PGE1-Containing Protocols Generate Mature (Leukemia-Derived) Dendritic Cells Directly from Leukemic Whole Blood

Dendritic cells (DCs) and leukemia-derived DC (DCleu) are potent stimulators of various immunoreactive cells and they play a pivotal role in the (re-) activation of the immune system. As a potential treatment tool for patients with acute myeloid leukemia, we developed and analyzed two new PGE1-containing protocols (Pici-PGE1, Kit M) to generate DC/DCleu ex vivo from leukemic peripheral blood mononuclear cells (PBMCs) or directly from leukemic whole blood (WB) to simulate physiological conditions. Pici-PGE1 generated significantly higher amounts of DCs from leukemic and healthy PBMCs when compared to control and comparable amounts as the already established protocol Pici-PGE2. The proportions of sufficient DC-generation were even higher after DC/DCleu-generation with Pici-PGE1. With Kits, it was possible to generate DCs and DCleu directly from leukemic and healthy WB without induction of blast proliferation. The average amounts of generated DCs and DCleu-subgroups were comparable with all Kits. The PGE1 containing Kit M generated significantly higher amounts of mature DCs when compared to the PGE2-containing Kit K and increased the anti-leukemic-activity. In summary PGE1-containing protocols were suitable for generating DC/DCleu from PBMCs as well as from WB, which reliably (re-) activated immunoreactive cells, improved the overall ex vivo anti-leukemic activity, and influenced cytokine-release-profiles.


Introduction
Acute myeloid leukemia (AML) is a clonal disease that is characterized by an uncontrolled proliferation and an impaired differentiation of myeloid progenitor cells (blasts) with overall five-year-survival rates of about 28.3 % [1,2]. In the last few decades, the therapy with antigen presenting cells (APCs), such as dendritic cells (DCs) revolutionized immunotherapy of AML [3][4][5].
APCs play a pivotal role in connecting the innate and the adaptive immune system with the properties to migrate into different tissues and activate different immune reactive cells. They internalize and process antigens, present antigen-fragments via major histocompatibility complex (MHC), and form immunological synapses with T cells, resulting in a clonally restricted and potent T cell-activation [6][7][8][9][10].
Two different DC-based immunotherapy strategies have been developed. Monocyte (CD14 + ) derived DCs can be generated with different response modifiers in cultures and they are loaded with leukemic-associated-antigens (LAA) by the electroporation of messenger ribonucleic acid (mRNA) or by peptide pulsing [11][12][13]. After expensive ex vivo manipulation and the production of cells under Good Manufacturing Practice (GMP), DCs can be (re-) administrated to patients as a vaccine [14,15].
Moreover, leukemic blasts can be converted ex vivo directly to leukemia derived DC (DC leu ), presenting the whole leukemic antigen repertoire. DC leu simultaneously express DC-antigens and -individual patients' blast markers (blast-antigens) [16].
The stimulation of T cell enriched immunoreactive cells with DC/DC leu ex vivo in a mixed lymphocyte culture (MLC) regularly results in the (re-) activation of T cells against leukemic blasts, although depending on the DC/DC leu protocol used for the generation of DC/DC leu [17,25,26]. We could already show that proliferating T cells (T prol CD71 + , CD69 + ), non-naïve T cells (T non-naïve , CD45RO + ), regulatory T cells (T reg , CD25 ++ CD127 low ), β-integrin + T cells and T cells with effector function, such as central-memory T cells (T cm , CD45RO + CCR7 + ), effector (memory) T cells (T eff-em , CD45RO + CCR7 − ) increase, while naïve T cells (T naive , CD45RO − ) decrease during MLC [27,28]. DC/DC leu probably contribute to stimulating and activating cells from the innate immune system and cells on the interface of the innate and the adaptive immune system, such as natural killer cells, invariant natural killer cells, or cytokine induced killer cells [29].
Prostaglandins are responsible and involved in different physiological functions, such as inflammation, regulation of renal-blood circulation, induction of fever, and the protection of the gastric mucosa from gastric acid [36][37][38][39]. PGE 2 and Prostaglandin E 1 (PGE 1 ) are arachidonic acid derivatives, which are synthesized via the cyclooxygenase 1 and 2 pathway (COX1 and COX2) [40]. The biochemical differences between PGE 1 and PGE 2 are caused by different amounts of double bounds in the side chain [41]. PGE 2 (e.g., Dinoproston) is approved by the US Food and Drug Administration (FDA) for the induction of labor in cases with medical or obstetrical indication. Drugs that contain PGE 1 are approved for the risk reduction of gastrointestinal ulcers during the treatment with nonsteroidal anti-inflammatory drugs (NSAIR) (e.g., Misoprostol) and to treat erectile dysfunction (e.g., Alprostadil). Furthermore, PGE 1 is used to treat peripheral arterial disease and to maintain the patency of the ductus arteriosus in patients with ductal-dependent cardiac lesions [42][43][44]. PGE 1 was also analyzed in a combination with heparin to prevent liver veno-occlusive disease (VOD), which is a life-threatening obliteration of hepatic venules, in patients with AML after bone marrow transplantation (BMT) [45,46].
To improve the DC/DC leu -treatment of patients with AML, we have developed minimalized Kits, containing combinations of at least two response modifiers. Following our hypotheses Kits should be able to convert leukemic blasts directly to DC/DC leu in WB cultures. With respect to clinical applications this could mean, that patients could be directly treated with Kits, thereby inducing DC/DC leu -generation in vivo, which would render an adoptive transfer of ex vivo generated DC/DC leu unnecessary.
The aim of this study was (1) to develop and to functionally evaluate a new PGE 1 -containing DC/DC leu generating protocol to produce DCs and DC leu from healthy and leukemic PBMCs; (2) to develop and to functionally evaluate an immunomodulatory Kit M (containing GM-CSF and PGE 1 ) to produce DCs and DC leu directly from healthy and leukemic WB, thereby simulating in vivo conditions; (3) to deduce an optimized protocol for the ex vivo generation of DC/DC leu which might be used for an adoptive cell transfer; and, (4) to deduce immunomodulatory Kits that might be able to convert myeloid leukemic blasts in vivo to DC/DC leu .

