Published online before print June 14, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,2,2
* Institute for Transfusion Medicine and Immunehaematology, Campus Virchow-Klinikum, Charité–Universitätsmedizin Berlin, Germany;
Memorial Sloan-Kettering Cancer Center, Sloan-Kettering Institute, New York, New York;
Robert-Koch-Institut, Berlin, Germany; and
Department of Pediatric Oncology, Charité, Campus Virchow-Klinikum, Humboldt-University Berlin, Germany
1 Correspondence: Institute for Transfusion Medicine and Immunhaemotology, Campus Virchow-Klinikum, Charité–Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin, Germany. E-mail: anja.moldenhauer{at}charite.de
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, granulocyte macrophage-colony stimulating factor, c-kit ligand, and soluble CD40 ligand; yet, addition of the HDAC inhibitor (Hdi) trichostatin A enhanced DC differentiation with retention of the fusion transcripts. These leukemic DC showed high-level CD83 and human leukocyte antigen (HLA)-DR expression and had a high allostimulatory potential. Only DC generated from these cell lines after Hdi induced blast-specific cytotoxic T cell responses in HLA-A-matched T cells with a cytotoxicity of 42% in parental Kasumi-1 and 83% in parental REH cells, respectively. This model system suggests that the Hdi supports the in vitro differentiation of DC from leukemic blasts with AML1-containing fusion proteins.
Key Words: leukemia • translocation • vaccination
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Recurrent cytogenetic abnormalities leading to the expression of leukemia-specific fusion proteins of transcription factors (TF) are found in a large percentage of patients with acute myeloid leukemia (AML) [8
] and acute lymphoblastic leukemia (ALL) [9
]. These fusion proteins represent unique, leukemia-specific antigens such as AML1-eight-twenty-one (ETO), which is found in 
40% of patients with AML FAB M2, and translocation ETS-like leukemia (TEL)-AML1, which is found in 20–30% of children with ALL. ETO and TEL recruit nuclear corepressor (N-CoR) molecules [10
] and HDAC [11
12
] to DNA transcriptional regulatory elements (Fig. 1
). In response to the deacetylation of core histone proteins, the accessibility of TF to DNA is disturbed [3
]. This leads to transcriptional repression of AML1-responsive genes [13
], thereby impairing normal hematopoiesis [15
]. Reversion of the transcriptional repression by histone deacetylase inhibition leads to cell differentiation [16
17
20
], which is also true for leukemic cells with non-AML1 dependent translocations [18
]; up-regulation of adhesion molecules has also been observed in some cell lines [19
].
 View larger version (33K): [in a new window] |
Figure 1. Schematic description of AML1 and AML1-ETO function. (A) In normal cell differentiation, AML1/core-binding factor (CBF)ß functions as an enhancer-organizing protein that interacts with several lineage-specific transcription factors (TF) to induce gene transcription. Binding of the AML1/CBFß complex to DNA usually leads to transcriptional activation of target genes whose transcription is regulated by AML1. Among these are myeloperoxidase (MPO), the receptor for colony-stimulating factor 1 (CSF-1R), the subunit of the T cell antigen receptor (TCR), interleukin (IL)-3 and granulocyte-macrophage colony stimulating factor (GM-CSF), and the cyclin-dependent kinase 4 (CDK4). (B) The AML1-ETO fusion protein retains the ability to bind the core-enhancer sequence and to heterodimerize with CBFß. In contrast to wild-type AML1, the ETO component of AML1-ETO binds a corepressor complex that includes mSin3, N-CoR, and Hdi. This latter interaction results in suppression of genes whose transcription is normally activated by AML1-CBFß, leading to a block of normal hematopoiesis. Application of histone deacetylase inhibitors (Hdi) (HDI in red) reverses transcriptional suppression, thereby supporting dendritic cell (DC) differentiation, while the leukemic fusion transcript is retained. Like AML1-ETO, TEL-AML1 also recruits N-CoR and histone deacetylase (HDAC), leading to a block in hematopoietic differentiation (adapted from refs. [13
, 14
]).
|
Several protocols for the differentiation of leukemic DC exist [23
24
25
]. A variety of cytokines, including granulocyte macrophage (GM)-CSF (G), tumor necrosis factor (TNF-; T), c-kit ligand (K; stem-cell factor), IL-4, and soluble CD40 ligand (s40L), stimulates leukemic cells to differentiate into DC [26
27
28
29
], and they retain the cytogenetic abnormality of the leukemic clone [30
31
32
33
]. CML blasts differentiated into DC were capable of inducing blast-specific cytotoxic T lymphocyte responses [34
35
]. However, others have demonstrated a lack of a clinical response despite the presence of a peptide-specific immune reaction in the patients treated [36
].
