Dendritic Cells of Leukemic Origin: Specialized Antigen-Presenting Cells as Potential Treatment Tools for Patients with Myeloid Leukemia

Abstract

The prognosis of elderly patients with acute myeloid leukemia (AML) and high-grade myelodysplastic syndrome (MDS) is limited due to the lack of therapy options and high relapse rates. Dendritic cell (DC)-based immunotherapy seems to be a promising treatment tool. DC are potent antigen-presenting cells and play a pivotal role on the interface of the innate and the adaptive immune system. Myeloid leukemia blasts can be converted to DC of leukemic origin (DC_leu), expressing costimulatory molecules along with the whole leukemic antigen repertoire of individual patients. These generated DC_leu are potent stimulators of various immune reactive cells and increase antileukemic immunity ex vivo. Here we review the generating process of DC/DC_leu from leukemic peripheral blood mononuclear cells as well as directly from leukemic whole blood with “minimized” Kits to simulate physiological conditions ex vivo. The purpose of adoptive cell transfer of DC/DC_leu as a vaccination strategy is discussed. A new potential therapy option with Kits for patients with myeloid leukemia, which would render an adoptive DC/DC_leu transfer unnecessary, is presented. In summary, DC/DC_leu-based therapies seem to be promising treatment tools for patients with AML or MDS but ongoing research including trials in animals and humans have to be performed.

State of the Art: Current Prognosis and Therapy Options for Patients with Acute Myeloid Leukemia

Acute myeloid leukemia (AML) and (high-grade) myelodysplastic syndromes (MDS) are hematopoietic disorders characterized by uncontrolled proliferation and impaired differentiation of leukemic blasts in the bone marrow (BM). The blasts replace the physiological hematopoiesis and cause typical symptoms such as anemia, bleeding, and infections [1]. The prognosis of AML patients, especially among the elderly (age >65 years), is unsatisfactory, with 5-year overall survival rates of 5–15% in older cohorts and 30% in younger cohorts. The outcomes of (most of the time elderly) MDS patients, characterized by hematopoietic failure and increasing blasts, are even worse [2, 3, 4].

High-dose induction chemotherapy with cytarabin ± anthracycline (e.g., 7 + 3 regimen) followed by allogeneic hematopoietic stem cell transplantation (HSCT) is the only potential curative treatment and it is the standard therapy, especially for young AML/MDS patients with fewer comorbidities [5, 6, 7]. Different side effects such as graft-versus-host disease (GVHD), which mainly effects the skin, the liver, and the gastrointestinal tract characterize and limit the therapy [8, 9]. Due to high morbidity and mortality rates, HSCT is no therapy option for patients older than 65 years, who represent the main cohort of patients affected by AML or MDS [2, 5]. Measurable residual disease (MRD), previously called minimal residual disease, i.e., a small reservoir of cells in the BM far below the 5% leukemic blast threshold or in tissues, resistant against chemotherapy, is one of the main reasons for relapse [10, 11]. HSCT and donor lymphocyte infusion can affect MRD via the graft versus leukemia effect mainly mediated by allogeneic effector T and natural killer (NK) cells [9, 12]. However, relapse of AML remains the main treatment failure [13]. Low-dose cytarabin or hypomethylating agents − with or without combined Venetoclax, a drug influencing apoptosis of leukemic blasts − are therapy options for patients with a low tolerance for intensive induction therapy or refractoriness to hypomethylating agents [3, 14].

In the last few years new (non-immunotherapeutic), treatment strategies for AML patients, i.e., antileukemic protein kinase inhibitors (e.g., FLT3 inhibitors), epigenetic modulators, new cytotoxic agents, and mitochondrial inhibitors including apoptosis-based therapies and therapies that target specific oncogenic proteins, have been in development. Their clinical value still has to be demonstrated for myeloid malignant diseases [15].

Dendritic Cell Based Immunotherapy for Patients with AML or MDS

Due to the high rates of relapse or persisting disease after treatment and the limited therapy options, especially for elderly patients, there is a need for alternatives such as immunotherapy to maintain stable remissions and/or stable disease. Due to the potential induction of an antileukemia immunity against MRD, different work groups have tried to utilize dendritic cells (DC) as a treatment tool for patients with AML [16]. DC are major antigen-presenting cells (APC) that internalize and process antigens (e.g., microbe fragments and necrotic tumor cell products) and present these fragments via major histocompatibility complex I and II (MHC I and II) molecules in lymphoid organs to cells of the innate and adaptive immune system [17, 18]. On the one hand, DC activate cells of the innate immune system, e.g., NK cells or invariant natural kill cells to avoid pathogen spread until cells of the adaptive immune system are activated [19]. On the other hand, DC form immunological synapses with T cells, resulting in a clonally restricted and potent T-cell activation against the presented antigens [20, 21, 22]. In that context, DC play a crucial role on the interface and crosstalk of the innate and the adaptive immune system [23, 24].

