Welcome international healthcare professionals

This site is no longer supported and will not be updated with new content. You are welcome to browse and download all content already included in the site. Please note you will have to register your email address to access the site.

You are here

High receptor tyrosine kinase (FLT3, KIT) transcript versus anti-apoptotic (BCL2) transcript ratio independently predicts inferior outcome in pediatric acute myeloid leukemia

Blood Cells, Molecules, and Diseases

Abstract

Objective

In acute myeloid leukemia (AML), simultaneous expression of proliferative (FLT3,KIT) and anti-apoptotic genes (BCL2) is unknown. The aim of the study was to prospectively evaluate proliferative and anti-apoptotic gene transcripts, their interrelationship and impact on the outcome in pediatric AML patients.

Methods

We assessed proliferative and anti-apoptotic gene transcripts by Q-polymerase chain reaction (TaqMan probe) in 64 consecutive pediatric AML patients. Survival data was analyzed by Kaplan–Meier curves followed by log rank test to compare statistical significance between groups. Stepwise multivariable Cox regression method was used to evaluate independent prognostic factors.

Results

In univariate analysis, transcript ratio ofFLT3/BCL2andFLT3 + KIT/BCL2significantly predicted event free survival (EFS) (< 0.01 and < 0.01 respectively) and overall survival (OS) (< 0.01 and < 0.01 respectively). In stepwise Cox-regression model, high white blood cell count and highFLT3 + KIT/BCL2ratio predicted EFS (HR: 2.2 and 2.3); high hemoglobin and highFLT3 + KIT/BCL2ratio predicted OS (HR: 0.45 and 3.85). Prognostic index (PI) was calculated using the hazard coefficient of independent prognostic factors; at 57.3 months, predicted OS of patients with the highest PI of 1.8 was 8% versus 73% for the lowest PI of − 0.3. The mean PI of patients who died was 1.8 ± 0.72 versus 0.54 ± 0.70 for those who are alive, P = 0.004.

Conclusions

This first study showed that individual expression of proliferative and anti-apoptotic transcripts is not as important in AML patients, rather their interrelationship and relative level probably determines the outcome.

Keywords: Real time PCR, FLT3, KIT, BCL2, Acute myeloid leukemia.

Introduction

Acute myeloid leukemia (AML) is characterized by increased myeloblasts in the bone marrow primarily because of the imbalance between proliferation and apoptosis. Proliferative markers such as receptor tyrosine kinases (RTKs) have a significant contribution in leukemogenesis[1], [2], [3], and [4]. The key proliferative RTKs for AML includeKITreceptor and its ligand stem cell factor, and Fms-like tyrosine kinase 3 (FLT3receptor) and its ligand. Mutations of these receptors have been extensively evaluated for their prognostic significance in AML[5], [6], and [7]. In AML, there is data to suggest that highFLT3expression is associated with poor prognosis but data onKITexpression is controversial with regard to the outcome[8], [9], and [10].FLT3andKITmutations are important in predicting the outcome of the patients and the same were recommended as validated prognostic factors in risk stratification and patient management [11] .

Programmed cell death is the inherent mechanism to check uncontrolled growth of cells hence many chemotherapeutic drugs kill cancer cells by inducing apoptosis. The balance of pro-apoptotic and anti-apoptotic proteins in a cell affects the apoptotic state of that cell.BCL2, an anti-apoptotic gene, has been studied extensively in different cancers and leukemia to determine the apoptotic state of the oncogenic cell. Previously,BCL2was evaluated in AML for its clinical, biological and pathological contributions in leukemia development, and has been variably linked to the outcome in AML[12], [13], and [14].

Logically one would expect that increased expression of all these three proteins would indirectly lead to increased proliferation of myeloblasts. Although they have been individually evaluated in AML patients, their simultaneous expression and as to how they vary with respect to each other and their relationship with the outcome is not known. In the current study, we prospectively assessed proliferative and anti-apoptotic gene transcripts, their interrelationship and impact on the outcome in pediatric AML patients.

Material and methods

Patient selection, treatment and sampling

Newly diagnosed consecutive AML patients (except acute promyelocytic leukemia) ≤ 18 years of age were enrolled from March 2008 till June 2010 prospectively. The study was approved by the institute ethics committee and informed consent was taken for the evaluation of the peripheral blood and/or bone marrow for the study. AML was classified into good, intermediate and poor risk cytogenetic groups [15] . Uniform treatment protocol was followed for all patients (Daunorubicin 60 mg/m2for 3 days with Cytarabine 100 mg/m2/day over 24 h infusion for 7 days); patients who were not in morphological complete remission (CR) post first induction received ADE protocol (Cytarabine 100 mg/m2slow intravenous push twice a day for 10 days, Daunorubicin 50 mg/m2daily for 3 days and Etoposide 100 mg/m2daily for 5 days) [16] . Patients who were not in CR after 2 inductions were declared refractory. After achieving remission patients received three cycles of high dose Cytarabine at 18g/m2/cycle.

