Abstract
Objectives
Proliferative capacity of acute myelogenous leukaemia (AML) blasts is important for leukaemogenesis, and we have investigated whether proliferative capacity of primary human AML cells could be used for subclassification of patients.
Materials and methods
In vitro proliferative capacity of AML cells derived from two independent groups was investigated. Cells were cultured under highly standardized conditions and proliferation assayed by 3H‐thymidine incorporation after seven days culture. Patients were subclassified by clustering models, and gene expression profile was examined by microarray analyses.
Results
Based on proliferative capacity of the AML cells, three different patient clusters were identified: (i) autocrine proliferation that was increased by exogenous cytokines; (ii) detectable proliferation only in presence of exogenous cytokines; and (iii) low or undetectable proliferation even in presence of exogenous cytokines. Patients with highest proliferative capacity cells had no favourable prognostic impact by NPM‐1 mutation. Analysis of gene expression profiles showed that the most proliferative cells generally had altered expression of genes involved in regulation of transcription/RNA functions, whereas patients with high proliferative capacity and internal tandem duplications (ITDs) in the FLT3 cytokine receptor gene had altered expression of several molecules involved in cytoplasmic signal transduction.
Conclusions
In vitro proliferative capacity of primary human AML cells was considerably variable between patients and could be used to identify biologically distinct patient subsets.
Introduction
Acute myelogenous leukaemia (AML) is a heterogeneous malignancy arising from transformed haematopoietic stem‐ and progenitor cells of the bone marrow 1, 2. Bone marrow infiltration of leukaemic cells with suppression of normal haematopoiesis is a hallmark of the disease. AML cell proliferation can be stimulated by several haematopoietic growth factors as well as by other cytokines. In vitro studies have also demonstrated that for some patients, AML cells have spontaneous proliferation due to autocrine growth factor release by leukaemic blasts 3, 4, whereas for other patients, leukaemic cells only proliferate in the presence of exogenous growth factors 3, 4, 5, 6. Cytokines may in addition, induce AML blast differentiation 7. Abnormalities of cytokine‐ and growth factor‐induced intracellular signalling are characteristics of myeloid malignancies 8, and gene expression profiles have emerged as powerful means to further classify AML patients 9, 10. The aims of the present study were to investigate effects of various cytokines on AML blast proliferation, to use this proliferative responsiveness for subclassification of patients, and to compare global gene expression profiles for patient subsets with different proliferative capacity.
Materials and methods
AML cells
This study was approved by the local Ethics Committee (Regional Ethics Committee III, University of Bergen, Norway) and samples were collected after written informed consent. We included patients with AML without signs of acute promyelocytic leukaemia (APL) based on clinical, morphological and genetic criteria. AML cells were isolated from peripheral blood of patients with high blood blast counts (>7 × 109/l) by density gradient separation 11. Isolated cells included at least 95% leukaemia blasts 7, 11, 12. Two independent and consecutive patient groups were examined; they are referred to as group I and group II and their clinical and biological characteristics are summarized in Table 1. The two cohorts represent two different groups of consecutive patients that were analysed at two different time periods in two independent sets of experiments, but patients were included according to the same criteria for both groups.
Table 1.
Clinical and biological characteristics of the included AML patients. The Table shows the clinical and biological characteristics for the two patient groups (absolute numbers). The classification of cytogenetic abnormalities as favourable, unfavourable or intermediate is based on the recent classification system 30.
| Patient characteristics | Group I | Group II |
|---|---|---|
| Demographic data and disease history | ||
| Gender (numbers) | ||
| Male/female | 20/20 | 27/26 |
| Age (median, range) | 64 (29–84) | 61 (26–82) |
| History | ||
| De novo | 35 | 33 |
| Secondary | 2 | 13 |
| Relapse | 3 | 4 |
| AML cell differentiation | ||
| FAB classification | ||
| M0‐1 | 14 | 15 |
| M2 | 7 | 12 |
| M4‐5 | 19 | 22 |
| M6‐M7 | 0 | 1 |
| CD34 expression | ||
| 0–20% | 13 | 20 |
| 20–50% | 2 | 3 |
| 50–100% | 22 | 18 |
| Genetic abnormalities | ||
| Cytogenetics | ||
| Favourable | 3 | 3 |
| Unfavourable | 2 | 6 |
| Intermediate | 6 | 6 |
| Normal | 22 | 17 |
| FLT3 | ||
| ITD | 22 | 19 |
| TKD | 2 | 4 |
| Wt | 16 | 26 |
| NPM‐1 | ||
| Mutated | 13 | 9 |
| Wt | 26 | 31 |
Proliferation assay
AML cells (5 × 104 cells in 150 μl medium per well) were cultured in flat‐bottomed 96‐well plates (Nucleon™; Nunc, Roskilde, Denmark) in serum‐free Stem Span SFEM™ medium with gentamicin 100 μg/ml (Stem Cell Technologies, Vancouver, BC, Canada). Cells were cultured in medium alone and in presences of 20 ng/ml of either stem‐cell factor (SCF), granulocyte‐macrophage colony‐stimulating factor (GM‐CSF), G‐CSF, M‐CSF, FLT3‐ligand (FLT3‐L), interleukin 3 (IL‐3) or IL‐1β (all purchased from PeproTech Ltd, Rocky Hill, NJ, USA). Cultures were incubated at 37 °C in a humidified atmosphere of 5% CO2 for seven days before proliferation was assayed by 3H‐thymidine incorporation 13, 14.