Monocyte-Derived DC

Ex vivo DC can be generated from autologous or allogeneic CD14⁺ monocytes (monocyte-derived DC; mo-DC) [25]. Generated mo-DC have to be loaded with leukemia-associated antigens (LAA) which are overexpressed peptides or proteins on leukemic blasts compared to healthy tissue [26]. Wilms tumor antigen 1 (WT1), preferentially expressed antigen in melanoma (PRAME), and human telomerase reverse transcriptase (hTERT) are widely used LAA for the loading process of mo-DC [25, 27, 28]. Mo-DC can be pulsed with apoptotic leukemic cells, whole leukemic cell lysates or RNA electroporation of whole leukemic cell-derived RNA as well as messenger ribonucleic acid (mRNA) encoding one or even more defined LAA [16, 25, 27, 29, 30, 31, 32]. After manufacture of mo-DC under Good Manufacturing Practice (GMP), they are re-administered to the patient as an intradermal or intravenous vaccination.

Leukemia-Derived DC

In AML, MDS and chronic myeloid leukemia (CML) DC can be generated directly from the malignant leukemic cells (leukemia-derived DC; DC_leu) with different DC/DC_leu-generating protocols due to the fact that leukemic blasts and DC originate from the same precursor cells [30]. For the generation of DC_leu, peripheral blood mononuclear cells (PBMNC) from patients with AML or MDS are cultured in the presence of different combinations of response modifiers [33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51]. Leukemic blasts thereby gain a typical DC morphology with an increased expression of costimulatory molecules as well as a higher antigen presentation capacity.

Proof of Leukemic Derivation of DC_leu and Characterization of Subtypes

DC_leu are characterized by the expression of individual patients' whole leukemic antigen repertoire including known as well as unknown leukemic antigens [35, 40, 52, 53, 54]. The leukemic origin of DC_leu can be confirmed by Western blot as well as fluorescence in situ hybridization (FISH) analyses by the detection of leukemia-specific numeric or structural chromosomal aberrations in the nucleolus of the generated DC_leu [35, 44, 55].

It has to be taken into account that leukemic blasts, CD14⁺ monocytes, or CD34⁺ stem cells in PBMNC obtained from peripheral blood (PB) or BM from patients with AML can be differentiated to DC/DC_leu during cell culture [40, 56]. To quantify only generated DC_leu and to differentiate them from nonleukemic DC or unconverted blasts, a special flow cytometric gating strategy has been developed [53]. Cells have to be stained with patient-specific blast-staining antibodies (e.g., CD15, CD34, CD65, and CD117) in combination with DC-staining antibodies (e.g., CD80, CD83, CD86, CD206, and CD209), which were not expressed on leukemic blasts before DC/DC_leu culture [40, 53, 57, 58]. With that strategy DC_leu can be quantified in the total cell fraction (DC_leu/PBMNC), in the DC fraction (DC_leu/DC⁺), or in the blast fraction to quantify the amount of blasts converted to DC_leu (DC_leu/bla⁺) [53].

Only mature DC/DC_leu can activate immune reactive cells. Therefore, quantification of mature DC/DC_leu is important. These cells are characterized by expression of the chemokine receptor 7 (CCR7), which acts as a lymph node-homing receptor and is crucial for the migratory capacity of DC/DC_leu [59, 60]. Furthermore, expression of CD83, a member of the immunoglobulin superfamily, and secretion of IL 12 are also considered markers for maturation of DC/DC_leu. Membrane-bound CD83 acts as a potent T-cell activating signal and plays an important role in the generation of thymocytes as well as in modulation of the immune response [61, 62, 63]. IL (interleukin)-12 is responsible for the induction of differentiation of CD8⁺ T cells toward cytotoxic T lymphocytes (CTL), activates NK cells, and mediates TH-1 development [64, 65, 66].

Protocols to Produce DC/DC_leu from Blast-Containing PBMNC

Various attempts have been made to improve DC/DC_leugeneration from isolated PBMNC from patients with AML, which might be used for an adoptive cell transfer. For generation of DC/DC_leu from leukemic PBMNC, different protocols have been published so far and they are summarized in Table 1. The main differences in the protocols are caused by the use of different combinations of response modifiers, cell culture media (serum free or serum enriched), and culture times − resulting in different efficiencies of sufficient DC/DC_leu generation − along with the lack of comparability of different studies.

Table 1.