Prolog
In the first part of this manuscript, we present a new PGE 1 -containing protocol (Pici-PGE1 ) for the generation of DC/DC leu from healthy and leukemic PBMCs.
In the second part, we simulated physiological conditions and generated DC/DC leu with the DC/DC leu -generating protocols Pici-PGE1 and Pici-PGE2 and immunomodulatory Kits directly from healthy and leukemic WB. The compositions of Picis and Kits are shown in Table 1 We correlated data with the DC/DC leu stimulatory potential of T cell enriched immunoreactive cells and with the potential to generate anti-leukemia directed T cells as well as with cytokine-release-profiles. With Pici-PGE1 and Pici-PGE2 we generated on average significantly*** higher amounts of DCs from healthy PBMCs as compared to controls (n = 9) (Pici-PGE1 : 17.4 ± 4.7% DC + /PBMC, p < 0.00003; Pici-PGE2 : 15.6 ± 5.1% DC + /PBMC, p < 0.0003; control: 6.0 ± 2.2% DC + /PBMC). Although differences were not significant, we found, on average, higher amounts of DC + /PBMC after the stimulation of healthy PBMCs with Pici-PGE1 when compared to Pici-PGE2 . No significant differences were found in amounts of DC mig /DC + with Pici-PGE1 and Pici-PGE2 (26.8 vs. 25.1% DC mig /DC + , p < 0.77) ( Figure 1A).
In summary, we conclude that DCs and DCmig can be generated with Pici-PGE1 and Pici-PGE2 in comparable amounts from healthy and leukemic PBMCs.  shows the average amounts ± standard deviation of generated dendritic cells (DCs) in the PBMC-fraction and mature DCs in the DC-fraction [CD197 + DC + , (DC mig /DC + )] from healthy PBMCs with Pici-PGE1 , Pici-PGE2 and control without added cytokines. (B) presents the average amounts ± standard deviation of generated DCs in the PBMC-fraction, DC leu -subgroups [including DC leu in the DC-fraction (DC leu /DC + ), DC leu in the blast-fraction (to quantify amounts of leukemic blasts converted to DC leu ) (DC leu /Bla + ), DC leu in the PBMC-fraction (DC leu /PBMC)] and DC mig in the DC-fraction (DC mig /DC + ) from leukemic PBMCs with Pici-PGE1 , Pici-PGE2 and control without added cytokines. (C) and (D) show the percentages of sufficient DC-generation from healthy (C) and leukemic (D) PBMCs with Pici-PGE1 , Pici-PGE2 , Pici-PGE1 or Pici-PGE2 and control without added response modifiers according to cut-off-values (≥10% DC + /PBMC). Each dot ( ) represents DC-proportions generated from each individual healthy volunteer or AML-patient. DCs dendritic cells; DC leu leukemic derived dendritic cells; PBMCs peripheral blood mononuclear cells. The differences were considered as significant*** with p values <0.005.
In summary, we conclude that DCs and DC mig can be generated with Pici-PGE1 and Pici-PGE2 in comparable amounts from healthy and leukemic PBMCs.
2.2.2. Efficiency of Sufficient DC-Generation is Higher with Pici-PGE1 Compared to Pici-PGE2 from Leukemic PBMCs In healthy and leukemic control groups, we found, in every given case, less than 10% DC + /PBMC. Therefore, we defined a cut-off value of ≥10% DC + /PBMC as a successful DC-generation from healthy and leukemic PBMCs. According to this cut-off value a successful DC-generation from healthy PBMCs was possible in 100% of cases (nine of nine cases) with Pici-PGE1 and in 89% of cases (eight of nine cases) with Pici-PGE2 ( Figure 1C).
In summary, the efficiencies of a sufficient DC-generation from leukemic PBMCs are higher with Pici-PGE1 as compared to Pici-PGE2 and comparable to healthy PBMCs. In four cases, no sufficient DC-generation was possible with both protocols. After DC/DC leu -culture from leukemic PBMCs, we found on average comparable amounts of proliferating blasts that were not converted to DC leu (Bla prol-CD71 ) with Pici-PGE1 or Pici-PGE2 as compared to control: Pici-PGE1 : 26.8 ± 19.9%, p < 0.29; Pici-PGE2 : 25.6 ± 16.8%, p < 0.39; control: 21.3 ± 14.7%. Comparable distributions were found for Bla prol-Ipo-38 (data not shown).
We conclude that neither Pici-PGE1 nor Pici-PGE2 induce proliferation of blasts not converted to DC leu .

