtained after stimulation and immunization of a female dog with DLA-identical male PBMCs (Table 3.1.1-1;in vivo (Bund et al., 2001)).Table 3.1.1-1Female dog T-cells stimulated with UTY (W248)-peptides loaded on autologous DC in vitro or male DLA-identical PBMC in vivo specifically recognized 'male' target cells Taken together we could demonstrate, that female dog effector T-cells could be specifically stimulated against male, UTY-gene product specific cells meaning in turns, that UTY seems to be a promising candidate antigen to improve GvL reactions after SCT.

LAA-specific CD8+ T-cells (Groupleaders: Busch, Borkhardt, Kolb; scientists: Doessinger, Steger, Schuster)
Between 60-90% of AML cases overexpress leukemia-associated antigens (LAA), that means antigens that are absent or only weakly expressed in normal tissues (e.g.WT1, PR1 or PRAME (Greiner et al., 2006;Steger et al., 2011)).Therefore T-cell based immunotherapeutic strategies addressing those LAA-expressing cells could be promising.Alternatively mHAs, preferentially expressed on hematopoietic cells, could qualify as T-cell targets in a GvL reaction (Mutis & Goulmy, 2002)         In summary this means that the cytokine release of IFN-or GM-CSF is no reliable tool to detect specificity in T-cell fractions.We could demonstrate however, that CD4+ T-cells in general can mediate cytotoxic reactions and that certain enriched CD4+ subtypes might be even more promising candidates for GvL-without GvH reactions.Therefore we wanted to study the antileukemic activity of T-cells stimulated with DC derived from leukemic blasts (DC leu ), as those DC bear the advantage of potentially presenting the whole (including as yet unknown) and patient-typical antigen pool of the leukemic cells.We have already established the methods to generate sufficient amounts of DC and especially of DC leu in every given case with AML: in a minimalized assay we cultured DC in three different media and choose the DC generation method with the quantitatively highest DC/DC leu counts (Kremser et al., 2010;Schmetzer et al., 2007).In a next step we study the functional profiles of DC/DC leu -stimulated T-cells.We could demonstrate, that isolated unstimulated T-cells were able to lyse blasts in 47% of cases, whereas only 26% of those T-cells showed antileukemic activity after a 10 days culture with blasts -pointing to an establishment of a T-cell inhibitory microenvironment in the presence of blasts.However, stimulation of T-cells with blasts after their conversion to DC leu resulted in a blast lytic activity in 58% of all cases.These data suggest, that the inhibitory (blastinduced) atmosphere could be abolished after blast-conversion to DC leu , although not completely: even a DC/DC leu stimulation of T-cells was not effective to induce antileukemic T-cells in every case (Schmetzer et al., 2011;Grabrucker et al., 2010).Therefore we analysed potential conditions and factors being responsible for