RNA preparation, labelling and microarray hybridization
Patient group I
Microarray experiments on AML blasts in vitro were performed using the Illumina iScan Reader, based on fluorescence detection of biotin‐labelled cRNA. For each sample, 300 ng of total RNA was reversely transcribed, amplified and biotin‐16‐UTP–labelled by using the Illumina TotalPrep RNA Amplification Kit (Applied Biosystems/Ambion, Austin, USA). Amount and quality of biotin‐labelled cRNAs were controlled both by NanoDrop spectrophotometer and by Agilent 2100 Bioanalyzer. Thereafter, 750 ng of biotin‐labelled cRNA was hybridized to HumanHT‐12 V4 Expression BeadChip according to the manufacturer's instructions. HumanHT‐12 V4 BeadChip targets 47 231 probes were derived primarily from genes of the NCBI RefSeq database (Release 38).
Patient group II
Agilent Human Whole Genome Oligo Microarray (Agilent Technologies, Inc., Palo Alto, CA, USA) was used to analyse 16 AML cell samples with FLT3‐ITD mutations. RNA isolation, cRNA synthesis and hybridization were then performed as previously described 15. These oligonucleotide microarrays were scanned (Agilent Scanner G2505B) and features automatically extracted using Agilent Feature Extraction v.7.5.
Bioinformatic analyses of microarray data
IlluminaiScan Reader data were investigated in GenomeStudio and J‐Express 2012 for quality control measures 16. Before being compiled into an expression profile data matrix, all arrays within the experiment were quantile normalized to be comparable. Analyses were performed using the J‐Express 2012 analysis suite (MolMine AS, Bergen, Norway) 17. Unsupervised hierarchical clustering was performed with Euclidian correlation as distance measure and complete linkage. Agilent data were imported and analysed in J‐Express software (Molmine, http://www.molmine.com). Mean spot signals were used as intensity measure and expression data normalized by using median over entire array. Analysis of variance (ANOVA) was used to identify differentially expressed genes in patients with FLT3‐mutations.
Results
Cytokine‐dependent AML cell proliferation showed wide variation between patients
Cytokine‐dependent proliferation was investigated by using 3H‐thymidine incorporation assay for both patient groups, and amounts of DNA‐incorporated 3H‐thymidine were then determined as counts per minute (cpm) 11. Overall results are summarized in Table 2. Only one subset of patients had spontaneous AML cell proliferation in medium alone, but all cytokines caused a statistically significant increase of AML cell proliferation (P < 0.001 for all cytokines, Wilcoxon's signed rank test) comparing overall results. However, for one subset of patients, leukaemia cells had no or only minimal proliferation even in the presence of exogenous cytokines.
Table 2.
Cytokine‐dependent AML cell proliferation. AML cells were cultured in the presence of 20 ng/ml of each cytokine (right column) for seven days before proliferation was assayed in the 3H‐thymidine incorporation assay. Results are given as median cpm ± standard error (SER) for each cytokine and for each of the two independent groups described in Table 1.
| Cytokine added | Group I (n = 40) | Group II (n = 53) |
|---|---|---|
| FLT3‐L | 6992 (±2982) | 2326 (±1964) |
| GM‐CSF | 6336 (±2941) | 4550 (±3247) |
| SCF | 5884 (±1288) | 3754 (±3247) |
| IL‐3 | 8433 (±2356) | 5760 (±2880) |
| IL‐1 | 3800(±1057) | 3079 (±2736) |
| G‐CSF | 6990 (±2571) | 6483 (±3829) |
| M‐CSF | 5119 (±1869) | 953 (±1487) |
| Medium alone | 2767 (±1507) | 430 (±726) |
Cytokine‐dependent proliferation of primary human AML cells can be used to subclassify patients
Values for proliferation were log10 converted before unsupervised hierarchal clustering analysis was performed (Euclidean as distance metrics and complete linkage). On the basis of this analysis, we identified three distinct patient clusters that could be seen in both patient groups (Figs 1 and 2): (i) one cluster showed high proliferation both in medium alone and in presence of exogenous cytokines (blue, Figs 1 and 2); (ii) a second cluster showed high proliferation only in the presence of exogenous cytokines (green); and (iii) the final cluster had absent or low proliferation both with and without exogenous cytokines (brown). The cluster model also allowed us to identify different cytokine clusters. Even though cytokine clustering varied between the two patient groups, GM‐CSF and IL‐3 clustered close to each other and FLT‐3 and SCF clustered together for both groups. Observations indicate that proliferative responsiveness showed similarities between cytokines that cluster together or close to each other.