Comparable DC-Amounts can be Generated with Immunomodulatory Kits and Picis
We compared the DC + /WB values generated with Kit M, Kit K, and Kit I (Kits n = 27) and DC + /WB values that were generated with Pici-PGE1 and Pici-PGE2 from leukemic WB (Picis n = 18).
We summarize that DC/DC leu -generation (including subgroups) is possible in comparable amounts with Kits from leukemic and healthy WB when compared to Picis. We summarize that DC/DCleu-generation (including subgroups) is possible in comparable amounts with Kits from leukemic and healthy WB when compared to Picis.

Significantly Higher Amounts of DC + /WB Generated from Healthy and Leukemic WB with Immunomodulatory Kit M, Kit K and Kit I Compared to Control
We generated DCs from leukemic WB and found, on average, significantly*** higher amounts of DC + /WB after WB-DC/DCleu-cultures with Kit M, Kit K, and Kit I as compared to control (n = 25) (Kit M: 9.5 ± 3.6% DC + /WB, p < 0.0004; Kit K: 9.5 ± 3.6% DC + /WB, p < 0.0004; Kit I: 10.1 ± 4.4% DC + /WB, p < 0.0003; control: 6.3 ± 1.9% DC + /WB) ( Figure 2B). Comparable distributions were found after DCcultures from healthy WB (n = 9, data not shown). In the comparison of Kit M, Kit K, and Kit I we found, on average, comparable amounts of generated DCs in the WB-fraction from healthy and leukemic WB. In DCleu-subgroups, including DCleu/Bla + , DCleu/DC + , as well as DCleu/WB no significant differences were found ( Figure 2C).
In summary, significantly higher amounts of DC + /WB can be generated with immunomodulatory Kits from healthy and leukemic WB as compared to control. The amounts of shows average amounts ± standard deviation of DC-and DC leu -proportions [including DC leu -subgroups: DC leu in the DC-fraction (DC leu /DC + ), DC leu in the blast-fraction (to quantify amounts of leukemic blasts converted to DC leu ) (DC leu /Bla + ) and DC leu in the WB-fraction (DC leu /WB)] from leukemic WB with Kits (including Kit M, Kit K, Kit I) compared to protocols Picis (including Pici-PGE1 , Pici-PGE2 ). (B) shows average amounts ± standard deviation of generated DCs with Kit M, Kit K and Kit I compared to control. (C) presents average amounts ± standard deviation of generated DC leu subgroups [including DC leu -subgroups DC leu in the DC-fraction (DC leu /DC + ), DC leu in the blast-fraction (to quantify amounts of leukemic blasts converted to DC leu ) (DC leu /Bla + ) and DC leu in the WB-fraction (DC leu /WB)] with Kit M, Kit K and Kit I. (D) shows average amounts ± standard deviation of DC mig /DC + generated with Kit M, Kit K and Kit I. DC mig are characterized by the expression of CCR7. DCs dendritic cells; DC leu leukemic derived dendritic cells; WB whole blood. The differences were considered as significant*, with p values between 0.1 and 0.05 and as significant*** with p values <0.005.

Significantly Higher Amounts of DC + /WB Generated from Healthy and Leukemic WB with Immunomodulatory Kit M, Kit K and Kit I Compared to Control
We generated DCs from leukemic WB and found, on average, significantly*** higher amounts of DC + /WB after WB-DC/DC leu -cultures with Kit M, Kit K, and Kit I as compared to control (n = 25) (Kit M: 9.5 ± 3.6% DC + /WB, p < 0.0004; Kit K: 9.5 ± 3.6% DC + /WB, p < 0.0004; Kit I: 10.1 ± 4.4% DC + /WB, p < 0.0003; control: 6.3 ± 1.9% DC + /WB) ( Figure 2B). Comparable distributions were found after DC-cultures from healthy WB (n = 9, data not shown). In the comparison of Kit M, Kit K, and Kit I we found, on average, comparable amounts of generated DCs in the WB-fraction from healthy and leukemic WB. In DC leu -subgroups, including DC leu /Bla + , DC leu /DC + , as well as DC leu /WB no significant differences were found ( Figure 2C).
In summary, significantly higher amounts of DC + /WB can be generated with immunomodulatory Kits from healthy and leukemic WB as compared to control. The amounts of generated DC and DC leu -subgroups are comparable with all three Kits, with the exception of significantly higher frequencies of DC mig /DC + after DC/DC leu -culture with Kit M as compared to Kit K.
In summary, Kits do not induce blast proliferation during DC/DC leu -culture from leukemic WB when compared to control. In general, we found a significantly higher activation status of immunoreactive T cells after MLC WB-DC Kits as compared to MLC WB and uncultured T cells. The main findings were characterized by increased amounts of CD3 + CD8 + /CD3 + and the corresponding decrease of CD3 + CD4 + /CD3 + after MLC WB-DC Kits when compared to MLC WB . We found after MLC WB-DC Kits as well as after MLC WB increased amounts of proliferating T cells when compared to uncultured cells, and also a shift from naive to non-naïve T cell subsets. Detailed results are shown in Table 2A and in Figure 3A. Since IL-2 was added to all MLC-experiments (including MLC WB ), amounts of T cell subsets also increased after MLC WB .
Furthermore, we pooled all the results obtained with the cytotoxicity assay after MLC WB-DC Kit (including MLC WB-DC Kit M , MLC WB-DC Kit K , MLC WB-DC Kit I ) and compared the results to MLC WB after 3 or 24 h (n = 24) of incubation of effector with target cells. Anti-leukemic activity was defined as lysis of blast-target-cells obtained either after 3 or 24 h. Moreover, we analyzed the anti-leukemic activity in each individual patient (n = 13).
After MLC WB-DC lysis of blasts target cells could be improved in 75.0% of cases (18 of 24 cases) as compared to MLC WB after 3 or 24 h. Furthermore, in 92.3% of patients (12 of 13 patients) we could select at least one MLC WB-DC Kit in each individual patient, which improved the blasts lysis after 3 h or 24 h when compared to MLC WB ( Figure 3B).
In summary, we regularly demonstrate an increase of T cells' anti-leukemic activity after stimulation with DC/DC leu generated with Kits from leukemic WB.