these impaired immune reactions.We could demonstrate that the quality of DC -especially with respect to proportions of mature DC and DC leu -is predictive for their activation capacity for antileukemic T-cells.Those DC/DC leu induce an ‚antileukemically effective' T-cell composition of DC-stimulated T-cells, characterized by higher proportions of CD4+ and non-naive T-cells (Grabrucker et al., 2010;Liepert et al., 2010).A detailed analysis of T-cells' compositions in cases with compared to those without antileukemic activity revealed significantly higher proportions of naïve (T naive ) and central memory T-cells (T cm ) and lower proportions of effector memory regulatory (T eff/em reg ) as well as CD8+ regulatory T-cells (CD8+ T reg ) (Schick et al., 2011;Vogt et al., 2011).Moreover we could define soluble factors that are predictive for the mediation of antileukemic activity of DC/DC leu -stimulated T-cells: A higher release of 'inflammatory' chemokines (CXCL8, CCL2) in DC culture supernatants or of 'T-cell-promoting' cytokines (IFN-, IL-6) in mixed lymphocyte culture (MLC) supernatants of T-cells with DC clearly correlated with antileukemic activity of DC-stimulated T-cells (Fischbacher et al., 2011;Merle et al., 2011;Schmetzer et al., 2011).Detailed studies of T-cell subsets via spectratyping analysis could verify that especially after DC-stimulation CD4+ as well as CD8+ T-cells were characterized by a highly restricted Vß T-cell-receptor (TCR) repertoire (Schuster et al., 2008).Interestingly, in one patient studied comprehensively in vitro stimulation with DC/DC leu resulted into an identical TCR  chain restriction pattern which could be identified in vivo in the patient's T-cells 3 months after allo-SCT (Reuther et al., 2011).In summary, DC leu are promising candidates to stimulate and enrich antileukemic T-cells without knowledge of defined antigen targets which is attended with the creation of an 'antileukemic cellular microenvironment' and could contribute to develop strategies to overcome immunological resistances.With our experimental in vitro models combining culture methods and functional flow cytometry with spectratyping we can moreover provide explanations for clinical observations and might provide predictive information about T-cellular response patterns in vivo.Further studies of selected T-cells (e.g. by their Vß type) for their phenotype and function will allow to understand clinical responses and to prepare T-cells for treatment.had a more than 75% chance to gain antileukemic ex vivo activity.In addition we could demonstrate, that clinical responders to immunotherapy were characterized by a higher ex vivo generability of DC leu and mature DC, a better ex vivo T-cell proliferation and CD4:CD8 and T non-naive : T naive ratios>1 and in addition by a high release of CCL2, IFN-and IL-6 (Liepert et al., 2010;Schmetzer et al., 2011;Fischbacher et al., 2011;Merle et al., 2011;Grabrucker et al., 2010).

