If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Medicine and Biosystemic Science, Graduate School of Medical Sciences, Kyushu University Graduate School of Medicine, Fukuoka, JapanCenter for Cellular and Molecular Medicine, Graduate School of Medical Sciences, Kyushu University Graduate School of Medicine, Fukuoka, Japan
Offprint requests to: Prof. Koichi Akashi, Kyushu University Graduate School of Medicine, Department of Medicine and Biosystemic Sciences, 3-1-1 Maidashi, Higashi-Ku, Fukuoka City, Fukuoka 812-0054, Japan
Department of Medicine and Biosystemic Science, Graduate School of Medical Sciences, Kyushu University Graduate School of Medicine, Fukuoka, JapanCenter for Cellular and Molecular Medicine, Graduate School of Medical Sciences, Kyushu University Graduate School of Medicine, Fukuoka, Japan
The cellular properties of leukemia stem cells (LSCs) are achieved at least through Class I and Class II mutations that generate signals for enhanced proliferation and impaired differentiation, respectively. Here we show that in t(8;21) acute myelogenous leukemia (AML), hematopoietic stem cells (HSCs) transform into LSCs via definitively-ordered acquisition of Class II (AML1/ETO) and then Class I (c-KIT mutant) abnormalities. Six t(8;21) AML patients with c-KIT mutants maintaining > 3 years of complete remission were analyzed. At diagnosis, all single LSCs had both AML1/ETO and c-KIT mutations. However, in remission, 16 out of 1,728 CD34+CD38− HSCs and 89 out of 7,187 single HSC-derived myeloerythroid colonies from these patients had AML1/ETO, whose breakpoints were identical to those found in LSCs. These cells had wild-type c-KIT, which expressed AML1/ETO at a low level, and could differentiate into mature blood cells, suggesting that they may be the persistent preleukemic stem cells. Microarray analysis suggested that mutated c-KIT signaling provides LSCs with enhanced survival and proliferation. Thus, in t(8;21) AML, the acquisition of AML1/ETO is not sufficient, and the subsequent upregulation of AML1/ETO and the additional c-KIT mutant signaling are critical steps for transformation into LSCs.
Acute myelogenous leukemia (AML) is characterized by deregulated proliferation and impaired differentiation of immature hematopoietic cells and originates from leukemia stem cells (LSCs). Leukemia stem cells have cellular properties, such as self-renewal activity, impairment of full maturation, and reinforced survival, which may cooperatively play a role in advantageous growth compared with normal hematopoietic stem cells (HSCs). Such cellular properties of LSCs result from multiple genetic abnormalities that are presumably accumulated within the long-surviving, self-renewing HSCs [
]. Recent mouse studies have suggested that these genetic abnormalities could be categorized into at least two classes. Class I mutations confer a proliferative and/or survival advantage against hematopoietic progenitors and are exemplified by constitutively activated tyrosine kinases such as BCR-ABL, FLT3 internal tandem duplication (FLT3-ITD), and mutated c-KIT. On the other hand, Class II mutations impair hematopoietic differentiation, which includes the core binding factor (CBF) mutations such as AML1/ETO [
]. Several mouse models have demonstrated that the combined effects of enhanced proliferation (by Class I abnormalities) and differentiation block (by Class II abnormalities) result in AML development [
] because AML1/ETO inhibits CBF complexes that can transactivate multiple myeloid-related genes (e.g., CEBPA, MPO, and IL3), in a dominant negative fashion. Frequently, t(8;21) AML patients possess constitutively active Class I mutation of c-KIT and FLT3 [
Mutations in the receptor tyrosine kinase pathway are associated with clinical outcome in patients with acute myeloblastic leukemia harboring t(8;21)(q22;q22).
]. These data strongly suggest that the acquisition of AML1/ETO fusion alone is not sufficient, and some additional oncogenic events are needed for the development of t(8;21) AML. However, these studies were based on mouse models, where the expression of Class I and Class II genes was artificially enforced. The critical questions are whether these multistep oncogenic events involve human AML and, if so, whether they occur at random or in a definitive order.