T cell anergy, which is often present in leukemic patients [37 ] may lead to disparities between laboratory test results and clinical efficacy. Although leukemic DC have reportedly been differentiated from more than 60% of AML patients studied, only a median 3.5% of the leukemic DC developed a DC-typical immune phenotype, and only a median 1.2% of those from ALL patients became CD1a(+)14(–) [38 ]. Moreover, leukemic DC are less potent T cell stimulators than monocytic DC [26 39 ]. This may explain why another group found anti-leukemic cytolytic T lymphocyte (CTL) responses in only one-fifth of its patients [32 ].
Given these facts, we hypothesized that Hdi would improve the differentiation of leukemic blasts into DC, while the cells retain their fusion proteins. These may then be presented as "tumor antigens." The AML1-ETO-positive cell lines Kasumi-1 and SKNO-1 as well as the TEL-AML1-positive cell line REH were used to test this hypothesis.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-positive acute leukemia cell line NB-4 was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ GmbH, Braunschweig, Germany). The T-ALL cell line SR-2 was kindly provided by Shahin Rafii (New York Hospital, New York, NY). All cell lines except SupB15 were cultured in RPMI supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, and 100 U/mL penicillin/streptomycin. SKNO-1 cells were kept in the above-mentioned medium supplemented with 80 U/ml GM-CSF (Immunex Corp., Seattle, WA). SupB15 cells were cultured in McCoy’s medium supplemented with 10% FBS and 100 U/mL penicillin/streptomycin.
Leukemia cells (0.5–1x106) were cultured in six-well plates with 100 ng/mL GM-CSF (specific activity, 2.5x107 U/mg, Immunex Corp.) and 100 ng/mL TNF- (specific activity, 1x108 U/mg, Genetech, San Francisco, CA) with variable combinations of the following agents: c-kit ligand (20 ng/mL stem-cell factor, specific activity, 5x106 U/mg, Amgen, Thousand Oaks, CA), IL-4 (20 ng/mL, specific activity, 1x107 U/mg, Intergen, New York, NY), recombinant human (rh)s40L (500 ng/mL, specific activity, 2x104 U/mg, PeproTech, St. Katharinen, Germany), and/or Hdi (trichostatin A, Sigma Chemical Co., St. Louis, MO). When used, the Hdi was dissolved in 98% ethanol and added to the cells on the first day of culture at a concentration of 60 nM.
Morphology
The morphology and maturation features of the cells in the aqueous cultures were documented by phase-contrast microscopy as described previously [41
]. Cell samples for scanning and transmission electron microscopy were collected 24–48 h after cytokine exposure. The samples were fixed in a mixture of 4% paraformaldehyde, 2.5% glutaraldehyde, and 0.02% picric acid in 0.1 M sodium cacodylate buffer (pH 7.3) for up to 2 days and then rinsed with 0.06 M phosphate buffer. Samples for scanning and transmission electron microscopy were prepared as described previously [42
43
] and then investigated using a DSM 982 Gemini scanning microscope and an EM 906 transmission microscope (Carl Zeiss Vision, Oberkochen, Germany).
Flow cytometry
Cells with detectable morphological changes were analyzed for DC-specific markers as described previously [41
] using a FACScan (Becton Dickinson, San Jose, CA) or Elite Profile II flow cytometer (Coulter, Hialeah, FL) and the following directly conjugated anti-human monoclonal antibodies (mAb): CD83-phycoerythrin (PE; Immunotech, Marseille, France), CD1a-PE (Becton Dickinson), CD80-PE (PharMingen, San Diego, CA), CD86-fluorescein isothiocyanate (FITC; PharMingen), human leukocyte antigen (HLA)-DR-FITC (Becton Dickinson), and CD14-FITC (Immunotech). Isotype-matched immunoglobulin G (IgG) antibodies served as negative controls.