Figure 1.
Unsupervised hierarchical clustering analyses based on proliferation of different cytokines among 40 AML patients. Primary human AML cells derived from 40 consecutive patients (Group I, Table 1) were cultured for 7 days before proliferation was assayed using 3H‐thymidine DNA incorporation. Values were log10 transformed and used in unsupervised hierarchical cluster modeling with Euclidian correlation distance measure with complete linkage. Patients could be divided into three distinct clusters: (i) high proliferation both in medium alone (autocrine proliferation) and in the presence of exogenous cytokines (indicated by a blue bar in the figure), (ii) high proliferation only in the presence of exogenous cytokines, but low/undetectable proliferation in medium alone (green bar), and (iii) low/undetectable proliferation both in medium alone and in the presence of exogenous cytokines (brown bar). Patient characteristics are indicated in the right column.
Figure 2.
Unsupervised hierarchical clustering analysis based on proliferation of different cytokines among 54 AML patients. Blasts derived from 54 consecutive AML patients (Group II, Table 1) were cultured for seven days before proliferation was assayed using the 3H‐thymidine proliferation assay. Values were log10 transformed and used in unsupervised hierarchical cluster modeling with Euclidian correlation distance measure with complete linkage. Patient could be divided into three distinct clusters: (i) high proliferation both in medium alone and in the presence of exogenous cytokines (indicated by a blue bar in the figure), (ii) high proliferation only in the presence of exogenous cytokines but undetectable/low proliferation in medium alone (green bar), (iii) low/undetectable proliferation both in medium alone and in the presence of exogenous cytokines (brown bar).
Proliferative AML cell profile was associated with biological characteristics – no patients with high proliferative capacity had favourable prognostic impact of NPM‐1 mutations
As proliferation patterns of AML blasts have been associated with prognosis for patients who receive intensive chemotherapy 3, we investigated whether any associations could be detected between proliferation and clinical or biological characteristics of the patients. No significant associations could be detected for FAB classification, cytogenetic abnormalities or FLT3 mutations. However, favourable impact of NPM‐1 mutations could not be detected for any patients with both high spontaneous and cytokine‐dependent proliferation (Figs 1 and 2, blue); these patients did not have NPM‐1 mutations (group II) or NPM‐1 mutations occurred either together with FLT3‐internal tandem duplications (FLT3‐ITD) or high‐risk cytogenetic abnormalities (group I, 4 patients). Furthermore, generally higher CD34 expression corresponding to at least 50% positive cells, by standard flow cytometric analyses, was seen for patient cluster with high spontaneous as well as cytokine‐dependent proliferation, but this difference reached statistical significance only for group II patients (Chi‐square test, P = 0.014).
AML cells derived from patients with leukaemia relapse were heterogeneous with regard to proliferative capacity but early relapse seemed to be associated with high proliferative responses
Our study included 7 patients with AML relapse after intensive chemotherapy (Figs 1 and 2); duration of first remission varied between 4 and 42 months and only two of these patients had a duration of first remission longer than 12 months (16 and 37 months, respectively). The possibility of achieving a second haematological remission is higher if patients have a long duration of their first remission, and 1 year is often used as the cut‐off point when this prognostic parameter is analysed in clinical studies 18, 19. This was therefore used in further analysis of our relapsed patients. Both patients with long‐lasting remission clustered in the subset showing low/absent spontaneous and cytokine‐dependent AML cell proliferation (Figs 1 and 2; brown), whereas the other 5 patients with short first remission clustered among patients, showing high spontaneous and/or cytokine‐dependent proliferation (blue or green). This difference in subset distribution is statistically significant (Chi‐square test, P = 0.0082). Finally, one of the patients with 9 months duration first remission was tested both at first time diagnosis and at time of relapse; the first time, he clustered among patients with low proliferative capacity, whereas by time of relapse, he was included in patient subset showing high cytokine‐dependent proliferation.
