MiR-15a-5p, c-MYB, and circHIPK3 Axis as a Potential Diagnostic and Clinicopathological-Related Network in Acute Myeloid Leukemia

Abstract

Background:

Due to heterogeneity of acute myeloid leukemia (AML) in prognosis and response to treatment, understanding the pathophysiology of AML helps to find new factors as diagnostic and clinicopathological-related biomarkers and therapeutic targets.

Materials and Methods:

Using quantitative real-time polymerase chain reaction on AML patients, the expression of the miR15a-5p, c-MYB, and circ-HIPK3 gene network was measured, and the diagnostic performance and clinical application value of this gene network were also investigated.

Results:

In AML patients compared with healthy controls, the expression of mir-15a-5p and circ-HIPK3 significantly decreased, and the expression of c-MYB was significantly upregulated. Furthermore, c-MYB correlates with circ-HIPK3 positively. The areas under receiver operating characteristic curves (AUCs) of miR15a-5p, c-MYB, and circ-HIPK3 were 0.675, 0.885, and 0.762, respectively. Also, peripheral blood, as a noninvasive and cost-effective diagnostic sample, has good diagnostic value for the investigation of the mir15a-5p, c-MYB, and circ-HIPK3 genes. Finally, the change in circ-HIPK3 expression and red blood cell count showed a significant relationship.

Conclusion:

The results can contribute to a better pathophysiology understanding of AML, lead to the discovery of new diagnostic biomarkers, and develop treatment goals for patients. Also, the relationship between the genes and the clinicopathological characteristics of the patients helps patient monitoring.

INTRODUCTION

Because acute myeloid leukemia (AML) has a lot of different phenotypes and genotypes, it is important to find the different causes of this cancer at the molecular and cellular levels in order to diagnose patients, figure out the risk of the cancer in each patient, start the right treatment, check for recurrence, and see how well the treatment is working.[1,2] Various studies have investigated the pathophysiology of AML through the methylation evaluation of different genes (such as SOX17 and RUNX3 and so on) and examination of noncoding RNAs changing expression.[3,4,5,6]

CircRNAs belong to the group of long ncRNAs and regulatory RNAs. Unlike linear RNAs, circRNAs lack 5’ caps and 3’ tails and are circular single-stranded RNAs with covalent binding, which makes them resistant to digestion by RNase and thus more stable. Splicing mechanisms in different regions of the primary RNA transcript (exon, intron, 3’ and 5 untranslated regions, intergenic regions, sometimes from antisense RNAs) lead to the formation of circular RNAs.[7,8,9] CircRNAs have four main biological functions, including acting as protein adaptors, sponging with miRNAs, coding some amino acids, and acting as transcriptional regulators.[9] New research in many diseases, mainly blood cancers, has shown that various circular RNAs can be very important (especially when sponging with microRNAs) as oncogenes or tumor suppressors in intracellular processes. Any change in their expression can cause cells to stop working normally, which can then lead to tumor growth.[7,9,10] On chromosome 11p13, the homeodomain-interacting protein kinase 3 gene (HIPK3) is found on the short arm. circHIPK3 (has-circ-0000284) is created from exon 2 of the HIPK3 gene’s primary mRNA transcript[11] [Figure 1]. In many studies, the expression of circHIPK3 has been measured in various cancers, such as lung cancer, colorectal cancer, stomach cancer, bladder cancer, prostate cancer, chronic myeloid leukemia, and so on.[12,13] Also, studies have shown that circHIPK3 has two important roles: one is as an oncogene and the other as a tumor suppressor. This shows how important circHIPK3 is for controlling the functions inside cells.[12,13] One of the parameters that have a significant effect on the function of hematopoietic stem cells (HSCs) is microRNAs (miRs), and many of them have been identified as effective on HSCs.[14] MiRs can also be used as biomarkers for diagnostics, disease grading, and prognosis in various diseases, including malignancies.[15] MiR-15a and miR-16-1 are located in a gene cluster on the long arm of chromosome 13 (13q).[16] In different studies, the importance of miR-15a in the pathogenesis of various cancers such as chronic lymphocytic leukemia, acute myeloid leukemia, chronic myeloid leukemia, ovarian cancer, and so on has been mentioned. MiR-15a can be considered an important tumor suppressor in the regulation of cell cycle, apoptosis, differentiation, and angiogenesis.[17]

Figure 1.