DC/DC_leu-generating protocols from leukemic PBMNC

Response modifiers	Protocol	Sources of cells	Culture time, days	Results	Ref.
Cytokine-based DC/DCleu-generating protocols
GM-CSF TNF-α SCF IL-6	Serum enriched	PB	n.g.	First published protocol for the generation of mature DCleu	33a
GM-CSF				80% of cases with a typical DC morphology	34
TNF-α				Generation of mature DCleu Potent antigen-presentation capacity Leukemic origin confirmed
GM-CSF	Serum free	PB/BM	10–15	↑ DC morphology of blasts	35
IL-4	Serum enricheda			Generation of mature DCleu	36a
TNF-α				↑ lysis of autologous leukemic cells ↑ IL-12 production Leukemic origin confirmed	37a
GM-CSF	Serum free	PB/B	10–15	↑ DC morphology of blasts	35
IL-4				Generation of mature DCleu
CD40L				↑ lysis of autologous leukemic cells Leukemic origin confirmed
GM-CSF	Serum enriched	PB	12–14	77% successful DCleu generation	38a
IL-4	Serum free			↑ DC morphology of blasts	109
TNF-α				Generation of mature DCleu	40
FLT3 L				↑ antileukemic immunity of T cells Leukemic origin confirmed
GM-CSF	Serum enriched	PB/BM	8	68% successful DCleu generation	41a
TNF-α				↑ DC morphology of blasts
FLT3 L				Generation of mature DCleu
TGF-β				↑ antileukemic immunity of T cells
SCF				Leukemic origin confirmed
GM-CSF	Serum enriched	PB/BM	3–5	100% successful DCleu generation	42a
SCF				↑ DC morphology of blasts
FLT3 L				Generation of mature DCleu ↑ Stimulating potential on T cells Leukemic origin confirmed
GM-CSF	Serum enriched	PB/BM	n.g.	↑ DC morphology of blasts	43a
SCF				Generation of mature DCleu
TNF-α				↑ Stimulating potential on T cells
+/– IL-4				Leukemic origin confirmed
GM-CSF	Serum enriched	PB	10	↑ DC morphology of blasts	44a
IL-4				Generation of mature DCleu	45a
TNF-α				↑ IL 12 production
CD40L				↑ Stimulating potential on T cells Leukemic origin confirmed
GM-CSF	Serum enriched	PB	8	80% successful DCleu generation	46 a, 47 a
IL-4	Serum free			Generation of mature DCleu	40
+/– TLR 3 A				↑ IL 12 production (especially with TLR4 and TLR 7/8 agonists)
+/– TLR 4 A				↑ Stimulating potential on T cells
+/– TLR 7/8 A				↑ anti-leukemic immunity of T cells
GM-CSF	Serum free	PB	9–11	Generation of mature DCleu	40
IL-4				↑ Stimulating potential on T cells
Picibanil PGE2				↑ anti-leukemic immunity of T cells
GM-CSF IL-4 Picibanil PGE1	Serum free	PB	7–10	Higher amounts as PGE2 containing protocols Generation of mature DCleu	48
GM-CSF IL-4 FLT3 L TNF-α IL-1β IL-6 PGE2	Serum free	PB	10–12	Generation of mature DCleu ↑ Stimulating potential on T cells ↑ anti-leukemic immunity of T cells	40
GM-CSF TNF-α IL-3 SCF FLT3 L IL-4	Serum free	PB/BM	14	83% DCleu generation possible No induction of T-cell-mediated cytotoxicity	49
GM-CSF TNF-α IFN-α	Serum free	PB	8–11	DC/DCleu less successful compared to other cytokine-based protocols ↑ antileukemic immunity of T cells	50
Cytokine free DC/DCleu-generating protocols
IL-4 A23187	Serum free	PB/BM	2	70% successful DCleu generation ↑ DC morphology of blasts ↑ mature DC compared to cytokine-based protocols ↑ antileukemic immunity of T cells	49 40
A23187	Serum enriched	PB	5–7	↑ DC morphology of blasts ↑ DCleu compared to cytokine-based protocols (GM-CSF, IL 4, TNF-α) I induction and proliferation of T cells	51a

↑, increase in cell amounts; +/–, with or without; n.g., not given.

^aSerum-enriched protocols.

For an adoptive cell transfer it is recommended to generate DC/DC_leu in serum-free cell culture media to avoid unspecific immunoreactions and anaphylactic or immunoreactions against xenoantigens [67]. Fetal calf or bovine serum or healthy human serum contains a large number of identified and unidentified factors (including different growth factors) that might influence DC/DC_leu generation in an unknown way [40, 67]. In general, DC/DC_leu generation has been found to be comparable in serum-free as well as serum-enriched cell culture media [67].