High cytokine‐dependent proliferation was associated with distinct gene expression signature
To further explore differences between patients with high and low AML cell proliferation, we analysed global gene expression profile for 33 patients in group I; 13 of these had high spontaneous and cytokine‐dependent proliferation. Class comparison analysis using ANOVA between the 13 patients with high proliferation and 20 patients belonging to the other clusters (Fig. 1) showed that 140 genes differed significantly between these two groups with a P‐value <0.005; 65 of these genes were upregulated, whereas 75 were downregulated in high‐proliferation patients (Table S1). We then performed new hierarchical clustering analysis based only on these 140 genes; all high‐proliferation patients then formed a separate cluster without any overlap with other patients (Fig. 3). Thus, the AML patient subset having leukaemic cells with high in vitro proliferative responses could also be identified by analysis of expression profile of a limited number of genes.
Figure 3.
Hierarchal clustering based on differentially expressed genes. Gene expression data were obtained for 33 patients belonging to the group I cohort; 13 of the 33 patients belonged to the high‐ proliferation cluster (indicated by the blue bar in Fig. 1) and ANOVA analysis was used to compare these 13 patients with the other patients. A total of 140 genes were then found to differ significantly between the two groups (i.e. showing P < 0.005); 65 genes then showed higher and 75 genes showed lower expression in the highproliferation patients. These 140 selected genes were used to perform a new hierarchical clustering analysis, Euclidian correlation distance measure with complete linkage. All high‐proliferation patients clustered separately and are identified by the blue bar at the top of the figure and the other patients are indicated by the purple colour bar. Up‐regulated genes are marked with red colour in the heat map and down‐regulated genes with green colour.
Global gene expression profiles of primary AML cells with high and low proliferative capacity – two subsets showing major differences in expression of genes involved in regulation of transcription
Gene ontology (GO) mapping is used to analyse how differences in gene expression profiles affect molecular functions, cell compartments or biological processes. To explore differentially expressed GO‐terms that are important for proliferative capacity of primary human AML cells, we selected those genes from the above study of Group I patients that were differently expressed with a P‐value <0.05 by ANOVA. We then identified 1694 genes that fulfilled these criteria; 849 were upregulated and 845 were downregulated for high‐proliferative cluster compared to other patients. We next used these genes in an overrepresentation analysis to identify those GO‐terms that had >30 genes in the annotation and, in addition, differed significantly between AML patients with high and low proliferative capacity. Fourteen different annotations were detected and these results are present in Table 3. A majority of the differentially expressed genes were downregulated in high‐proliferation patients (Chi‐square test, p<0.0001), and this was true for all 14 annotations. It can be seen that (i) 6 of these GO terms describe functions of DNA/RNA, (ii) two additional terms describe macromolecular metabolic processes, whereas (iii) 5 terms describe cationic binding/function and in particular zinc ion binding. Zinc is important for regulation of gene expression 20, 21, 22, 23. Thus, difference in proliferative capacity between AML patients seems to be determined mainly by differences in regulation of gene expression, whereas differences in expression of growth factors/growth factor receptors/downstream intracellular mediators seem to be less important. The complete list of genes that belong to at last one of these 13 annotations in Table 3 is given in Table S2.
Table 3.
Gene annotations that were significantly altered in patients with high proliferative (autocrine‐ and cytokine‐dependent) capacity compared with patients with low capacity patients (patient group I). The expression of 1694 genes differed significantly (ANOVA, P < 0.05) when comparing patients with high and low proliferative capacity, and an overrepresentation analysis was performed based on these genes. We then identified those GO‐terms that differed significantly (P < 0.05) between the two groups and included at least 30 genes in the annotation; the 14 annotations that fulfilled these criteria are presented in the Table. A majority of the 1695 genes showed a lower expression in patients with high proliferative capacity compared with the low‐capacity patients, and this difference was seen for all 14 annotations (Chi‐squared test P < 0.0001, for all presented data).