Purification of CD83(+) cells
At peak frequencies of CD83(+)DR(+) positivity, the cells were magnetically sorted using the magnetic cell sorter system (Miltenyi Biotec, Auburn, CA) and then washed with Hank’s buffered salt solution (HBSS). Cells (1x107) were resuspended in 170 µl HBSS and incubated with purified mouse anti-human CD83 mAb (Immunotech, 30 µl) for 15 min at 4°C. After removing the excess antibodies, the labeled cells were incubated with 30 µl immunomagnetic beads (Miltenyi Biotec) conjugated with goat anti-mouse IgG1 mAb (15 min, 4°C). The CD83(+) population was then separated on magnetic stainless-steel wool columns according to the manufacturer’s instructions. The enriched population was >75% CD83(+), as confirmed by flow cytometry.
Expression of the AML1-ETO and TEL-AML1 fusion proteins
After cytokine treatment and CD83 sorting, the Kasumi-1 cells were lysed in a sodium dodecyl sulfate (SDS) buffer containing the following protease inhibitors: 125 mM TrisCl, 4% SDS, 200 mM dithiothreitol, 20% glycerol, 1 µg/mL aprotinin, 5 µg/mL leukopeptin, and 1 µg/mL pepstatin. Cells (1x105) were then loaded onto SDS-polyacrylamide gel electrophoresis (10%) and transferred to nitrocellulose filters. REH cell nuclear extracts were prepared according to the Dignam-Roeder protocol [44
], after 5 x 105 cells/mL had been cultured in variable concentrations of trichostatin A (0, 20, 50, and 100 ng/mL) for 24 and 48 h. The following agents were used to identify the fusion proteins: AML1-ETO and ETO, an affinity-purified anti-ETO antibody (kindly provided by Paul Erickson, University of Colorado Health Sciences Center, Denver, CO); AML1, anti-AML1 (RUNT; Calbiochem, San Diego, CA); and TEL, anti-TEL (PNT; kindly provided by Dr. Peter Marynen, Flanders Interuniversity Institute for Biotechnology, Leuven, Belgium). Acetylated histones were identified using antiserum to acetylated histone H3 and H4 (Upstate Biotech Inc., Lake Placid, NY). Anti--tubulin (Sigma Chemical Co.) served as control. Enhanced Chemiluminescence Plus was used for chemiluminescence detection (Amersham Pharmacia Biotech, Piscataway, NJ).
Real time-polymerase chain reaction for TEL/AML1
The content of TEL/AML1 mRNA in CD83(+)-isolated REH-DC was quantified as described previously [45
] with the following modification: Instead of ß-actin, ß2-microglobulin transcripts were used as reference.
Mixed lymphocyte reaction (MLR) and stimulation of HLA-matched T cells
T cell isolation
Peripheral blood mononuclear cells (PBMC) were isolated from buffy coat samples or from heparinized blood of HLA-matched healthy donors by Ficoll-Paque separation (Pharmacia, Piscataway, NJ) and used as fresh PBMC. For some experiments, CD3(+) and CD8(+) cells were isolated using immunomagnetic microbeads (Miltenyi Biotec), according to the manufacturer’s instructions. When the highest purity of leukemic CD83(+)DR(+) cells was achieved (usually by day 6), an aliquot of the differentiated cells was irradiated (Kasumi-DC at 120 Gy, SKNO-1 and REH-DC at 30 Gy), and an allogeneic MLR was performed using 3H-thymidine incorporation, as described previously [41
].
Kasumi cells
Autologous stimulation was achieved by culturing HLA-A-matched T cells from two healthy volunteers with naive Kasumi cells or Kasumi-derived DC produced following exposure to various cytokines with or without Hdi. CD3(+) or CD8(+) cells (1x106) per well were cultured with 0.4 –1 x 107-irradiated stimulators [yielding a stimulator-to-responder (S/R) ratio of 1:4–1:10] in 24-well plates with RPMI 1640 supplemented with 5% autologous plasma, 2 mM L-glutamine, 100 mM nonessential amino acid solution, 1 mM sodium pyruvate, 100 U/mL penicillin, 100 mg/L streptomycin, and 55 mM 2-mercaptoethanol (all supplied by Gibco-BRL, Life Technologies, Great Island, NY). rhIL-2, 20 U/mL (Genzyme, Cambridge, MA), was added after the second stimulation. T cells were isolated after the cells had been stimulated at least three times.
