Evaluation of the HOXA9 and MEIS1 genes as a potential biomarker in adult acute myeloid leukemia

Acute myeloid leukemia (AML) is a heterogeneous disorder encompassing a set of hematopoietic tumors that develop when the myeloid precursor cells undergo disproportionate clonal proliferation. Homeobox A 9 (HOXA9) is a pioneer transcription factor in AML pathogenesis along with its cofactor myeloid ecotropic integration site 1 (MEIS1). Our work aimed to evaluate the different expression levels of HOXA9 and MEIS1 genes and their diagnostic and prognostic significance in adult Egyptian patients with de novo AML. The study was carried out on 91 de novo AML Egyptian patients and 41 healthy individuals. Bone marrow samples were obtained from both patients and controls and then tested by reverse transcription-quantitative polymerase chain reaction to assess the mRNA expression in the studied genes. HOXA9 and MEIS1 gene expression levels were significantly elevated in AML patients compared to controls (p < 0.001). There was a statistically significant positive correlation between HOXA9 and MEIS1 gene expression in AML patients. However, there was no association between HOXA9 and MEIS1 gene expression levels and disease-free survival (DFS) and overall survival (OS) (p = 0.264 and 0.351, respectively). HOXA9 and MEIS1 genes are highly expressed in Egyptian AML patients, suggesting their interesting pathogenic role in AML. They could be used as markers for the diagnosis of AML, but not for the disease prognosis.


Background
Acute myeloid leukemia (AML) is a malignant hematological disorder in which immature clonal myeloid cells excessively proliferate and show arrested differentiation [1]. There is a continuous increase in AML incidence. The risk of AML is higher in males than in females [2]. AML is the most common type of acute leukemia in adults and is associated with the highest leukemia-related mortality rate [3].
A deep understanding of AML biology has shown some progress, but it is sluggish to formulate this knowledge into reliable therapy protocols. Therefore, a comprehensive study of molecular mechanisms associated with the proliferation, carcinogenesis, and recurrence of AML is crucial for developing effective therapeutic and prognostic strategies [4].
The transcription factor genes necessary for normal blood cell development are the most frequently targeted sites of genetic alterations in leukemia. Among these transcription factors are the homeobox A9 (HOXA9) and the myeloid ecotropic integration site 1 (MEIS1) genes. They control progenitor cell regeneration during hematopoiesis and leukemogenesis [5,6]. The expression of HOXA9 and MEIS1 genes leads to the transition into the cluster of differentiation 34 (CD34) stage of initial differentiation of progenitors [7]. AML is commonly featured by HOXA9 and MEIS1 gene expression, which are dominant collaborating oncoproteins leading to AML development [8,9]. Despite blocking the differentiation of non-leukemogenic myeloid progenitors by the overexpressed HOXA9 gene [10], MEIS1 expression is mandatory to produce AML-initiating progenitors [11]. Thus, investigating the relationship between the transcription level of these genes and AML could help to identify a new targeted therapy for the disease.
The aim of the study was to evaluate HOXA9 and MEIS1 gene expression levels and their diagnostic and prognostic roles in de novo AML adult Egyptian patients.