Our previous t(8;21) AML patient studies have proven that t(8;21) is acquired in long-term HSCs but it is not sufficient for AML development [
]. We found that t(8;21) AML patients maintaining remission long term (>10 years) always possessed a small amount of AML1/ETO mRNA in their blood and bone marrow cells. These AML1/ETO+ cells may be derived from HSCs having a low level of AML1/ETO mRNA, the frequency of which was estimated to be approximately 1% of HSCs [
]. Thus, the acquisition of AML1/ETO is not sufficient for leukemic transformation in humans. Therefore, we proposed that such AML1/ETO+ HSCs are preleukemic clones that have achieved a precondition for leukemic transformation by additional oncogenic hits [
]. Wiemels et al. have reported that a fraction of t(8;21) AML children had AML1/ETO+ clones in their blood samples from neonatal Guthrie blood spots [
], suggesting that t(8;21) translocation can be achieved in utero, and resultant AML1/ETO+ HSCs can form a reservoir for the preleukemic clone after birth [
Based on these data, we sequentially tracked the involvement of Class I and Class II mutations during clinical courses of t(8;21) AML patients. Here we show that, by single HSC and LSC analyses of AML1/ETO in patients with mutated c-KIT, all single AML1/ETO+ LSCs at diagnosis had c-KIT mutations, whereas they were never found within AML1/ETO+ HSCs in remission. Our data clearly show that AML1/ETO+ HSCs should belong to the preleukemic clone and are transformed into LSCs by subsequent acquisition of c-KIT mutation. This is, to our knowledge, the first clear-cut evidence that normal HSCs transform into LSCs via definitively-ordered acquisition of Class II and then Class I mutations in de novo human AML.
Methods
Patients and samples
Patients' characteristics are shown in Table 1. This study included bone marrow cells from 33 t(8;21) AML cases at diagnosis (Patients 1–33), 13 cases in remission (Patients 1, 3, 7–9, 11, 13, 21–23, 26–27, and 31), and 13 cases at relapse (Patients 2, 5–6, 10, 14, 16–17, 25, 28–30, and 32–33). Remission marrow samples were obtained at least 12 months from first remission, and all the patients remained in remission at the time of this report. Out of 33 cases, 20 obtained complete remission only by chemotherapies. Patients 2 and 10 further received allogeneic bone marrow transplantation, and Patient 5 further received cord blood transplantation. On AML cells, CD19 and CD56 are known as the prognostic markers associated with the possession of c-KIT mutation [
Mutations in the receptor tyrosine kinase pathway are associated with clinical outcome in patients with acute myeloblastic leukemia harboring t(8;21)(q22;q22).
Expression of the neural cell adhesion molecule CD56 is associated with short remission duration and survival in acute myeloid leukemia with t(8;21)(q22;q22).
]. The AML cells of all 13 cases with c-KIT mutation at diagnosis were CD19−CD56+. Human marrow was purchased from AllCells (Emeryville, CA). Informed consent was obtained from all patients. The Institutional Review Board of Kyushu University Hospital (Fukuoka, Japan) approved all research.
]. Cells were stained with APC-anti-CD34, FITC-anti-CD90, PE-anti-CD117 (c-KIT), Cy5-PE-lineage (Lin) mixture (anti-CD3, -CD4, -CD8, -CD10, -CD20, -CD256) (BD Pharmingen, San Jose, CA), and biotin-anti-CD38 (Caltag Laboratories, Buckingham, UK). Streptavidin-Cy7-allophycocyanin (BD Pharmingen) was also used.
