Significance

Long noncoding RNAs (lncRNAs) are involved in numerous biological roles including epigenetic regulation, apoptosis, and cell cycle. Whereas lncRNAs contribute to epigenetic gene regulation, metastasis, and prognosis in solid tumors, their role in acute myeloid leukemia (AML) has not been hitherto reported. Here, we show that lncRNA expression profiles are associated with recurrent mutations, clinical features, and outcome in AML. A fraction of these lncRNAs may have a functional role in leukemogenesis. Furthermore, lncRNAs could be used as biomarkers for outcome in AML. The identification of patients likely to achieve complete remission with standard therapy alone, based on lncRNA expression, is a significant advance potentially sparing such patients from other toxicities and focusing investigational approaches on postremission studies.

Abstract

Long noncoding RNAs (lncRNAs) are transcripts longer than 200 nucleotides, located within the intergenic stretches or overlapping antisense transcripts of protein coding genes. LncRNAs are involved in numerous biological roles including imprinting, epigenetic regulation, apoptosis, and cell cycle. To determine whether lncRNAs are associated with clinical features and recurrent mutations in older patients (aged ≥60 y) with cytogenetically normal (CN) acute myeloid leukemia (AML), we evaluated lncRNA expression in 148 untreated older CN-AML cases using a custom microarray platform. An independent set of 71 untreated older patients with CN-AML was used to validate the outcome scores using RNA sequencing. Distinctive lncRNA profiles were found associated with selected mutations, such as internal tandem duplications in the FLT3 gene (FLT3-ITD) and mutations in the NPM1, CEBPA, IDH2, ASXL1, and RUNX1 genes. Using the lncRNAs most associated with event-free survival in a training cohort of 148 older patients with CN-AML, we derived a lncRNA score composed of 48 lncRNAs. Patients with an unfavorable compared with favorable lncRNA score had a lower complete response (CR) rate [P < 0.001, odds ratio = 0.14, 54% vs. 89%], shorter disease-free survival (DFS) [P < 0.001, hazard ratio (HR) = 2.88] and overall survival (OS) (P < 0.001, HR = 2.95). The validation set analyses confirmed these results (CR, P = 0.03; DFS, P = 0.009; OS, P = 0.009). Multivariable analyses for CR, DFS, and OS identified the lncRNA score as an independent marker for outcome. In conclusion, lncRNA expression in AML is closely associated with recurrent mutations. A small subset of lncRNAs is correlated strongly with treatment response and survival.
Acute myeloid leukemia (AML) is diagnosed most often in older adults (age ≥ 60 y) who often have a worse prognosis, with long-term overall survival (OS) rates of only 5–16% (13). The reasons for the poor outcome of older patients relate to higher frequencies of secondary disease (e.g., AML diagnosed after hematologic disorders and/or therapy-related disease), adverse cytogenetics (e.g., complex karyotype), comorbid conditions, and poor performance status (37). Nonrandom chromosomal abnormalities (e.g., deletions, translocations) are identified in 50–55% of all older patients with primary AML (38). In contrast, about 45–50% of all AML cases are cytogenetically normal (CN-AML) when assessed using conventional banding analysis (38). Recent work has identified novel genetic alterations, including gene mutations and changes in gene expression in CN-AML that have improved the classification and risk stratification of this large subgroup of patients (9, 10). Whereas most of these studies were performed in younger patients with CN-AML, some studies have investigated the prognostic significance of genetic alterations in older patients (1119). Our group reported that nucleophosmin (nucleolar phosphoprotein B23, numatrin) (NPM1) mutations are associated with a more favorable outcome in older patients with CN-AML (11), whereas fms-related tyrosine kinase 3 (FLT3) internal tandem duplications (FLT3-ITDs) and mutations in the additional sex combs like transcriptional regulator 1 (ASXL1), runt-related transcription factor 1 (RUNX1) and DNA (cytosine-5-)-methyltransferase 3 alpha (DNMT3A), which affect arginine codon 882 (R882-DNMT3A), are independently associated with worse disease-free survival (DFS) and OS (1215). We further showed that high expression levels of the brain and acute leukemia, cytoplasmic (BAALC), v-ets avian erythroblastosis virus E26 oncogene homolog (ERG), meningioma (disrupted in balanced translocation) 1 (MN1), and DNA (cytosine-5-)-methyltransferase 3 beta (DNMT3B) genes are associated with unfavorable outcome in older patients with CN-AML (1618).
In addition to genetic alterations in protein coding genes, aberrant expression of noncoding RNAs plays a critical role in leukemia initiation and outcome prediction (19). MicroRNAs (miRNAs) constitute the first class of noncoding RNAs, whose expression was found widely dysregulated in AML and associated with clinical features including outcome (20, 21). We reported unique miRNA expression signatures associated with NPM1, RUNX1, and tet methylcytosine dioxygenase 2 (TET2) mutations in older patients with CN-AML (11, 15, 22). High levels of miR-155 and miR-3151 were also found to be associated with worse outcome in older patients with CN-AML (23, 24).
More recently, another class of noncoding RNAs named long noncoding RNAs (lncRNAs) was discovered (25). LncRNAs are trancripts longer than 200 nucleotides, located within the intergenic stretches or overlapping antisense transcripts of protein coding genes (25, 26). LncRNA genes are typically shorter than protein-coding genes and are predominantly transcribed by RNA polymerase II (26). LncRNAs have emerged as important regulators of gene expression, showing cell-specific expression patterns and subcellular localization and are involved in many biological roles including imprinting, epigenetic regulation, apoptosis, and cell cycle (26, 27). The well-characterized lncRNA HOX transcript antisense RNA (HOTAIR) has been found to be up-regulated in multiple cancers including breast, colorectal, hepatic, gastric, and pancreatic cancers (2833). HOTAIR modulates chromatin structures by serving as a scaffold and recruiting histone modifiers (28). In breast cancer, HOTAIR promotes cancer cell invasion and metastasis in vivo through epigenetic silencing of metastasis suppressor genes (2933). In patients with nonsmall cell lung cancer, high expression of the lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) was associated with metastases and poor survival, indicating its importance as a prognostic biomarker (34).
Whereas lncRNAs contribute to epigenetic gene regulation, metastasis, and prognosis in solid tumors, their role in AML has not been reported to date. In this study, we analyzed a large set of older patients with CN-AML using custom lncRNA microarrays and RNA sequencing (RNA-seq) to investigate whether lncRNA expression is associated with clinical features, molecular abnormalities, and outcome.