Study group
Ninety-one de novo AML adult Egyptian patients and 41 healthy bone marrow donors were enrolled in the study.
The study was conducted between 2015 and 2018 at the Clinical Pathology Department at the National Cancer Institute, Cairo University, Egypt. The study was approved by the ethical scientific committee of Benha University and followed the Helsinki declaration guidelines.
AML was diagnosed and classified by the French-American-British (FAB) criteria [12]. The Medical Research Council Classification (MRC-C) was used to classify the cytogenetic risk in AML patients [13].
The inclusion criterion was adult patients with de novo AML. Patients ˂18 years, patients with secondary AML and patients with promyelocytic leukemia, and patients presented with AML with myelodysplasia were excluded from the study. Informed written consent was obtained from each patient included in the study.
Patients and controls were tested by RT-qPCR to determine the expression of messenger ribonucleic acid (mRNA) of HOXA9 and MEIS1 genes within the bone marrow samples.
Chemotherapy was induced in AML patients using the standard intensive 3 + 7 regimen (30 mg/m 2 doxorubicin on days 1-3 and 100 mg/m 2 cytarabine on days 1-7). However, dose reduction may be given for patients more than 60 years old. According to the cytogenetic risk and availability of a matched BM donor, the patients undergoing complete remission (CR) have received consolidation therapy. The patients with favorable risk were given high doses of cytarabine-based consolidation chemotherapy, while the intermediate-risk patients received either high doses of cytarabine-based consolidation chemotherapy or allogeneic BMT. Allogeneic BMT was considered the treatment of choice for adverse-risk patients. There were 9 patients who underwent allogeneic BMT.
CR was defined as regaining the BM's normal cellularity with fewer than 5% leukemic blasts, an absolute neutrophil count above 1.0*10 9 /L, and a platelet count above 100*10 9 /L. Resistance was defined as failure to fulfill CR, CR with incomplete recovery, or partial remission (PR) in a patient alive ≥ 7 days after completion of initial therapy, with evidence of persistent leukemia in peripheral blood and/or bone marrow examination [14,15]. Disease-free survival (DFS) was defined as the duration between gaining CR and relapse or death. Overall survival (OS) was defined as the time from involvement in the study to death from any cause; patients not known to have died at the last follow-up are censored on the date they were last known to be alive [16].

RNA extraction and cDNA synthesis
The BM samples (1 ml) were collected from all patients and controls on potassium ethylenediaminetetraacetic acid (K-EDTA) containing tubes. Ribonucleic acid (RNA) extraction from BM cells was performed using QIAamp ® RNA Mini Kit (QIAGEN, Austin, Texas, USA, Catalog no. 52304) according to the manufacturer's instructions.
The concentration and purity of the extracted RNA were assessed by spectrophotometer NanoDrop (Quawell, Q-500, Scribner, USA) and then stored at -80° C till used. The extracted RNA was reverse-transcribed to complementary DNA (cDNA) using the available cDNA Reverse Transcription Kit (Applied Biosystems, Thermo Fisher Scientific, USA; Catalog no. 4368814) as recommended by the manufacturer's instructions. The synthesized cDNA was stored at -20° C until performing quantitative real-time PCR.

Gene expression analysis
We used the quantitative real-time PCR to determine the level of expression of HOXA9, MEIS1 genes, and the reference gene (β-Actin). The RT-qPCR reactions were . The data were presented by cycle threshold (Ct). The Ct was obtained for HOXA9 or MEIS1 as the target gene (Ct-target) and for the reference gene β-Actin (Ctreference). Data were analyzed using relative quantitation (RQ). RQ and results were expressed as 2 −∆∆Ct according to Schmittgen and Livak [17].

Statistical methods
The obtained data were statistically analyzed using version 24 of the SPSS statistical software (SPSS Inc., Chicago, IL, USA). Categorical data analysis was performed using the chi-square and Fisher's exact test. The Mann-Whitney test compared quantitative data between groups. The performance of the tested genes as AML biomarkers was tested using receiver operating characteristic (ROC) analysis. The correlation between HOXA9 and MEIS1 gene expression was analyzed using the Spearman-rho test. Cumulative DFS and OS were compared by Kaplan-Meier survival analysis and the logrank test. p values < 0.05 were considered statistically significant. Patients were classified according to HOXA9 and MEIS1 gene expression into low and overexpression groups based on the cutoff values obtained by ROC curve analysis.

Patients' characteristics
The main demographic, clinical, and laboratory characteristics of the studied patients are summarized in Table 1.

Determination of delta Ct cutoff value
ROC analysis was applied to determine the best cutoff value for HOXA9 and MEIS1 gene expression for the diagnosis of AML. The curve value showed that the best cutoff for HOXA9 mRNA expression was 1.06, with an AUC of 0.910. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were 87.9%, 92.9%, 98.8%, 54.2%, and 88.6%, respectively, while for MEIS1 mRNA expression, the best cutoff value was 1.3, with an AUC of 0.831. The sensitivity, specificity, positive predictive value (PPV), negative (NPV) predictive value, and accuracy were 82.4%, 100%, 100%, 46.7%, and 84.8%, respectively. This indicated that HOXA9 and MEIS1 mRNA expression levels could be used in AML diagnosis ( Fig. 1 and Additional file 1: Table S1).