Quantitative real-time polymerase chain reaction
We isolated RNA from 5,000 cells using Isogen reagent (Nippon gene). We reverse transcribed RNA to cDNA using TaKaRa RNA polymerase chain reaction (PCR) kit (Takara Shuzo, Shiga, Japan). The mRNA levels were quantified by real-time PCR (Applied Biosystems, Carlsbad, CA). β2-microglobulin (B2MG) was used for internal control. The primer and probes for B2MG, c-KIT, C-X-X chemokine receptor type 4 (CXCR4), B-cell lymphoma 2 (BCL2), myeloid cell leukemia 1 (MCL1) and nuclear factor kappa B1 (NFKB1) were purchased from Applied Biosystems.
Reverse transcription polymerase chain reaction
To examine the AML1/ETO and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA expression, reverse transcription-PCR (RT-PCR) was performed as previously reported [
]. All of the myeloid colonies were picked up and separated to extract the RNA and genomic DNA.
Identification of the gene mutations
Genomic DNA was extracted by Micro Kit (QIAGEN, Hilden, Germany). The presence of FLT3-ITD, NRAS, and c-KIT mutations was examined as previously descried [
Modeling interactions between leukemia-specific chromosomal changes, somatic mutations, and gene expression patterns during progression of core-binding factor leukemias.
]. The primers for the c-KIT mutation are shown in Supplementary Table E1 (online only, available at www.exphem.org). The clonal PCR product was purified by QIAquick Spin (QIAGEN) and directly sequenced by ABI 3730 Genetic analyzer (Applied Biosystems).
Identification of breakpoint of AML1/ETO of genomic DNA
Patients' breakpoints were determined by sequencing the PCR products of long-distance inverse PCR (LDI-PCR) and conventional long distance PCR (LD-PCR), as previously described [
]. We used AML1/ETO external primers (Supplementary Table E3, online only, available at www.exphem.org). Single-cell genomic PCR was performed by nested PCR utilizing external and internal primers (Supplementary Table E1, online only, available at www.exphem.org). The method of nested PCR for genomic DNA was the same as the RT-PCR method. For single-cell quantitative nested PCR, we performed first round of RT-PCR with external primers for these diluted, pre-amplified cDNA. The protocol of the first round of PCR was same as that for RT-PCR (thermal cycling setting was 16). Nested PCR was performed by using a BioMark 48 × 48 Dynamic Array system with internal primers (Supplementary Table E1, online only, available at www.exphem.org).
Microarray analysis
Eighteen wild-type c-KIT LSC and 13 mutated c-KIT LSC samples were investigated with Sentrix Bead Chip Assay, Human-6 V2 (Illumina, San Diego, CA) as previously reported [
]. Microarray data were analyzed with Gene Spring GX11.01 (Agilent Technologies, Santa Clara, CA).
Cytokine stimulation assays
The c-KIT signaling repercussion for AML1/ETO expression level by addition of stem cell factor was evaluated after 24-hour serum-free liquid culture. The details were previously described [
c-KIT mutation was found in approximately 40% of patients with t(8;21) acute myelogenous leukemia
Thirty-three t(8;21) AML patients were enrolled in this study. Previous studies have shown that Class I abnormalities, such as c-KIT, NRAS, and FLT3 mutations, are frequently found in t(8;21) AML [
Mutations in the receptor tyrosine kinase pathway are associated with clinical outcome in patients with acute myeloblastic leukemia harboring t(8;21)(q22;q22).
]. As shown in Table 1, 13 out of 33 t(8;21) AML patients had c-KIT mutation, one patient had FLT3-ITD, and no NRAS mutations were observed. In all cases, involvement of Class I mutation was heterozygous.
