Association of miR-93 expression with clinical and molecular characteristics in the chemotherapy-only group
The clinical and molecular characteristics of the patients who received the chemotherapy-only were shown in Table 1. Median age was 66 (range, 22–88), with 61 cases older than 60. Fifty were men. The medians of WBC counts, BM blast, and PB blast percentages were 15.6 × 109/L, 72%, and 28.5%, respectively. FAB subtypes were mainly M1, M2, and M4 (72.2%). Forty-six patients had abnormal karyotypes. The proportion of good, intermediate, and poor cytogenetic risk patients were 14.8%, 62.5%, and 22.7%, respectively. NPM1 had the highest mutation frequency (n = 29, 32.2%), followed by DNMT3A (n = 25, 27.8%), FLT3-ITD (n = 16, 17.8%), IDH1/2 (n = 16, 17.8%), NRAS/KRAS (n = 13, 14.4%), TP53 (n = 11, 12.2%), TET2 (n = 12, 13.3%), and RUNX1 (n = 8, 8.9%). Then all patients who underwent chemotherapy-only were divided into either miR-93high group or the miR-93low group based on miR-93 median expression levels. In miR-93high group, there were fewer patients with complex karyotype, more patients with RUNX1-RUNX1T1, more with good cytogenetic risk and fewer poor-risk, less frequent NPM1 and TP53 mutations (all P < 0.05). No significant differences were identified in gender distribution, peripheral WBC count, BM blast and PB blast percentage, and the frequencies of other genetic mutations (FLT3, DNMT3A, RUNX1, TET2, IDH1/IDH2, NRAS/KRAS) between the two groups.
Table 1
Comparison of clinical and molecular characteristics in the chemotherapy-only group
Characteristics | Total | miR-93 | P |
High (n = 45) | Low (n = 45) |
Age/years, median (range) | 66 (22–88) | 64 (22–83) | 67 (31–88) | 0.309* |
Age group/n (%) | | | | |
≥60 years | 61 (67.8) | 28 (62.2) | 33 (73.3) | 0.259§ |
<60 years | 29 (32.2) | 17 (37.8) | 12 (26.7) | |
Gender/n (%) | | | | 0.671§ |
Male | 50 (55.6) | 26 (57.8) | 24 (53.3) | |
Female | 40 (44.4) | 19 (42.2) | 21 (46.7) | |
WBC/×109/L, median (range) | 15.6 (0.7-298.4) | 14.3 (1-114.5) | 16 (0.7-298.4) | 0.305* |
BM blasts/%, median (range) | 72 (30–99) | 72 (30–95) | 74 (32–99) | 0.997* |
PB blasts/%, median (range) | 28.5 (0–98) | 46 (0–97) | 22 (0–98) | 0.207* |
FAB subtypes/n (%) | | | | |
M0 | 8 (8.9) | 3 (6.7) | 5 (11.1) | 0.459§ |
M1 | 20 (22.2) | 11 (24.4) | 9 (20.0) | 0.612§ |
M2 | 21 (23.3) | 13 (28.9) | 8 (17.8) | 0.213§ |
M4 | 24 (26.7) | 11 (24.4) | 13 (28.9) | 0.634§ |
M5 | 13 (14.4) | 6 (13.3) | 7 (15.6) | 0.764§ |
M6 | 1 (1.1) | 1 (2.2) | 0 (0.0) | 0.315§ |
M7 | 3 (3.3) | 0 (0.0) | 3 (6.7) | 0.078§ |
Cytogenetics/n (%) | | | | |
Normal | 44 (50.0) | 20 (45.4) | 24 (54.5) | 0.394§ |
Complex | 12 (13.6) | 2 (4.5) | 10 (22.7) | 0.013§ |
inv(16)/CBFβ-MYH11 | 7 (8.0) | 5 (11.4) | 2 (4.5) | 0.237§ |
t(8;21)/RUNX1-RUNX1T1 | 6 (6.8) | 6 (13.6) | 0 (0.0) | 0.011§ |
11q23/MLL | 3 (3.4) | 2 (4.5) | 1 (2.3) | 0.557§ |
-7/7q- | 3 (3.4) | 0 (0.0) | 3 (6.8) | 0.078§ |
t(9;22)/BCR-ABL1 | 1 (1.1) | 0 (0.0) | 1 (2.3) | 0.315§ |
Others | 12 (13.6) | 9 (20.5) | 3 (6.8) | 0.062§ |
Risk/n (%) | | | | |
Good | 13 (14.8) | 11 (25.0) | 2 (4.5) | 0.007§ |
Intermediate | 55 (62.5) | 27 (61.4) | 28 (63.6) | 0.826§ |
Poor | 20 (22.7) | 6 (13.6) | 14 (31.8) | 0.042§ |
FLT3-ITD/n (%) | | | | 0.098§ |
Positive | 16 (17.8) | 11 (24.4) | 5 (11.1) | |
Negative | 74 (82.2) | 34 (75.6) | 40 (88.9) | |
NPM1/n (%) | | | | 0.042§ |
Mutation | 29 (32.2) | 10 (22.2) | 19 (42.2) | |
Wild type | 61 (67.8) | 35 (77.8) | 26 (57.8) | |
DNMT3A/n (%) | | | | 0.239§ |
Mutation | 25 (27.8) | 10 (22.2) | 15 (33.3) | |
Wild type | 65 (72.2) | 35 (77.8) | 30 (66.7) | |
IDH1/IDH2/n (%) | | | | 0.098§ |
Mutation | 16 (17.8) | 11 (24.4) | 5 (11.1) | |
Wild type | 74 (82.2) | 34 (75.6) | 40 (88.9) | |
NRAS/KRAS/n (%) | | | | 0.134§ |
Mutation | 13 (14.4) | 4 (8.9) | 9 (20.0) | |
Wild type | 77 (85.6) | 41 (91.1) | 36 (80.0) | |