REH and SKNO-1 cells
REH-DC and SKNO-1 DC sensitized HLA-matched T cells pretreated with IL-2 (100 U/ml) at a ratio of 1:1 according to a previously published stimulation protocol [30
]. The major histocompatibility complex (MHC) class I antigens of the leukemia cell lines used as well as the HLA profile of HLA-A-matched and -mismatched T cells are summarized in Table 1.
|
View this table: [in a new window] |
Table 1. HLA Profile of Leukemia Cells As Well As the HLA-A-Matched and Mismatched T Cells
|
(IFN-) enzyme-linked immunospot (ELISPOT) assay antibodies in phosphate-buffered saline (PBS; 10 µg/mL; clone 1-D1K, Mabtech, Sweden), incubated overnight at 4°C, washed with PBS to remove excess antibodies, and blocked with 10% autologous plasma in RPMI for 1 h at 37°C. Purified Kasumi or Kasumi-DC-stimulated CD8(+) cells were plated at a concentration of 1 x 105 cells/well. For antigen presentation, 5 x 104 irradiated, naive Kasumi cells or Kasumi-derived DC were added to yield a final volume of 100 µl/well. The plates were incubated at 37°C for 20 h, then washed thoroughly with 0.05% Tween in PBS; 100 µl biotinylated anti-IFN- antibodies (2 µg/mL; clone 7-B6-1, Mabtech) were added to each well. The plates were incubated for an additional 2 h at 37°C, and spot development was performed as described previously [46
]. The spots were automatically counted using a computer-assisted video image analyzer with KS ELISPOT 4.0 software (Carl Zeiss Vision). Each sample was tested in duplicate or triplicate.
Cytotoxicity assay
51Chromium (Cr) release
Kasumi cells (4x106) and SupB15 cells, matching one HLA-A with donor T cells but mismatched with Kasumi-1, were labeled with 300 µCi Na2CrO4 (NEN Life Science Products Inc., Boston, MA) for 1 h at 37°C. After washing by centrifugation, target cells were resuspended in complete medium at 5 x 104 cells/mL and plated in 96-well round-bottom plates at 5 x 103 cells per well. CD8(+) effector cells were added in effector/target (E/T) ratios ranging from 50:1 to 12.5:1. Plates were incubated for 5 h (37°C, 5% CO2). Radioactivity of the supernatants was measured using a -counter.
Lactate dehydrogenase (LDH) release
Target cells (5x103) were cultured with sensitized T cells at E/T ratios of 50:1, 25:1, and 12.5:1 for 5 h in 96-well round-bottom plates containing phenol red-free RPMI-1640 medium plus 5% heat-inactivated, pooled human AB serum. Cytotoxicity was measured using a LDH release assay, according to the manufacturer’s instructions (Promega, Madison, WI). Light absorption (490 nm) was determined by a microplate reader (MR 5000, Dynatech Laboratories, Denkendorf, Germany). Non-matched leukemic cells were used as HLA-mismatched target controls. T cells stimulated with parental leukemic cells or with IL-2 alone served as effector control. The percentage of specific lysis was determined using the following formula: 100 x (experimental release–spontaneous release target–spontaneous release effector)/(target maximum release–spontaneous release). HLA-mismatched T cells stimulated with REH- and Kasumi-DC, respectively, were used as additional controls. In some experiments, target Kasumi and REH cells were incubated with anti-MHC I or II antibodies (Dako, Glostrup, Denmark) for 20 min prior to the addition of effector T cells.
Statistical analysis
Student’s t-test on Microsoft Excel 97 was used to evaluate the differences between two groups. P < 0.05 was considered to be statistically significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
plus GM-CSF with or without c-kit ligand, Kasumi cells, like SKNO-1 cells, developed dendrites, grew in clusters, and demonstrated the typical morphological features of DC; these features were also present after Hdi application (Fig. 2
). Electron microscopy studies showed that these cells contained numerous small, dense granules, microvilli, and polymorphic pockets, characteristic of mature DC (Fig. 3
). In contrast to Kasumi cells, REH cells tend to grow in small clusters with spindle-shaped cells lying on the bottom of the well. Treatment with Hdi-TNF- + GM-CSF + c-kit ligand (TGK) led to an increase in cell clustering and the development of hair-like spikes within 1 week.