Of the patients with c-KIT mutations, six had D816V mutation, five had N822K, and two had D816Y (Table 1). The expression levels of c-KIT mRNA and protein in the CD34+CD38− LSC fraction of t(8;21) AML were equal in all cases, regardless of the involvement of c-KIT mutations, and their levels were identical to those of normal CD34+CD38− HSCs (Supplementary Figure E1, online only, available at www.exphem.org). Out of 13 patients with c-KIT mutations, seven patients relapsed. Six out of seven relapsed patients had mutations identical to those found at diagnosis, whereas Patient 14 acquired an independent de novo c-KIT mutation at relapse (N822K at diagnosis and D816Y at relapse) (Table 1). This intriguing case suggests that the acquisition of c-KIT mutation is the second event that is independent of t(8;21). These data led us to test whether c-KIT mutation is involved in preleukemic AML1/ETO+ HSCs in remission [
All single leukemia stem cells possess both AML1/ETO and c-KIT mutations at diagnosis
Six cases of t(8;21) AML with c-KIT mutations, including D816V, N822K, and D816Y, who maintained complete remission for > 3 years (Patients 3, 7, 8, 11, 21, and 23) were investigated to track AML1/ETO and c-KIT status at both diagnosis and during remission.
We first tested the presence of AML1/ETO mRNA and c-KIT mutation in LSCs at diagnosis at the single cell level. As shown in Figure 1A, genomic DNA and mRNA were extracted from single CD34+CD38− leukemic marrow cells and were subjected to PCR to test for the presence of AML1/ETO mRNA. The c-KIT gene was amplified from single cell–derived genomic DNA and analyzed by direct sequencing to identify c-KIT mutations.
Figure 1A fraction of single preleukemic HSCs in remission expressed a low level of AML1/ETO, whose breakpoints were identical to LSCs at diagnosis. (A) The experimental method of single cell analysis. Genomic DNA and mRNA were extracted from fluorescence activated cell sorting-purified single CD34+CD38− cells. These extracted genomic DNA and mRNA were preamplified and were then analyzed by nested PCR and direct sequence to detect c-KIT mutations, as well as by single-cell quantitative PCR for AML1/ETO transcripts. (B) The representative single-cell quantitative PCR analysis at diagnosis and in remission (Patient 8). Each lane represents the level of AML1/ETO mRNA in single cells. Almost all single CD34+CD38− cells in the bone marrow that were detectable at the first round of PCR at diagnosis expressed AML1/ETO at a high level. In contrast, a small fraction of single CD34+CD38− cells that were detectable only by the second round of PCR in remission expressed AML1/ETO at a very low level. The existence of sorted single cells was confirmed by successful detection of B2MG mRNA. (C) Detection of the breakpoint of the AML1/ETO fusion gene specific to each patient. All single AML1/ETO+ cells at remission had breakpoints identical to those at diagnosis in all three patients tested. Representative data are shown.
Figure 1B shows representative results of AML1/ETO mRNA analysis of single LSCs (Patient 8). At diagnosis, nearly all LSCs expressed AML1/ETO mRNA at a high level, whereas, in remission, only a few percent of HSCs expressed AML1/ETO mRNA, whose levels were so low they were only detectable after the second round of PCR (Fig. 1B).
Summarized data are shown in Table 2. In the analysis of six cases at diagnosis, 1,608 (98.9%) out of 1,626 single LSCs that were analyzed had AML1/ETO mRNA, and c-KIT mutations specific to each patient were observed in all 1,608 AML1/ETO mRNA+ cells. In contrast, the remaining 18 CD34+CD38− cells that did not express AML1/ETO mRNA had the wild-type c-KIT, indicating AML1/ETO mRNA and c-KIT mutation always coexist at diagnosis in all single LSCs.
Table 2Summary of detection of AML1/ETO mRNA+ cells in single CD34+CD38− cells in diagnostic and remission marrow
All single preleukemic AML1/ETO+ hematopoietic stem cells in remission lacked c-KIT mutation
We then tested whether AML1/ETO+ HSCs in remission had c-KIT mutation. In each patient maintaining remission for >3 years, single CD34+CD38− HSCs were sorted from the bone marrow and were subjected to PCR to evaluate the presence of AML1/ETO mRNA and c-KIT mutations, as shown in Figure 1A. Summarized data are shown in Table 2.