DC/DC_leu Differentiation and Maturation

In general, 3 steps are needed for successful generation of DC/DC_leu from leukemic PBMNC, i.e., induction of hematopoietic differentiation, danger signaling, and DC/DC_leu maturation [40, 48]. In a first step, substances (especially cytokines) such as granulocyte macrophage colony-stimulating factor (GM-CSF), ligand of the FLT3 receptor (FLT3 L), stem cell factor (SCF), interferon (IFN)-α, and IL-4 induce the differentiation of myeloid leukemia blasts to immature DC/DC_leu. In a second step molecules such as the bacterial lysate Picibanil, toll-like receptor (TLR) agonists, nucleic acids, or lipopolysaccharides or transforming growth factor (TGF)-β mediate a danger signaling that is crucial for the activation of DC/DC_leu. In the last step, response modifiers such as tumor necrosis factor (TNF)-α, prostaglandin (PG)E₂, PGE₁, and INF-α are used for the induction of maturation [35, 40, 48, 50]. The combination of response modifiers can have synergistic effects on danger signaling and DC/DC_leu maturation.

The Most Relevant Response Modifiers, Their Combination, and Their Role in DC/DC_leu Generation in Detail

Picibanil (OK-432) a lysis product of Streptococcus pyogenesis, acts as a TLR agonist, and induces danger signaling as well as maturation of generated mo-DC from CD14⁺ progenitor cells in cancer patients [68]. Picibanil-containing protocols have been shown to be usable for generation of DC/DC_leu from PBMNC from patients with AML [40, 48]. The efficiency of sufficient DC/DC_leu generation was higher with PGE₁-containing protocols (Pici-_PGE-1) in a direct comparison to PGE₂-containing protocols (Pici-_PGE-2), suggesting that PGE₁ is not just responsible for the maturation of DC/DC_leu but it is also involved in the induction of differentiation of leukemic blasts toward DC/DC_leu [48]. We could already show that antileukemic activity is improved after stimulation of T-cell-enriched immune reactive cells with DC/DC_leu generated with PGE₁-containing protocols compared to PGE₂-conaining ones [48]. This effect might be explained by the different mode of action of PGE₂ and PGE₁ on prostaglandin receptors (EP receptors) with different down streaming effects [69]. This effect might influence the expressions of indoleamin 2,3-dioxgenase-1 (IDO1; see below), which might not impair antigen presentation capacity of DC but could activate regulatory T cells [70].

Another parallel comparison of DC/DC_leu generation showed that under serum-free conditions 3 out of 5 methods qualified most to generate DC/DC_leu. With at least 1 of the 3 methods, i.e., MCM-Mimic (containing: GM-CSF, IL-4, FLT3 L, TNF-α, IL-1β, IL-6, and PGE₂), Pici-_PGE-2(containing: GM-CSF, IL 4, Picibanil, and PGE₂), and Ca-Inophore (containing: IL 4 and calcium inophores [A23187]), it was possible in every given AML patient, whereas the protocols “Cytokines” (containing: GM-CSF, TNF-α IL-4, IL-3, SCF, and FLT3 L) or “Poly-I:C” (containing: GM-CSF, IL-4, and Poly I:C) were less efficient [40].

The combination of GM-CSF, IL-4, TNF-α, and FLT3 L generated significantly more DC/DC_leu compared to cultures without added FLT3 L − as shown in serum-free and serum-enriched cell cultures [38, 39, 40, 71]. FLT3 L plays a crucial role as a growth and differentiation factor in the hematopoietic system [72]. Different TLR agonists, affecting TLR3 (e.g., Poly I:C), TLR4 (e.g., lipopolysaccharide), and TLR7/8 (e.g., R848), were used to increase the maturation of generated DC/DC_leu from leukemic PBMNC. Maturation was increased especially after the addition of TLR7/8 [46]. The combination of TLR4 and TLR7/8 agonists seems to have a synergistic effect on the generation of mature DC/DC_leu [47]. IFN-α, a protein with stimulating effects on the immune system was analyzed in a combination with GM-CSF +/– TNF-α for DC/DC_leu generation from patients with AML. Sufficient DC/DC_leu generation was possible, although amounts of generated DC/DC_leu were decreased in some settings compared to other established protocols [50]. Furthermore, the CD40 ligand (CD40L, also known as CD154), an important costimulatory molecule for the activation and maturation of DC, and SCF, which regulates death, apoptosis, cell differentiation, and migration of hematopoietic cells, have been used in different combinations for DC/DC_leu generation with various levels of efficiency [33, 35, 42, 43, 45, 73]. The addition of bryostatin-1, a substance that interferes with protein kinase C activity, increased the percentage of cases with sufficient DC/DC_leu generation when combined with different cytokine-based DC/DC_leu generating protocols [74].