Results

LncRNA Signatures Associated with Recurrent Mutations in CN-AML.

To identify lncRNAs associated with recurrent mutations in CN-AML, we compared mutated and wild-type (WT) samples in the training set of 148 older patients with de novo CN-AML. For all of the signatures, we used P < 0.001 and fold change (FC) ≥ 1.5 (Table S1 and SI Methods and Data show patient clinical and molecular features).

NPM1 LncRNA Signature.

Patients with NPM1-mutated CN-AML (n = 84) have a strong and distinctive lncRNA signature composed of 205 probes corresponding to 180 lncRNAs (Fig. 1A and Table S2). Among the up-regulated lncRNAs, several were antisense transcripts of HOX genes (e.g., HOXB-AS3, MEIS1-AS2) that have been reported to be up-regulated in NPM1-mutated AML (35). The plasmacytoma variant translocation 1 (PVT1) lncRNA, which is expressed at higher levels in NPM1-mutated patients, is associated with poor prognosis in colorectal cancer via apoptosis inhibition (36). Another study reported that PVT1 silencing in breast and ovarian cancer cell lines resulted in a strong proliferation and apoptotic response (37). The coiled-coil domain containing 26 (CCD26) lncRNA, also up-regulated in the NPM1-mutated subset, is a retinoic acid-dependent modulator of myeloid cell differentiation and death (38). It has been reported that the genotype “mutant NPM1 without FLT3-ITD” appears to be a marker predictive for response to all-trans retinoic acid given as an adjunct to intensive chemotherapy (39).
Fig. 1.
Heat map of the lncRNA-expression signature associated with NPM1, FLT3-ITD, and CEBPA mutations in older patients with primary CN-AML. (A) The heat map shows expression levels of 205 probe sets found differentially expressed between NPM1-mutated (mut) (n = 84) vs. NPM1 wild-type (WT) (n = 64) cases. (B) The heat map shows expression levels of 140 probe sets found differentially expressed between FLT3-ITD (n = 48) vs. FLT3-WT (n = 100) cases. (C) The heat map shows expression levels of 57 probe sets found differentially expressed between CEBPA-mut (n = 18) vs. CEBPA-WT (n = 130) cases. Rows represent probe sets and columns, individual patients. Patients are grouped according to NPM1, FLT3-ITD, and CEBPA mutation status, and genes are ordered by hierarchical cluster analysis. Expression values of the probe sets are represented by color: green represents expression less than the median value for the given probe set, and red, expression greater than the median value for the given probe set.

FLT3-ITD LncRNA Signature.

We identified 119 lncRNAs (from 140 probes) associated with the presence of FLT3-ITD mutations (n = 48) compared with FLT3-WT (n = 100) (Fig. 1B and Table S3). Because NPM1 and FLT3-ITD mutations frequently coexist, there was considerable overlap in their lncRNA signatures (65 probes) (9, 10). Among the 75 probes that were present only in the FLT3-ITD signature, there were two probes corresponding to Wilms tumor 1 antisense RNA (WT1-AS) lncRNA. WT1-AS has been reported to interact with WT1 sense RNA resulting in WT1 protein up-regulation (40). Interestingly, WT1 RNA expression correlates with FLT3-ITD expression in patients with AML (41). Alternative splicing of WT1-AS has been found in AML. Whereas the functional significance of this finding is unknown, aberrations in alternative splicing have been suggested as contributing factors in the development of various diseases including cancer (42).

CEBPA LncRNA Signature.

We identified a signature of 51 lncRNAs (from 57 probes) in patients with CCAAT/enhancer binding protein (C/EBP), alpha (CEBPA)-mutated CN-AML (n = 18) (Fig. 1C and Table S4). The most down-regulated lncRNAs in CEBPA-mutated leukemia samples corresponded to HOXB-AS3, already identified as strongly associated with NPM1 mutation. This finding was expected, because CEBPA mutations occur preferentially (>90%) in patients with wild-type NPM1 (9, 10). Among the up-regulated lncRNAs (n = 50) in patients with CEBPA-mutated CN-AML is urothelial cancer associated 1 (UCA-1), which has been related to more aggressive disease or metastatic potential in bladder and tongue squamous cell carcinoma (43, 44).

IDH1 and IDH2 LncRNA Signatures.