Five milliliter peripheral blood [N = 47 patients] (if peripheral blast count was more than 30%) or otherwise 5 mL bone marrow [N = 17 patients] was collected using ethylenediaminetetraacetic acid (EDTA) as anticoagulant in EDTA coated sterile vacutainer (BD).

RNA isolation and reverse transcription

Total RNA was isolated from 10 million mononuclear cells using TRIzol® method. RNA was evaluated for quality and quantity by spectrophotometry. cDNA was synthesized from 1 μg aliquots of total RNA in a 20 μL standard reaction mixture using reverse transcription (RT) kit [Roche Applied Sciences, Mannheim, Germany] according to the manufacturer's instructions. We assessed cDNA quality by assessing its conversion by RTPCR for a long span of beta actin housekeeping gene of 5 exon length starting from exon 2 to exon 6 (1100 bp). cDNA was processed further for the absolute quantification of transcripts [14] .

Absolute quantification by TaqMan probe based real-time polymerase chain reaction

For the absolute quantification of transcripts by TaqMan chemistry, probes forFLT3andBCL2were conjugated with Yakima Yellow (YAK) as reporter dye at 5′ and BlackBerry Quencher (BBQ) at 3′, whileKITwas conjugated with 6-carboxyfluorescein (FAM) as reporter dye and BlackBerry Quencher (BBQ) at 5′ and 3′ ( Table 1 ) [17] . Standards were prepared for all 3 genes (FLT3,KITandBCL2) using respective cloned quantified plasmids. Ten billion (10 × 109) copies of each cloned plasmid were reconstituted in 100 μL of PCR grade water to produce 108 copies/μL. Different concentrations ranging from 1 × 102to 1 × 108of plasmid were prepared using serial dilution and 4 different concentrations were used in duplication for standard curve preparation. We quantified the absolute transcript number using labeled pre-quantified plasmids and the same was used as normalizing control and internal control. Standard curves were prepared with four different concentrations and run each in duplication. Whereas each patient sample was quantified by single run along with pre-quantified (5 × 104copies) plasmids for normalization and as internal control. Standard curve was prepared in the beginning and saved, and the same was imported for analysis. Quantitative RT-PCR analysis was performed in a Light Cycler 2.0 (Roche Applied Sciences, Mannheim, Germany).

Table 1 Primers and probes for proliferative and anti-apoptotic genes used in Q PCR.

Gene Amplicon size PubMed gene ID Sequence (primer length)
FLT3-forward 117 NM_004324 5′CACCGGATACTACACTTGTTCCTCT3′, (25)
FLT3-probe NM_004324 YAK 5′TTTGGTTACCATCGTAGGAAAGGGATTTATAA3′BBQ, (32)
FLT3-reverse NM_004324 5′TTGGTCAATTTCATAATCTTCACTTGA3′, (27)
KIT-forward 74 NM_000222 5′GAGGCAACTGCTTATGGCTTAATTA3′, (25)
KIT-probe NM_000222 FAM 5′TTACAGCGACAGTCATGGCCGCAT3′BBQ, (24)
KIT-reverse NM_000222 5′GGCACTCGGCTTGAGCAT3′, (18)
BCL2-forward 72 NM_000633 5′CAGATGGACCTAGTACCCACTGAGA3′, (25)
BCL2-probe NM_000633 YAK 5′TTCCACGCCGAAGGACAGCGAT3′BBQ, (22)
BCL2-reverse NM_000633 5′CCTATGATTTAAGGGCATTTTTCC3′, (24)

Statistical analysis

Outcome analysis was assessed based on CR rate, event free survival (EFS) and overall survival (OS). EFS were defined as the time between the diagnosis and first event such as failure to achieve complete remission, relapse or death. OS was defined as the time between the diagnosis and death or last follow-up. The end point of study was December 31st 2012. Data was expressed as median (range) and mean ± SD; the differences between the values were determined by using Kruskal–Wallis test and independent Students t test respectively. Median values of quantitative variables were used as a cut-off point for categorization into high or low expression. For hemoglobin 8 g/dL and for WBC and platelets 50,000/mm3were taken as cut-off for high and low values. The mean values of quantitative variables were compared with baseline patient characteristics. Log mean of transcripts (copy number) were compared with baseline patient characteristics as the range was wide. The Pearson correlation coefficient was a measure of linear relationship between transcript copy numbers. Kaplan–Meier curves were obtained for survival analysis followed by log rank test to compare statistical significance between groups. Stepwise multivariable Cox regression method was employed to evaluate independent prognostic factors. Prognostic index (PI) of individual patients was calculated with hazard coefficient and using this index, nomogram for predicted EFS and OS of patients was estimated based on their differential PI. P value < 0.05 was considered significant. All statistical analysis was done using STATA 11.0.