To make the DC/DC_leu generation process more cost efficient, reduction of the culture time is an important factor. In general, DC/DC_leu cell cultures are conducted for 7–14 days. Two studies analyzed DC/DC_leu generation in 2 days (using calcium ionophores plus IL-4), respectively days 3–5 (using GM-CSF, FLT3 L, SCF, and TNF-α). In both cases, DC/DC_leu generation was possible in sufficient amounts, although there is a lack of comparability [42, 49].

As a cytokine-free generating protocol, calcium inophores (A23187) were analyzed with or without a combination of IL-4 for generation of DC/DC_leu. It was shown that the culture time could be reduced to 2 days, maturation of DC/DC_leu could be induced, and the T-cell stimulatory capacity could be preserved, although the viability of the generated DC/DC_leu was reduced compared to other established cytokine-based DC/DC_leu-generating protocols [49, 51].

DC/DC_leu Generation in Leukemic Subtypes

The generated DC/DC_leu are perfect APC due to the fact that they coexpress a variety of costimulatory molecules, MHC I and II complexes, and the whole leukemic antigen repertoire [75]. In general, ex vivo DC/DC_leu generation is possible from AML and MDS samples, independently of the patient's subtype, age, sex, cytogenetic risk group, and blast counts [39, 40, 48, 71, 76]. The expression of CD14, CD34, CD120, as well as CD86 on leukemic blasts seems to be a preliminary predictor of a successful cytokine-based DC/DC_leu generation [76, 77, 78]. Therefore, different protocols have been analyzed to increase the expression of CD14 on leukemic blasts before DC/DC_leu culture to influence the efficiency of sufficient DC/DC_leu generation [79]. Internal tandem duplication in the gene of the Fms-like tyrosin kinase 3 (FLT3) receptor, occurring in around 24 % of patients with AML, seems to impair differentiation of leukemic blasts towards DC_leu [80].

Ex vivo Immune Stimulating Effect of Generated DC/DC_leu

Stimulation of T-cell-enriched immune reactive cells with DC/DC_leu, generated from leukemic PBMNC, in a mixed-lymphocyte culture (MLC) increased T-cell activation and shifted T-cell subsets to a higher activation status [81, 82]. Proliferating T cells (T_prol CD71⁺, CD69⁺), non-naive T cells (T_non-naive, CD45RO⁺), β-integrin⁺ T cells, and T cells with an effector function such as central-memory T cells (T_cm, CD45RO⁺CCR7⁺) and effector (memory) T cells (T_eff-em, CD45RO⁺CCR7−) increased while naive T cells (T_naive CD45RO⁻) decreased during MLC [81, 82]. Furthermore, decreased amounts of regulatory T cells (T_reg, CD25⁺⁺CD127^low) could be found after MLC with DC/DC_leu compared to the control groups, although the general expression of CD25⁺⁺CD127^low molecules increased in both settings [82]. This might be explained as mentioned above by the potential induction of indoleamin 2,3-dioxgenase-1 (IDO1) [83, 84].

DC/DC_leu stimulation was shown to induce regularly antileukemic activity against leukemic blasts after MLC, though not in every given case, pointing to a specific induction of antileukemic immunity [40, 85, 86].

These findings confirm that generated DC/DC_leu can help to overcome anergy of immune reactive cells in AML. Due to the DC/DC_leu concept, it can be postulated that DC/DC_leu prime antileukemic T cells against several antigens [40]. This could be confirmed by the finding that increased frequencies of WT1 and PRAME-specific T cells could be detected after stimulation of T-cell-enriched immune reactive cells with generated DC/DC_leu from leukemic PBMNC compared to the control [Klauer et al., pers. commun.].