RUNX1/n (%) | | | | 1.000§ |
Mutation | 8 (8.9) | 4 (8.9) | 4 (8.9) | |
Wild type | 82 (91.1) | 41 (91.1) | 41 (91.1) | |
TET2/n (%) | | | | 1.000§ |
Mutation | 12 (13.3) | 6 (13.3) | 6 (13.3) | |
Wild type | 78 (86.7) | 39 (86.7) | 39 (86.7) | |
TP53/n (%) | | | | 0.004§ |
Mutation | 11 (12.2) | 1 (2.2) | 10 (22.2) | |
Wild type | 79 (87.8) | 44 (97.8) | 35 (77.8) | |
Abbreviations: WBC, white blood cell; BM, bone marrow; PB, peripheral blood; FAB, French American British; |
‘*’ denotes Mann-Whitney U test; ‘§’ denotes chi-square test. |
In order to further evaluate the prognostic value of miR-93, we constructed the multivariate Cox proportional hazard models (Table 2). The variables included the expression levels of miR-93 (high vs. low), age (≥ 60 vs. <60 years), peripheral WBC count (≥ 15 × 109/L vs. <15 × 109/L), BM blasts (≥ 70% vs. <70%), PB blasts (≥ 70% vs. <70%), FLT3-ITD (positive vs. negative) and common AML mutations (NPM1, DNMT3A, TET2, RUNX1 and NRAS/KRAS, mutated vs. wild type).
Table 2
Multivariate analysis of EFS and OS in the chemotherapy-only group
Variables | EFS | OS |
HR (95%CI) | P-value | HR (95%CI) | P-value |
miR-93 (high vs. Low) | 0.487 (0.282–0.844) | 0.010 | 0.450 (0.257–0.791) | 0.005 |
Age (≥ 60 vs. <60 years) | 4.987 (2.619–9.495) | 0.000 | 4.051 (2.163–7.584) | 0.000 |
WBC (≥ 15 vs. <15 × 109/L) | 0.971 (0.565–1.670) | 0.916 | 1.018 (0.592–1.751) | 0.949 |
BM blasts (≥ 70 vs. <70%) | 1.848 (1.058–3.226) | 0.031 | 1.942 (1.105–3.414) | 0.021 |
PB blasts (≥ 70 vs. <70%) | 3.247 (1.622–6.502) | 0.001 | 2.459 (1.204–5.023) | 0.014 |
FLT3-ITD (positive vs. negative) | 1.386 (0.705–2.726) | 0.344 | 1.484 (0.728–3.028) | 0.278 |
NPM1 (mutated vs. wild) | 0.548 (0.269–1.117) | 0.098 | 0.419 (0.204–0.859) | 0.018 |
DNMT3A (mutated vs. wild) | 1.738 (0.981–3.080) | 0.058 | 1.872 (1.065–3.289) | 0.029 |
RUNX1 (mutated vs. wild) | 1.273 (0.559–2.897) | 0.566 | 1.431 (0.630–3.253) | 0.392 |
TET2(mutated vs. wild) | 0.701 (0.333–1.474) | 0.349 | 0.564 (0.270–1.177) | 0.127 |
NRAS/KRAS (mutated vs. wild) | 0.675 (0.307–1.488) | 0.330 | 0.847 (0.390–1.841) | 0.675 |
Abbreviations: EFS, Event-free survival; OS, Overall survival; HR, hazard ratio; CI, confidence interval; WBC, white blood cell; |
BM, bone marrow; PB, peripheral blood |
Multivariate analysis reported that high miR-93 expression was an independent favorable prognostic factor for both EFS and OS (all P < 0.05). In addition, age ≥ 60 years, BM blasts ≥ 70% and PB blasts ≥ 70% were also independent risk factors for EFS and OS (all P < 0.05). NPM1 mutation was an independent favorable factor for OS (P = 0.018), while DNMT3A mutation played an independent negative role for OS (P = 0.029).
Associations between genome-wide gene expression profile and miR-93 expression
To further assess the role of miR-93 in AML, the high-throughput sequencing information from TCGA database was used to summarize the miR-93-associated gene expression profiles. By using a conservative cutoff for significance (P < 0.05), 1479 genes were identified to be positively correlated with miR-93 expression, and 985 genes were negatively associated with the expression of miR-93 (Fig. 2A). Genome-wide expression profile analysis showed that genes positively related to miR-93 expression included tumor suppressors (SDPR, EHD2, KIAA0125, ZFP36L1, LILRA3, ABI3, SERPINA1, C5aR1, FCN1, BASP1, FCER1G and HK3). However, there was a negative correlation between the expression of miR-93 and FGR(a tumorigenesis promoter in multiple leukemia) (Fig. 2B). Furthermore, GO enrichment analysis suggested that the genes related to miR-93 expression were mainly concentrated in “staphylococcus aureus infection”, “complement and coagulation cascades”, “pertussis”, “osteoclast differentiation”, “B cell receptor signaling pathway” and “phagosome” signaling pathways (Fig. 2C).