The expression levels of HOXA9 and MEIS1 mRNA in AML and control group
The relative expression levels of HOXA9 mRNA and MEIS1 mRNA in newly diagnosed AML patients were significantly higher than in the healthy controls (p < 0.001). A broad range of HOXA9 mRNA expression (range = 0.22-921.1; median = 14.6) and MEIS1 mRNA expression (range = 0-1716; median = 33.9) were observed in AML patients (

Correlation between the expression of HOXA9 and MEIS1 genes
As shown in Fig. 4, we have found a significant positive intermediate correlation between HOXA9 and MEIS1 gene expression.

Correlation between elevated HOXA9 and MEIS1 gene expression and clinicopathological characteristics of AML patients
The AML patients were divided based on the median expression level of the HOXA9 gene into two groups: low gene expression (< 14.6) and high gene expression (> 14.6). There were no significant differences in age, total leukocytic count (TLC), hemoglobin (Hb) level, platelet count, and blast percentage in the peripheral blood or the bone marrow between the low and high HOXA9 gene expression groups (p > 0.05) ( Table 3). Additionally, there were no significant differences regarding the distribution of FAB subtypes (M1, M2, M4, M5, M7), sex, organomegaly (hepatomegaly, splenomegaly, lymphadenopathy), BM cellularity, and IPT characteristics (p > 0.05). Furthermore, cytogenetic risk (favorable, intermediate, and adverse) and FLT3-ITD gene mutation did not significantly differ (p > 0.05) ( Table 3).  Patients were classified based on the median expression level of the MEIS1 gene into low gene expression (< 33.9) and high gene expression (> 33.9) groups.
The results revealed no significant differences in age, TLC, Hb level, platelet count, and blast percentage in peripheral blood and BM between the low and high expression groups (p > 0.05) ( Table 4). Furthermore, no significant differences were observed regarding the distribution of FAB subtypes (M1, M2, M4, M5, M7), sex, organomegaly (hepatomegaly, splenomegaly, lymphadenopathy), bone marrow cellularity, and IPT characteristics (p > 0.05). However, patients with low gene expression had aberrant markers with a percentage of (43.8%), while patients with high gene expression had aberrant markers with a percentage of (14.7%), (p = 0.014). Aberrant IPT markers were significantly associated with low MEIS1 gene expression. No significant differences were reported regarding cytogenetic risk (favorable, intermediate, and adverse) and FLT3-ITD gene mutation (p > 0.05) ( Table 4).

Impact of HOXA9 and MEIS1 gene expression on response to initial treatment
Regarding the response to initial treatment, 68 patients (74.7%) showed CR, 23 patients (25.3%) showed no CR, and 20 patients relapsed from those who achieved CR. No significant differences were detected regarding response to treatment (CR and no CR) between the high and low HOXA9 and MEIS1 gene expression groups (Tables 3 and 4).

Impact of HOXA9 and MEIS1 gene expression on the prognosis of AML patients
Follow-up was performed for 40 months; there were 19 patients died during the induction: 15 patients died from infection and 4 died from bleeding. The median follow-up period was 6.7 months (range, 0.3-40 months).
Kaplan-Meier analysis and log-rank test showed that there was no significant association between low expression and over-expression of the HOXA9 gene and OS (p = 0.799). The median OS was 11.070 and 11.4 months, respectively, while the cumulative OS after 3 years was 15.9% and 30.4%, respectively. In addition, there was no association between low expression and over-expression of the HOXA9 gene and DFS (p = 0.264). The median DFS was 9.230 and 16.8 months, respectively, while the cumulative DFS after 3 years was 0.0% and 44.1%, respectively, in AML patients (Fig. 5A,B and Table 3, Additional file 1: Table S2).
Regarding the MEIS1 gene, there was no association between low expression and over-expression of the gene and OS (p = 0.695). The median OS was 9.0 and 12.47 months, respectively, while the cumulative OS after 3 years was 21.1% and 31.8%, respectively. Also, there was no association between low expression and overexpression of the MEIS1 gene and the DFS (p = 0.351). The median survival was 14.070 and 25.630, respectively, while the cumulative DFS after 3 years was 25.4% and 44.7%, respectively, in AML patients (Fig. 5C,D