Vaccination Strategies with ex vivo Generated DC/DC_leu

As shown above, DC/DC_leu generation from isolated leukemic PBMNC is regularly possible ex vivo. To go one step further, different working groups tried to use these generated leukemic APC as a potential treatment tool for patients with AML [52]. DC/DC_leu were produced ex vivo and transferred to the patients as a subcutaneous vaccine. Irradiation of vaccinated cells is recommended to avoid transfer of not-to-DC_leu-converted leukemic blasts. This strategy would render the complicated loading process of generated mo-DC with apoptotic leukemic cells or whole leukemic cells lysates or by RNA electroporation unnecessary. Furthermore, DC/DC_leu have the advantage that they express the whole leukemic antigen repertoire in comparison to mo-DC (except mo-DC pulsed with apoptotic leukemic cells/lysates) and, therefore, the in vivo inducted antileukemic immunity is not just restricted to a single presented antigen [53]. The results of 3 preliminary phase 1/2 trials with autologous DC/DC_leu are summarized in Table 2. Currently, there are no ongoing clinical trials with DC/DC_leu for patients with AML [87]. In general, vaccinations with DC/DC_leu are well tolerated (as also already shown in other trials). Only in 1 patient did extensive eczema with an increased antinuclear factor occur, possibly pointing to an induction of autoimmunity [88, 89, 90]. In the trial of Roddie et al. [90], 5 patients (after achieving complete remission with intensive chemotherapy) were vaccinated weekly with an escalating dose of generated DC/DC_leu. It could be shown that IFN-γ-secreting antileukemic CTL increased and WT1-specific CTL could be detected. An (ongoing) clinical benefit of the DC/DC_leu vaccinating strategy could not be ascertained due to the lack of a control group [90]. Li et al. showed comparable results. After biweekly vaccination of patients with a constant dose of DC/DC_leu, 3 out of 5 patients achieved complete remission, whereas the other patients died due to rapid progression of the AML. As an immunological response PRAME-specific CD8⁺ T cells, a TH-1 cytokine release, and a higher intracellular IFN-γ concentration in CD4⁺ cells could be detected after the vaccination [89]. Dong et al. [88] combined vaccination of autologous DC with the administration of generated autologous cytokine-induced killer cells (CD3⁺CD56⁺ cells). In a direct comparison to the control group, treated with low-dose chemotherapy alone, significantly higher complete and partial remission rates could be achieved after DC/DC_leu vaccination (although the effects mediated by CIK-cells or DC/DC_leu could not be differentiated) [88]. In summary, the immunological phenomena found in the different trials point to a specific induction of antileukemic immunity after vaccination with DC/DC_leu.

Table 2.

Clinical trials with autologous DC/DC_eu generated from leukemic PBMNC

Patients included in the study, n	Stage of disease	Source of DC	Cytokines used for DC generation	Vaccination protocol	Immunological effects	Clinical effects	Ref.
5	AD	Autologous DC/DCleu	GM-CSF	s.c.	↑ PRAME (LAA)-specific CD8+ T cells	CR 3/5	89
			IL-4	Four vaccinations	↑ TH-1 cytokine release	PD 2/5
			TNF-α	biweekly	↑ IFN-γ CD4+ cells	No major side effects
				5 × 106 DC/DCleu
5	After achieving CR	Autologous DC/DCleu	GM-CSF	s.c.	↑ IFN-γ secreting antileukemic CTL	No evidence of a clinical benefit	90
			IL-4	4 vaccinations	↑ WT1-specfic CTL	(no control group)
			TNF-α	weekly	no change in amounts of Treg	Induction of autoimmunity (1/5)
			INF-γ	escalating doses:
			poly I:C	0.125 × 106 to 1 × 106 DC/DCleu
21	AD	Autologous DC/DCleu	GM-CSF	i.v.	↑ CD3+, CD4+CD3+, CD8+CD3+ compared	CRa 6/21	88
		+	INF-α	5 vaccinations	to before culture	PRa 9/21
		autologous CIK cells	FLT3 L	7.36±0.48 ×107 DC/DCleu	f IL 12, IL 2, IL 7, INF-γ, and TNF-α	NR 6/21
		+	SCF		compared to before culture	Mild side effects
		low dose CTX	TGF-β

AD, advanced disease; CR, complete remission; CIK cells, cytokine-induced killer cells (CD3+CD56+ cells); CTX, chemotherapy; s.c., subcutaneous; i.v., intravenous; CR, complete remission; PD, persisting disease; PR, partial remission; ↑, increase; CTL, cytotoxic T lymphocyte; T_reg, regulatory T cells.

^aSignificantly higher compared to the control group

Furthermore, an allogeneic DC/DC_leu vaccine was developed from an AML cell line with expression of a wide range of different LAA (DCP-001). These generated DC/DC_leu were analyzed in a phase 1 clinic trail as a post-HSCT therapy in 12 elderly AML patients and could show induction of cellular as well as humoral immunity in vivo with low side effects [91]. However, due to a lack of a control group the role of allogeneic effects alone could not be evaluated.

Different clinical trials with mo-DC vaccines (from autologous or allogeneic CD14⁺ monocytes) loaded with a variety of different antigens and methods for the treatment of AML have already been conducted and previously reviewed [16, 25]. Due to the “multi-antigen” concept mo-DC pulsed with apoptotic leukemic cells/lysates should be comparable with DC/DC_leu generated from leukemic PBMNC. In preliminarily clinical trials with mo-DC pulsed with apoptotic leukemic cells/lysates the stimulatory capacity of T cells was increased and CD8⁺ T cell responses to WT1 and hTERT could be detected [92, 93].