 View larger version (63K): [in a new window] |
Figure 2. (A–C) Suspension cultures examined by light microscopy (200x original magnification). Morphological changes observed in Kasumi-1 cells (A, grown without cytokines) cultured in DC-inducing cytokines (TNF-, GM-CSF, and K) for 48 h without (B) and with (C) trichostatin A. (C) The image was taken using phase-contrast. Original size bar = 30 µm.
|
 View larger version (107K): [in a new window] |
Figure 3. Electron microscopic characterization of Kasumi cell-derived DC. In contrast to naive Kasumi cells (A, B), Kasumi-derived DC, generated in TNF-, GM-CSF, and K, developed dendrite-like microvilli, small, dense granules, lipid drops, and polymorphic pockets (C, D). Similar morphologic changes were seen after Hdi and addition of the same cytokine combination (E, F), as demonstrated by transmission microscopy (6000x original magnification; A, C, and E, original size bar=5 µm) and scanning electron microscopy (5000x original magnification; B, D, and F, original size bar=10 µm).
|
and GM-CSF, which was further enhanced by application of Hdi (14.9 vs. 22%), although this was not generally the case. Significant up-regulation of CD83(+) with dual expression of CD83(+)DR(+) was observed in 37% of the cells stimulated with TGK. The addition of Hdi to TGK (Hdi-TGK) further increased the expression of CD83(+)DR(+) cells from 37% to 49%. After culturing in TGK plus s40L (TGK-s40L), nearly 70% of the Kasumi cells became CD83(+)DR(+) and expressed costimulatory molecules, such as CD80 and CD86. Up-regulation of CD1a or CD14 did not occur in any of the culture conditions. The addition of IL-4 to TGK did not enhance the generation of DC.
 View larger version (48K): [in a new window] |
Figure 4. (A) Up-regulation of CD83 and HLA-DR in Kasumi cells by various cytokine combinations. Flow cytometry revealed an up-regulation of CD83 and HLA-DR with a slight increase in CD80 and CD86 expression, after culturing Kasumi cells in cytokines for 6 days. No change in CD1a nor CD14 was observed. The numbers in each box represent the frequency of positive cells. Hdi (trichostatin A). (B) Up-regulation of CD83 and HLA-DR in REH cells. After 6 days in culture, the frequency of CD83(+)HLA-DR(+) cells in REH cells treated with cytokines plus the Hdi was significantly higher than in those treated with cytokines alone. Unlike in Kasumi cells, s40L did not up-regulate CD83 dramatically.
|
In REH cells, TGK, Hdi-TGK, and TGK-s40L also stimulated the TEL-AML1-positive REH cell line to differentiate into DC. Hdi significantly increased the expression of CD83, regardless of whether s40L was used (Fig. 4 B) . Hdi increased the percentage of CD83(+)HLA-DR(+) cells from 1% to 66.3% without s40L and from 14.7% to 58.2% with s40L.
In contrast to the Kasumi-1 and REH cells, neither the K562 chronic myelogenous leukemia cell line, which contains a t(9;22) translocation and expresses the BCR-ABL fusion protein, nor the lymphoblastic cell lines SD-1 and SR-2 could be induced to differentiate into mature DC, even after Hdi. No DC-like morphologic changes nor up-regulation of surface markers typical for DC were observed. PML-RAR-positive NB-4 cells did not repeatedly demonstrate a significant up-regulation of CD83 after Hdi (on average control 1.4%, TGK 12.2%, Hdi-TGK 13.8%).
Expression of fusion proteins
Western blot analysis demonstrated that the differentiation of Kasumi cells into DC did not alter their expression of the AML1-ETO fusion protein (Fig. 5A
). As Hdi can affect the expression of a variety of genes, we analyzed the REH-derived DC for changes in the level of the TEL-AML1 and the wild-type AML1 protein. As shown in Figure 5B
, Hdi did not alter TEL-AML1 levels significantly. However, we observed the expected accumulation of the acetylated histones H3 and H4 at 24 and 48 h.