Comparison of T Cell Amounts, Phenotypes and Blast Lytic Activity after MLC WB-DC Kit M and MLC WB-DC Kit K
DC/DC leu -generation was possible with Kits from leukemic WB and the stimulation of T cell enriched immunereactive cells with these DC/DC leu resulted in T cell activation. The premise for this analysis was that MLC WB-DC Kit M and MLC WB-DC Kit K (n = 6) were performed in parallel in patients with the corresponding leukemic WB-samples.
We conclude that DC/DC leu in a MLC WB-DC have an impact on the amounts of different T cell subsets after stimulation as compared to MLC WB and uncultured cells, however T cell compositions after MLC WB-DC Kits were comparable with Kits used to generate DC/DC leu. Furthermore, we conclude that blast-lysis was improved in all cases (except one) after MLC WB-DC Kit M when compared to MLC WB-DC Kit K after 24 h.

Cytokine-Release-Profiles in Serum and after WB DC/DC leu -Culture
Cytokine releases by blasts, DCs, as well as by immunoreactive cells are known to influence immunological as well as immune-escape-reactions. Therefore, we correlated cytokine releases in DC/DC leu -culture-supernatants as well as in serum.
2.5.1. Significantly Higher Concentrations of the Inflammatory Cytokine and Antitumor Response Related Cytokine Found after WB-DC/DC leu -Culture with Kits Compared to Serum We studied the cytokine release levels after WB-DC/DC leu -culture with Kits and in serum. Therefore, we pooled all of the results after WB-DC/DC leu -culture with Kits (Kit M, Kit K and Kit I) (n = 27) from leukemic WB and compared it to cytokine concentrations in serum (n = 8).

Significantly Higher Concentrations of MCP-1 (CCL-2) Found after DC/DC leu -Culture with Kit M and Kit K Compared to Control
We compared cytokine secretion after DC/DC leu -culture of leukemic WB with Kits when compared to control without added response modifiers. We found significantly** higher median concentrations of MCP-1 (CCL-2) after DC/DC leu -culture with Kit M and Kit K as compared to control (n = 9) [Kit M: 5.4 ng/mL (range: 1.1-11.0), p < 0.06; Kit K: 5.3 ng/mL (range: 1.0-10.9), p < 0.04; control: 2.3 ng/mL (range: 0.4-6.3)]. For the cytokines IL-17A and IL-10, no significant differences were found as compared to controls ( Figure 4B). Comparable results were found for the Kit I (data not shown).
We conclude that the addition of Kits to leukemic WB influences the cytokine release profiles when compared to control. Kit M and Kit K produce comparable cytokine-release-profiles.

DC and DC leu Based Immunotherapy for Patients with AML
DCs play a crucial role in the (re-) activation of the immune system and in linking the innate and adaptive immune system. These professional APCs have the ability to migrate into different tissues and to induce an immunological memory. During the last decades, different strategies have been developed to utilize DCs as a treatment tool for patients with AML. DCs can be generated ex vivo from CD14 + monocytes, loaded with different LAAs or tumor antigens, and they can be re-administrated to the patient as vaccine. It was already reported that vaccination with ex vivo generated monocyte-derived DCs increased the amounts of leukemia specific T cells in AML-patients and stabilized complete remission [15,48,49]. Unfortunately, the production of monocyte-derived DCs ex vivo is time-consuming, expensive, has to be performed under GMP-conditions, and the cell production is limited by the selection of LAA [11,14,15,50].
On the other hand, we and others could already show that clonal leukemic blasts can be regularly converted to DC leu with different DC/DC leu -generating protocols, independent from age, FAB-classification, mutation or hematopoietic stem cell transplantation (HSCT) status, mutations of the disease, and FAB classification [17,51]. The clonal leukemic origin of DC leu was already confirmed with fluorescence in situ hybridization (FISH) analysis [52].