Results

Baseline patient characteristics and outcome

Sixty-four consecutive pediatric AML patients with a median age of 10 years were recruited in the study period; baseline characteristics and expression ofFLT3,KITandBCL2by real time PCR are shown in ( Table 2 ). Out of 64 patients, 53 (82.8%) patients achieved CR at the end of induction chemotherapy. At median follow-up time of 18.3 months (range: 0.5–57.3), EFS was 28.6% ± 5.7 (18.1–40.02) and OS 37.5% ± 6.3 (25.3–49.7).

Table 2 Baseline patient characteristics.

  AML patients (N = 64)
Variables  
 Median age in years (range) 10 (1–18)
 Sex (male:female) 3: 1
 Median hemoglobin (g/dL) (range) 6.7 (3.3–14.5)
 Median WBC (/mm3) (range) 22,550 (700–3,50,000)
 Median platelets (/mm3) (range) 30,000 (1900–2,72,550)
Cytogenetics (n = 48)  
 Good risk 14 (29.1%)
 Intermediate risk 24 (50%)
 Poor risk 10 (20%)
FAB AML subtype  
 M0 6 (9.3%)
 M1 6 (9.3%)
 M2 35 (54.6%)
 M4 10 (15.6%)
 M5 5 (7.8%)
 M7 2 (3.1%)
FLT3 ITD  
 Positive 12 (18.75%)
 Negative 52 (81.25%)
 
Expression of FLT3, KIT and BCL2 [transcript copies/μg of RNA; N = 64]
  FLT3 KIT BCL2
Median (range) 3.49 × 106

(0.01 × 106–306 × 106)
0.40 × 106

(0.0015 × 106–21 × 106)
0.49 × 106

(0.0018 × 106–12.3 × 106)

WBC: white blood cells, FAB: French–American–British,FLT3ITD:FMSlike tyrosine kinase 3 internal tandem duplication.

Association of patient characteristics with the outcome

The association of baseline patient characteristics with the outcome is shown in ( Table 3 ). High WBC count predicted inferior EFS (P = 0.014). Cytogenetics andFLT3ITD did not predict EFS or OS.

Table 3 Association of study variables with the outcome.

Variable   CR

N (%)
P EFS a P OS a P
Hemoglobin

(g/dL)
< 8.0 (39) 30 (77) 0.109 18.5 ± 6.3 (8.2–32.1) 0.026 27.2 ± 7.7 (13.6–42.7) 0.091
≥ 8.0 (25) 23 (92) 44.0 ± 9.9 (24.5–61.9) 52.0 ± 10 (31.3–69.2)
WBC

(/mm3)
< 50,000.0 (44) 39 (89) 0.073 36.4 ± 7.3 (22.6–50.3) 0.014 44.3 ± 7.6 (29.3–58.3) 0.178
≥ 50,000.0 (20) 14 (70) 10.7 ± 7.2 (1.8–28.8) 19.8 ± 10.2 (4.9–41.9)
Platelets

(/mm3)
< 50,000.0 (45) 35 (80) 0.286 27.3 ± 6.7 (15.2–40.8) 0.397 35.7 ± 7.6 (21.3–50.3) 0.498
≥ 50,000.0 (19) 17 (89) 33.6 ± 11.2 (13.8–54.8) 44.6 ± 11.7 (21.7–65.3)
FLT3 ITD ITD negative (52) 41 (79) 0.081 31.5 ± 6.5 (19.4–44.3) 0.691 40.0 ± 7.1 (26.3–53.4) 0.781
ITD positive (12) 12 (100) 16.7 ± 10.8 (2.7–41.3) 27.5 ± 13.5 (6.6–54.2)
Cytogenetics Good risk (14) 14 (100) 0.084 50.0 ± 13.4 (22.9–72.2) 0.127 64.3 ± 12.8 (34.3–83.3) 0.149
Intermediate (24) 19 (79) 17.5 ± 7.9 (5.5–35.1) 30.6 ± 10.4 (12.6–50.9)
Poor risk (10) 7 (70) 20.0 ± 12.7 (3.1–47.5) 23.3 ± 14.3 (3.6–52.9)

a Survival at 57.3 months.

WBC: White blood cells,FLT3ITD:FMSlike tyrosine kinase 3 internal tandem duplication.

Relationship of proliferative and apoptotic transcripts

The transcripts ofFLT3,KITandBCL2showed significant linear association with each other by Pearson correlation [FLT3vsKIT: R = 0.7008, P < 0.001;FLT3vsBCL2: R = 0.6774, P < 0.001;KITvsBCL2: R = 0.5138, P < 0.001] ( Fig. 1 A–C).

gr1

Fig. 1 Relationship of proliferative and apoptotic transcripts, A:FLT3andKIT, B:FLT3andBCL2, and C:BCL2andKIT.