Biogenesis of circ-HIPK3

When HSCs differentiate, especially in the early stages, c-MYB is a transcription factor that is important. It is also a proto-oncogene that is very important for HSCs to divide and differentiate.[18,19] Two transcription factors, c-MYB and C/EBPα, play a big part in how FLT3 is expressed in AML cells.[20] It is also mentioned that the MYB transcription factor plays a key role in the stability and function of AML fusions involving MLL, which could be a therapeutic target.[18] So finally, it is clearly established that c-MYB is essential for the initiation and/or maintenance of many cancers, especially in AML patients.[18,21] Articles have looked into the role of miR-15a and how it affects hematopoietic stem cells and other cells. MiR-15a-5p can be thought of as a negative regulator of c-MYB.[22,23,24,25,26] A number of studies have also shown that the c-MYB transcription factor connects to the HIPK3 gene’s promoter and manages the production of circHIPK3.[27,28] This study is the first in the world to look at how the miR-15a-5p, c-MYB, and circHIPK3 gene networks relate to each other, how their expression changes, and how they help with diagnosis in AML patients. We did this to learn more about the role of this gene network in the development of AML.

MATERIALS AND METHODS

Patients and sample collection

Between July and September 2023, 40 and 32 bone marrow (BM) and peripheral blood (PB) samples were taken from newly diagnosed AML patients who did not receive treatment (from Taleghani Hospital in Teheran, Iran) and healthy people as controls, respectively. All participants provided informed consent, and the Zanjan University of Medical Sciences ethics committee approved the study (No. IR.ZUMS.REC.1402.094). Inclusion criteria for patients were 1) diagnosed and classified based on guidelines of World Health Organization (WHO) and French-American-British (FAB) criteria; 2) Patients who were above 18 years old; 3) patients who had not received any form of treatment. Also, exclusion criteria were 1) exposed to radiotherapy or chemotherapy before being diagnosed with AML; 2) history of infections, bleeding, solid tumors, or other hematological malignancies; 3) bone marrow failure syndromes; 4) pregnant or lactating women. Clinical and laboratory data as well as risk stratification based on ELN 2017[29] were recorded.

RNA Isolation and cDNA synthesis

The extraction of total RNA from BM and PB white blood cells was performed using Trizol reagents (Thermo Fisher Scientific, Waltham, MA, USA). A nanodrop device (Micro UV-VIS Spectrophotometer, BOECO N-1C, Germany) was used to check the quality and purity of the extracted total RNA and then was kept at −70°C until it was needed. Synthesis of single-stranded cDNA was carried out using the ExcelRT™ Reverse Transcriptase Kit (SMOBio, South Korea) based on the instructions of the manufacturer and a universal stem-loop primer (Metabion, Germany) for microRNA.

