Distinct clinical and biological characteristics of acute myeloid leukemia with higher expression of long noncoding RNA KIAA0125

Expression of long non-coding RNA KIAA0125 has been incorporated in various gene expression signatures for prognostic prediction in acute myeloid leukemia (AML) patients, yet its functions and clinical significance remain unclear. This study aimed to investigate the clinical and biological characteristics of AML bearing different levels of KIAA0125. We profiled KIAA0125 expression levels in bone marrow cells from 347 de novo AML patients and found higher KIAA0125 expression was closely associated with RUNX1 mutation, but inversely correlated with t(8;21) and t(15;17) karyotypes. Among the 227 patients who received standard chemotherapy, those with higher KIAA0125 expression had a lower complete remission rate, shorter overall survival (OS) and disease-free survival (DFS) than those with lower expression. The prognostic significance was validated in both TCGA and GSE12417 cohorts. Subgroup analyses showed that higher KIAA0125 expression also predicted shorter DFS and OS in patients with normal karyotype or non-M3 AML. In multivariable analysis, higher KIAA0125 expression remained an adverse risk factor independent of age, WBC counts, karyotypes, and mutation patterns. Bioinformatics analyses revealed that higher KIAA0125 expression was associated with hematopoietic and leukemic stem cell signatures and ATP-binding cassette transporters, two predisposing factors for chemoresistance.


Introduction
Long non-coding RNAs (lncRNAs) are non-protein coding RNAs that are longer than 200 nucleotides. Comparing to other classes of ncRNAs, lncRNAs exhibit a wide range of structures and functions [1]. Recently, lncRNAs have emerged as important regulators for gene expression via remodeling nuclear architecture, modulating mRNA stability and translation, and post-translational modifications [1][2][3][4]. Besides, some lncRNAs are dysregulated and harbor prognostic relevance in several types of cancers [5][6][7][8]. However, the roles of lncRNAs in tumorigenesis are still largely unknown.
KIAA0125 (also named as FAM30A), a hematopoietic stem cell gene localized on chromosome 14, is unique because it is the only non-coding gene and is expressed in humans but not in mice (From the UniProt database, https://www.uniprot. org/uniprot/Q9NZY2). Additionally, KIAA0125 expression was integrated into a recently proposed 17-gene stemness score, which could predict outcomes in AML patients [19].
This study aimed to investigate the association of KIAA0125 expression with clinical and biological characteristics in AML patients. We first profiled the expression levels of KIAA0125 in bone marrow (BM) cells from AML patients and normal controls and demonstrated that AML patients had higher KIAA0125 expression than normal controls. Higher expression of KIAA0125 was associated with distinct clinical and biological characteristics and served as an independent poor prognostic biomarker for AML patients in ours and two other publicly annotated cohorts. Further bioinformatics analyses showed that higher expression of KIAA0125 in AML was closely associated with hematopoietic stem cell (HSC) and leukemic stem cell (LSC) signatures and several important ATP-binding cassette transporters (ABC transporters); these factors are regarded responsible for chemoresistance in AML. Further functional studies are needed to unravel its underlying mechanism and pathogenetic role in AML.

Patients
We recruited 347 adult patients with de novo AML diagnosed in the National Taiwan University Hospital (NTUH) from 1996 to 2011 who had enough cryopreserved BM cells for tests. The diagnoses were based on the French-American-British (FAB) and the 2016 World Health Organization classifications [25,26]. Among them, 227 patients received standard chemotherapy. Non M3 (acute promyelocytic leukemia, APL) patients received idarubicin 12 mg/m 2 per day days 1-3 and cytarabine 100 mg/m 2 per day days 1-7, and then consolidation chemotherapy with 2-4 courses of high-dose cytarabine 2000 mg/m 2 q12h for total 8 doses, with or without an anthracycline (Idarubicin or Mitoxantrone), after achieving complete remission (CR) as described previously [27]. APL patients received concurrent all-trans retinoic acid and chemotherapy. The remaining 120 patients received supportive care and/or reduced-intensity anti-leukemia therapy due to underlying comorbidities or based on the decision of the physicians or patients. BM samples from 30 healthy donors of hematopoietic stem cell transplantation (HSCT) were collected as normal controls. This study was approved by the Research Ethics Committee of NTUH with informed consent obtained from all participants.