The PGE 1 -Containing Protocol Pici-PGE1 is More Reliable to Generate DCs in Sufficient Amounts from Healthy and Leukemic PBMCs Compared to the PGE 2 -Containing Protocol Pici-PGE2
It was already shown that PGE 2 produced by the enzyme cyclooxygenase 2 (COX-2) has a crucial role in tumor genesis in colorectal cancers in vivo [53,54]. It was reported that PGE 2 influences the apoptosis, induces the angiogenesis, and activates the proliferation of cancer cells [55][56][57]. Whereas, in highly metastatic melanoma cells, it was shown that PGE 1 promotes anti-tumor effects, such as the inhibition of invasion and cell growth [58]. Moreover, PGE 1 increased the differentiation of tumor cells and decreased the expression levels of metalloproteinase (MMP) 2 and 9 [59]. Furthermore, it was shown that PGE 1 increased the concentrations of Cis-Platin in cancers cells after treating rats with peritoneal carcinomatosis [60]. Transferring these findings into the AML treatment, we hypothesize that PGE 1 could influence anti-cancer reactions in a positive way. Therefore, we analyzed the potential of PGE 1 -containing DC/DC leu -generating protocols to produce DCs and DC leu from healthy and leukemic PBMCs as well as from WB. In the past few years it was shown, that it is possible to generate mature DCs and DC leu from leukemic and healthy PBMCs with the PGE 2 -containing protocol Pici-PGE2 [17,20]. We could confirm this and we add, in addition, that the new protocol Pici-PGE1 produced comparable amounts of DCs and DC leu with a higher efficiency of sufficient DC generation from healthy and leukemic PBMCs when compared to Pici-PGE2 . These results suggest, that PGE 1 mediates the maturation of DCs and DC leu with comparable efficiency as PGE 2 , but might be also involved in the differentiation of myeloid progenitor cells and the production of DC/DC leu . This effect might be explained by the different mode of action of PGE 1 and PGE 2 on prostaglandin receptors (EP receptors) with different down streaming effects due to their different biochemical reactivities [61]. The generation of DC/DC leu was possible with both protocols, independent of patients' FAB-classification, stage (first diagnosis vs. relapse), or status (primary vs. secondary AML) of the disease. Moreover, we could confirm that neither Pici-PGE1 nor Pici-PGE2 induced the proliferation of non-converted blasts during DC/DC leu -cultures [19]. In four AML-cases, no sufficient DC/DC leu -generation was possible from leukemic PBMCs with Pici-PGE2 and Pici-PGE1 , which might be due to various expressions of cytokine-receptors (e.g., GM-CSF receptors, stem cell-receptors) or prostaglandin-receptors (EP-receptors) on leukemic blasts [62,63]. These patients did not receive chemotherapy before conducting DC/DC leu culture experiments, however spontaneous apoptosis or cell death of blood cells has to be discussed. Technical limitations were excluded.

Successful Generation of DCs and DC leu from Healthy and Leukemic WB Cultures
We established a WB-model to generate DC/DC leu from leukemic WB containing all soluble and cellular factors of the individual patient to simulate physiological conditions. According to our hypothesis, DCs and DC leu could also be generated in vivo by modulating myeloid blasts in their natural microenvironment after administrating a combination of different response modifiers (e.g., Kit M, Kit K, Kit I) to the patient. Therefore, well known DC/DC leu generating protocols (e.g., Pici-PGE1 , Pici-PGE2 ) were applied in WB settings and results that were obtained with Kit M, Kit K, or/and Kit I compared: our results show that DCs and DC leu (including subgroups) could be generated in comparable amounts with Kits and Picis directly from WB [17,20]. Additionally, DC leu subgroups did not significantly differ. We conclude that DCs and DC leu can be generated with Kits from leukemic WB ex vivo. The fact that significantly more DC/DC leu could be generated with Kit M, Kit K, and Kit I when compared to control without added response modifiers points to the fact that the combination of two response modifiers is necessary to generate DCs and DC leu in sufficient amounts directly from leukemic WB: the induction of hematopoietic differentiation is induced by GM-CSF, danger signaling, and/or maturation signaling of DC/DC leu is caused by PGE 1 , PGE 2 , or Picibanil. When compared to PBMC-cultures, response modifiers, such as IL-4 and other/unknown cytokines, are physiological components of the microenvironment in WB, and therefore they have not to be added or included in DC/DC leu -generating protocols [35]. The average amounts of generated DCs (including DC leu subgroups) were comparable with all three Kits, although amounts of matured DCs were significantly higher after WB-treatment with the PGE 1 -containing Kit M as compared to the PGE 2 -containing Kit K. This finding might also be explained by the different mode of action of PGE 1 and PGE 2 on EP receptors with different down streaming effects [61]. Especially, for the in vivo use of DC/DC leu , the expression of the lymph-node-homing-receptor CCR7 (marker for mature DCs) is crucial for the migratory capacity of DCs and DC leu to the lymph node, where they activate T cells and other immunoreactive cells and induce anti-tumor/anti-leukemic activity [64][65][66]. Comparable results were already found under hypoxic conditions [67]. The proliferation of blasts (not converted to DC leu ) was not induced with Kits during DC/DC leu -cultures. Therefore, we conclude that Kits might be safe tools for an in vivo treatment.
We could already show that, after the stimulation of T cell, enriched immunoreactive cells with DCs and DC leu generated from leukemic PBMCs composition of T cells could be shifted and influenced in a positive way and that anti-leukemic activity could be induced [17,25,27,28]. T cells directly kill cancer-cells via the Granzyme B and Perforin pathway and/or indirectly through the secretion of IFN-γ or tumor-necrosis-factor-alpha (TNF-α). We found after MLC WB DC-Kits a higher activation status with increased amounts of proliferating T cells, as well as a shift from naive to non-naïve T cell subsets when compared to cells before culture. When compared to MLC WB , amounts of CD3 + CD8 + significantly increased. After MLC WB higher amounts of different T cell subsets could also be found and they can be explained by IL-2 added to all MLC, which is a potent endogenous T cell and NK cell activating cytokine [68]. The induction of an immunological memory can be postulated due to the DC concept. We found comparable amounts of central memory T cells after MLC WB-Kits , MLC WB and in uncultured cells. However, we found significantly more effector memory T cells after MLC WB-Kits , pointing to an induction of these specialized memory cells, which can enter different tissues to initiate inflammation and cytotoxicity [69]. We speculate that, in our setting, the incubation time to produce memory T cells after DC/DC leu -stimulation is too short, therefore only effector memory T cells are produced. The induction of memory T cells is crucial for the induction of long term remission in patients with AML.
The composition of Kits to produce DC/DC leu , used for the stimulation of T cell enriched immunoreactive cells has no impact on the average amounts and phenotypes of T cell subsets after MLC, but on the anti-leukemic activity. The most important result of our study was that we could demonstrate that anti-leukemic activity could be improved after MLC WB-DC-Kits compared to MLC WB . Increased production of leukemia-specific cells after MLC WB-DC-Kits as compared to MLC WB was shown in the proof of concept-assays. Extended leukemia-specific evaluations are part of our ongoing research. These results suggest that DC/DC leu generated with Kits from leukemic WB induce the immune system and can activate specific anti-leukemic activity against leukemic blasts after MLC. Furthermore, we could show that the anti-leukemic activity is comparable and it might be superior after MLC DC/DC leu generated with the PGE 1 -containing Kit M when compared to the PGE 2 -containing Kit K, especially after 24 h of simultaneous incubation of the effector and target cells. This finding might be explained by the fact that PGE 2 could induce the expression of indoleamin 2,3-dioxgenase-1 (IDO1), an immunoregulatory enzyme that activates immunosuppressive T reg , but does not impair the antigen presentation capacity of DCs [70][71][72][73][74].
Moreover, during WB DC/DC leu -cultures with Kits as well as Picis cytokine-release-profiles were influenced compared to serum as well as compared to control. We can conclude that Kits induce the production of anti-tumor response related cytokines as well as inflammatory cytokines during DC/DC leu cultures, and therefore improve anti-leukemic reactivity of T cells after stimulation with Kit treated WB.