Real-Time Quantitative PCR

Expression analysis was performed using SYBR green-based quantitative real-time polymerase chain reaction (PCR). In a 20-μl reaction, 10 μl of Master Mix 2x (Real Q Plus Master Mix High ROX, Denmark), 0.5 μl of each primer, 1.5 μl of the synthesized cDNA, and 7.5 μl of nuclease-free water were mixed and run in Applied Biosystems Step One Plus (Thermo Fisher Scientific, CA, and USA). The cycling program was performed as following: initial denaturation at 95°C for 15 min, 38 cycles of denaturation at 95°C for 15 s, annealing at 53–54°C for 15 s, extension at 72°C for 55 s, and 25 min for the melting curve. To show the specificity of the PCR product as a single peak and the absence of contamination with gDNA, melting curve analysis, a negative test control (NTC) for all reactions, and also nRT (no reverse transcriptase) control were conducted. Also, in order to avoid bias when performing cDNA synthesis and the real-time step, blinding was done on patient and control samples. The B2-M (Beta 2 Microglobulin) gene as a reference gene for circHIPK3 and c-MYB, as well as SNORD47 as a reference gene for miR-15a-5p, was used. We found the fold change of genes using the formula (E target)Δ^{CT target (control-sample)}/(E reference)Δ^{CT reference (control-sample)}. We also used Linreg software[30] to find the specific efficiency of reactions [Table 1]. The sequence and optimum annealing temperature of the primers for each gene, along with the sequence of the stem loop primer, are shown in Table 1.

Table 1.

Primer sequence of gene expression analysis and stem-loop primer for cDNA synthesis

Gene Names	Sequences of primers (5' → 3')	Optimum annealing™	Mean PCR efficiency
c-MYB	F: AGAAGAAGATCAGAGAGTGATAG	61.5	1.869
	R: CCAGTGGTTCTTGATAGCATTA
B2-M	F: AGGCTATCCAGCGTACTCCA	60	1.938
	R: TCATCCAATCCAAATGCGGC
Circ-HIPK3	F: TTCAACATATCTACAATCTCGGT	61.5	1.954
	R: ACCATTCACATAGGTCCGT
Mir15a-5p	F: GACAGGTAGCAGCACATAATG	61	1.941
Snord47	F: CCAATGATGTAATGATTCTGCCA	62	1.905
Universal reverse	CGAGGAAGAAGACGGAAGAAT
USLP	GAAAGAAGGCGAGGAGCAGATCGAGGAA GAAGACGGAAGAATGTGCGTCTCGCCTTCTTTCNNNNNN

Bioinformatics and statistical analysis

The correlation between miR-15a-5p and c-MYB was predicted using databases such as Mirbase.org, Mirdb.org, and Targetscan.org. Also, circ-HIPK3 divergent primers were designed and checked using the circinteractome.nia.nih.gov database and CircPrimer 2.0 software.[31] All the data analyses were done on GraphPad Prism (San Diego, California, U.S.A., V 8) and South Texas Art Therapy Association STATA software (College Station, Texas, USA, V 14.2). Differences between two groups were measured via Student’s t-test and the Mann–Whitney U test. To find differences in categorical variables (looking at how gene expression is related to clinicopathological features), the odds ratio (OR) and the Chi-square test (χ2) were used. Correlation evaluation was conducted using Spearman’s rank correlation test. Receiver operating characteristic (ROC) analysis was performed to evaluate the ability of genes to distinguish AML patients from healthy donors. P value ≤ 0.05 was considered statistically significant, and a P value between 0.05< and <0.1 was considered marginally significant.

RESULTS

Participants in the patient group were 42.5% male and 57.5% female, and in the controls, they were 56.25% male and 43.75% female.

Clinical and laboratory characteristics of AML patients

In Table 2, the differences in clinicopathological characteristics between the two patient and control groups are shown. On the other hand, Tables 1 and 2 in supplementary data S1 and Table 3 demonstrate the relationship between the changing of miR15a-5p, c-MYB, and circ-HIPK3 expression (the gene expression was classified into two high and low expression groups based on median expression as a cutoff value) and the clinicopathological characteristics of AML patients with two OR and χ² tests. The changes in circ-HIPK3 expression have a significant relationship between RBC count (χ² P value: 0.026, OR: 12.91; 95% CI: 0.63–263.68) and also a marginally significant relationship between WBC count (χ² P value: 0.056, OR: 7.5; 95% CI: 0.76–74.16) and hematocrit (χ² P value: 0.059, OR: 9.24; 95% CI: 0.44–195.69). Furthermore, in relation to other genes, there was no significant relationship between expression changes and clinicopathological characteristics.