Relationship of FLT3, KIT and BCL2 transcripts with baseline patient characteristics

Patients with higher platelet count were associated with increased transcript levels ofFLT3(P = 0.019). None of the other baseline clinical features were associated with transcripts ofFLT3,KITandBCL2( Table 4 ).

Table 4 Relationship ofFLT3,KITandBCL2transcripts with baseline patient characteristics.

  FLT3 a (mean ± SD) P KIT a (mean ± SD) P BCL2 a (mean ± SD) P
Sex            
 Male (48) 32.9 ± 63.7 0.758 1.3 ± 3.3 0.394 1.74 ± 2.95 0.835
 Female (16) 27.4 ± 54.2 2.03 ± 3.4 1.55 ± 3.41
Hemoglobin (g/dL)            
 < 8 (39) 22.6 ± 47.5 0.146 0.6 ± 1.4 0.128 1.31 ± 2.22 0.216
 ≥ 8 (25) 45.4 ± 76.8 1.8 ± 4.3 2.28 ± 3.99
WBC (/mm3)            
 < 50,000 (44) 32.2 ± 62.3 0.889 1.36 ± 3.46 0.385 1.75 ± 3.19 0.805
 ≥ 50,000 (20) 29.9 ± 60.0 0.67 ± 1.20 1.55 ± 2.78
Platelets (/mm3)            
 < 50,000 (45) 20.0 ± 43.0 0.019 6.9 ± 1.58 0.063 1.45 ± 2.84 0.330
 ≥ 50,000 (19) 59.0 ± 86.8 0.22 ± 4.78 2.28 ± 3.54
FLT3 ITD            
 Positive (12) 52.61 ± 106.5 0.128 0.27 ± 0.50 0.256 1.40 ± 2.98 0.721
 Negative (52) 43.02 ± 91.92 1.3 ± 3.2 1.76 ± 3.08
Cytogenetics            
 Good risk (14) 50.8 ± 70.5 0.838 1.33 ± 1.96 0.821 2.69 ± 4.10 0.693
 Intermediate risk (24) 13.6 ± 38.8 0.66 ± 1.14 0.68 ± 0.84
 Poor risk (10) 19.2 ± 31.7 0.58 ± 1.21 1.51 ± 3.6
AML subtype            
 M2 (35) 35.4 ± 70.4 0.658 1.47 ± 3.75 0.574 1.74 ± 2.86 0.492
 M4 (10) 34.1 ± 47.4 1.44 ± 2.18 1.85 ± 3.76
 Others (19) 22.9 ± 49.7 0.39 ± 0.85 1.53 ± 3.15

a Multiplication factor for transcripts is 1 × 106.

WBC: White blood cells,FLT3ITD: FMS like tyrosine kinase 3 internal tandem duplication.

Relationship of proliferative and anti-apoptotic markers with the outcome

In univariate survival analysis, none of the proliferative and anti-apoptotic transcripts was significant for CR ( Table 5 ). The ratios ofFLT3/BCL2andFLT3 + KIT/BCL2significantly predicted EFS (P < 0.01 and < 0.01 respectively) and OS (< 0.01 and < 0.01 respectively) ( Table 5 ) ( Fig. 2 A–F; Fig. 3 A–F).

Table 5 Relationship of proliferative and anti-apoptotic transcripts with the outcome.

Variables

(N = 64)
Median

(N)
CR

N (%)
P EFS a  ± SE

(CI)
P OS a  ± SE

(CI)
P
FLT3 < 3.4 × 106 (32) 27 (84) 0.500 31.3±8.2 (16.4–47.3) 0.441 46.9±9.5 (27.8–63.9) 0.114
≥ 3.4 × 106 (32) 26 (81) 25.9±7.9 (12.3–42.0) 29.1±8.2 (14.6–45.4)
KIT < 0.17 × 106 (32) 25 (78) 0.255 25.0±7.7 (11.8–40.7) 0.637 38.3±9.1 (21–55.5) 0.984
≥ 0.17 × 106 (32) 28 (87) 32.4±8.4 (17–48.8) 36.8±8.8 (20.2–53.5)
BCL2 < 0.40 × 106 (32) 27 (84) 0.500 25.0±7.7 (11.8–40.7) 0.568 34.8±9.0 (18.2–51.9) 0.672
≥ 0.40 × 106 (32) 26 (81) 32.4±8.4 (17–48.8) 40.1±9.0 (22.9–56.8)
FLT3 + KIT < 3.9 × 106 (32) 27 (84) 0.500 31.3±8.2 (16.4–47.3) 0.441 46.9±9.5 (27.8–63.9) 0.114
≥ 3.9 × 106 (32) 26 (81) 25.9±7.9 (12.3–42) 29.1±8.2 (14.6–45.4)
FLT3/BCL2 < 12.0 (31) 27 (87) 0.293 38.7±8.8 (22–55.2) < 0.01 53.3±9.5 (33.3–69.7) < 0.01
≥ 12.0 (33) 26 (79) 18.9±6.9 (7.7–33.8) 22.9±7.6 (10.1–38.7)
KIT/BCL2 < 0.5 (32) 26 (81) 0.500 34.4±8.4 (18.8–50.6) 0.271 43.2±9.2 (25.2–60) 0.237
≥ 0.5 (32) 27 (84) 22.7±7.5 (10–38.5) 31.5±8.7 (15.9–48.5)
FLT3 + KIT/BCL2 < 24.0 (31) 27 (87) 0.293 38.7±8.8 (22–55.2) < 0.01 53.3±9.5 (33.3–69.7) < 0.01
≥ 24.0 (33) 26 (79) 18.9±6.9 (7.7–33.8) 22.9±7.6 (10.1–38.7)