Sample Collection
After obtaining written informed consent, in accordance with the Helsinki protocol and the local Ethic Committee (Pettenkoferstraße 8a, 80336 Munich, Germany, Ludwigs-Maximilians-University-Hospital in Munich; VoteNo 33905), the heparinized peripheral WB samples were taken from patients in acute phases of AML and from healthy volunteers. The University-Hospitals of Oldenburg, Tuebingen, Munich and Augsburg provided samples. Anticoagulation was performed with Lithium-heparin-tubes (7.5 mL, Sarstedt, Nuernberg, Germany) containing standardized concentrations of Heparin.
PBMCs were isolated from WB-samples by density gradient centrifugation while using the Ficoll-Hypaque-Technique (Biocoll-Separating-solution, Biochrom, Berlin, Germany) with a density gradient of 1.077 g/mL. PBMCs were washed and then suspended in phosphate-buffered saline (PBS, Biochrom, Berlin, Germany). CD3 + T cells were enriched while using the MACS-technology (Milteney Biotech, Bergisch Gladbach, Germany). The purity of the viable T cells was, on average, 87.9% (range 77.4-96.1%). The viable cells were quantified while using Trypan Blue (Biochrom, Berlin, Germany) and they were counted with Neubauer-counting-chambers. PBMCs were directly used to set up DC/DC leu -cultures. T cells and the remaining PBMCs were used for subsequent experiments, therefore cells were frozen at −80 • C (using DMSO) and then thawed according to standardized protocols. Furthermore, the serum was frozen for subsequent ELISA-experiments.

Patients' Characteristics and Diagnostics
DC/DC leu were generated from PBMC-and WB-samples that were obtained from AML patients (n = 29) and healthy volunteers (n = 10). The average age of AML patients was 58.6 years (range 21-79) and of healthy volunteers 28.1 years (range 20-56). The female to male ratio of AML patients was 1:1.2 and of healthy 1:0.4. The ages of healthy volunteers and AML patients were not age-matched, since no direct comparisons of those two groups were needed.
The cellular composition of AML-and healthy-peripheral-WB-samples as well as of AMLand healthy-PBMCs-samples are shown in Table 4. In cases with the aberrant expression of T-, B-, or monocytoid-antigens or CD56 on leukemic blasts, proportions of the corresponding cells were not included in the analysis.

DC/DC leu -Generation from Isolated PBMCs
DC/DC leu were generated from 3-4 × 10 6 isolated healthy and leukemic PBMCs with the DC/DC leu -generating protocols Pici-PGE1 and Pici-PGE2 [17,20]. Therefore, cells were pipetted into 12-multiwell-tissue-culture-plates (ThermoFisher Scientific, Darmstadt, Germany) and they were diluted in 2 mL serum-free X-Vivo-15-medium (Lonza, Basel, Swiss). Cytokines were added, as described below. Half medium exchange was carried out after 3-4 cell culture days. A culture without added response modifiers served as a control.

DC/DC leu -Generation from WB
DC/DC leu were generated from healthy and leukemic WB (presenting the physiological cellular and soluble composition of the individual samples) with the DC/DC leu -generating protocols Pici-PGE2 , Pici-PGE1 , Kit M, Kit K and Kit I [47]. Therefore, 500 µL WB were pipetted in 12-multiwell-plates and then diluted 1:2 in X-Vivo-15-medium (Lonza, Basel, Swiss) to imitate the physiological conditions. Response modifiers and immune-modulating factors were added to cultures, as described below. A culture without added response modifiers served as a control. All of the response modifiers used for the DC/DC leu -generation are approved for human treatment. Table 1 provides the compositions of DC/DC leu -generating protocols.
All of the cell-culture-experiments were conducted at standard laboratory conditions comprising 37 • C, 21% O 2 and 5% CO 2 . At the day of harvest, cell culture supernatants were collected from DC/DC leu -cultures and MLC-cultures and frozen, according to the standardized protocols for subsequent ELISA-experiments.