Molecules, Cells, and Particles Influencing and Regulating DC/DC_leu Functions

As demonstrated above, DC/DC_leu generation is possible ex vivo. In principle, the same factors are necessary in vivo to induce successful (DC-based) activation of the immune system against different tumor cells and to avoid autoimmune reactivity. Regulatory mechanisms, i.e., regulatory T cells, regulatory cytokines (e.g., IL-10) as well as exosomes (various nanoparticles secreted by several cells, including DC and lymphocytes or even from tumor cells), which mediate cell-cell communication [81, 82, 94, 95, 96], are necessary and well known. Exosomes could qualify to mobilize the immune system against tumors, they could mediate tolerance, or they could be used for DC pulsing as demonstrated in preliminary experiments [97, 98]. Indoleamin 2,3-dioxgenase-1 (IDO1) is an immunoregulatory enzyme that is responsible for tumor-related immunosuppression, e.g., mediated by regulatory T cells [99]. It is currently being discussed whether certain response modifiers (e.g., PGE₂) might be responsible for an induction of IDO1 expression leading to activation of immunosuppressive regulatory T cells or, vice versa, whether inhibition of IDO1 might increase antitumor reactions ex vivo or in vivo [99]. Monitoring of these molecules and cells under the influence of response modifiers and/or in the course of the leukemic disease (under treatment) is recommended.

From Bench to Bedside: from ex vivo to in vivo DC/DC_leu Generation

Ex vivo production of mo-DC and generation of DC/DC_leu from leukemic PBMNC to be used for vaccinations is a challenging process. The manufacturing process is time consuming, expensive, and has to be performed under GMP conditions. In summary, the whole vaccination process is logistically complicated. Cell products have to undergo quality (e.g., for infectious contamination) as well as quantity testing to control purity and amounts of generated DC/DC_leu before re-administration to the patients.

Therefore, it would be highly preferable to activate an antileukemic immune response in vivo, thereby circumventing an adoptive cell transfer. DC/DC_leu cultures with leukemic PBMNC are artificial cell culture models and do not represent the physiological situation in vivo. Therefore, our group established in a first step an ex vivo culture system thereby simulating physiological conditions, i.e., from individual AML or MDS patients' whole blood (WB), presenting all soluble and cellular components (including known and unknown activating or inhibitory factors) were cultured using a combination of at least 2 cytokines or response modifiers (“Kits”) to induce DC/DC_leu generation. After an intensive comparative analysis of 12 different Kits, the best 3 Kits could be selected for further examination and are shown in Table 3 [Kugler et al., pers. commun.] (European patent No. EP 3 217 975 B1; MODIBLAST GmbH; (inventor Helga Schmetzer)). Kits represent a “minimalized” DC/DC_leu-generating protocol (utilizing in addition patients' individual hematopoietic cellular and soluble blood background influencing immune reactions) and contain GM-CSF, responsible for the induction of differentiation of myeloid leukemia blasts and one of the response modifiers PGE₁, PGE₂, and Picibanil (OK-432) for a danger signaling effect and induction of maturation. It could be shown that generation of DC/DC_leu is possible directly from leukemic WB with Kits in comparable amounts as with already established DC/DC_leu-generating protocols [Kugler et al., pers. commun.] [48]. In a direct comparison of all 3 Kits, comparable amounts of DC/DC_leu could be generated, although significantly more mature DC could be generated with PGE₁-containing Kit M compared to PGE₂-containing Kit K. Stimulation of leukemic WB with Kits did not induce proliferation of not-to-DC/DC_leu-converted blasts [48, 100]. To imitate the in vivo situation even better, DC/DC_leu culturesfrom leukemic WB were conducted in parallel under hypoxic conditions (10% oxygen) and normoxic conditions (21% oxygen). In both settings, the same amounts of DC/DC_leu as well as the same functional features could be achieved [101].

Table 3.

Kits for the generation of DC/DC_leu from leukemic WB

Kit	Composition	Concentration	Culture Time, days	Reference
Ia	GM-CSF Picibanil	800 U/mL 10 µg/mL	7–10	Kugler et al., pers. commun.
Ka	GM-CSF PGE2	800 U/mL 1 µg/mL	7–10	Kugler et al., pers. commun.
Ma	GM-CSF PGE1	800 U/mL 1 µg/mL	7–10	Kugler et al., pers. commun.

^aEuropean patent (No. EP 3 217 975 B1; developed by Helga Maria Schmetzer, MODIBLAST GmbH).

Kit-pretreated WB (containing DC/DC_leu) improved the activation of autologous T-cell-enriched immune reactive cells in mixed MLC compared to the control; proliferating, nonnaive T cells, as well as CD8⁺ T cells, increased significantly during MLC [48][Ugur et al., pers. commun.]. Kit-generated DC/DC_leu from leukemic WB might also influence cells on the interface of the innate and adoptive immune system and cells of the innate immune system. We could, moreover, show that NK cells (CD56⁺CD3⁻), CIK cells (CD56⁺CD3⁺), and invariant NK cells (6B11⁺) increased after stimulation with DC/DC_leu [102] [Klauer et al., pers. commun.].