| Significantly differing GO term | Total number of genes included in the GO term | Number of downregulated genes | Number of upregulated genes |
|---|---|---|---|
| Nucleic acid binding | 3364 | 142 | 60 |
| Gene expression | 2979 | 131 | 43 |
| Regulation of RNA and metabolic processes | 1542 | 102 | 42 |
| DNA binding | 2369 | 108 | 44 |
| Regulation of RNA biosynthetic processes | 1542 | 80 | 28 |
| Transcription, DNA dependent | 2256 | 115 | 38 |
| Ion binding | 4014 | 149 | 81 |
| Cation binding | 4005 | 148 | 81 |
| Metal ion binding | 1478 | 62 | 35 |
| Transition metal ion binding | 2250 | 103 | 36 |
| Zinc ion binding | 2014 | 95 | 31 |
| Regulation of macromolecule metabolic processes | 2262 | 101 | 49 |
| Macromolecule metabolic processes | 3020 | 134 | 43 |
| Intracellular organelles | 7979 | 286 | 140 |
FLT3‐ITD+ patients with high and low proliferative capacity of their AML cells differed in gene expression profiles and ITD length
For group II, global gene expression profiles were available for 16 consecutive patients with FLT3‐ITD (10 with high and 6 with low proliferative capacity). FLT3‐ITD + patients have a genetic abnormality in the same region of the gene and these mutations are associated with adverse prognosis and distinct gene expression profiles in leukaemic cells 1, 9. However, despite these similarities, these patients are heterogeneous with regard to molecular structure of the ITD 24, 25, prognosis (there are long‐time survival also among these patients) 25, 26 and proliferative capacity (Figs 1 and 2); FLT3‐ITDs cause constitutive activation of this surface‐expressed cytokine receptor 11. Comparison of global gene expression profiles for FLT3‐ITD + patients with high and low proliferative capacity showed that these two patient groups differed (especially in upregulated genes) in their expression of genes important for communication between cells and intracellular signal transduction. This is illustrated by genes presented in Table 4; it can be seen from their GO annotations that a major part of the differentially expressed genes with fold change >2 and ANOVA P‐values <0.004 were important for cytoplasmic signal transduction (for example, the ubiquitin‐proteasome pathway, JAK signalling or signalling downstream to surface receptors). Thus, the FLT3‐ITD + patients differ from other AML patients, in that variations in signal transduction seem more important than variation in gene expression regulation for proliferative capacity for them.
Table 4.
Differentially expressed genes when comparing FLT3‐ITD+ patients (group II) with AML cells showing high (autocrine and cytokine‐dependent) and low proliferative capacity. We identified 17 genes that were differently expressed (at least 2‐fold altered and with P < 0.005); 10 GENES were upregulated in the high‐proliferation group and seven were downregulated compared with the low‐proliferation patients.
| Upregulated genes in patients with high proliferative capacity | |||
|---|---|---|---|
| Gene | Complete name | Fold change | GO‐annotations |
| GPR27 | G protein‐coupled receptor 27 | 7.7 | G‐protein coupled receptor signalling pathway |
| SORT1 | Sortilin 1 | 3.9 | G‐protein coupled receptor signalling pathway |
| HSPA1A | HSP 70 kDa protein 1A | 3.4 | Ubiquitin‐proteasome pathway |
| RAB33A | Ras‐related protein Rab‐33A | 3.4 | Small GTPase‐mediated signal transduction |
| EIF4E3 | EIF4E3 eukaryotic translation initiation factor 4E | 3.2 | Cytokine‐mediated signalling pathway |
| PHACTR1 by similarity | Phosphatase and actin regulator 1 | 2.8 | Phosphatase and actin regulation |
| TTC39C | TTC39C tetratricopeptide repeat domain 39C | 2.7 | Ubiquitin‐proteasome pathway |
| TOR4A | Torsin family 4, member A | 2.2 | Ubiquitin‐proteasome pathway |
| CTNNB1 | Beta‐catenin | 2.2 | Transcription coactivator activity Wnt signalling |
| ENT1 | Equilibrative nucleoside transporter 1 | 2.1 | Nucleobase‐containing compound metabolic process |
| Downregulated genes in patients with high proliferative capacity | |||
|---|---|---|---|
| Gene | Full name | Fold change | GO‐annotations |
| TSC22D1 | TSC22 domain family, member 1 | 4.1 | Regulation of apoptosis and cell proliferation |
| FLJ13197 | Homo sapiens hypothetical protein FLJ13197 | 3.7 | Not characterized |
| JAMIP2 | Janus kinase and microtubule interacting protein 2 | 3.3 | Jak and microtubule interacting protein |
| ADA | Adenosine deaminase | 2.5 | Negative regulation of migration |
| L1TD1 family | Member of the L1 transposable element family | 2.4 | ES cell‐associated and transposase domain‐containing protein |
| ZBED1 by similarity | Ac‐like transposable element | 2.3 | Transposase activity |
| TJP3 | Tight junction protein 3 | 2 | Cell junction |
Association between length of FLT3‐ITD and prognosis in AML has previously been described 25, 27. We therefore investigated the association between size of the ITD and proliferation for group II patients (Fig. 2). Data regarding size of the ITD were available for 17 patients, five with high spontaneous proliferation (blue bars, Fig. 2) and 12 patients with low proliferation rates. FLT3‐ITD patients with high spontaneous proliferation had significantly longer ITDs (mean 100 bp, range 66–153 bp) compared to ITD patients with low or undetectable spontaneous proliferation (mean 51 bp, range 24–90) (P = 0.0392, Mann–Whitney U‐test). FLT3‐ITD: FLT3‐wt ratio did not differ between the two groups (data not shown).