Discussion
AML therapeutic options are still limited due to its heterogeneity and complex biologic conduct. Despite the continuous efforts to develop novel specified therapy, the rates of AML-associated morbidity and mortality remain high [18]. The current advancement in molecular techniques is essential to provide an in-depth comprehension of leukemia pathogenesis and to offer novel molecular markers for diagnosis and prognosis [18].
Considering their documented role in hematopoiesis, aberrations of HOXA9 and its cofactor MEIS1 have been assessed in AML, focusing on their clinical intimation [19,20]. Their synergetic expression leads to aggressive leukemia in mice [21].
HOXA9 gene expression pattern has been investigated in AML [22]. Yet, there is still a lack of evidence concerning its association with MEIS1 gene expression [23].
The present study has shown significantly higher expression of HOXA9 mRNA and MEIS1 mRNA levels in de novo AML patients compared to the control group, suggesting its pathogenic role in AML. These data are consistent with Gao et al. [24], who demonstrated higher HOXA9 and MEIS1 expression in AML patients than in healthy BM donors. Additionally, Kandel et al. [25] examined HOXA9 in sixty samples of AML patients and reported higher expression than in healthy controls.
Ward et al. [18] observed the highest differential expression of HOXA9 and MEIS1 genes in AML patients. This finding has insightful repercussions since these genes are fundamentals in preserving and progressing variable myeloid disorders, and their combination results in AML occurrence [26].
An interesting finding in the present study was the significant positive intermediate correlation between HOXA9 and MEIS1 gene expression in AML cases. Similarly, Chen et al. [30] stated that the HOXA9 gene expression positively correlates with the MEIS1 gene expression.
In the current study, no significant differences were observed between the high and low HOXA9 gene expression groups regarding age, sex, organomegaly, complete blood picture, bone marrow examination, immunophenotypic characteristics, cytogenetic risk, and FLT3-ITD gene mutation. Similar results were reported between the high and low MEIS1 gene expression groups. However, the low-expression group demonstrated a significantly higher percentage of aberrant IPT markers than the high-expression group.
In agreement with our results, Gao et al. [24] concluded no significant associations of the baseline demographic, clinical, and molecular patterns with HOXA9 and MEIS1 gene mutations. However, lower transcript levels are shown in patients with M3 than in those with other FAB subtypes. On the other hand, Kandel et al. [25] demonstrated that HOXA9 gene expression is associated with hepatosplenomegaly in AML patients. According to Ismail et al. [27], HOXA9 gene expression is significantly associated with a larger size of the spleen, a higher breakpoint cluster region-Abelson proto-oncogene (BCR-ABL) expression, and a higher Sokal score.    The current study reported no significant differences between subjects with high and low HOXA9 and MEIS1 gene expression regarding response to initial treatment (CR and no CR). In contrast, Gao et al. [24] demonstrated a significant difference between patients with high and low HOXA9 gene expression levels regarding complete remission after one cycle of chemotherapy and 1-year relapse. This contradiction may be due to different treatment protocols or may be related to different geographical distributions.
Regarding the association of HOXA9 and MEIS1 gene expression levels and AML outcome, there was no association between the expression levels of HOXA9 and MEIS1 genes and DFS and OS. In contrast, Gao et al. [24] stated that the high expression of HOXA9 gene in AML is related to shorter relapse-free survival and OS compared to its low expression and acts as a powerful prognostic marker for response to therapy and survival. Similar results from previous studies demonstrated that high HOXA9 gene expression has a shorter OS in AML [31,32].
This discrepancy may be caused by the small size of our sample. Additionally, the MEIS1 gene expression behavior in malignant cells is influenced by many factors, including the immune status of the patients and the tumor microenvironment [33].

Conclusion
HOXA9 and MEIS1 genes are highly expressed in AML cases than in controls. However, their expression level was not associated with treatment response or survival rate of AML patients. Therefore, based on the results of the current study, HOXA9 and MEIS1 could be used as diagnostic biological markers for AML, but could not be used for assessment of the disease prognosis. Analyzing HOXA9 and MEIS1 gene mutations in further studies, including larger sample sizes, is recommended to assess the mechanism of aberrant HOXA9 and MEIS1 gene expression and to explore the molecular pathway linking these genes to leukemia progression. This will offer novel treatment modalities targeting the affected genes. Finally, longer follow-up may reveal other associations not depicted in the present study.