Due to the DC/DC_leu concept induction of a specific immunological memory can be postulated. In preliminarily experiments significantly higher amounts of effector or central memory T cells could be found after stimulation of T-cell-enriched immune reactive cells in MLC with Kit-generated DC/DC_leu compared to the control [48, 103].

One of the most important findings evaluated with functional assays is that the antileukemic activity could be significantly improved compared to the control. In a direct comparison of Kits, DC/DC_leu generated with Kit M increased the antileukemic cytotoxicity of T cell-enriched immune reactive cells the most [Ugur et al., pers. commun.]. Studying the provision of leukemia-specific cells significantly increased the frequencies of IFN-y secreting T, NK, and CIK cells that could be detected with the cytokine secretion assay (CSA) after MLC [Klauer et al., pers. commun.].

Moreover, cytokine release profiles were shifted to a higher release of inflammatory cytokines and antitumor response-related cytokines after WB-DC/DC_leu culture with Kits compared to patients' serum [48].

What Next? Kits as a Treatment Tool for Patients with AML?

Transforming the idea of ex vivo generation of DC/DC_leu from leukemic WB from bench to bedside, our hypothesis is that the administration of Kits to patients with AML or MDS can convert leukemic blasts to DC/DC_leu in vivo, thereby activating an antileukemic immunoreaction. This would render an adoptive transfer of ex vivo generated and manipulated mo-DC or DC/DC_leu unnecessary. The concept of in vivo Kit therapy in comparison to an ex vivo adoptive cell transfer of DC/DC_leu is illustrated in Figure 1. All Kit substances are approved for human treatment and they are already being used in the clinical routine; GM-CSF is used for the treatment of neutropenia in patients after chemotherapy or HSCT [104]. PGE₁ is already being used for the reduction of gastrointestinal ulcers during treatment with nonsteroidal anti-inflammatory drugs, to treat veno-occlusive disease after HSCT, erectile dysfunction, or to maintain the patency of the ductus arteriosus in newborns with ductal-dependent cardiac lesions [105, 106, 107]. PGE₂ is used for induction of labor in cases with a medical or obstetrical indication. Sclerotherapy of congenital lymphatic malformation in the head and neck is possible with Picibanil (OK 432) [108].

Fig. 1.

DC/DC_leu-mediated antileukemic treatment strategies. Two potential DC/DC_leu-generating strategies for the induction of antileukemic reactivity and immunity in vivo are shown. Left: vaccination with ex vivo generated DC/DC_leu. The patient's blasts (enriched by leukapheresis) are cultured ex vivo in the presence of different combinations of “response modifiers” (e.g., GM-CSF, IL-4, and TNF-α) under GMP conditions. After conversion to DC_leu and different quality as well as quantity analyses, generated cells are readministered to the patient as an intradermal vaccination. Irradiation of vaccinated cells is recommended to avoid transfer of not-to-DC_leu-converted leukemic blasts. Right: Patients are treated directly with Kits (e.g., Kit M, containing GM-CSF, PGE₁) that convert leukemic blasts to DC/DC_leu in vivo. Also cells of nonleukemic origin such as monocytes could be converted to DC (these cells could also phagocyte, process, and present leukemic particles), thereby supporting the DC_leu-based strategy. The Kit strategy would render the logistically complicated adoptive transfer of ex vivo generated DC/DC_leu unnecessary. Both strategies (left and right) give rise to DC/DC_leu that can migrate to tissues and induce antileukemic reactivity and immunity in vivo.

Our findings contribute to the development of a new potential therapy option for patients with AML or high-grade MDS, especially with the aim of stabilizing remissions or at least the disease. We could show that DC/DC_leu generation is possible with Kits from leukemic WB, without the induction of blast proliferation, independently of the patient's age, gender, MHC or mutation status, or cytogenetic risk profile or any subtype of AML. The administration of Kits to patients with AML or MDS might generate DC/DC_leu in vivo, which might activate the innate and adaptive immune system and especially leukemia-specific T cells followed by an immunoreaction against residual leukemic blasts. To prove our hypothesis, ongoing research with trials in animals and humans has to be performed.

Conflict of Interest Statement

All of the authors declare that there are no financial conflicts with regards to this work. “Use of immunomodulatory effective compositions for the immunotherapeutic treatment of patients suffering from myeloid leukemias” has a European patent (No. EP 3 217 975 B1) and was developed by Helga Maria Schmetzer, MODIBLAST GmbH.

Funding Sources

No funding was needed for this paper.

Author Contributions

D.C.A. and H.S.M. contributed to the writing, review and discussion of the manuscript.