Patients’ samples with R132 mutations in isocitrate dehydrogenase 1 (NADP+), soluble (IDH1R132) (n = 14) had only two differentially expressed lncRNAs: DLEU2 (up-regulated in IDH1R132) and RP11-147N17.1 (down-regulated in IDH1R132). Regarding isocitrate dehydrogenase 2 (NADP+), mitochondrial (IDH2) mutations, 22 lncRNAs were differentially expressed between IDH2-mutated (IDH2R140 in 31 cases and IDH2R172 in 9 cases) and IDH2-WT cases (n = 108) (Table S5). Among these lncRNAs, 19 were differentially expressed in IDH2R172 and 8 in IDH2R140 cases. Altogether, the data indicate that there are lncRNAs associated with both types of IDH2 mutations, whereas some lncRNAs are characteristic of only one type of mutation. For example, LOC100505854 is overexpressed in the IDH2R172 subset, but not in the IDH2R140 samples. This lncRNA associates with and represses the tumor suppressor CDKN1A/p21 promoter by recruiting polycomb complex proteins (45).

RUNX1 LncRNA Signature.

In our patient cohort, RUNX1 mutations were almost mutually exclusive with NPM1 mutations (Fig. S1), which themselves have strong, characteristic lncRNA gene-expression signatures. To avoid confounding effects due to the NPM1 mutations, we performed a two-class analysis (RUNX1-mutated vs. WT) blocking for NPM1 mutation cases and using a randomized block design, as described in SI Methods and Data. Using this approach, we identified a signature comprising 76 lncRNAs from 83 probes (Table S6). Among the lncRNAs up-regulated in RUNX1-mutated AMLs, several were located in the proximity of genes characteristic of lymphoid cells such as the B-cell linker (BLNK) and the immunoglobulin heavy locus (IGH) complex. We previously reported that patients with RUNX1-mutated CN-AML exhibit up-regulation of lymphoid genes (15). Another up-regulated lncRNA in mutated RUNX1, vault RNA 1-1 (VTRNA1-1) has been associated with multidrug resistance (46). This could be relevant because patients with RUNX1-mutated CN-AML have a particularly poor outcome (15).

ASXL1 LncRNA Signature.

Because ASLX1 mutations were found exclusively in NPM1-WT cases (Fig. S1) (9, 10, 15), we again performed the lncRNA analysis blocking for NPM1-mutated cases as described above. We identified only two differentially expressed lncRNAs in ASXL1-mutated samples; PR11-1055B8.4 was up-regulated 22-fold, whereas the ankyrin repeat domain 20 family, member A5, pseudogene (ANKRD20A5P) was down-regulated twofold in patients with ASXL1 mutations.

LncRNAs Profiles Associated with AML with Tyrosine Kinase Domain Mutations in the FLT3 Gene, Partial Tandem Duplications of the MLL Gene, and DNMT3A, TET2, and WT1 Mutations.

We did not identify significant lncRNAs associated with tyrosine kinase domain mutations in the FLT3 gene (n = 16), DNMT3A (n = 47) or TET2 (n = 38) mutations. For WT1 mutations (n = 7) and partial tandem duplications of the MLL gene (n = 9) there were not enough cases to derive lncRNA signatures with confidence.

Building the Prognostic LncRNA Score.

To develop a prognostic lncRNA score, we used pretreatment frozen bone marrow (BM) samples from 148 older patients with de novo CN-AML (Table S1). All patients were treated with intensive cytarabine/daunorubicin-based frontline therapy on Cancer and Leukemia Group B (CALGB) treatment protocols and did not receive hematopoietic stem cell transplant (HSCT) in first complete response (CR) (Table S1 and SI Methods and Data). First, we identified 48 lncRNAs that were associated with (P < 0.001) event-free survival (EFS) by univariable Cox analysis (Table S7). The lncRNA score was derived as a linear combination of the expression of the 48 lncRNAs as described in Patients and Methods. We then divided the lncRNA score into quartiles. After creating the Kaplan–Meier plots for the quartiles, it was evident that the first quartile separated from the remaining quartiles and had a significantly better EFS. We therefore classified patients with scores in the first quartile as “favorable” and the rest as “unfavorable.” We then evaluated the lncRNA score using a dichotomized approach comparing unfavorable with favorable patients (Table 1). Interestingly, the majority of the 48 lncRNAs associated with survival do not associate with known prognostic gene mutations in older patients with CN-AML. Only three of them do: LOC728606 (up-regulated in NPM1-mutated patients) and CTA-392E5.1 and BC016361 (both down-regulated in patients with RUNX1-WT).
Table 1.
Treatment outcomes according to the lncRNA score status in 148 older (aged ≥60 y) patients with de novo CN-AML
End pointFavorable lncRNA scoreUnfavorable lncRNA score
(n = 37)(n = 111)P*OR/HR (95% CI)
Complete response, no. (%)33 (89)60 (54)<0.0010.14 (0.05–0.43)
Disease-free survival  <0.0012.88 (1.74–4.76)
 Median, y1.50.7 
 % disease-free at 3 y (95% CI)39 (23–55)7 (2–15) 
 % disease-free at 5 y (95% CI)36 (32–52)2 (0–8) 
Overall survival  <0.0012.95 (1.89–4.61)
 Median, y2.20.9 
 % alive at 3 y (95% CI)43 (27–58)10 (6–17) 
 % alive at 5 y (95% CI)41 (25–56)5 (2–10) 
CI, confidence interval; HR, hazard ratio; OR, odds ratio.
*
P values for categorical variables are from Fisher’s exact test, P values for the time to event variables are from the log-rank test.

Associations of the LncRNA Score with Clinical and Molecular Characteristics.