Picibanil-PGE2 (Pici-PGE2 )
DC/DC leu were generated from PBMCs and WB with the Pici-PGE2 DC/DC leu -generating protocol, with the same composition, as given above, for Pici-PGE1 , however substituting PGE 1 by PGE 2 (PeproTech, Berlin, Germany) [17,20]. Surface marker combinations for the analysis of DC and DC leu (including subsets) and T cell subtypes after flow cytometric staining with fluorochrome-labelled-antibodies are given. Cells were analyzed before and after different cultures.
To quantify generated DC leu , the cells were stained with patient-specific blast-staining antibodies (e.g., CD15, CD34, CD65, and CD117), according to diagnostic reports in combination with DC-staining antibodies (e.g., CD80, CD83, CD86, CD206, and CD209), which were not expressed on blasts before culture. For the analysis and quantification of DCs and DC leu in the total-or in subtype-cell fractions after DC/DC leu -cultures, we used a refined gating strategy [16,17]. DC/DC leu subgroup analyses were only conducted in cases with ≥10% DCs in the PBMC-and WB-fractions (premise for analysis). DC leu were quantified in the total cell fraction (DC leu /PBMC or WB), 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 + ). The workflow and FACS analysis of DC/DC leu -generation with Pici-PGE1 as compared to control is exemplarily illustrated in Figure 5 The amount of mature DCs [DC co-expressing the migration marker CCR7 (CD197)] in the DC fraction after culture (DC mig /DC + ) was quantified in the cases with ≥10% DC + /cells. A refined gating strategy was used to detect non-converted proliferating blasts in the PBMC-or WB-fractions (Bla prol /PBMC or WB) after DC/DC leu -culture [19]. The proliferating blasts were characterized by the co-expression of CD71 or IPO-38 without the co-expression of DC-markers. (Table 5). PBMCs were cultured in 12-multiwell-tissue-culureplates and diluted in 2 mL serum-free X-vivo-15-medium. For the generation of DC/DCleu with Pici-PGE1 500 U/mL GM-CSF and 250 U/mL IL-4 were added on day 0. After 6-7 days, 10µg/mL Picibanil a lysis product from Streptococcus pyogenes, which has unspecific immune modulatory effects and 1 µg/mL PGE1 were added. No response modifiers were added to the control culture. Cells were harvested after 7-10 days of incubation. Half medium exchange was carried out after 3-4 days. x-axis: CD80; y-axis: CD34.

Mixed-Lymphocyte-Culture (MLC) of T Cell-Enriched Immunoreactive Cells with WB-Stimulator-Cell-Suspensions, Preincubated or Not Preincubated with DC/DCleu-Generating Protocols
1 × 10 6 CD3 + T cells (effector cells) from AML patients were co-cultured in 24-multiwell-tissueculture-plates (ThermoFisher Scientific, Darmstadt, Germany) with a stimulator cell suspension containing approximately 2.5 × 10 5 DC/DCleu (MLC WB-DC ), which were generated with different DC/DCleu-generating protocols from leukemic WB or PBMCs. A MLC of T cell enriched immunoreactive cells with a stimulator cell suspension without preincubation with different DC/DCleu-generating protocols (MLC WB ) severed as a control. The total volume of the cell culture was adjusted to 1 mL with RPMI-1640 medium (Biochrom, Berlin, Germany) containing 1% Penicillin (Biochrom, Berlin, Germany) and 50 U/mL Interleukin 2 (IL-2, PeproTech, Berlin, Germany). After 2-3 days, 50 U/mL IL-2 was added to all cultures. PBMC-cultures contained 15% human serum (Healthcare Europa GmbH, Vienna, Austria). The cells were harvested after 6-7 days and they were used for cytotoxicity-fluorolysis-assay, as described below. Before and after culture different T cell subsets in MLC were quantified by flow cytometry (Table 5).

Cytotoxicity (Fluorolysis) Assay
A Fluorolysis assay was performed to analyze the blast lytic activity of T cell-enriched . Workflow and FACS analysis of DC/DC leu generated with Pici-PGE1 compared to control from leukemic PBMCs in a case of AML. 3-4 × 10 6 PBMCs were cultured in 12-multiwell-tissue-culure-plates and diluted in 2 mL serum-free X-vivo-15-medium. For the generation of DC/DC leu with Pici-PGE1 500 U/mL GM-CSF and 250 U/mL IL-4 were added on day 0. After 6-7 days, 10µg/mL Picibanil a lysis product from Streptococcus pyogenes, which has unspecific immune modulatory effects and 1 µg/mL PGE 1 were added. No response modifiers were added to the control culture. Cells were harvested after 7-10 days of incubation. Half medium exchange was carried out after 3-4 days. x-axis: CD80; y-axis: CD34. 4.6. Mixed-Lymphocyte-Culture (MLC) of T Cell-Enriched Immunoreactive Cells with WB-Stimulator-Cell-Suspensions, Preincubated or Not Preincubated with DC/DC leu -Generating Protocols 1 × 10 6 CD3 + T cells (effector cells) from AML patients were co-cultured in 24-multiwell-tissueculture-plates (ThermoFisher Scientific, Darmstadt, Germany) with a stimulator cell suspension containing approximately 2.5 × 10 5 DC/DC leu (MLC WB-DC ), which were generated with different DC/DC leu -generating protocols from leukemic WB or PBMCs. A MLC of T cell enriched immunoreactive cells with a stimulator cell suspension without preincubation with different DC/DC leu -generating protocols (MLC WB ) severed as a control. The total volume of the cell culture was adjusted to 1 mL with RPMI-1640 medium (Biochrom, Berlin, Germany) containing 1% Penicillin (Biochrom, Berlin, Germany) and 50 U/mL Interleukin 2 (IL-2, PeproTech, Berlin, Germany). After 2-3 days, 50 U/mL IL-2 was added to all cultures. PBMC-cultures contained 15% human serum (Healthcare Europa GmbH, Vienna, Austria). The cells were harvested after 6-7 days and they were used for cytotoxicity-fluorolysis-assay, as described below. Before and after culture different T cell subsets in MLC were quantified by flow cytometry (Table 5).