In the six patients analyzed at remission, AML1/ETO mRNA was detected in 16 (0.9%) out of 1,728 single cells of CD34+CD38− HSC fraction in the remission marrow. The frequency of AML1/ETO+ HSCs in remission is consistent with previous estimation based on limit-dilution analysis [
]. All 16 AML1/ETO mRNA+ CD34+CD38− cells had wild-type c-KIT. Furthermore, no c-KIT mutations were found in the remaining 1,712 CD34+CD38− cells, which did not have AML1/ETO mRNA, suggesting that c-KIT mutations never precede the acquisition of t(8;21).
To confirm that AML1/ETO+ mutant c-KIT+ LSCs at diagnosis and AML1/ETO+ HSCs in remission belong to a common clone, we tested whether their AML1/ETO breakpoints were identical. We amplified specific breakpoints of the AML1/ETO fusion gene using a long PCR method [
] in three of these patients (Patient 8, 11, and 23) and prepared PCR primers to detect the breakpoint of the AML1/ETO fusion gene specific to each case. As shown in Figure 1C, in all of these three patients, single AML1/ETO+ cells at remission always had breakpoints identical to those at diagnosis, indicating that AML1/ETO+ HSCs in remission and the original AML LSCs share their origin. Collectively, these results strongly suggested that acquisition of c-KIT mutations in pre-leukemic AML1/ETO+ HSCs may be a critical event for the transformation of t(8;21) preleukemic HSCs into LSCs.
Preleukemic AML1/ETO+ hematopoietic stem cells without c-KIT mutation can differentiate into myeloerythroid cells in vitro
The main leukemogenic function of AML1/ETO may be to block differentiation by abrogating the CBF function through dominant-negative inhibition of AML1 [
]. However, because the expression of AML1/ETO is very low in remission (Fig. 1B), such a low level of AML1/ETO may not be able to inhibit differentiation of AML1/ETO+ HSCs. In fact, AML1/ETO+ mRNA is detectable in a small fraction of mature granulocytes and lymphoid cells in remission [
]. Thus, we wished to confirm that AML1/ETO+ HSCs with the wild-type c-KIT in remission differentiate into mature blood cells. Single CD34+CD38− HSCs purified from remission marrow were cultured in methylcellulose, and each colony was picked up (Fig. 2A) and tested for the presence of AML1/ETO and c-KIT mutations (Fig. 2B). As summarized in Table 3, of 7,187 total myeloid colonies from six patients, 89 (1.2%) were positive for AML1/ETO mRNA, and all of these colonies had wild-type c-KIT. These data confirm that AML1/ETO+ HSCs in remission with wild-type c-KIT are capable of differentiating into a variety of myeloerythroid cells and contribute toward maintaining normal hematopoiesis.
Figure 2Single preleukemic HSCs in remission do not have c-KIT mutation and can differentiate into mature myeloid cells. (A) Morphology of AML1/ETO-positive myeloid colonies derived from single CD34+CD38− cells in remission. The representative results of Patient 11 are shown. (B) PCR analyses for AML1/ETO mRNA and c-KIT genes of cells picked from single HSC-derived myeloid colonies. AML1/ETO mRNA can be detected in a fraction of myeloid colonies only by nested PCR. Simultaneously, genomic DNA from these colonies was subjected to PCR amplification for the c-KIT gene to evaluate the presence of c-KIT mutation by direct sequencing. Results were summarized in Table 3.
Mutated c-KIT signaling endows leukemic stem cells with growth advantages through upregulation of several key molecules
In these patients, c-KIT mutations have been shown to constitutively provide active c-KIT signaling and may therefore contribute to proliferation and survival of leukemic cells [
]. To understand the function of mutant c-KIT signaling, we compared the gene expression profile of the CD34+CD38− LSC fraction purified from 18 t(8;21) patients with wild-type c-KIT with that of 13 patients with c-KIT mutants using microarray analysis. As shown in Figure 3A, the clustering analysis showed t(8;21) AML LSCs with c-KIT mutation had a distinct expression pattern, regardless of their type of c-KIT mutation. Genes upregulated or downregulated by > twofold in patients with mutant c-KIT are listed in Supplementary Table E4 (online only, available at www.exphem.org). For example, MCL1, BCL2, NFKB1A and CXCR4 were significantly upregulated in AML LSCs with c-KIT mutations (Fig. 3B). These data are consistent with those in previous reports, in which the c-KIT signaling effectively upregulated these genes to enhance their LSC activity [
], was upregulated in LSCs with c-KIT mutations (Fig. 3B and C). This may be reasonable because FLT3-ITD, which is another mutation of the receptor-type tyrosine kinase, is known to upregulate MCL1 to promote AML LSC survival [
]. These data collectively suggest that the acquisition of c-KIT mutation may at least contribute to reinforce proliferation and survival of t(8;21) AML LSCs.