Patients with an unfavorable lncRNA score had higher white blood cell (WBC) counts (P = 0.004), and were more likely to have extramedullary involvement (P = 0.01; Table S1). Also, the unfavorable lncRNA score was associated with higher expression of MN1 (P = 0.003) and miR-3151 (P = 0.01; Table S1).

Prognostic Value of the LncRNA Score.

Compared with patients with a favorable lncRNA score, those with an unfavorable score had a lower CR rate [P < 0.001, odds ratio (OR) = 0.14, 54% vs. 89%)], and shorter DFS [P < 0.001, hazard ratio (HR) = 2.88; Fig. 2A and Table 1]. Three years after CR, only 7% of patients with unfavorable lncRNA were disease-free compared with 39% of patients with favorable lncRNA. The patients with an unfavorable lncRNA score also had shorter OS (P < 0.001, HR = 2.95; Fig. 2B and Table 1). The OS rate at 3 y was only 10% for patients with unfavorable lncRNA compared with 43% for patients with favorable lncRNA. In multivariable analyses, after adjusting for BAALC and miR-155 expression status, patients with an unfavorable lncRNA score had lower CR rates (P = 0.007, OR = 0.19) (Table 2). They also had shorter DFS (P < 0.001, HR = 3.23), after adjusting for the European LeukemiaNet (ELN) genetic group status (8), and shorter OS (P < 0.001, HR = 3.62), after adjusting for miR-155 expression and the ELN genetic group status (Table 2). All statistical analyses were also performed using the continuous lncRNA score, with similar results (SI Methods and Data shows the methods and results of the analyses of continuous lncRNA score).
Fig. 2.
Comparison of disease-free survival (DFS) and overall survival (OS) according to the lncRNA score in two datasets of older patients with CN-AML. Training dataset (n = 148) DFS (A) and OS (B). Validation dataset (n = 71) DFS (C) and OS (D).
Table 2.
Multivariable analysis of outcome according to the lncRNA score in older patients with CN-AML
  Complete responseDisease-free survivalOverall survival
VariableCategoriesOR95% CIPHR95% CIPHR95% CIP
LncRNA scoreUnfavorable vs. favorable0.190.06–0.630.0073.231.93–5.43<0.0013.622.26–5.78<0.001
BAALC expression groupHigh vs. low0.170.07–0.40<0.001
miR-155 expression groupHigh vs. low0.340.14–0.810.011.891.29–2.770.001
ELNFavorable vs. intermediate-I*0.50.32–0.790.0030.420.28–0.63<0.001
ELN, European LeukemiaNet; HR, hazard ratio; OR, odds ratio.
*
Favorable Genetic Group is defined by CEBPA mutated or FLT3-ITD negative and NPM1 mutated status; Intermediate-I Genetic Group is defined by patients that are not in the favorable category: CEBPA wild-type and either FLT3-ITD positive and NPM1 mutated, FLT3-ITD negative and NPM1 wild-type, or FLT3-ITD positive and NPM1 wild-type status (8).

Validation of the LncRNA Score by RNA-Seq.

To validate the lncRNA score, we performed lncRNA profiling using RNA-seq in pretreatment BM and peripheral blood (PB) samples from 71 older patients with de novo CN-AML. All patients were treated with intensive cytarabine/daunorubicin-based frontline therapy on CALGB protocols and did not receive HSCT in first CR. There were no significant baseline clinical and molecular differences between the training (n = 148) and validation datasets (n = 71) (Table S8). To validate the impact of lncRNA scores on CR, DFS, and OS, we considered 46 of the 48 survival-associated lncRNAs that were used in the training set, because two lncRNAs were not detected by RNA-seq. The RNA-seq score was derived as a linear combination of the expression of the 46 lncRNAs as described in Patients and Methods. Because we used a different platform (RNA-seq), we explored different approaches to select the best cut point (quartiles vs. median). After creating the Kaplan–Meier plots for quartiles, it was evident that the first and second quartiles grouped together and separated well from the third and fourth quartiles and had a significantly better EFS. We therefore classified patients with scores above the median as favorable and patients below the median as unfavorable. Patients with an unfavorable lncRNA score had a lower CR rate (P = 0.03, 63% vs. 86%) compared with patients with favorable lncRNA score and had shorter DFS (P = 0.009) and OS (P = 0.009) (Fig. 2 C and D).