 View larger version (48K): [in a new window] |
Figure 5. Persistence of AML1-ETO and TEL-AML1 expression. (A) Western blot (WB) analysis using an anti-ETO antibody demonstrated continuous expression of AML1-ETO protein (kD) in Kasumi-DC derived in cytokines plus Hdi after immunomagnetic isolation of CD83-positive cells. No differences between CD83(+) and CD83(–) cells were found. A/E = AML1-ETO protein. (B) In REH cells, trichostatin A (ng/mL) led to an increase in H3 acetylated protein after 24 h, and the expression of TEL-AML1, AML1, and H4 acetylated proteins did not change. Increased H4 expression was observed at 48 h. -Tubulin served as control for equivalent loading. Results of one of three series of measurements are shown.
|
Allostimulatory capacity
As shown in Figure 6A
, the allostimulatory capacity of Kasumi-derived DC cultured in TGK was generally two to five times higher than that of the undifferentiated, parental Kasumi cells (38.5±9.6 vs. 15.7±1.7x103 cpm, P<0.05). The difference between the TGK and Hdi-TGK groups became significant at lower S/R ratios, e.g., 7.0 ± 0.98 versus 13.4 ± 1.6 x 103 cpm (P=0.012) at an S/R ratio of 1:480, but was not significant at an S/R ratio of 1:30 (38.5±9.6 vs. 52.5±10.3x103 cpm; P=0.37). The highest level of immunogenicity was observed in the TGK-s40L group (75.6±3.6x103 cpm, S/R=1:30). In SKNO-1 cells, the highest immunostimulatory potential was achieved, if the cytokines had been applied in combination with trichostatin A.
 View larger version (38K): [in a new window] |
Figure 6. T cell stimulatory capacity of Kasumi and REH cell-derived DC. (A) TGK-s40L achieved the highest thymidine uptake in Kasumi DC, followed by Hdi-TGK. S/R ratios of 1:30–1:480 were achieved by incubating 5000–313 stimulators with 150,000 responders. The MLR was repeated three times in triplicate. (B) The highest allostimulatory capacity in REH-DC was observed in cells cultured in the triple cytokine combination after Hdi.
|
Kasumi-, SKNO-1-, and REH-DC derived after Hdi primed HLA class-I-matched T cells to induce a specific cytotoxic response against the parental cell line
ELISPOT
As the Kasumi-DC generated with TGK-s40L and Hdi-TGK had the highest allostimulatory potential, we compared their potential to induce a specific immune response in HLA-matched T cells against the non-treated Kasumi cells. Only the Kasumi-DC generated by the triple cytokine combination plus trichostatin A sensitized HLA class I-matched T cells to release detectable levels of IFN- in the presence of the undifferentiated, parental Kasumi cells. The mean number of spots detected in T cells primed with Kasumi-derived DC from cytokines plus Hdi (Hdi-TGK) was 90.8 ± 21.2 compared with 20.2 ± 9.6 in T cells primed with the parental Kasumi cells (P=0.02) and only 4.3 ± 1.5 in T cells primed with Kasumi-derived DC from the TGK-s40L group (P=0.006; Table 2
). Kasumi-derived DC generated with cytokines only sensitized T cells to Kasumi-derived DC but not to untreated Kasumi cells.
|
View this table: [in a new window] |
Table 2. Results of ELISPOT Assay Using HLA-A-Matched T Cells
|
 View larger version (23K): [in a new window] |
Figure 7. Specific cytotoxic T cell responses. Kasumi-derived DC cultured in cytokines plus the Hdi primed HLA-A-matched cytotoxic T cells, inducing specific lysis of the parental Kasumi cells (A). Less than 7% of control cells were lysed. Kasumi-DC generated by three (TGK) or four cytokines alone (TGK-s40L) did not stimulate a Kasumi-specific T cell response. Similar to the Kasumi cells, REH-DC obtained after Hdi induced a specific cytotoxic T cell response to parental REH cells (Hdi, B). SD-1 cells served as controls. REH-DC, generated with cytokines alone (TGK), stimulated a non-specific cytolysis of REH and control cells. When HLA-mismatched T cells stimulated with REH-DC were used as effectors, no cytolysis of REH cells was demonstrated. The figure lists the mean values ± SEM of three representative experiments in duplicate or quadruplicate (mean results±SEM of three series of measurements in triplicate or quadruplicate).