Table 2.

Comparison between the AML patients and healthy controls based on different parameters

Characteristics	AML	Control	P	test
Gender
Male	17	18	0.2461	χ2
Female	23	14
Age(years)
Min-Max	23-67	20-67	0.0625	t
Median	45.5	40.5
Mean±SD	45.5±13.3	39.93±11.2
WBCcounts (x109/L)
Min-Max	8-220	4.4-11.2	<0.0001	u
Median	38	6.35
Mean±SD	59.1±64.68	6.6±1.64
RBCcounts (x1012/L)
Min-Max	1.18-4.52	4.09-6.5	<0.0001	t
Median	2.56	5.32
Mean±SD	2.67±0.95	5.30±0.63
HB(g/dl)
Min-Max	3.6-13.2	11.9-18.6	<0.0001	t
Median	7.6	14.55
Mean±SD	8.21±2.47	14.8±1.6
HCT(%)
Min-Max	10-39.3	36.4-51.9	<0.0001	t
Median	23.5	44.7
Mean±SD	24.36±7.37	44.8±4.06
Plateletcounts (x109/L)
Min-Max	4-357	138-340	<0.0001	u
Median	56	219.5
Mean±SD	73.38±74.69	221.43±48.7
Mir-15a-5p
Min-Max	0.23-1.43	0.06-6.65	<0.0001	u
Median	0.73	0.97
Mean±SD	0.77±0.3	1.49±1.52
c-MYB
Min-Max	0.42-647	0.04-233	<0.0001	u
Median	20.67	0.53
Mean±SD	73.92±131.5	11.38±41.79
Circ-HIPK3
Min-Max	0.007-334.3	0.01-27.09	<0.0001	u
Median	0.28	1.005
Mean±SD	14.61±61.63	3.07±6.44

χ²: Chi-square test, t: Student's t test, u: Mann–Whitney U test

Table 1.

Correlation between mir15a-5p expression and different clinicopathological features in AML patients

Characteristics	Number	Mir-15a-5p		OR (95% CI)	P*
		Low expression (n=20)	High expression (n=20)
Age (years)	40
≥46	20	11	9	1.49 (0.43-5.19)	0.527
<46	20	9	11
Gender	40
Male	17	10	7	1.85 (0.52-6.61)	0.337
Female	23	10	13
BM blasts (%)	28
≥79	21	8	13	0.24 (0.03-1.58)	0.125
<79	7	5	2
WBC counts (×109/L)	31
≥10	25	12	13	1.84 (0.28-11.97)	0.516
<10	6	2	4
RBC counts (×1012/L)	31
<4	27	13	14	2.78 (0.25-30.27)	0.385
≥4	4	1	3
HB (g/dl)	31
<12.5	30	14	16	2.63 (0.09-69.88)	0.356
≥12.5	1	0	1
HCT (%)	31
<36	28	13	15	1.73 (0.14-21.38)	0.664
≥36	3	1	2
Platelet counts (×109/L)	31
<50	12	8	6	2.44 (0.57-10.44)	0.223
≥50	17	6	11
FAB classification	40
Non-M3	31	15	16	0.75 (0.16-3.33)	0.705
M3	9	5	4
CRP (mg/L)	20
>0.3	19	10	9	3.31 (0.120-91.60)	0.304
≤0.3	1	0	1
ESR (mm/hr)	22
>15	21	8	13	0.21 (0.008-5.76)	0.218
≤15	1	1	0
ALT (IU/L)	26
>36	6	4	2	3.71 (0.53-25.59)	0.168
≤36	20	7	13
AST (IU/L)	26
>33	11	5	6	1.25 (0.25-6.02)	0.780
≤33	15	6	9
Risk stratification	22
Adverse and Intermediate	11	7	4	2.1 (0.38-11.58)	0.391
Favorable	11	5	6

*χ² test

Table 2.