Figure 3The expression of molecules that enforce the survival of LSCs. (A) Results of microarray analysis of t(8;21) LSCs with wild-type c-KIT and mutated c-KIT. t(8;21) AML LSCs with c-KIT mutation had a distinct expression pattern, irrespective of their types of c-KIT mutation. (B) Representative molecules that upregulated greater than twofold in AML LSCs with mutated c-KIT, as compared with those with the wild-type c-KIT. (C) The quantitative PCR analysis of MCL-1 in N, WT and M cell types. M = LSCs with c-KIT mutation; N = CD34+CD38− normal HSCs; WT = LSCs with wild-type c-KIT.
Upregulation of AML1/ETO may also constitute a critical step for transformation into leukemia stem cells
In LSCs in remission, AML1/ETO transcripts become detectable only after the second round of PCR, whereas they are easily detected in LSCs at diagnosis by single PCR (Fig. 1B). This suggests that the increase in AML1/ETO expression may also be important in LSC development. Therefore, we quantified AML1/ETO transcripts in CD34+CD38− cells at diagnosis and remission using a single cell quantitative PCR method.
Figure 4A shows the amount of AML1/ETO transcripts in 16 single AML1/ETO+ CD34+CD38− cells in remission relative to those in single AML1/ETO+ LSCs at diagnosis in the six patients listed in Table 1. In every case, regardless of the c-KIT mutant type, the amount of AML1/ETO transcripts per cell in remission was more than one hundredfold less than that in LSCs at diagnosis. Taken together, LSCs at diagnosis had approximately five hundred times more AML1/ETO transcripts compared with AML1/ETO+ HSCs in remission at the single cell level. We could not conduct a similar analysis for AML1/ETO+c-KIT wild-type patients because of the lack of sufficient samples.
Figure 4The level of AML1/ETO transcripts in single CD34+CD38− preleukemic HSCs and LSCs at diagnosis. (A) Results of quantitative PCR analysis of AML1/ETO mRNA in single preleukemic HSCs and LSCs at diagnosis. In all cases, LSCs expressed over one hundredfold higher levels of AML1/ETO than preleukemic HSCs in remission, irrespective of their c-KIT mutant types. (B) AML1/ETO mRNA expression in human t(8;21) LSCs with wild type c-KIT and those with mutated c-KIT in the presence or absence of SCF. The amounts of AML1/ETO transcripts were not affected by c-KIT signaling.
We hypothesized that, in AML1/ETO+ LSCs with c-KIT mutations, constitutively active c-KIT signaling may stimulate the expression of AML1/ETO transcripts. Therefore, we quantified the levels of AML1/ETO transcripts in 5,000 cells of CD34+CD38− LSCs from 10 patients with wild-type c-KIT and in 5,000 cells from patients with c-KIT mutations. However, as shown in Figure 4B, the AML1/ETO transcript level was not significantly different, regardless of the presence of c-KIT mutation. Furthermore, the ligation of c-KIT by addition of stem cell factor (SCF) in culture did not affect the AML1/ETO levels in each group (Fig. 4B). Thus, c-KIT signaling may not stimulate AML1/ETO transcription, suggesting that the acquisition of c-KIT mutation and the upregulation of AML1/ETO transcription are independent events.