Discussion

Despite the advances in understanding both the biology and risk stratification of adult AML, outcome is still poor, particularly in older patients aged 60 y or older (26). Whereas recent studies have reported that clinical outcome of older patients with CN-AML is influenced by several recurrent mutations, patients with apparently identical cytogenetic and molecular “makeup” do not have uniform outcomes (1119). This may result from additional undiscovered genetic alterations, which may identify smaller, clinically meaningful patient groups for which it is possible to identify specific therapeutic targets.
In this work, we investigated the associations of lncRNA expression with clinical characteristics, recurrent mutations, and outcome in older patients with CN-AML. We identified distinctive lncRNA expression signatures associated with the most common recurrent mutations in CN-AML. Patients with NPM1 mutations exhibit one of the strongest lncRNA signatures, as evidenced by higher lncRNA fold change and lower false discovery rates. It has been reported that NPM1-mutated CN-AML exhibits unique and strong mRNA and miRNA expression profiles characterized by HOX gene and miR-10 family overexpression and CD34 negativity (11, 47). Because ∼40% of lncRNAs intersect protein-coding loci and exhibit a strong pattern of coexpression (26), we found many up-regulated lncRNAs overlapping antisense transcripts of HOX genes (e.g., HOXB-AS3) in the NPM1-mutated lncRNA signature. A previous study has shown that the HOX antisense transcript, known as HOTAIR, is transcribed antisense to the HOXC locus and despite its genomic location, it has little effect on the regulation of the sense transcript (28). Instead, HOTAIR was shown to function in trans to negatively regulate HOXD via increased polycomb repressive complex 2 (PRC2) occupancy at the HOXD locus (28). A second group reported that HOX antisense intergenic RNA myeloid-specific 1 (HOTAIRM1) is transcribed antisense to the HOXA genes (between HOXA1 and HOXA2), shows myeloid-specific expression and is up-regulated during granulocytic differentiation (48). Knockdown of HOTAIRM1 quantitatively blunted retinoic acid-induced expression of HOXA1 and HOXA4 during the myeloid differentiation of NB4 cells and selectively attenuated induction of transcripts for the myeloid differentiation genes ITGAM (CD11b) and ITGB2 (CD18) (48). These findings suggest that HOTAIRM1 plays a role in myelopoiesis through modulation of gene expression in the HOXA cluster. Whereas the functional role of these HOX antisense transcripts associated with NPM1 mutations is unknown, we hypothesize that these transcripts may regulate overlapping HOX protein coding genes or other HOX genes located in different loci.
In addition, we report here lncRNA expression profiles associated with other frequent mutations in CN-AML, such as FLT3-ITD, CEBPA, IDH2, RUNX1, and ASXL1 mutations. Remarkably, no lncRNA profiles were found associated with two frequent mutations in older patients with CN-AML, DNMT3A and TET2, both of which are involved in epigenetic regulation (14, 22). The fact that lncRNA profiles were “mutually exclusive” with these mutations is suggestive of a shared effect within the epigenetic pathway. The association of the WT1-AS lncRNA with FLT3-ITD is interesting because the expression of the near WT1 protein-coding gene is also increased in this molecular subtype of CN-AML, and splicing alterations have been described for WT1-AS (41, 42). It remains to be investigated whether this event represents just a coexpression phenomenon or has a truly functional role.
Regardless of whether lncRNAs are functional or not, lncRNAs might have prognostic applications in AML. Several lncRNAs have been associated with outcome in solid tumors (2934). However, to our knowledge, there are no published reports for hematological malignancies, including AML. To minimize the impact of type of treatment in AML outcome interpretation, we analyzed a cohort of patients that were all treated with cytarabine/daunorubicin-based induction chemotherapy and did not receive HSCT in first CR. Using the lncRNAs most associated with EFS in a training cohort of 148 older patients with CN-AML, we derived a score of 48 lncRNAs. Patients with a favorable lncRNA score had a higher CR rate than those with an unfavorable score (89% vs. 54%; P < 0.001). These results were reproduced in the validation cohort, which was studied using a different technical platform. The identification of patients likely to achieve CR with standard therapy (7 + 3) alone represents a significant advance, potentially sparing them from other toxicities associated with experimental induction regimens and focusing investigational approaches on postremission (allogeneic HSCT and eradication of minimal residual disease) studies. Furthermore, patients with a favorable lncRNA score exhibit longer DFS and OS than patients with an unfavorable lncRNA score in both the training and validation cohorts. Multivariable analyses for CR, DFS, and OS identified the lncRNA score as an independent marker for outcome. It is remarkable how the different prognostic mutations associate with distinct lncRNA profiles. On the other hand, expression levels of the prognostic lncRNAs mostly do not associate with prognostic mutations. Indeed, the prognostic lncRNA score has divided older patients with CN-AML into two groups, which, despite having very similar frequencies of FLT3-ITD and NPM1 and CEBPA mutations (all strong prognosticators in CN-AML), have had significantly different outcomes. Overall, these data indicate that lncRNAs could be used to predict treatment response and outcome in older patients with CN-AML. Further studies will be needed to investigate the prognostic impact of lncRNAs in younger adults (aged 15–59 y) and children (under 15 y) with de novo CN-AML.
In summary, here we describe lncRNA profiles in older patients with CN-AML and their associations with clinical features, mutations, and outcome. Distinctive lncRNA signatures were found associated with recurrent mutations in CN-AML. Future work is needed to investigate whether these lncRNAs are functional and whether they are regulated directly or indirectly by biologic changes promoted by these specific mutations. Additionally, we identified a lncRNA score that is strongly associated with outcome, including treatment response to standard intensive chemotherapy in older patients with CN-AML. This score could be used to prospectively identify those patients who are likely to have high CR rates with standard therapy, while at the same time treating the remaining patients with novel combination treatment approaches using targeted therapies to improve their CR rates.

Patients and Methods

Patient Set.

Pretreatment frozen BM samples from 148 older (aged ≥60 y) patients with de novo CN-AML (training set) and 71 pretreatment frozen samples (52 BM and 19 PB) from older patients with de novo CN-AML (validation set) were obtained from the CALGB/Alliance leukemia tissue bank. Informed consent was obtained from the patients in accordance with the Declaration of Helsinki to procure and bank the cells for future research according to the institutional review boards of each of the participating institutions. Patient characteristics are shown in Tables S1 and S8. Cytogenetic analyses were performed in CALGB/Alliance approved institutional laboratories and confirmed by central karyotype review (49), and the diagnosis of normal karyotype was based on ≥20 metaphase cells analyzed in BM specimens subjected to short-term (24 h or 48 h) unstimulated cultures. Molecular studies shown in Tables S1 and S8 were performed as described in SI Methods and Data and elsewhere (1118). The patients in the training and validation sets were treated with intensive cytarabine/daunorubicin-based therapy on one of the CALGB frontline treatment protocols as described in SI Methods and Data.