Antileukemic
That means, that we can not only associate cellular subtypes of T-cells and DC or cytokine/chemokine release patterns with antileukemic functions of T-cells in ex vivo settings and in the context of a clinical response to immunotherapies, but in addition define predictive 'cut-off values' , that means proportions of cells with certain cellular subtypes or concentrations of soluble factors, that allow a correlation with cellular responses or the clinical course of the disease after immunotherapy.

Clinical use of donor T-cells for prevention and treatment of AML relapse after allogeneic SCT ( Schmetzer, Schmid)
Donor lymphocyte infusion (DLI) for treatment of leukemic relapse after allogeneic hematopoietic stem cell transplantation (SCT) has been introduced in the early nineties (Kolb et al., 1990).Being extremely effective in chronic myeloid leukemia, the procedure was less successful in AML, although remissions were observed in selected cases (Kolb et al., 1995;Collins et al., 1997).Therefore, on behalf of the Acute leukemia Working Party of the European Group of Blood and Marrow Transplantation (EBMT) our group performed a retrospective analysis of patients who had been transplanted for AML in complete remission, and had suffered from leukemia relapse post SCT (Schmid et al., 2007).The analysis was based on the EBMT transplant registry, and included 399 adult patients who had received (n=171) or not (n=228) DLI as part of their treatment.With a median follow up of 27 and 40 months in the both groups, overall survival (OS) at two years was 21±3% for patients receiving, and 9±2% for patients not receiving DLI.After adjustment for differences between the groups, better outcome was associated with younger age (p= 0.008), remission duration >5 months after SCT (p<0.0001), and use of DLI (p=0.04).Among DLI recipients, a lower tumour burden at relapse (<35% of bone marrow blasts; p=0.006), female gender (p=0.02),favorable cytogenetics (p=0.004) and remission at time of DLI (p<0.0001) were predictive for survival in a multivariate analysis.Two-year survival was 5610%, if DLI was performed in remission or with favourable karyotype, and 153% if DLI was given in aplasia or with active disease.Therefore, an algorithm for the clinical use of DLI in the treatment of relapsed AML after allogeneic SCT was developed, comprising the sequence of cytoreductive chemotherapy for disease control or induction of complete remission, followed by DLI for long term control of the leukemia based on cellular immune effects (Schmid et al., 2011).
In an approach to increase the antileukemic efficacy of donor T-cells against myeloid leukemias, systemic application of GM-CSF was studied after DLI for relapse of AML or MDS after SCT (Schmid et al., 2004).GM-CSF was chosen due to its capacity to contribute in vitro to the generation of antigen-presenting cells (APC) from leukemic blasts (Woiciechowsky et al., 2001;Kufner et al., 2005;Kremser et al., 2010;Dreyssig et al., 2011).
As described above, blasts from myeloid leukemias should have the full genetic repertoire for effective antigen presentation, but might be ineffective stimulators or even induce specific anergy, due to inferior or aberrant expression of co-stimulatory molecules, such as CD80 or CD86.GM-CSF has been shown to induce up-regulation of these molecules on the surface of leukemia blasts and to improve cytotoxic efficacy of autologous and allogeneic T-cells.In a clinical pilot trial for AML relapse after SCT, mild chemotherapy with low dose AraC, infusion of donor T-cells together with stem cells for reconstitution of haematopoiesis, and s.c. or i.v.application of GM-CSF, was studied.Overall response rate was 67% among evaluable patients', overall survival at 2 years was 29%.Long term survival was associated with longer remission post transplant, disease control by low dose AraC and development of chronic GvHD.These results confirm the proposed strategy of initial cytoreduction by chemotherapy and induction of a GvL reaction for long term disease control.Systemic application of GM-CSF was safe in this setting, however, its clinical efficacy remains to be evaluated in randomised studies.Nevertheless, in accompanying ex vivo experiments we could demonstrate, that cases in which DC could be generated ex vivo using GM-CSF-based protocols showed a more favourable outcome after in vivo immunotherapy (Freudenreich et al., 2011).Since overall, the outcome of patients with AML who relapse after allogeneic SCT is poor, strategies to prevent occurrence of overt haematological relapse are of increasing interest.
Intensive monitoring of minimal residual disease and donor chimerism in different cellular compartments (CD34+, CD3+) of bone marrow or peripherial blood has gained in importance by allowing early interventions (chemotherapy, DLI, second SCT) before haematological relapse has occurred (Bornhauser et al., 2009).Our group has developed a protocol for the use of prophylactic or preemptive DLI (pDLI) for patients with high-risk AML.Starting from day +120 after SCT, patients in haematologic remission, free of immunosuppression for at least 30 days without clinically evident GvHD, and free of infections receive up to 3 courses of DLI in 4 weeks' intervals, using an escalating cell dose schedule.Patients receiving prophylactic DLI have been compared to a control group of high-risk AML patients, who were treated according to the same transplant protocol, would have fulfilled the criteria for pDLI, but did not receive the cells since their transplant centres did not take part at this part of the study (Schmid et al., 2005;Schleuning et al., unpublished results).Hence, patients receiving pDLI showed a significantly lower incidence of relapse and achieved a longer overall survival as compared to controls without pDLI.The treatment was also safe and induction of severe GvHD was a rare event in this setting.In summary, although less effective as in CML, the clinical use of DLI for AML patients after SCT is an effective therapeutic tool for prevention or as part of treatment of relapses after SCT in AML.