a Survival at 57.3 months.

gr2

Fig. 2 Event free survival (EFS) forFLT3transcript (A);KITtranscript (B);BCL2transcript (C);KIT/BCL2transcript ratio (D);FLT3/BCL2transcript ratio (E); andFLT3 + KIT/BCL2transcript ratio (F).

gr3

Fig. 3 Overall survival (OS) forFLT3transcript (A);KITtranscript (B);BCL2transcript (C);KIT/BCL2transcript ratio (D);FLT3/BCL2transcript ratio (E); andFLT3 + KIT/BCL2transcript ratio (F).

Multivariable analysis

In multivariable analysis with stepwise Cox regression model, high WBC count and high ratio ofFLT3 + KIT/BCL2emerged as independent predictors for inferior EFS with hazard ratios of 2.2 and 2.3 respectively ( Table 6 ). Similarly using the same model, high hemoglobin and high ratio ofFLT3 + KIT/BCL2independently predicted inferior OS with hazard ratios of 0.45 and 3.85 respectively ( Table 6 ).

Table 6 Multivariate Cox proportional-hazards regression model.

Variables EFS a OS a
Hazard ratio Standard error P 95% Confidence interval Hazard ratio Standard error P 95% Confidence interval
Hemoglobin < 8.0 1.0 0.161 0.027 0.223 0.912
≥ 8.0 0.45
WBC < 50,000.0 1.0 0.706 < 0.01 1.20 4.17
≥ 50,000.0 2.2
FLT3 + KIT/BCL2 < 24.0 1.0 0.707 < 0.01 1.20 4.21 1.00 1.38 < 0.01 1.90 7.811
≥ 24.0 2.3 3.85

a Survival at 57.3 months.

Predicted survival and prognostic index (PI)

Hazard coefficient was determined for each of the independently significant variables in step wise Cox regression multivariate analysis. Based on the hazard coefficient, PI for predicting EFS was determined for each patient using the following formula (0.81 × WBC score) + (0.83 × FLT3 + KIT/BCL2score) [where score was 1 if WBC count was < 50,000/mm3and 2 if ≥ 50,000/mm3, and score forFLT3 + KIT/BCL2was 1 if the same was less than median and 2 if more than or equal to median]. The generated nomogram and EFS in relation to PI is shown in ( Fig. 4 A). The mean PI of patients who incurred an event versus those who did not was 2.44 ± 0.52 vs 2.08 ± 0.58, P = 0.020.

gr4

Fig. 4 Survival with prognostic index (PI): (A): predicted EFS based on PI, (B): predicted OS based on PI, (C): hazard ratio (HR) as function of PI, and (D): nomogram of predicted survival (%) based on PI.

Similarly, PI for predicting OS was determined for each patient using the following formula (− 0.8 × hemoglobin score) + (1.3 × FLT3 + KIT/BCL2score) [where score was 1 if hemoglobin was < 8 g/dL and 2 if ≥ 8 g/dL, and score forFLT3 + KIT/BCL2was 1 if the same was less than median and 2 if more than or equal to median] ( Fig. 4 B). At 57.3 months, the predicted OS of patients with the highest PI of 1.8 was 8% as compared to 73% for patients with the lowest PI of − 0.3. The hazard ratio increased exponentially with increase in PI reaching up to a maximum of 11.4 for OS ( Fig. 4 C). The mean PI of patients who died was 1.8 ± 0.72 versus 0.54 ± 0.70 for those who are alive, P = 0.004.