Discussion
In vitro proliferative capacity of primary human AML cells has wide variation between patients 3, 7. In our present study, we used a highly standardized experimental model 28, 29 to analyse proliferative capacity of non‐APL AML cells for two independent and unselected/consecutive patient groups. Based on ability of their AML cells to undergo spontaneous or autocrine proliferation and responsiveness of leukaemic cells to a limited number of haematopoietic growth factors, patients could be classified into three major subsets: (i) cells with both autocrine proliferation and responsiveness to a wide range of growth factors; (ii) no autocrine proliferation, but generally strong proliferative responses to exogenous growth factors; and (iii) no autocrine proliferation and no or only weak responsiveness to all or most growth factors.
Proliferative capacity of AML cells showed no (for example, cytogenetic abnormalities, FLT3 mutations) or only weak (CD34 expression) associations with other clinical or biological patient characteristics, the only exception being that favourable prognostic impact of NPM‐1 mutations 30 could not be detected in patients with highest proliferative capacity as NPM‐1 mutations then were either not present or combined with dominant adverse prognostic parameters. This may explain why previous clinical studies have shown association between adverse prognosis and autocrine in vitro proliferation of AML cells 3, 4.
Our study included 7 patients with AML relapse, and AML cells at the time of relapse also differed in their proliferative capacity (Figs 1 and 2). The possibility of achieving second haematological remission is higher if patients have had long duration of their first remission, and 1 year is often used as a cut‐off point, when this prognostic parameter is analysed in clinical studies 18, 19. Here, patients with short (<1 year) duration of first remission, differed significantly from patients with long‐lasting remission with regard to clustering, based on proliferative capacity; short duration was then associated with high proliferative capacity. Proliferative capacity is thus associated with prognosis also for patients with AML relapse. However, even though this difference reached statistical significance, data should be interpreted with great care as number of patients was small and verification in a larger study is definitely needed.
Patients with cells of different proliferative capacity also differed in their global gene expression profiles, and patient subset with high proliferative capacity could then be identified based on expression of a limited number of genes. This gene expression profile is more suitable for identification of this patient subset than analysis of in vitro proliferation, and such gene expression analyses can also be easily included in future clinical studies. Even though our subclassification is based on proliferative responsiveness to autocrine cytokine release or addition of exogenous cytokines, global gene expression analysis of unselected patients in patient group I did not show any major differences between patients from different clusters with regard to expression of cytokines, cytokine receptors or corresponding downstream mediators for cytoplasmic signal transduction. Rather, patients with high proliferative capacity differed with regard to expression of genes encoding regulators of DNA transcription/RNA synthesis. This is probably the explanation why patients either respond to a wide range of growth factors or show general low proliferative responsiveness; proliferative capacity is not determined by differences in expression of individual growth factor receptors or their downstream mediators, but rather by differences in common regulatory mechanisms that integrate signalling from various cytokine‐activated pathways at DNA/RNA level.
We performed a more detailed analysis of proliferative capacity of FLT3‐ITD+ AML cells as (i) these patients have similar genetic abnormality of this cell surface receptor 11, 30 although the prognostic impact of ITDs differs between patients and seems to depend on the ITD structure 25; and (ii) FLT3‐L activates receptor signalling and functions as an AML growth factor, whereas FLT3‐ITDs cause constitutive activation of the receptor 11. For these reasons, we analysed FLT3‐ITD:FLT3‐wt ratio, ITD length and global gene expression profiles for FLT3‐ITD+ patients in one of our cohorts (group II, see Fig. 2). Thus, FLT3‐ITDs are important for chemosensitivity and have adverse prognostic impact for patients receiving intensive chemotherapy 26, 27. When comparing FLT3‐ITD+ patients with high and low proliferative capacity, the two groups differed with regard to both ITD length and expression of molecules involved in intracellular signal transduction. Our hypothesis is that additional mechanisms at signal transduction levels (for example, the ubiquitin‐proteasome pathway) are important for regulation of proliferative capacity in a minority of FLT3‐ITD+ AML patients in addition to the general difference in regulation of DNA transcription/RNA synthesis (Table 3).
To conclude, AML patients can be subclassified based on proliferative capacity of their leukaemic cells. For most patients, this classification reflects differences in regulation of cell proliferation at the DNA/RNA level, but a minority of FLT3‐ITD+ patients represents an exception where additional differences at the signal transduction level are also important (for example, proteosomal function). Proliferation is important for disease progression, and our observations suggest that described differences between patients may translate into differences in susceptibility to new targeted therapies, such as proteasome inhibition, that may be most effective only for certain subsets of FLT3‐ITD+ patients, whereas targeting gene expression through modulation of epigenetic regulation may be most effective in other patients.