Statistical Analyses.

Definitions of clinical end points are provided in SI Methods and Data. The associations of the lncRNA score with baseline clinical, demographic, and molecular features were compared using the Wilcoxon rank sum and Fisher’s exact tests for continuous and categorical variables, respectively. Estimated probabilities of DFS and OS were calculated using the Kaplan–Meier method, and the log-rank test evaluated differences between survival distributions. Logistic regression was used for CR to calculate univariable ORs and the Cox proportional hazards model was used for DFS and OS to calculate univariable HRs. All statistical analyses were performed by the Alliance for Clinical Trials in Oncology Statistics and Data Center using SAS 9.3 and TIBCO Spotfire S+ 8.2.

Multivariable Analyses.

Multivariable logistic regression models were generated for attainment of CR, and multivariable proportional hazards models were constructed for DFS and OS, using a limited forward selection procedure. Variables significant at α = 0.20 from the univariable analyses were considered for multivariable analyses (see SI Methods and Data for the list of variables considered for model inclusion). For the time-to-event end points, the proportional hazards assumption was checked for each variable individually.

Data Availability

Data deposition: The data reported in this paper have been deposited in the Gene Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession nos. GSE63646 and GSE63614).

Acknowledgments

This work was supported in part by the National Cancer Institute Grants P50-CA140158-05 (to R.G., W.B., G.M., J.C.B., and C.D.B.), CA101140 (to G.M., C.D.B., and M.A.C.), CA180861 (to G.M., and C.D.B.), Brigham and Women's Hospital BWH 108415 and 110414 (to M.A.C. and G.M.), the Coleman Leukemia Research Foundation, and the Pelotonia Fellowship Program (A.-K.E.). S.V. was supported by Associazione Italiana Ricerca sul Cancro (IG 13585), and Ministero della Istruzione Università e Ricerca (PRIN 2010).

Supporting Information

Supporting Information (PDF)
Supporting Information
pnas.1422050112.st01.docx
pnas.1422050112.st02.docx
pnas.1422050112.st03.docx
pnas.1422050112.st04.docx
pnas.1422050112.st05.docx
pnas.1422050112.st06.docx
pnas.1422050112.st07.docx
pnas.1422050112.st08.docx