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In our efforts to optimize leukemic DC generation, we found that TNF- plus GM-CSF (TG) alone were able to induce a dendritic immunophenotype but that the combined use of TG plus c-kit ligand (TGK) increased the absolute number of DC. Mature DC are characterized by a paucity of intracellular organelles, the prominence of endosomes and lysosomes [47
48
] and the expression of CD83. All of these characteristics were present in our leukemic DC. IL-4 plays an important role in the generation of DC from monocytes [49
], as it can down-regulate the monocytic marker CD14 [50
51 ] and suppress macrophage development [52
]. Nonetheless, Kasumi cells did not become CD14-positive, and the lack of an IL-4 effect suggests that Kasumi cells may differentiate into DC through a pathway that does not involve monocytic maturation.
The combined use of TGK plus s40L (TGK-s40L) dramatically increased the frequency of CD83(+)DR(+) Kasumi cells, which were also highly positive for CD86. s40L, a member of the TNF superfamily present on the surface of activated T cells [53 ], can induce the maturation of human DC from CD34(+) cells [54 ], monocytes [55 ], and leukemic cells [56 57 ]. Nevertheless, s40L did not increase the absolute number of Kasumi-DC despite their purity, likely due to negative effects of s40L on cell proliferation. The high level of CD83, HLA-DR, and costimulatory molecule expression most likely accounts for the significant allostimulatory capacity of these Kasumi-DC.
Hdi has been found to induce cell differentiation and apoptosis in Kasumi cells [5 16 ] and to up-regulate costimulatory molecules such as CD80 and CD86 in several AML cell lines. However, neither we nor Maeda and colleagues [20 ] were able to observe the latter effect in Kasumi-1.
In REH cells, the number of CD83(+)DR(+) cells increased significantly in response to treatment with TGK plus Hdi. Treatment with TGK plus s40L led to up-regulation of CD83 expression by almost 15%, but the effect of Hdi on DC maturation in response to this cytokine combination was not as profound as with TGK. Hdi also affected the allostimulatory capacity of REH-DC, as the proliferation of allogeneic T cells was up to five times higher in REH-DC derived in Hdi-TGK. Like Kasumi-DC, CD83(+) REH-DC retained expression of the AML1-containing fusion protein despite DC differentiation, suggesting an impairment of TEL-AML1 function. The increased amount of AML1 noted 48 h after Hdi most likely reflects a release in the block of AML1 expression caused by TEL-AML1.
It is important to note that Kasumi-, SKNO-1-, and REH-derived DC induced a leukemia-specific CTL when generated after Hdi. In contrast to other studies [32 58 ], Hdi-generated, leukemic DC sensitized HLA-matched T cells to induce very high levels of specific target cytolysis (42% in the case of Kasumi-DC and 83% in REH-DC), and HLA-mismatched T cells were not stimulated. Fabre [59 ] postulated that optimal allogeneic DC would be incompatible at one HLA-DR allele, while sharing as many MHC class I and II alleles as possible for tumor peptide presentation. This might explain why the cytotoxicity observed in REH cells, which matched their effector T cells in one HLA-B and two HLA-A alleles, was almost twice as high as in Kasumi cells. Although leukemic DC generated in the absence of Hdi also sensitized matched T cells, the cytotoxic response to the parental cell line was lower than that induced by Hdi-generated DC and equal to the response to control cell lines.
The cytolysis of target leukemic cells by HLA-matched T cells was significantly reduced by blocking MHC class I (and, to a lesser extent, class II) molecules. This implies that the observed cytotoxicity is predominantly a CD8(+) T cell response, but it might involve cytolytic CD4(+) T helper cells.
CD83 is known to be a surface marker of mature DC [60 ]. In murine and human monocyte-derived DC, a strong correlation between the degree of CD83 positivity and the allostimulatory capacity as well as their antigen-presentation capacity has been demonstrated [61 62 ]. However, the importance of CD83 in leukemic DC is still unknown. CD83(–) leukemic DC can have a high, allostimulatory capacity, as demonstrated by Narita et al. [37 ]. In our study, we saw a good correlation between CD83 expression and the allostimulatory capacity of Kasumi-DC, SKNO-1-DC, and REH-DC; yet, in Kasumi-DC, there was no correlation with their ability to induce a leukemia-specific CTL response. This is consistent with the findings published by Harrison et al. [32 ]. Their results and ours suggest that CD83 is a good, functional marker for the allostimulatory potential of leukemic DC but not for their ability to induce a specific, antileukemic CTL response.