Cytotoxicity (Fluorolysis) Assay
A Fluorolysis assay was performed to analyze the blast lytic activity of T cell-enriched immunoreactive cells after MLC WB-DC and MLC WB [25]. Therefore, effector cells were co-cultured 1:1 with thawed blast-containing target cells for 3 and 24 h in standard laboratory conditions, as described above. As a control effector-and target-cells were cultured separately for the same time and mingled on ice shortly before flow cytometric analyses were carried out. Before culture, the target-cells were stained for 15 min. with FITC-, PE-, or APC-conjugated blast specific target cell antibodies. To evaluate viable cells and the lytic activity of effector cells, the cultures were harvested after 3 h and 24 h and then resuspended in PBS containing 7AAD (Becton Dickson, Heidelberg, Germany) and a defined number of Fluorosphere beads (Beckman Coulter, Krefeld, Germany). A refined gating was used to analyse blast lytic activity (anti-leukemic activity) [25]. Therefore, viable target cells were gated in a Forward Scatter (FSC) 7AAD − gate. The cells were analyzed with a fluorescence-activated cell sorting Flow-Cytometer (FACSCalibur TM ) and Cell-Quest-data-acquisition and analysis software (Becton Dickson, Heidelberg, Germany). The lytic activity of effector cells was calculated and defined as the difference in the percentage of viable target cells in the culture with co-cultured effector and target cells (for 3 h and 24 h) as compared to control.

Statistical Methods
Data were presented as mean ± standard-deviations. The cytokine-concentrations were presented as median and the corresponding range. Statistical comparisons of two groups were performed while using the two-tailed t test (in cases with data normally distributed) and the Mann-Whitney-Wilcoxon-Test (in cases with data not normally distributed). Statistical analyses were performed with Microsoft Excel 2013 ® (Microsoft, Redmond, Washington, USA) and SSPS Statistic 24 software © (IBM, Armonk, USA). The differences were considered as 'not significant' in cases with p values >0.1, as 'borderline significant' (significant*), with p values between 0.1 and 0.05, as 'significant' (significant**) with p values between 0.05 and 0.005 and as 'highly significant' (significant***) with p values <0.005. The figures were created with GraphPad Prism7 © (GraphPad Software, California, USA).

Conclusion: DC/DC leu Based Treatment Protocols for AML-Patients
We developed a DC/DC leu -generating protocol Pici-PGE1 and demonstrated (by replacing PGE 2 with PGE 1 ) that the efficiency of sufficient DC/DC leu generation is superior from leukemic PBMCs. Furthermore, the PGE 1 -containing Kit M is a reliable tool for generating ex vivo mature DC/DC leu from healthy and leukemic WB, containing the complete patient individual soluble and cellular microenvironment. These 'new PGE 1 generated' DC/DC leu reliably (re-) activate immunoreactive cells, improve the overall ex vivo anti-leukemic activity, and influence cytokine-release-profiles compared to the PGE 2 -containing Kit K.
We conclude that (1) PGE 1 -containing protocols qualify to generate and to maturate monocyte derived DCs from healthy or even patients' PBMCs ex vivo that could, in consequence, be manipulated (e.g., pulsed with LAA) and (re-) administrated to the patients as a vaccine. (2) PGE 1 -containing protocols qualify to produce ex vivo DC leu from leukemic PBMCs.
These DC/DC leu could be (re-) administered to the patient in the course of an adaptive cell transfer. (3) GM-CSF, PGE 1 , and Picibanil are drugs that are approved for human treatment, and so we conclude that, for example, the PGE 1 -containing Kit M qualify to convert (residual) myeloid blasts to DCs and DC leu in vivo after the application of Kits to AML-patients. This could contribute to stabilize remission or the disease by presentation of the complete leukemic antigen repertoire to T cells and other immunoreactive cells independent of mutation or transplantation status, cytogenetic markers, FAB-classification, as well as sex or age of the patients. (4) In vivo trials with PGE 1 -containing Kits (in animals and humans with AML) have to be performed to study safety, the efficiency of DC/DC leu -generation, the mediation of anti-leukemic reactions, and the establishment of immunological effects in vivo.

Patent
HMS is the inventor of the European Patent 15 801 987.7-1118 'Use of immunomodulatory Kits for immunotherapeutic treatment of patients with myeloid leukemia's'. No financial conflicts of interest have to be declared.