References

1
Ries LAG, et al. (eds) (2008) SEER Cancer Statistics Review, 1975–2005. National Cancer Institute, Bethesda, MD. Available at seer.cancer.gov/csr/1975_2005/, based on November 2007 SEER data submission, posted to the SEER website, 2008. Accessed October 17, 2014.
2
CP Leith, et al., Acute myeloid leukemia in the elderly: Assessment of multidrug resistance (MDR1) and cytogenetics distinguishes biologic subgroups with remarkably distinct responses to standard chemotherapy. A Southwest Oncology Group study. Blood 89, 3323–3329 (1997).
3
SS Farag, et al., Pretreatment cytogenetics add to other prognostic factors predicting complete remission and long-term outcome in patients 60 years of age or older with acute myeloid leukemia: Results from Cancer and Leukemia Group B 8461. Blood 108, 63–73 (2006).
4
S Fröhling, et al., Cytogenetics and age are major determinants of outcome in intensively treated acute myeloid leukemia patients older than 60 years: Results from AMLSG trial AML HD98-B. Blood 108, 3280–3288 (2006).
5
H Kantarjian, et al., Results of intensive chemotherapy in 998 patients age 65 years or older with acute myeloid leukemia or high-risk myelodysplastic syndrome: Predictive prognostic models for outcome. Cancer 106, 1090–1098 (2006).
6
E Estey, Acute myeloid leukemia and myelodysplastic syndromes in older patients. J Clin Oncol 25, 1908–1915 (2007).
7
T Büchner, et al., Age-related risk profile and chemotherapy dose response in acute myeloid leukemia: A study by the German Acute Myeloid Leukemia Cooperative Group. J Clin Oncol 27, 61–69 (2009).
8
H Döhner, et al., 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 (2010).
9
RF Schlenk, et al., Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. N Engl J Med 358, 1909–1918 (2008).
10
K Mrózek, G Marcucci, P Paschka, SP Whitman, CD Bloomfield, Clinical relevance of mutations and gene-expression changes in adult acute myeloid leukemia with normal cytogenetics: Are we ready for a prognostically prioritized molecular classification? Blood 109, 431–448 (2007).
11
H Becker, et al., Favorable prognostic impact of NPM1 mutations in older patients with cytogenetically normal de novo acute myeloid leukemia and associated gene- and microRNA-expression signatures: A Cancer and Leukemia Group B study. J Clin Oncol 28, 596–604 (2010).
12
SP Whitman, et al., FLT3 internal tandem duplication associates with adverse outcome and gene- and microRNA-expression signatures in patients 60 years of age or older with primary cytogenetically normal acute myeloid leukemia: A Cancer and Leukemia Group B study. Blood 116, 3622–3626 (2010).
13
KH Metzeler, et al., ASXL1 mutations identify a high-risk subgroup of older patients with primary cytogenetically normal AML within the ELN Favorable genetic category. Blood 118, 6920–6929 (2011).
14
G Marcucci, et al., Age-related prognostic impact of different types of DNMT3A mutations in adults with primary cytogenetically normal acute myeloid leukemia. J Clin Oncol 30, 742–750 (2012).
15
JH Mendler, et al., RUNX1 mutations are associated with poor outcome in younger and older patients with cytogenetically normal acute myeloid leukemia and with distinct gene and MicroRNA expression signatures. J Clin Oncol 30, 3109–3118 (2012).
16
S Schwind, et al., BAALC and ERG expression levels are associated with outcome and distinct gene and microRNA expression profiles in older patients with de novo cytogenetically normal acute myeloid leukemia: A Cancer and Leukemia Group B study. Blood 116, 5660–5669 (2010).
17
S Schwind, et al., Low expression of MN1 associates with better treatment response in older patients with de novo cytogenetically normal acute myeloid leukemia. Blood 118, 4188–4198 (2011).
18
C Niederwieser, et al., Prognostic and biologic significance of DNMT3B expression in older patients with cytogenetically normal primary acute myeloid leukemia. Leukemia, 2014).
19
R Garzon, G Marcucci, CM Croce, Targeting microRNAs in cancer: Rationale, strategies and challenges. Nat Rev Drug Discov 9, 775–789 (2010).
20
G Marcucci, et al., MicroRNA expression in cytogenetically normal acute myeloid leukemia. N Engl J Med 358, 1919–1928 (2008).
21
R Garzon, et al., MicroRNA signatures associated with cytogenetics and prognosis in acute myeloid leukemia. Blood 111, 3183–3189 (2008).
22
KH Metzeler, et al., TET2 mutations improve the new European LeukemiaNet risk classification of acute myeloid leukemia: A Cancer and Leukemia Group B study. J Clin Oncol 29, 1373–1381 (2011).
23
G Marcucci, et al., Clinical role of microRNAs in cytogenetically normal acute myeloid leukemia: miR-155 upregulation independently identifies high-risk patients. J Clin Oncol 31, 2086–2093 (2013).
24
A-K Eisfeld, et al., miR-3151 interplays with its host gene BAALC and independently affects outcome of patients with cytogenetically normal acute myeloid leukemia. Blood 120, 249–258 (2012).
25
P Kapranov, et al., RNA maps reveal new RNA classes and a possible function for pervasive transcription. Science 316, 1484–1488 (2007).
26
T Derrien, et al., The GENCODE v7 catalog of human long noncoding RNAs: Analysis of their gene structure, evolution, and expression. Genome Res 22, 1775–1789 (2012).
27
I Ulitsky, DP Bartel, lincRNAs: Genomics, evolution, and mechanisms. Cell 154, 26–46 (2013).
28
JL Rinn, et al., Functional demarcation of active and silent chromatin domains in human HOX loci by noncoding RNAs. Cell 129, 1311–1323 (2007).
29
RA Gupta, et al., Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 464, 1071–1076 (2010).
30
R Kogo, et al., Long noncoding RNA HOTAIR regulates polycomb-dependent chromatin modification and is associated with poor prognosis in colorectal cancers. Cancer Res 71, 6320–6326 (2011).
31
YJ Geng, SL Xie, Q Li, J Ma, GY Wang, Large intervening non-coding RNA HOTAIR is associated with hepatocellular carcinoma progression. J Int Med Res 39, 2119–2128 (2011).
32
ZY Xu, et al., Knockdown of long non-coding RNA HOTAIR suppresses tumor invasion and reverses epithelial-mesenchymal transition in gastric cancer. Int J Biol Sci 9, 587–597 (2013).
33
K Kim, et al., HOTAIR is a negative prognostic factor and exhibits pro-oncogenic activity in pancreatic cancer. Oncogene 32, 1616–1625 (2013).
34
P Ji, et al., MALAT-1, a novel noncoding RNA, and thymosin β4 predict metastasis and survival in early-stage non-small cell lung cancer. Oncogene 22, 8031–8041 (2003).
35
RGW Verhaak, et al., Mutations in nucleophosmin (NPM1) in acute myeloid leukemia (AML): Association with other gene abnormalities and previously established gene expression signatures and their favorable prognostic significance. Blood 106, 3747–3754 (2005).
36
Y Takahashi, et al., Amplification of PVT-1 is involved in poor prognosis via apoptosis inhibition in colorectal cancers. Br J Cancer 110, 164–171 (2014).
37
Y Guan, et al., Amplification of PVT1 contributes to the pathophysiology of ovarian and breast cancer. Clin Cancer Res 13, 5745–5755 (2007).
38
W Yin, A Rossin, JL Clifford, H Gronemeyer, Co-resistance to retinoic acid and TRAIL by insertion mutagenesis into RAM. Oncogene 25, 3735–3744 (2006).
39
RF Schlenk, et al., Gene mutations and response to treatment with all-trans retinoic acid in elderly patients with acute myeloid leukemia. Results from the AMLSG Trial AML HD98B. Haematologica 94, 54–60 (2009).
40
K Moorwood, et al., Antisense WT1 transcription parallels sense mRNA and protein expression in fetal kidney and can elevate protein levels in vitro. J Pathol 185, 352–359 (1998).
41
BV Spassov, et al., Wilms’ tumor protein and FLT3-internal tandem duplication expression in patients with de novo acute myeloid leukemia. Hematology 16, 37–42 (2011).
42
AR Dallosso, et al., Alternately spliced WT1 antisense transcripts interact with WT1 sense RNA and show epigenetic and splicing defects in cancer. RNA 13, 2287–2299 (2007).
43
Y Wang, et al., Long non-coding RNA UCA1a(CUDR) promotes proliferation and tumorigenesis of bladder cancer. Int J Oncol 41, 276–284 (2012).
44
Z Fang, et al., Increased expression of the long non-coding RNA UCA1 in tongue squamous cell carcinomas: A possible correlation with cancer metastasis. Oral Surg Oral Med Oral Pathol Oral Radiol 117, 89–95 (2014).
45
M Negishi, et al., A new lncRNA, APTR, associates with and represses the CDKN1A/p21 promoter by recruiting polycomb proteins. PLoS ONE 9, e95216 (2014).
46
R Garzon, et al., Distinctive microRNA signature of acute myeloid leukemia bearing cytoplasmic mutated nucleophosmin. Proc Natl Acad Sci USA 105, 3945–3950 (2008).
47
X Zhang, et al., A myelopoiesis-associated regulatory intergenic noncoding RNA transcript within the human HOXA cluster. Blood 113, 2526–2534 (2009).
48
VA Kickhoefer, et al., Vaults are up-regulated in multidrug-resistant cancer cell lines. J Biol Chem 273, 8971–8974 (1998).
49
K Mrózek, et al., Central review of cytogenetics is necessary for cooperative group correlative and clinical studies of adult acute leukemia: The Cancer and Leukemia Group B experience. Int J Oncol 33, 239–244 (2008).