Supporting information
Table S1. List of genes that are differentially expressed (p<0.005) when comparing patients with high and low proliferative capacity (Group I patient cohort).
Table S2. List of genes that are differentially expressed (p<0.05) when comparing patients with high and low proliferative capacity (Group I patient cohort) and each of these genes belonging to at least one of the G0‐annotation listed in Table 3.
Acknowledgments
The study received financial support from the Norwegian Cancer Society. Bioinformatic analysis was performed as cooperation project with the Norwegian Bioinformatics platform, funded by the FUGE programme, the Norwegian Research Council.
Conflicts of interest
The authors report no potential conflict of interest.
References
- 1. Estey EH (2012) Acute myeloid leukemia: 2012 update on diagnosis, risk stratification, and management. Am. J. Hematol. 87, 89–99. [DOI] [PubMed] [Google Scholar]
- 2. Welch JS, Ley TJ, Link DC, Miller CA, Larson DE, Koboldt DC et al (2012) The origin and evolution of mutations in acute myeloid leukemia. Cell 150, 264–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Lowenberg B, van Putten WL, Touw IP, Delwel R, Santini V (1993) Autonomous proliferation of leukemic cells in vitro as a determinant of prognosis in adult acute myeloid leukemia. N. Engl. J. Med. 328, 614–619. [DOI] [PubMed] [Google Scholar]
- 4. Hunter AE, Rogers SY, Roberts IA, Barrett AJ, Russell N (1993) Autonomous growth of blast cells is associated with reduced survival in acute myeloblastic leukemia. Blood 82, 899–903. [PubMed] [Google Scholar]
- 5. Bruserud O, Frostad S, Foss B (1999) In vitro culture of acute myelogenous leukemia blasts: a comparison of four different culture media. J. Hematother. 8, 63–73. [DOI] [PubMed] [Google Scholar]
- 6. Bruserud O, Gjertsen BT, Foss B, Huang TS (2001) New strategies in the treatment of acute myelogenous leukemia (AML): in vitro culture of aml cells–the present use in experimental studies and the possible importance for future therapeutic approaches. Stem Cells 19, 1–11. [DOI] [PubMed] [Google Scholar]
- 7. Bruserud O, Gjertsen BT (2000) New strategies for the treatment of acute myelogenous leukemia: differentiation induction–present use and future possibilities. Stem Cells 18, 157–165. [DOI] [PubMed] [Google Scholar]
- 8. Van Etten RA (2007) Aberrant cytokine signaling in leukemia. Oncogene 26, 6738–6749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Valk PJ, Verhaak RG, Beijen MA, Erpelinck CA, Barjesteh van Waalwijk van Doorn‐Khosrovani S, Boer JM et al (2004) Prognostically useful gene‐expression profiles in acute myeloid leukemia. N. Engl. J. Med. 350, 1617–1628. [DOI] [PubMed] [Google Scholar]
- 10. Bullinger L, Dohner K, Bair E, Frohling S, Schlenk RF, Tibshirani R et al (2004) Use of gene‐expression profiling to identify prognostic subclasses in adult acute myeloid leukemia. N. Engl. J. Med. 350, 1605–1616. [DOI] [PubMed] [Google Scholar]
- 11. Bruserud O, Hovland R, Wergeland L, Huang TS, Gjertsen BT (2003) Flt3‐mediated signaling in human acute myelogenous leukemia (AML) blasts: a functional characterization of Flt3‐ligand effects in AML cell populations with and without genetic Flt3 abnormalities. Haematologica 88, 416–428. [PubMed] [Google Scholar]
- 12. Bruserud O, Ryningen A, Wergeland L, Glenjen NI, Gjertsen BT (2004) Osteoblasts increase proliferation and release of pro‐angiogenic interleukin 8 by native human acute myelogenous leukemia blasts. Haematologica 89, 391–402. [PubMed] [Google Scholar]
- 13. Stapnes C, Doskeland AP, Hatfield K, Ersvaer E, Ryningen A, Lorens JB et al (2007) The proteasome inhibitors bortezomib and PR‐171 have antiproliferative and proapoptotic effects on primary human acute myeloid leukaemia cells. Br. J. Haematol. 136, 814–828. [DOI] [PubMed] [Google Scholar]
- 14. Reikvam H, Hatfield KJ, Ersvaer E, Hovland R, Skavland J, Gjertsen BT et al (2012) Expression profile of heat shock proteins in acute myeloid leukaemia patients reveals a distinct signature strongly associated with FLT3 mutation status ‐ consequences and potentials for pharmacological intervention. Br. J. Haematol. 156, 468–480. [DOI] [PubMed] [Google Scholar]