Validation cohort (microarray gene expression dataset)

All three transcripts (FLT3,KITandBCL2) were assessed in two different microarray mRNA expression datasets of 104 pediatric AML cases  [18] and 237 pediatric patients [19] . The linear relationships of the transcripts were similar to our cohort. The linear relationship ofFLT3,KITandBCL2transcripts in validation transcriptome database was, (Dataset GSE43176 ) N = 104 pediatric AML patients [FLT3vsKIT: R = 0.2475, P < 0.011;FLT3vsBCL2: R = 0.3038, P = 0.001;KITvsBCL2: R = 0.4044, P < 0.001] ( Fig. 5 A–C); Dataset GSE17855 of pediatric AML cases (N = 237) [FLT3vsKIT: R = 0.0755, P = 0.247;FLT3vsBCL2: R = 0.4024, P < 0.001;KITvsBCL2: R = 0.2694, P < 0.001] ( Fig. 6 A–C).

gr5

Fig. 5 Relationship of proliferative and apoptotic transcripts, A:FLT3andKIT, B:FLT3andBCL2, and C:BCL2andKIT[microarray dataset GSE43176 ].

gr6

Fig. 6 Relationship of proliferative and apoptotic transcripts, A:FLT3andKIT, B:FLT3andBCL2, and C:BCL2andKIT[microarray dataset GSE17855 ].

Discussion

In our cohort, highFLT3transcript did not predict survival. This is consistent with the previous report by Müller-Tidow et al. wherein the significance ofFLT3transcript on survival was not observed [20] . We could not find any impact ofKITtranscript level on either EFS or OS while previous two studies had shown contradicting results; Graf et al. [10] had indicated the association of high transcripts with poor survival and another study by Müller-Tidow et al. [20] had shown favorable impact of higherKITtranscripts in 65 AML cases.

All three transcripts (FLT3,KITandBCL2) were assessed in another microarray mRNA expression datasets of pediatric AML cases. Microarray dataset used relative gene expression and so normalization factor affected the final up-fold and down-fold expression. On the contrary, our analysis was based on the absolute quantification of the genes with cloned quantified transcript as normalizing factor which gives more accurate values in the sample. Hence, higher Pearson correlation between the transcripts was observed in our cohort. Despite the relative gene expression of microarray datasets, similar and consistent significant relationship was observed among the transcripts as in our study.

With regard toBCL2also, we did not find any impact of its level of expression on either EFS or OS.BCL2has shown variable relationship with the outcome in AML. In a previous study, higher expression of theBCL2transcripts resulted in poor CR rate and inferior long term OS (49% vs 10% P = 0.028) and disease free survival (71% vs 15% P < 0.001) [21] . In another study, high expression ofBCL2was associated with poor outcome in good and intermediate risk patients, and good outcome in those with poor risk cytogenetics [12] .

It is difficult to understand how the same transcript could have such a variable effect on the outcome in the same disease in different studies.FLT3andKITassist in proliferation whereasBCL2reduces apoptosis. Thus, in an individual patient there is a balance of proliferation and apoptosis. One possible explanation for the variable effect of the same transcript in AML patients may be related to the interplay of the proliferative and anti-apoptotic factors in an individual patient and the relative ratio of these could be responsible for the differential proliferative potential of the blasts which in turn could be related to the outcome as well.

The above study is the first of its kind wherein the effect of two tyrosine kinases andBCL2have been evaluated simultaneously in AML patients in order to understand the interrelationship between proliferative and anti-apoptotic transcripts. We observed that there was a direct linear correlation betweenFLT3,KITandBCL2with each other. This would imply that if all three increased in the same patient, there would be increased proliferation and reduced apoptosis; but the relative presence of each of these could vary between patients and affect outcome. We evaluated a combination and ratio of these three transcripts whereby we observed thatFLT3 + KITandKIT/BCL2were not significant for survival, but highFLT3/BCL2andFLT3 + KIT/BCL2ratios predicted inferior EFS and OS. Yet in multivariate analysis highFLT3 + KIT/BCL2ratio emerged as an independent prognostic factor for poor EFS and OS. This suggests that it is the relative ratio ofFLT3,KITandBCL2in a particular patient that is responsible for the outcome rather than their levels in isolation.

The synergistic role of both RTKs and anti-apoptotic transcript ofBCL2is suggested by the multivariable analysis whereinFLT3 + KIT/BCL2ratio emerged as the strongest independent factor for predicting EFS and OS with hazard ratios of 2.3 and 3.85 respectively. Additionally, the PI that was generated based on the factors predicting EFS and OS independently in multivariate analysis is useful for predicting survival; however, this needs to be validated in different settings. In our cohort, with increase in PI the hazard ratio for OS increased exponentially in a multiplicative manner rather than an additive manner ( Fig. 4 C).

One of the limitations of our study was that instead of bone marrow samples, we analyzed peripheral blood samples in patients with peripheral blast count more than 30%. Notably there was no difference inFLT3,KITandBCL2mean transcript expression in patients wherein the sample was obtained from peripheral blood versus those wherein the sample was obtained from the bone marrow [FLT3,P = 0.612;KIT, P = 0.376,BCL2, P = 0.272]. Another limitation was that the analysis was not performed on enriched blasts. However in the group wherein the samples were obtained from the peripheral blood for analysis, there was no significant linear relationship between the transcripts and peripheral blast percentage [KIT, R = − 0.2529, P = 0.086;FLT3, R = − 0.1574, P = 0.290;BCL2, R = − 0.1941, P = 0.191]. Likewise in the group wherein the samples were obtained from the bone marrow, there was no relationship between transcripts and bone marrow blast percentage [KIT, R = − 0.3662, P = 0.163;FLT3, R = − 0.3072, P = 0.246;BCL2, R = − 0.2049, P = 0.446]. This suggests that even though the blasts were not enriched yet the values of transcripts obtained did not vary with the blast percentage.