Correlation between c-MYB expression and different clinicopathological features in AML patients

Characteristics	Number	c-MYB		OR (95% CI)	P*
		High expression (n=20)	Low expression (n=20)
Age (years)	40
≥46	20	9	11	0.66 (0.19-2.32)	0.527
<46	20	11	9
Gender	40
Male	17	11	6	2.85 (0.77-10.46)	0.109
Female	23	9	14
BM blasts (%)	28
≥79	21	13	8	2.16 (0.38-12.30)	0.377
<79	7	3	4
WBC counts (×109/L)	31
≥10	25	13	12	5.41 (0.55-53.27)	0.118
<10	6	1	5
RBC counts (×1012/L)	31
<4	27	12	15	0.80 (0.09-6.54)	0.834
≥4	4	2	2
HB (g/dl)	31
<12.5	30	14	16	2.63 (0.09-69.88)	0.356
≥12.5	1	0	1
HCT (%)	31
<36	28	13	15	1.73 (0.14-21.38)	0.664
≥36	3	1	2
Platelet counts (×109/L)	31
<50	14	7	7	1.42 (0.34-5.94)	0.623
≥50	17	7	10
FAB classification	40
Non-M3	31	15	16	0.75 (0.16-3.33)	0.705
M3	9	5	4
CRP (mg/L)	20
>0.3	19	9	10	2.71 (0.09-74.98)	0.353
≤0.3	1	0	1
ESR (mm/hr)	22
>15	21	9	12	0.25 (0.009-6.93)	0.262
≤15	1	1	0
ALT (IU/L)	26
>36	6	5	1	7.50 (0.73-76.77)	0.062
≤36	20	5	12
AST (IU/L)	26
>33	11	7	4	2.62 (0.52-13.06)	0.233
≤33	15	6	9
Risk stratification	22
Adverse and Intermediate	11	4	7	0.68 (0.12-3.78)	0.664
Favorable	11	5	6

*χ² test

Table 3.

Correlation between circ-HIPK3 expression and different clinicopathological features in AML patients

Characteristics	Number	Circ-HIPK3		OR (95% CI)	P*
		Low expression (n=20)	High expression (n=20)
Age (years)	40
≥46	20	10	10	1.0 (0.29-3.45)	>0.999
<46	20	10	10
Gender	40
Male	17	6	11	0.35 (0.09-1.28)	0.109
Female	23	14	9
BM blasts (%)	28
≥79	21	10	11	2.27 (0.35-14.45)	0.377
<79	7	2	5
WBC counts (×109/L)	31
≥10	25	15	10	7.5 (0.75-74.15)	0.056
<10	6	1	5
RBC counts (×1012/L)	31
<4	27	16	11	12.9 (0.63-263.86)	0.026
≥4	4	0	4
HB (g/dl)	31
<12.5	30	16	14	3.41 (0.12-90.49)	0.293
≥12.5	1	0	1
HCT (%)	31
<36	28	16	12	9.24 (0.43-195.68)	0.059
≥36	3	0	3
Platelet counts (×109/L)	31
<50	14	6	8	0.52 (0.12-2.20)	0.376
≥50	17	10	7
FAB classification	40
Non-M3	31	17	14	2.42 (0.51-11.51)	0.256
M3	9	3	6
CRP (mg/L)	20
>0.3	19	12	7	5.0 (0.18-139.15)	0.208
≤0.3	1	0	0
ESR (mm/hr)	22
>15	21	12	9	0.43 (0.01-12.01)	0.394
≤15	1	1	0
ALT (IU/L)	26
>36	6	4	2	2.0 (0.29-13.51)	0.472
≤36	20	10	10
AST (IU/L)	26
> 33	11	7	4	2.0 (0.40-9.83)	0.391
≤ 33	15	7	8
Risk stratification	22
Adverse and Intermediate	11	5	6	1.2 (0.21-6.67)	0.835
Favorable	11	5	6

*χ² test

Expression analysis

The expression of the gene network according to the real-time results was analyzed. Compared to controls (median: 0.97) [Figure 2a], the level of miR-15a-5p was significantly lower in people with AML (median: 0.73, P value: 0.0103). It was also seen that c-MYB levels were significantly higher in AML patients (median: 20.67, P value: <0.0001) than in the control group (median: 0.53) [Figure 2b]. On the other hand, circ-HIPK3 levels were significantly lower in AML patients (median: 0.28, P value: <0.0001) than in the control group (median: 1.006) [Figure 2c].