Although s40L-generated Kasumi DC had the highest allostimulatory capacity and the highest frequency of CD83(+)DR(+) cells, they did not induce a specific, cytotoxic T cell response against the Kasumi cells from which they had originated, as demonstrated by ELISPOT and cytotoxicity assays. One possible explanation could be the nonspecific up-regulation of antigen-presenting molecules, which could dilute a specific response. Another explanation could be that Hdi enhanced the expression of the translocation protein or other specific leukemic markers on the cell surface, thereby promoting their presentation as antigens.
In spite of its toxicity, trichostatin A promoted the generation of functional, antigen-presenting cells, but it did not seem to have a universal effect on the differentiation of leukemic DC. Other cell lines, which lack an AML1-dependent translocation, did not show induction of CD83/HLA-DR expression by the addition of Hdi. For example, the development of leukemic DC from NB-4 cells, which are positive for the fusion transcript PML-RAR, was not enhanced after histone deacetylation. Although PML-RAR does recruit HDAC, PML-RAR-positive leukemic cells could only be differentiated after a combined use of trichostatin A and all-trans retinoic acid [19
]. Therefore, in these cells, the sole use of trichostatin A in combination with cytokines did not induce DC differentiation.
This emphasizes the potential benefit of using Hdi in generating leukemic DC in patients with AML1-dependent leukemias. On the one hand, Hdi enhances the "dendriticity" of leukemic DC. On the other hand, Hdi-induced apoptosis of leukemic cells [63 64 65 ] might further support the induction of a leukemia-specific CTL. Spisek et al. [58 ] obtained similar results by cross-presenting apoptotic blasts to patients’ nonleukemic DC.
In conclusion, the combination of cytokines with Hdi improved DC differentiation from two t(8;21)-positive AML and one t(12;21)-positive ALL cell line. These leukemic DC continued to express the leukemic-fusion proteins, AML1-ETO and TEL-AML1, respectively. They displayed a typical DC morphology and immunophenotype and evoked high, blast-specific cytotoxic responses in HLA-matched T cells to the leukemia cell lines from which they originated. These cell line-derived DC may be useful as new immunotherapeutic tools for patients with identical AML1 translocations and matching HLA profiles. More practically, Hdi may be able to increase the therapeutic efficacy of leukemic DC applied in the vaccination against AML1-associated leukemias.
Received November 23, 2003; revised December 19, 2003; accepted May 9, 2004.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
in response to HLA-A2.1-binding melanoma and viral peptide antigens J. Immunol. Methods 191,131-142[CrossRef][Medline] for use in cancer immunotherapy Ann. Surg. 226,6-16[CrossRef][Medline]
B activation in hematopoietic progenitor cells J. Immunol. 160,1224-1232This article has been cited by other articles:
 |
|
 |
|
  K. Schmidt, K. Seeger, C. Scheibenbogen, R. Bender, M. Abdulla, S. Sussmilch, A. Salama, and A. Moldenhauer Histone deacetylase inhibition improves differentiation of dendritic cells from leukemic blasts of patients with TEL/AML1-positive acute lymphoblastic leukemia J. Leukoc. Biol., March 1, 2009; 85(3): 563 - 573. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Nencioni, J. Beck, D. Werth, F. Grunebach, F. Patrone, A. Ballestrero, and P. Brossart Histone Deacetylase Inhibitors Affect Dendritic Cell Differentiation and Immunogenicity Clin. Cancer Res., July 1, 2007; 13(13): 3933 - 3941. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Yin, L. Gonzales, V. Kolla, N. Rath, Y. Zhang, M. M. Lu, S. Kimura, P. L. Ballard, M. F. Beers, J. A. Epstein, et al. Hop functions downstream of Nkx2.1 and GATA6 to mediate HDAC-dependent negative regulation of pulmonary gene expression Am J Physiol Lung Cell Mol Physiol, August 1, 2006; 291(2): L191 - L199. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Gialitakis, A. Kretsovali, C. Spilianakis, L. Kravariti, J. Mages, R. Hoffmann, A. K. Hatzopoulos, and J. Papamatheakis Coordinated changes of histone modifications and HDAC mobilization regulate the induction of MHC class II genes by Trichostatin A Nucleic Acids Res., February 1, 2006; 34(3): 765 - 772. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