Analysis of gene expression in next-generation sequencing datasets
We analyzed gene expression data of 141 AML samples profiled with Illumina Genome Analyzer RNA Sequencing in the TCGA database [30] to investigate the absolute gene expression levels.

Gene set enrichment analysis
The preranked Gene Set Enrichment Analysis (GSEA) implemented by R package clusterProfiler was performed using the stem cell-related gene sets from the MSigDB databases. The genes were ranked based on the Spearman's correlation coefficient between the given gene and KIAA0125.

Statistical analysis
We used the Mann-Whitney U test and ANOVA test, where appropriate, to compare continuous variables and medians/ means of distributions. The Fisher exact test or the χ2 test was performed to examine the difference in discrete variables, including gender, cytogenetic changes, and genetic alterations between patients with lower and higher KIAA0125 expression. Overall survival (OS) was the duration from the date of initial diagnosis to the time of last follow-up or death from any cause, whichever occurred first. Disease-free survival (DFS) was the duration from the date of attaining a leukemia-free state until the date of AML relapse or death from any cause, whichever occurred first. The survival prediction power of KIAA0125 expression was evaluated by both the log-rank test and the univariate Cox proportional hazards model. We plotted the survival curves with Kaplan-Meier analysis and calculated the statistical significance with the log-rank test. To find the optimal cutoff for separating patient groups, we used maximally selected rank statistics implemented in the maxstat R package. The Cox proportional hazards model was used in multivariable regression analysis. P values < 0.05 were considered statistically significant. All statistical analyses were performed with BRB-ArrayTools (version 4.5.1; Biometric Research Branch, National Cancer Institute, Rockville, MD), and IBM SPSS Statistics 23 for Windows.

Results
The median age of the 347 AML patients was 57 years. Among the 331 patients who had cytogenetic data at diagnosis, 165 (49.8%) had clonal chromosomal abnormalities. Sixty patients (18.1%) had favorable cytogenetics; 223 (67.2%), intermediate-risk cytogenetics; and 14.8% unfavorable cytogenetics (Supplement Table 1) based on the refined British Medical Research Council (MRC) classification [35]. The clinical and laboratory characteristics of these patients at diagnosis are summarized in Table 1.

Comparison of clinical characteristics and genetic alterations between patients with higher and lower KIAA0125 expression
The distribution of KIAA0125 expression of 347 AML patients is shown with dot plots in Supplement Fig. 1. We first compared the BM KIAA0125 expression between the 30 healthy controls and 347 AML patients. The expression of KIAA0125 was significantly higher in AML samples than healthy controls (p < 0.001, Fig. 1a). Then, the 347 AML patients were divided into two groups by the median value of the KIAA0125 expression. The comparison of clinical and laboratory features between the two groups is shown in Table 1. The higher-KIAA0125 group had higher circulating blasts at diagnosis (p = 0.021) and higher incidence of FLT3-ITD in the absence of NPM1 mutation (NPM1-/FLT3-ITD+) (p = 0.002) and RUNX1 mutation (p = 0.034), but lower incidence of t(8;21) and t(15;17) (both p < 0.001), compared with the lower-KIAA0125 group (Table 1). From another perspective, patients with t(8;21) or t(15;17) had lower KIAA0125 expression, whereas those with RUNX1 mutation, ASXL1 mutation, NPM1-/FLT3-ITD+, or unfavorable karyotypes had higher expression of KIAA0125 (F = 15.124, p < 0.001, Fig. 1b, Supplement Table 1 and Supplement Table 2). Furthermore, the association of higher-KIAA0125 with lower frequencies of t(8;21) and t(15;17) was observed in both the NTUH cohort (both p < 0.001, Supplement Table 3) and TCGA cohort (p = 0.006 and p < 0.001, respectively, Supplement Table 3). The higher-KIAA0125 patients more frequently had FLT3-ITD (p = 0.048) and mutations in DNMT3A (p = 0.015) and RUNX1 (p = 0.034) (Supplement Table 4). Compatible with this finding, patients with DNMT3A or RUNX1 mutation had higher KIAA0125 expression than those without the mutation (p = 0.019 and 0.045, respectively, Supplement Fig. 2). Similarly, there was close association between higher KIAA0125 expression and DNMT3A (p = 0.001) and RUNX1 mutations (p = 0.017) in the TCGA cohort (Supplement Table 5). Among the 227 patients who received standard chemotherapy, 165 (72.7%) patients attained a complete remission (CR), while 42 (18.5%) patients had primary refractory diseases. Notably, the patients with higher KIAA0125 expression had a lower CR rate (61.2% vs. 84.7%, p < 0.001) than those with lower expression. In accordance with this finding, the patients who achieved CR after induction chemotherapy had lower expression of BM KIAA0125 at diagnosis than those who did not (p < 0.001, Fig. 1c).