Figure 2.

The relative expression of genes in AML patients in comparison to healthy controls and various subtypes. The downregulated expression of mir15a-5p (a), the upregulated expression of c-MYB (b), and the downregulated expression of circ-HIPK3 (c)

Correlations among the genes in network

According to Spearman correlation analysis, there was a negative correlation between miR15a-5p and c-MYB, but a nonsignificant P value (0.87) was shown (r = -0.025, 95% CI: -0.34 to 0.29) [Figure 3b]. In addition, between c-MYB and circ-HIPK3, a significant positive correlation with a P value of 0.0091 was observed (r = 0.4073, 95% CI: 0.1003 to 0.64) [Figure 3a]. Finally, there was no correlation between circ-HIPK3 and miR-15a-5p.

Figure 3.

The correlation between gene networks. The correlation between c-MYB and circ-HIPK3 (a) and the correlation between mir15a-5p and c-MYB (b)

Diagnostic ability of the gene network

The diagnostic indicators of the gene network are shown in supplementary data S1, Table 3. Based on ROC curve analysis, miR-15a-5p had an AUC of 0.675 (95% CI: 0.54–0.80, P value: 0.01), a Sen of 72.50%, a Spe of 59.38%, and a DOR of 3.85 [Figure 4, a1]. Also, AUC: 0.885 (95% CI: 0.79–0.97, P value: 0.0001), Sen: 92.50%, Spe: 75%, and DOR: 37 for the c-MYB gene were obtained [Figure 4, a2], and finally, diagnostic indicators for circ-HIPK3 were AUC: 0.762 (95% CI: 0.63–0.88, P value: 0.0001), Sen: 87.50%, Spe: 71.88%, and DOR: 17.89 [Figure 4, a3]. Furthermore, the ROC curve analysis was performed based on sample type (BM and PB) [Figure 4, b and c, respectively], and the results demonstrated good diagnostic indices for the investigation of genes in PB samples.

Table 3.

Diagnostic indicators of gene network

Diagnostic indicators	AUC (95% CI)	P	Sen % (95% CI)	Spe % (95% CI)	DOR (95% CI)	PLR (95% CI)	NLR (95% CI)	PPV % (95% CI)	NPV % (95% CI)	Cut-off point
Mir15a-5p	0.675 (0.54-0.80)	0.01	72.50 (57.17-83.89)	59.38 (42.26-74.48)	3.85 (1.43-10.37)	1.78 (1.13-2.83)	0.46 (0.26-0.83)	69.05 (58.47-77.95)	63.33 (9.19-75.50)	≤0.95
c-MYB	0.885 (0.79-0.97)	0.0001	92.50 (80.14-97.42)	75 (57.89-86.75)	37.00 (8.92-153.5)	3.70 (2.02-6.79)	0.10 (0.03-0.30)	82.22 (71.60-89.45)	88.89 (72.57-96.03)	≥3.23
Circ-HIPK3	0.762 (0.63-0.88)	0.0001	87.50 (73.89-94.54)	71.88 (54.63-84.44)	17.89 (5.32-60.18)	3.11 (1.77-5.48)	0.17 (0.07-0.41)	79.55 (68.83-87.26)	82.14 (66.33-91.48)	≤0.84

Figure 4.