Information & Authors

Information

Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 111 | No. 52
December 30, 2014
PubMed: 25512507

Classifications

Data Availability

Data deposition: The data reported in this paper have been deposited in the Gene Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession nos. GSE63646 and GSE63614).

Submission history

Published online: December 15, 2014
Published in issue: December 30, 2014

Keywords

  1. lncRNAs
  2. acute myeloid leukemia
  3. outcome

Acknowledgments

This work was supported in part by the National Cancer Institute Grants P50-CA140158-05 (to R.G., W.B., G.M., J.C.B., and C.D.B.), CA101140 (to G.M., C.D.B., and M.A.C.), CA180861 (to G.M., and C.D.B.), Brigham and Women's Hospital BWH 108415 and 110414 (to M.A.C. and G.M.), the Coleman Leukemia Research Foundation, and the Pelotonia Fellowship Program (A.-K.E.). S.V. was supported by Associazione Italiana Ricerca sul Cancro (IG 13585), and Ministero della Istruzione Università e Ricerca (PRIN 2010).

Authors

Affiliations

Ramiro Garzon2,1 [email protected]
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Stefano Volinia1
Department of Morphology, Surgery, and Experimental Medicine, University of Ferrara, Ferrara, FE 44121, Italy;
Dimitrios Papaioannou
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Deedra Nicolet
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Alliance for Clinical Trials in Oncology Statistics and Data Center, Mayo Clinic, Rochester, MN 55905;
Jessica Kohlschmidt
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Alliance for Clinical Trials in Oncology Statistics and Data Center, Mayo Clinic, Rochester, MN 55905;
Pearlly S. Yan
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Krzysztof Mrózek
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Donna Bucci
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Andrew J. Carroll
Department of Genetics, University of Alabama at Birmingham, Birmingham, AL 35294;
Maria R. Baer
Greenebaum Cancer Center, University of Maryland, Baltimore, MD 21201;
Meir Wetzler
Roswell Park Cancer Institute, Buffalo, NY 14263;
Thomas H. Carter
Holden Comprehensive Cancer Center, University of Iowa, Iowa City, IA 52242;
Bayard L. Powell
The Comprehensive Cancer Center of Wake Forest University, Winston-Salem, NC 27157;
Jonathan E. Kolitz
North Shore Cancer Institute, Lake Success, NY 11042;
Joseph O. Moore
Department of Medicine, Division of Hematology-Oncology, Duke University and Durham VA Medical Centers, Durham, NC 27701; and
Ann-Kathrin Eisfeld
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
James S. Blachly
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
William Blum
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Michael A. Caligiuri
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Richard M. Stone
Dana-Farber Cancer Institute, Harvard University, Boston, MA 02215
Guido Marcucci
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Carlo M. Croce2 [email protected]
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
John C. Byrd
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;
Clara D. Bloomfield2 [email protected]
The Ohio State University Comprehensive Cancer Center, Columbus, OH 43210;

Notes

2
To whom correspondence may be addressed. Email: [email protected], [email protected], or [email protected].
Author contributions: R.G., S.V., D.P., D.N., J.K., M.A.C., C.M.C., J.C.B., and C.D.B. designed research; R.G., S.V., D.P., D.N., J.K., P.S.Y., D.B., A.-K.E., J.S.B., and C.D.B. performed research; A.J.C., M.R.B., M.W., T.H.C., B.L.P., J.E.K., J.O.M., W.B., R.M.S., and G.M. contributed new reagents/analytic tools; R.G., S.V., D.P., D.N., J.K., K.M., A.J.C., M.R.B., A.-K.E., J.S.B., W.B., M.A.C., C.M.C., J.C.B., and C.D.B. analyzed data; and R.G. wrote the paper.
Reviewers: E.H.E., Seattle Cancer Center Alliance; and F.L., Ospedale Bambino Gesù, Roma and Università di Pavia.
1
R.G. and S.V. contributed equally to this work.

Competing Interests

The authors declare no conflict of interest.

Metrics & Citations

Metrics

Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.


Citation statements




Altmetrics

Citations

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by

    Loading...

    View Options

    View options

    PDF format

    Download this article as a PDF file

    DOWNLOAD PDF

    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Expression and prognostic impact of lncRNAs in acute myeloid leukemia
    Proceedings of the National Academy of Sciences
    • Vol. 111
    • No. 52
    • pp. 18401-18799

    Media

    Figures

    Tables

    Other

    Share

    Share

    Share article link

    Share on social media