The impacts of the KIAA0125 expression on OS and DFS
Next, we divided patients into two groups with high and low KIAA0125 expression with cut points determined by the maximally selected rank statistics (7.72 in the NTUH cohort, 8.56 in the TCGA cohort, and 9.71 in GSE12417 cohort, respectively, Supplement Fig. 3). As expected, patients with higher KIAA0125 expression had an inferior DFS and OS than those with lower expression, no matter whether the survival was censored on the day of hematopoietic stem cell transplantation (HSCT) (median, 3.2 months vs. 31.7 months, p < 0.001; and 17 months vs. not reached (NR), p < 0.001, respectively; Fig.  2a and b) or not (p < 0.001 and p < 0.001, respectively; Supplement Fig. 4a and 4b). Subgroup analyses showed that the prognostic significance of KIAA0125 expression for DFS and OS remained valid in both non-APL and normal karyotype patients (Figs. 2c and d).
In multivariable analysis, we included clinically relevant parameters and variables with a p value < 0.05 in univariate Cox regression analysis (Supplement Table 4) as covariates, including age, white blood cell counts at diagnosis, karyotypes, mutation statuses of NPM1/FLT3-ITD, CEBPA double mutations , RUNX1, MLL-PTD, and TP53, and KIAA0125 expression. Higher KIAA0125 expression, either divided by the selected cut-point (Table 2) or calculated as continuous values (Supplement Table 5), was an independent adverse prognostic factor for DFS (p < 0.001 and p < 0.001, respectively) and OS (p = 0.003 and p = 0.001, respectively). To verify the prognostication power of the KIAA0125 expression, we analyzed the expression of KIAA0125 and its prognostic significance in the TCGA cohort and the GSE12417-GPL96 cohort. Consistent with the findings in the NTUH cohort, patients with higher KIAA0125 expressions had a significantly shorter OS