Diagnostic value of the gene network. Main ROC curves for mir-15a-5p, c-MYB, and circ-HIPK3 (a 1, 2, and 3, respectively); PB sample ROC curves for mir-15a-5p, c-MYB, and circ-HIPK3 (b 1, 2, and 3, respectively); and BM sample ROC curves for mir-15a-5p, c-MYB, and circ-HIPK3 (c 1, 2, and 3, respectively)

DISCUSSION

AML is one of the most common hematological malignancies, with a very variable prognosis.[32,33] Also, high heterogeneity occurs at different ages, especially in the elderly.[33,34] Despite the progress of science and extensive research in the field of treatment methods and the diagnosis of AML, the survival rate of AML patients is still poor.[35,36] Today, AML diagnosis relies on the cytomorphology, flow cytometry, cytogenetics, and molecular genetics assessment on PB or BM.[37] On the other hand, evaluation and monitoring of patients’ prognosis are done using their cytogenetic abnormalities, gene mutations, age, white blood cell count, and so on.[32] Therefore, expanding our understanding of the pathophysiology of AML helps us find new factors as diagnostic and prognostic biomarkers and therapeutic targets. In several studies, the role of genes in the pathophysiology of AML and the importance of these biomarkers have been investigated.[7,38,39,40,41,42,43] Based on these factors and the importance of genes (especially noncoding RNAs) in prognosis, diagnosis, therapeutic targets, and personalized medicine, the study’s goal was to look into the expression, correlation, and diagnostic value of the miR15a-5p, c-MYB, and circ-HIPK3 network. It was also supposed to find out how the expression of the genes in the network related to the clinicopathological features of AML patients.

Here, for the first time, the expression of miR-15a-5p and circ-HIPK3 was measured in AML. The expression of miR-15a-5p (1.5-fold) and circ-HIPK3 (2.5-fold) decreased in AML patients, which can be considered miR-15a-5p and circ-HIPK3 as tumor suppressors. On the other hand, c-MYB expression with an oncogenic role had increased about 20-fold in AML patients. Furthermore, the expression of genes in different subgroups was evaluated (based on FAB classification, sample type, WBC count, and risk stratification). The results indicate the changes in gene expression in the FAB classification, WBC count, and risk stratification subgroups were not significantly different, while the changes in c-MYB and circ-HIPK3 expression in the sample type subgroup were significantly different from each other. Therefore, it can be said that the amount of change in the c-MYB and circ-HIPK3 expressions is different according to the sample type, and on the other hand, the amount of change in the miR15a-5p, c-MYB, and circ-HIPK3 expressions is not related to the FAB classification, WBC count, or risk stratification subgroups. Bioinformatic analysis showed that miR15a-5p, as a tumor suppressor gene, has a negative regulatory role in relation to the c-MYB gene. So based on correlation analysis, this negative correlation was observed, but likely due to a limitation in the sample size (effect on the statistical power of the study), this negative correlation was nonsignificant. Furthermore, according to the interpretation areas of the correlation coefficient,[44] a low positive correlation between c-MYB and circHIPK3 was seen, which confirms the findings of previous studies about the positive role of c-MYB in increasing the expression of circHIPK3,[27,28] but interestingly, despite the significant increase in the expression of c-MYB, which has a positive effect on the circHIPK3 gene promoter, the expression of circHIPK3 is still downregulated, which confirms this hypothesis that the expression of circHIPK3 in AML patients is influenced by other powerful factors.