Discussion
It has been suggested that genetic abnormalities are accumulated in self-renewing, long-surviving HSCs and that these abnormalities cooperatively transform normal HSCs into LSCs [
]. Our intensive analysis of human t(8;21) AML with c-KIT mutation revealed that, at diagnosis, all single LSCs had both AML1/ETO and c-KIT mutants (Table 2). During remission, a small fraction (∼1%) of HSCs or their myeloerythroid colonies had AML1/ETO, and such single AML1/ETO+ HSCs always had wild-type c-KIT. Hematopoietic stem cells with only the c-KIT mutation were never observed (Table 2, Table 3).
Furthermore, the breakpoint of AML1/ETO was identical at diagnosis and in remission in all three patients analyzed (Fig. 1C), indicating that AML1/ETO+ cells at diagnosis and those in remission originated from the same preleukemic clone. In addition, we found a patient who had N822K at diagnosis but newly obtained D816Y at relapse (Patient 14, Table 1), suggesting that acquisition of c-KIT mutation is a subsequent event. Thus, our sequential analysis provides definitive evidence that HSCs first acquire AML/ETO fusion (Class II) and then c-KIT mutation (Class I) for transformation into LSCs. The proposed developmental model for t(8;21) AML is schematized in Figure 5.
Figure 5The proposed multistep developmental scheme of human t(8;21) AML with c-KIT mutations. A normal HSC acquires t(8;21) and forms a reservoir of preleukemic AML1/ETO+ HSCs as the first step. These preleukemic AML1/ETO+ HSCs upregulate AML1/ETO transcripts (second step), and the acquisition of a c-KIT mutation (3rd step) finally transforms preleukemic AML1/ETOhigh HSCs into LSCs. The detailed mechanism of AML1/ETO upregulation at the second step remains unknown. The time course of the second and third steps was not confirmed in this study.
], the enhanced c-KIT signals may play a critical role in leukemic transformation. Our patients had D816V, D816Y, and N822K c-KIT mutants, and microarray analysis showed that c-KIT mutations at least caused upregulation of NFKB1A, BCL2, and MCL1, whose signals may promote proliferation/survival of leukemic cells [
]. Thus, it is possible that active c-KIT signaling itself does not decide the phenotype of leukemia, but it may contribute toward development of AML in the presence of CBF mutations such as AML1/ETO.
It is also important to note that the significant upregulation of AML1/ETO transcripts may also be a key event in leukemic transformation. Because AML1/ETO knock-in mice did not present any leukemia phenotype [
], AML1/ETO may not play a role in the inhibition of CBF function. In contrast, in LSCs, AML1/ETO mRNA levels were elevated up to one hundredfold at diagnosis, and this high level of AML1/ETO may be effective in inhibiting myeloid differentiation. Our data suggest that c-KIT signaling is independent of AML1/ETO upregulation (Fig. 4B). Collectively, an unknown, presumably epigenetic mechanism that elevates AML1/ETO mRNA levels may also be critical for the development of t(8;21) AML. This event may precede the acquisition of c-KIT mutations, since no AML1/ETO+ HSCs in remission had mutated c-KIT (Table 2).
The reason the ordered acquisition of Class II and Class I mutation is consistently observed in t(8;21) AML patients remains unclear. To achieve leukemic hematopoiesis, the single cell with the first oncogenic hit needs to have a clonal advantage for self-renewal compared with normal HSCs. Class I mutations, when their expression is enforced, are capable of conferring cytokine-independent growth activity to cell lines [
]. In mouse models, HSCs with a high level of BCR-ABL (Class I) enforced by retroviruses showed myeloproliferation, whereas, in healthy human adults, a low level of BCR-ABL transcripts is sometimes detectable [
The presence of typical and atypical BCR-ABL fusion genes in leukocytes of normal individuals: biologic significance and implications for the assessment of minimal residual disease.