Discussion
AML cells have abnormal genetic background, either mutations or aberrant expression of specific genes. In recent years, several gene expression scores have been proposed for prognostic prediction of AML patients. We previously developed a 11-gene mRNA expression signature, including AIF1L, CXCR7, DNTT, GPR56, H1F0, IFITM3, KIAA0125, MX1, STAB1, TM4SF1, and TNS3, for prognostication in AML patients [21]. Another group built a six-gene leukemia stem cell (LSC) score with the incorporation of DNMT3B, GPR56, CD34, SOCS2, SPINK2, and KIAA0125 expressions for pediatric AML [40]. Recently, Ng et al. proposed a 17-gene LSC score that incorporated expressions of 17 stemness-related genes, including KIAA0125, and showed the scoring system was powerful to predict prognosis in AML patients [19]. Among these prognostic-relevant genes, KIAA0125 is the only non-coding gene and expressed only in the Homo sapiens, but not in mice.
KIAA0125 is located on chromosome 14 of the human genome. It was reported to be upregulated in ameloblastoma but shown as a tumor suppressor gene in colorectal cancer [59,   In this study, we found that the expression level of KIAA0125 in BM was significantly higher in AML patients than normal HSC transplant donors. The expression of KIAA0125 was lower in patients with t(8;21) and t (15;17) which are associated with more differentiated AML subtypes, but higher in patients with RUNX1, ASXL1 mutations, NPM1-/FLT3-ITD+ or poor-risk karyotypes. It is interesting that the expression of KIAA0125 was high in patients with RUNX1 mutation but modest in those with RUNX1/ RUNX1T1 fusion consisting with the fact that AML patients with a RUNX1 mutation usually had poor outcomes while those with RUNX1/RUNX1T1 fusion had favorable prognosis. Recently, Hornung et al. identified that expression of CD109, HOPX, and KIAA0125 genes might be responsible for inferior survival in AML patients with RUNX1 mutations but, on the other hand, better outcome in RUNX1/RUNX1T1 fusion through a newly proposed statistical tool "mediation analysis." The three genes' expression levels were significantly higher in patients with RUNX1 mutant but lower in those with RUNX1/RUNX1T1 fusion [61]. Intriguingly, though there has been no study showing direct evidence that RUNX1 binds to KIAA0125 till now in the literature, RUNX1 has been reported to bind to TGTGG core sequences as a heterodimer of RUNX1 and CBFβ [62]. We downloaded and retrieved the DNA sequence of KIAA0125 from the UCSC Genome Browser (https://genome.ucsc.edu/) and found several sequences of TGTGG (Supplement Table 8) within the 3000 bp upstream sequence, which might be the potential binding sites of RUNX1. Further studies are needed to explore the effect of the possible interaction between RUNX1 domain and KIAA0125.
Interestingly, the expression levels of several ABC transporter genes, including ABCA2, ABCB1, ABCC1, and ABCG1, were also significantly higher in AML patients with higher KIAA0125 expression. The ABC transporter family consists of 48 proteins in subfamilies designated A to G and some of them are known to be associated with multidrug resistance via ATP-dependent drug efflux [53,54,57]. ABCB1, ABCC1, and ABCG1 were reported to be responsible for chemoresistance in AML [53,56]. The translational expression of ABCA2 was shown to be a prognostic marker for drug resistance in pediatric acute lymphoblastic leukemia [55,58]. The underlying mechanistic basis of the high correlation of these 4 genes to the expression of KIAA0125 warrants further studies. This study's limitations lie in its retrospective nature and, crucially, the unsorted BM sample, as many cells in BM may be differentiated cells of myeloid and erythroid lineages. The study could have been more informative if we could profile KIAA0125 expression of healthy CD34 + CD38-HSCs and more mature progenitors (CD34 + CD38-and CD34-CD117+, respectively) and compare those with leukemia blasts. Moreover, the putative oncogenic role of KIAA0125 expression in both the NTUH and TCGA cohorts (pink). The correlation measurement is based on the Spearman's correlation coefficient between the given gene and KIAA0125. The strongly correlated genes are defined as their correlation values at top 5% of all genes in both cohorts could be more strengthened were the expressions of KIAA0125 investigated in AML stem cells and bulk. Despite the limitations mentioned, to the best of our knowledge, this is by far the first study specifically addressing the expression of lncRNA KIAA0125 and its clinical and biological associations in AML patients. We found that higher KIAA0125 expression was closely associated with RUNX1 and DNMT3A1 mutations in both the NTUH and TCGA cohorts. Patients with higher KIAA0125 expression were more refractory to chemotherapy with a lower CR rate and higher refractory rate (Table 1). They had shorter OS and DFS among the total cohort and subgroups of patients with non-APL and those with normal karyotype. Based on its crucial clinical significance, further experimental studies are necessary to delineate how KIAA0125 participates in the stem cell biology of hematopoietic lineages and its role in the pathogenesis in AML. Authorship contributions YHW and CCL contribute equally to this study. YHW and CCL were responsible for data collection and management, statistical analysis and interpretation, literature research, and manuscript writing; SYH and CYY were responsible for data management and statistical analysis; CLH assisted in statistical analysis; SHL, CHT, and HAH were responsible for data collection and management; and WCC and HFT planned, designed, and coordinated the study over the entire period and wrote the manuscript.
Funding Data availability statement The raw and normalized microarray data reported in this article have been deposited in the Gene Expression Omnibus database (accession number GSE68469 and GSE71014).

Compliance with ethical standards
Conflict of interest The authors declare that they have no relevant competing financial interests.
Informed consent This study was approved by the Research Ethics Committee of NTUH with informed consent obtained from all individual participants included in the study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.