- 15. Oyan AM, Anensen N, Bo TH, Stordrange L, Jonassen I, Bruserud O et al (2009) Genes of cell‐cell interactions, chemotherapy detoxification and apoptosis are induced during chemotherapy of acute myeloid leukemia. BMC Cancer 9, 77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Stavrum AK, Petersen K, Jonassen I, Dysvik B (2008) Analysis of gene‐expression data using J‐Express. Curr. Protoc. Bioinformatics 7, 7 3. [DOI] [PubMed] [Google Scholar]
- 17. Dysvik B, Jonassen I (2001) J‐Express: exploring gene expression data using Java. Bioinformatics 17, 369–370. [DOI] [PubMed] [Google Scholar]
- 18. Lazarus HM, Perez WS, Klein JP, Kollman C, Bate‐Boyle B, Bredeson CN et al (2006) Autotransplantation versus HLA‐matched unrelated donor transplantation for acute myeloid leukaemia: a retrospective analysis from the Center for International Blood and Marrow Transplant Research. Br. J. Haematol. 132, 755–769. [DOI] [PubMed] [Google Scholar]
- 19. Ferrara F, Palmieri S, Mele G (2004) Prognostic factors and therapeutic options for relapsed or refractory acute myeloid leukemia. Haematologica 89, 998–1008. [PubMed] [Google Scholar]
- 20. Yang XJ, Ullah M (2007) MOZ and MORF, two large MYSTic HATs in normal and cancer stem cells. Oncogene 26, 5408–5419. [DOI] [PubMed] [Google Scholar]
- 21. Morrison AA, Viney RL, Ladomery MR (2008) The post‐transcriptional roles of WT1, a multifunctional zinc‐finger protein. Biochim. Biophys. Acta 1785, 55–62. [DOI] [PubMed] [Google Scholar]
- 22. Katsumoto T, Yoshida N, Kitabayashi I (2008) Roles of the histone acetyltransferase monocytic leukemia zinc finger protein in normal and malignant hematopoiesis. Cancer Sci. 99, 1523–1527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Razin SV, Borunova VV, Maksimenko OG, Kantidze OL (2012) Cys2His2 zinc finger protein family: classification, functions, and major members. Biochemistry 77, 217–226. [DOI] [PubMed] [Google Scholar]
- 24. Hatfield KJ, Hovland R, Oyan AM, Kalland KH, Ryningen A, Gjertsen BT et al (2008) Release of angiopoietin‐1 by primary human acute myelogenous leukemia cells is associated with mutations of nucleophosmin, increased by bone marrow stromal cells and possibly antagonized by high systemic angiopoietin‐2 levels. Leukemia 22, 287–293. [DOI] [PubMed] [Google Scholar]
- 25. Kayser S, Schlenk RF, Londono MC, Breitenbuecher F, Wittke K, Du J et al (2009) Insertion of FLT3 internal tandem duplication in the tyrosine kinase domain‐1 is associated with resistance to chemotherapy and inferior outcome. Blood 114, 2386–2392. [DOI] [PubMed] [Google Scholar]
- 26. Schlenk RF, Dohner K, Krauter J, Frohling S, Corbacioglu A, Bullinger L et al (2008) Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. N. Engl. J. Med. 358, 1909–1918. [DOI] [PubMed] [Google Scholar]
- 27. Stirewalt DL, Kopecky KJ, Meshinchi S, Engel JH, Pogosova‐Agadjanyan EL, Linsley J et al (2006) Size of FLT3 internal tandem duplication has prognostic significance in patients with acute myeloid leukemia. Blood 107, 3724–3726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Bruserud O, Gjertsen BT, von Volkman HL (2000) In vitro culture of human acute myelogenous leukemia (AML) cells in serum‐free media: studies of native AML blasts and AML cell lines. J. Hematother. Stem Cell Res. 9, 923–932. [DOI] [PubMed] [Google Scholar]
- 29. Bruserud O, Ulvestad E (1999) Effects of gamma‐irradiation on acute myelogenous leukemia blasts: in vitro studies of proliferation, constitutive cytokine secretion, and accessory cell function during T cell activation. J. Hematother. Stem Cell Res. 8, 431–441. [DOI] [PubMed] [Google Scholar]
- 30. Döhner H, Estey EH, Amadori S, Appelbaum FR, Buchner T, Burnett AK et al (2010) Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood 115, 453–474. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. List of genes that are differentially expressed (p<0.005) when comparing patients with high and low proliferative capacity (Group I patient cohort).
Table S2. List of genes that are differentially expressed (p<0.05) when comparing patients with high and low proliferative capacity (Group I patient cohort) and each of these genes belonging to at least one of the G0‐annotation listed in Table 3.