Perspectives for future therapies: Adoptive transfer or in vivo activation of antileukemic T-cells? (Schmetzer, Schmid)
We could clearly demonstrate, that it is possible to detect and monitor leukemia specific Tcells by LAA-peptid specific (HLA-A2 restricted) MHC-Multimer analyses or LAA-protein specific CD4+ T-cells, especially if combined with spectratyping and cellular subtype analyses.Concerning 'known' (leukemia-) specific antigens we can conclude from our data, that LAA-peptide specific CD8+ T-cells can be prepared by MHC multimer-technology, LAA-protein specific CD4+ T-cells by preparation of (enriched) proliferating CD40L+ or cloned CD4+ cells and both cell types could be used for adoptive therapies.Moreover we could work out in vitro as well as in a dog model, that male specific antigens might be promising candidate antigens for immunotherapies.In addition we could show, that (enriched) T-cells addressing known as well as unknown leukemia-associated antigens, as demonstrated after DC/DC leu -stimulation, can mediate cytotoxic reactions.Those cells could be promising candidates for adoptive immunotherapies in selected patients.
Since the manipulation and selection of antigen-specific T-cells is not only an oncological challenge, but has to be approved by special committees before a clinical application another strategy circumventing T-cell manipulations could be more promising: applicating immunemodulators and cytokines like GM-CSF or IFNin vivo could possibly induce the conversion of (residual) blasts in patients to DC leu .In a small patients' cohort we could already show, that patients receiving GM-CSF in the context of a DLI-relapse therapy had a better chance to respond to this relapse therapy compared to patients without additionally applicated GM-CSF (Freudenreich et al., 2011).Moreover we could show, that the convertibility of blasts to DC leu in ex vivo settings correlated with the clinical response and outcome to immunotherapies, what can be interpretated by an 'ex vivo simulation' of the DC-generating potential out of blasts.We could even demonstrate, that cases, in that higher proportions of DC leu could be generated ex vivo had a longer overall survival compared to cases with lower DC leu proportions.
Our ex vivo focus in the future will therefore be to thoroughly investigate and optimize in vivo strategies with allo SCT applying different donor transplants or ex vivo 'manipulated' grafts.We further want to develop and test different 'immune modulating cocktails' (Ansprenger et al., 2011;Deen et al., 2011) that can be applicated to patients with the aim to induce leukemia-derived DC in vivo and in consequence to stimulate the generation of leukemia-specific T-cells in vivo.In parallel we want to further enlighten the role of different (enriched, selected) effector cells -e.g.CD4+, CD8+, NK, NK-T-cells) in the mediation of antileukemic reactions in order to find promising candidates for adoptive Tcell transfer.

Bold: > 0 .
Fig. 3.1.2-1.LAA-specific T-cells can be isolated from AML-patients after SCT by MHC multimer technique.Selected T-cells are highly Vß restricted CD4+ T-cells were characterized by ELISPOTS, ELISAS or ICS for secretion of IFN-or GM-CSF before or after single cell cloning or sorting of CD40L+ CFSElow cells.Moreover some cases were tested for antileukemic activity in a chromium-release or fluorolysis assay.

Fig
Fig. 3.1.3-1.Preparation, stimulation (I) and characterisation (II) of (untouched) CD4+T-cell Although technically possible the preparation of LAA-specific T-cells is cumbersome, in case of MHC-Multimer technology HLA-A2 restricted and moreover restricted to defined LAAs.
. Principally T-cell based strategies could be based on a vaccination with LAA/mHAs or by identification, selection and transfer of LAA/mHAs specific T-cells already present in the donor.LAA/mHAs specific T-cells can be found at a low frequency in normal persons implying a low level of immunity.

Table 3
One patient could not be analysed under a clinical point of view since he died one month after sample acquisition of an infect(pt 1148(pt  , table 3.1.2-2,Stegeret al., 2012)).

Table 3 .
1.3-2 Antileukemic functionality (demonstrated by fluorolysis assay) of different CD4+ effector cells prepared before or after specific stimulation, enrichment or sorting

T-cell profiles to predict antileukemic reactions of DC/DC leu stimulated T-cells and prognosis of patients (Groupleaders: Schmetzer; Scientists: Fischbacher, Freudenreich, Grabrucker, Liepert, Merle, Schick, Vogt)
Since not every AML patient responds to immunotherapy (SCT, DLI) in vivo and since not every ex vivo T-cell stimulation with DC leu results in antileukemia effector T-cells we wanted to elucidate responsible cells or soluble factors.We could evaluate 'cut-off values' for DC-or T-cell subtypes in the cellular settings and in addition amounts of soluble factors that allow a prediction of the antileukemic function of T-cells in this cellular or microenvironmental context or a prediction of the clinical response to immunotherapy.We could demonstrate, that T-cells stimulated with DC leu in a 'favorable cellular and soluble chemokine/cytokine context' (with >45% mature DC and >65% DC leu with a release of >200pg CXCL8, >100pg CCL2, >10pg IFN-or >15pg IL-6, resulting in >65% CD4+, T non-naive and <60% CD8+ T-cells and especially >3% T naive , >11% T cm and low proportions of CD8+ T reg after the DC stimulation)