According to our findings in the previous meta-analysis regarding the more reliable diagnostic value in case-control studies, the control/patient ratio should be above 50%.[45] However, the ratio of control/patient in this study is 80%. A sensitivity and a specificity close to 1 indicate the high ability of the test for the diagnosis of patients and healthy people. In this study, the c-MYB and circ-HIPK3 genes showed higher sensitivity and specificity, respectively, but miR-15a-5p had better sensitivity than specificity (72.5 vs 59.38), which indicates miR-15a-5p is better for screening patients. As an essential index in diagnostic tests, higher DOR is related to better diagnostic value, and DOR shows that the odds of a positive test in patients are higher than the odds of a positive test in people without disease.[46] In our study, DOR related to c-MYB (DOR: 37) and circ-HIPK3 (DOR: 17.89) is very high and suggests that these genes are promising potential diagnostic tests for recognizing AML patients from healthy controls. In addition, the ROC curve and the AUC display the ability of the diagnostic test to differentiate patients from healthy controls, and the larger AUC shows a higher diagnostic value.[47] These genes have very good, good, and enough diagnostic accuracy based on the AUC interpretation area: c-MYB, circ-HIPK3, and miR-15a-5p. Positive and negative likelihood ratios and post-test probabilities are helpful indicators for medical professionals because they have a relationship to clinical application and provide information about the likelihood that a patient with a positive or negative test actually has AML or not. According to the interpretation area and the result of the likelihood ratio,[48] the NLR regarding miR-15a-5p shows a large and often conclusive shift in the probability of the disease, and the NLR regarding c-MYB and circ-HIPK3 shows a moderate shift in the probability of the disease. Moreover, PLR in association with c-MYB and circ-HIPK3 genes indicates a small shift in the probability of the disease, which in the total gene network is a suitable indicator for clinical application. In addition, to better understand the diagnostic power of genes, we also drew the ROC curves based on sample type (BM and PB samples). The ROC curve data clearly showed that PB samples are as good or better at diagnosing than BM samples (based on diagnostic interpretation areas). This means that changes in miR-15a-5p, c-MYB, and circ-HIPK3 expression can be used as strong diagnostic markers in patients’ PB samples (as noninvasive and cost-effective methods).

According to the pathophysiology of AML disease, an increase in WBC count along with a decrease in RBC and platelet count, Hb level, and HCT percentage can be seen in most patients, and this pathophysiology causes acute symptoms in patients.[33,49] Finally, in this study, by comparing the clinicopathological characteristics of patients and healthy controls, the aforementioned changes were observed in AML patients. On the other hand, the two patient and control groups did not have significant differences in age and gender characteristics, which indicates that age and gender matched between the two groups and helps to make the results more reliable. In this study, to better understand the association between changes in gene expression and clinicopathological features, we reported the OR (as an indicator of effect size) along with statistical significance (P value of the Chi-square test). The interpretation areas of OR are as follows: 1 to 1.49: trivial (inconsiderable); 1.49 to 3.45: small; 3.45 to 9: moderate; 9 or more: large.[50]

The results showed that there was a significant or marginally significant link between the change in circ-HIPK3 expression and the RBC and WBC counts and the HCT percentage. There was also a nonsignificant link between the change in c-MYB and miR15a-5p expression and the clinicopathological features. However, for a more accurate assessment of the effect size and the relationship between gene expression and clinicopathological characteristics, it is necessary to conduct meta-analysis studies according to the information of several primary studies (such as our study), which in our two previous meta-analysis studies related to AML[51] and MM diseases,[52] fully investigated these relationships. So finally, the results of this study can be combined with the results of other primary studies (whether statistically significant or not) to look at the subject with a larger statistical sample size.

CONCLUSION

Circ-HIPK3, c-MYB, and miR15a-5p can be considered as promising diagnostic biomarkers for AML patients. Due to the relationship between the gene network and clinicopathological characteristics of the patients and to better patient monitoring, we suggest use and study of the network in personalized medicine.

Graphical abstract.

Description of the relationship between mir15a-5p, c-MYB and circ-HIPK3 gene network.

Availability of data and materials

The data have included in the text and documented/reserved with the corresponding author.

Ethics approval and consent to participate

All procedures were in accordance with the ethical standards of the institutional and/or national research committee. The ethics committee of Zanjan University of Medical Sciences (Ethic Code No. IR.ZUMS.REC.1402.094) approved this study.

Consent for publication

All protocols and blood sampling were conducted after filling out the informed consent form by patients.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Zanjan University of Medical Sciences.