To conclude, we elucidated and correlated both proliferative and antiapoptotic genes in individual AML patients at transcript level.FLT3,KITandBCL2directly correlated with each other. HighFLT3 + KIT/BCL2ratio emerged as an independent prognostic factor for poor EFS and OS. This study for the first time has shown that individual expression of the proliferative and antiapoptotic transcripts may not be as important in an individual patient of AML, rather it is their interrelationship and relative level which may determine the outcome in pediatric AML patients.

Acknowledgment

Surender Kumar Sharawat is thankful to the Indian Council of Medical Research (ICMR) New Delhi, India for a Senior Research Fellowship.

Source of funding

This study was supported by the Institute Research Grant from the All India Institute of Medical Sciences, New Delhi, India (F.8-70/A-070/2011/RS).

References

  • [1] L.K. Ashman. The biology of stem cell factor and its receptor C-kit. Int. J. Biochem. Cell Biol.. 1999;31(10):1037-1051 Crossref
  • [2] D. Linnekin. Early signaling pathways activated by c-Kit in hematopoietic cells. Int. J. Biochem. Cell Biol.. 1999;31(10):1053-1074 Crossref
  • [3] C.E. Carow, M. Levenstein, S.H. Kaufmann, J. Chen, S. Amin, P. Rockwell, L. Witte, M.J. Borowitz, C.I. Civin, D. Small. Expression of the hematopoietic growth factor receptor FLT3 (STK-1/Flk2) in human leukemias. Blood. 1996;87(3):1089-1096
  • [4] D.G. Gilliland, J.D. Griffin. The roles of FLT3 in hematopoiesis and leukemia. Blood. 2002;100(5):1532-1542 Crossref
  • [5] H. Kiyoi, T. Naoe, Y. Nakano, S. Yokota, S. Minami, S. Miyawaki, N. Asou, K. Kuriyama, I. Jinnai, C. Shimazaki, H. Akiyama, K. Saito, H. Oh, T. Motoji, E. Omoto, H. Saito, R. Ohno, R. Ueda. Prognostic implication of FLT3 and N-RAS gene mutations in acute myeloid leukemia. Blood. 1999;93(9):3074-3080
  • [6] S. Schnittger, C. Schoch, M. Dugas, W. Kern, P. Staib, C. Wuchter, H. Löffler, C.M. Sauerland, H. Serve, T. Büchner, T. Haferlach, W. Hiddemann. Analysis of FLT3 length mutations in 1003 patients with acute myeloid leukemia: correlation to cytogenetics, FAB subtype, and prognosis in the AMLCG study and usefulness as a marker for the detection of minimal residual disease. Blood. 2002;100(1):59-66 Crossref
  • [7] R.S. Care, P.J.M. Valk, A.C. Goodeve, F.M. Abu-Duhier, W.M.C. Geertsma-Kleinekoort, G.A. Wilson, M.A. Gari, I.R. Peake, B. Löwenberg, J.T. Reilly. Incidence and prognosis of c-KIT and FLT3 mutations in core binding factor (CBF) acute myeloid leukaemias. Br. J. Haematol.. 2003;121(5):775-777 Crossref
  • [8] L.K. Ashman, M.M. Roberts, S.J. Gadd, S.J. Cooper, C.A. Juttner. Expression of a 150-kD cell surface antigen identified by monoclonal antibody YB5.B8 is associated with poor prognosis in acute non-lymphoblastic leukaemia. Leuk. Res.. 1988;12(11–12):923-928 Crossref
  • [9] M.A. Reuss-Borst, H.J. Bühring, H. Schmidt, C.A. Müller. AML: immunophenotypic heterogeneity and prognostic significance of c-kit expression. Leuk. Off. J. Leuk. Soc. Am. Leuk. Res. Fund UK. 1994;8(2):258-263
  • [10] M. Graf, K. Hecht, S. Reif, R. Pelka-Fleischer, K. Pfister, H. Schmetzer. Expression and prognostic value of hemopoietic cytokine receptors in acute myeloid leukemia (AML): implications for future therapeutical strategies. Eur. J. Haematol.. 2004;72(2):89-106 Crossref
  • [11] H. Döhner, E.H. Estey, S. Amadori, F.R. Appelbaum, T. Büchner, A.K. Burnett, H. Dombret, P. Fenaux, D. Grimwade, R.A. Larson, F. Lo-Coco, T. Naoe, D. Niederwieser, G.J. Ossenkoppele, M.A. Sanz, J. Sierra, M.S. Tallman, B. Löwenberg, C.D. Bloomfield. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood. 2010;115(3):453-474