]. This suggests that the acquisition of BCR-ABL in HSCs cannot directly provide clonal advantages against normal HSCs. Previous studies reported that mice having HSCs with FLT3-ITD showed expansion of hematopoietic progenitors resulting in myeloproliferation [
]. Therefore, in the light of de novo development of human AML, if a single HSC achieves FLT3-ITD, the FLT3-ITD+ HSCs may not be able to outgrow normal HSCs. It is possible that HSCs with a c-KIT mutation alone cannot exhibit the advantage of self-renewal against normal HSCs to become a dominant clone because c-KIT and FLT3 use similar signal transduction pathways [
]. This evidence indicates that HSCs achieved through AML1/ETO do not have advantages in self-renewal but can coexist with normal HSCs for a long period. This is probably because the expression level of AML1/ETO in t(8;21)+ HSCs is low, and it does not significantly block hematopoietic differentiation. It is assumed that the long-term coexistence of normal and t(8;21)+ HSCs allows the latter to acquire second or third oncogenic hits, including Class I mutations and some unknown abnormalities that can cause upregulation of AML1/ETO. It is critical to test whether this hypothesis can be applied to AML with other Class II mutations.
Collectively, the results of our intensive analysis of de novo t(8;21) human AML suggest that there are at least three independent leukemogenic steps in this type of leukemia (Fig. 5). First, the normal HSC acquires t(8;21), which generates a low level of AML1/ETO. Second, the long-term existence of such AML1/ETO+ HSCs allows them to obtain additional epigenetic or genetic abnormalities that upregulate AML1/ETO. Finally, the acquisition of Class I mutations, such as c-KIT mutants, transforms AML1/ETO+ preleukemic HSCs into AML LSCs. Thus, the original description of Class I and Class II mutations [
] is very useful in the understanding of the leukemogenesis of AML. In future studies, intensive tracking of mutational processes during clinical courses is critical to understand the step-wise leukemogenesis involved in de novo human AML.
Acknowledgements
Dr. Miyamoto and Prof. Akashi received a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.
Conflict of interest disclosure
No financial interest/relationships with financial interest relating to the topic of this article have been declared.
Supplementary data
Supplementary Table E1Primers for the genomic PCR analysis of c-KIT mutation
Supplementary Figure E1Analyses of CD34+CD38− fraction of normal, diagnostic and remission bone marrow. (A) The expression level of c-KIT in normal HSCs, t(8;21) AML with wild-type c-KIT LSCs, and t(8;21) AML with mutated c-KIT LSCs in a quantitative PCR analysis. Each bar shows the fold difference of the level of c-KIT mRNA in comparison to normal HSCs. There was no significant difference of c-KIT mRNA expression levels among these cells. (B) Phenotype of normal HSCs, t(8;21) AML with wild-type c-KIT LSCs, and t(8;21) AML with mutated c-KIT LSCs by five-color flow cytometer. The phenotype of normal HSCs was CD34+CD38−CD90+c-KIT+. LSCs displayed the CD34+CD38−CD90−c-KIT+ phenotype, irrespective of additional c-KIT mutation. Representative data (Patients 1 and 3) are shown. (C) The expression level of c-KIT did not differ among HSCs and LSCs with or without c-KIT mutant.
Mutations in the receptor tyrosine kinase pathway are associated with clinical outcome in patients with acute myeloblastic leukemia harboring t(8;21)(q22;q22).
Expression of the neural cell adhesion molecule CD56 is associated with short remission duration and survival in acute myeloid leukemia with t(8;21)(q22;q22).
Modeling interactions between leukemia-specific chromosomal changes, somatic mutations, and gene expression patterns during progression of core-binding factor leukemias.
The presence of typical and atypical BCR-ABL fusion genes in leukocytes of normal individuals: biologic significance and implications for the assessment of minimal residual disease.
00621-3&id=gr1.jpg){kind=link}
00621-3&id=gr2.jpg){kind=link}
00621-3&id=gr3.jpg){kind=link}
00621-3&id=gr4.jpg){kind=link}
00621-3&id=gr5.jpg){kind=link}
00621-3&id=fx1.jpg){kind=link}