  • [12] S.M. Kornblau, P.F. Thall, Z. Estrov, M. Walterscheid, S. Patel, A. Theriault, M.J. Keating, H. Kantarjian, E. Estey, M. Andreeff. The prognostic impact of BCL2 protein expression in acute myelogenous leukemia varies with cytogenetics. Clin. Cancer Res.. 1999;5(7):1758-1766
  • [13] L. Campos, J.P. Rouault, O. Sabido, P. Oriol, N. Roubi, C. Vasselon, E. Archimbaud, J.P. Magaud, D. Guyotat. High expression of bcl-2 protein in acute myeloid leukemia cells is associated with poor response to chemotherapy. Blood. 1993;81(11):3091-3096
  • [14] S.K. Sharawat, R. Bakhshi, S. Vishnubhatla, R. Gupta, S. Bakhshi. BAX/BCL2 RMFI ratio predicts better induction response in pediatric patients with acute myeloid leukemia. Pediatr. Blood Cancer. 2013;60(8):E63-E66 Crossref
  • [15] D. Grimwade, H. Walker, F. Oliver, K. Wheatley, C. Harrison, G. Harrison, J. Rees, I. Hann, R. Stevens, A. Burnett, A. Goldstone. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children's Leukaemia Working Parties. Blood. 1998;92(7):2322-2333
  • [16] J.A. Liu Yin, K. Wheatley, J.K. Rees, A.K. Burnett. Comparison of ‘sequential’ versus ‘standard’ chemotherapy as re-induction treatment, with or without cyclosporine, in refractory/relapsed acute myeloid leukaemia (AML): results of the UK Medical Research Council AML-R trial. Br. J. Haematol.. 2001;113(3):713-726 Crossref
  • [17] K.M. Clark-Langone, J.Y. Wu, C. Sangli, A. Chen, J.L. Snable, A. Nguyen, J.R. Hackett, J. Baker, G. Yothers, C. Kim, M.T. Cronin. Biomarker discovery for colon cancer using a 761 gene RT-PCR assay. BMC Genomics. 2007;8:279 Crossref
  • [18] M.E. Ross, R. Mahfouz, M. Onciu, H.C. Liu, X. Zhou, G. Song, S.A. Shurtleff, S. Pounds, C. Cheng, J. Ma, R.C. Ribeiro, J.E. Rubnitz, K. Girtman, W.K. Williams, S.C. Raimondi, D.C. Liang, L.Y. Shih, C.H. Pui, J.R. Downing. Gene expression profiling of pediatric acute myelogenous leukemia. Blood. 2004;104(12):3679-3687 Crossref
  • [19] B.V. Balgobind, M.M. Van den Heuvel-Eibrink, R.X. De Menezes, D. Reinhardt, I.H. Hollink, S.T. Arentsen-Peters, E.R. van Wering, G.J. Kaspers, J. Cloos, E.S. de Bont, J.M. Cayuela, A. Baruchel, C. Meyer, R. Marschalek, J. Trka, J. Stary, H.B. Beverloo, R. Pieters, C.M. Zwaan, M.L. den Boer. Evaluation of gene expression signatures predictive of cytogenetic and molecular subtypes of pediatric acute myeloid leukemia. Haematologica. 2011;96(2):221-230 Crossref
  • [20] C. Müller-Tidow, J. Schwäble, B. Steffen, N. Tidow, B. Brandt, K. Becker, E. Schulze-Bahr, H. Halfter, U. Vogt, R. Metzger, P.M. Schneider, T. Büchner, C. Brandts, W.E. Berdel, H. Serve. High-throughput analysis of genome-wide receptor tyrosine kinase expression in human cancers identifies potential novel drug targets. Clin. Cancer Res.. 2004;10(4):1241-1249
  • [21] T. Karakas, U. Maurer, E. Weidmann, C.C. Miething, D. Hoelzer, L. Bergmann. High expression of bcl-2 mRNA as a determinant of poor prognosis in acute myeloid leukemia. Ann. Oncol.. 1998;9(2):159-165 Crossref

Footnotes

a Department of Medical Oncology, Dr. B.R.A. Institute Rotary Cancer Hospital, All India Institute of Medical Sciences, New Delhi, India

b Department of Biomedical Sciences, Dr. B.R.A. Institute Rotary Cancer Hospital, All India Institute of Medical Sciences, New Delhi, India

c Department of Biostatistics, Dr. B.R.A. Institute Rotary Cancer Hospital, All India Institute of Medical Sciences, New Delhi, India

lowast Corresponding author at: Department of Medical Oncology, Dr. B. R.A. Institute Rotary Cancer Hospital, All India Institute of Medical Sciences, New Delhi 110 029, India. Fax: + 91 11 26588663.