Investigation of the expression and potential mechanistic role of BYSL in acute myeloid leukemia

Abstract

BYSL, located on chromosome 6p21.1, is implicated in tumor progression, but its role in acute myeloid leukemia (AML) remains unclear. BYSL expression was analyzed using public databases and AML clinical samples. A prognostic model incorporating BYSL was constructed and validated. Functional assays were performed by knocking down BYSL in AML cell lines to evaluate proliferation, apoptosis, and cell cycle regulation. To investigate the underlying mechanism, Gene set enrichment analysis (GSEA), Western blotting and chromatin immunoprecipitation-quantitative polymerase chain reaction (ChIP-qPCR) assays were performed. BYSL was significantly overexpressed in AML, with high expression correlating with severe anemia and poor overall survival. A risk model incorporating BYSL along with clinical factors (age, cytogenetic risk, transplantation, gene mutations) demonstrated strong predictive accuracy (c-index = 0.754). Functional assays showed that BYSL knockdown significantly inhibited cell proliferation, reduced colony-forming ability, and induced cell cycle arrest at the G0/G1 phase. Mechanistically, BYSL suppression notably decreased PI3K and AKT phosphorylation, implicating this signaling pathway. Additionally, ChIP-qPCR experiments confirmed that BYSL is transcriptionally regulated by c-MYC through direct promoter binding. Our findings support a role for BYSL in AML pathogenesis, potentially through modulation of the PI3K/AKT signaling pathway. Transcriptionally regulated by c-MYC, BYSL may serve as a prognostic biomarker and warrants further investigation as a potential therapeutic target.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10238-026-02122-6.

Introduction

Acute myeloid leukemia (AML) is a malignant clonal hematological disorder characterized by genetic and epigenetic alterations, dysplasia, and impaired apoptosis [1–3]. Despite significant improvements in prognosis over recent decades due to advancements in chemotherapy, supportive care, molecular targeted therapies, and hematopoietic stem cell transplantation, AML continues to exhibit high rates of recurrence and mortality [4–10]. The long-term survival rate for patients under 60 years is approximately 35% to 45%, but drops dramatically to 10% to 15% in those aged 60 and older [11]. Therefore, understanding the etiology of AML, identifying therapeutic targets, and developing new treatments remain critical research priorities in hematology.

The BYSL gene, located on human chromosome 6p21.1, encodes a protein that plays multiple critical roles in cellular processes [12, 13]. It was first demonstrated that BYSL protein mediates interactions between trophoblast cells and endometrial epithelial cells by forming complexes with Tastin and Trophin [14–16]. Additionally, BYSL is crucial in ribosome biosynthesis, collaborating with assembly molecules such as PNO1 and NOB1 in the maturation of the 40 S ribosomal subunit [17, 18]. BYSL is upregulated in astrocytes following brain injury or infection, and is also highly expressed in various solid tumors, including liver and prostate cancers, where it promotes cell survival and tumorigenesis [19–24].

Although BYSL’s oncogenic potential in solid malignancies is driven largely by transcriptional overexpression rather than mutations, its role in leukemia has remained largely unexplored. Beyond genetic mutations and chromosomal aberrations, accumulating evidence highlights transcriptional dysregulation as a critical non-mutational driver of AML pathogenesis and therapeutic resistance [25–28]. Recent studies have identified several oncogenes, such as MEF2C, MN1, and PRDM16, that are overexpressed at the transcriptional level without underlying mutations, each with independent prognostic significance in AML [25, 26, 28].

Building on this paradigm, the present study supports a role for BYSL in AML, demonstrating its overexpression driven by c-MYC and its functional association with the PI3K/AKT signaling pathway, suggesting potential oncogenic relevance and therapeutic interest.

Methods

Databases resource

Microarray data of AML patients and normal controls were obtained from the GSE114868 and GSE7186 datasets in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). BYSL expression and clinical information for AML patients were retrieved from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov) and Therapeutically Applicable Research to Generate Effective Treatments (TARGET, https://ocg.cancer.gov/programs/target) databases.

The DepMap database (https://depmap.org/) was used to analyze the correlation between c-MYC and BYSL expression in 55 AML cell lines. The JASPAR database (http://jaspar.genereg.net) was utilized to predict transcription factors potentially regulating BYSL.

Patient samples

A total of 93 newly diagnosed AML patients were enrolled at West China Hospital of Sichuan University. AML diagnosis was established using comprehensive assessments, including morphology, immunophenotyping, cytogenetics, and molecular profiling (MICM). Additionally, 10 healthy volunteers were included as controls.

Construction of a risk model for the OS of patients with AML

In the univariate Cox regression analysis of OS using the TCGA AML cohort, prognostic factors with a P-value of < 0.1 were included in the multivariable analysis. Final variable selection was performed using backward elimination based on the Akaike Information Criterion (AIC), retaining only variables with P < 0.05 in the final model. These variables were subsequently incorporated into a prognostic nomogram constructed using the “rms” package in R. The predictive performance of the model was evaluated using the concordance index (c-index) and receiver operating characteristic (ROC) curve analysis. The internal validation set was derived from the TCGA AML cohort, and external validation was conducted using our independently collected clinical patient data. In the validation cohort, ROC curves and calibration plots were used to assess the predictive accuracy of the model.

Cell culture

HEK-293T cells were obtained from the Cell Bank of the Chinese Academy of Sciences, along with AML cell lines including HL-60, K562, Kasumi, Molm-13, NB4, THP-1, and U937. HEK-293T cells were cultured in DMEM medium (Hyclone, USA, SH30228), while all AML cell lines were cultured in RPMI-1640 medium (Hyclone, USA, SH30809). All culture media were supplemented with 10% fetal bovine serum (Gemini, USA, 900108) and 1% penicillin-streptomycin (Hyclone, USA, SV30010). Cells were incubated in a 5% CO₂ incubator at 37 °C.

RNA extraction and real-time quantitative polymerase chain reaction (RT-qPCR)

Monocytic cells were isolated from peripheral blood using a density gradient centrifugation assay and used for subsequent RNA extraction and gene quantification. Total RNA was extracted from human monocytic cells or cultured AML cells using TRIzol reagent (Invitrogen). RNA was reverse-transcribed into cDNA using the PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Dalian, China). RT-PCR was performed using the SYBR Premix Ex Taq II kit (Takara, Dalian, China) following the manufacturer’s protocol. Relative expression of BYSL was calculated using the 2^−ΔΔCt method, with GAPDH as internal control. Primer sequences for qPCR are listed in Table 1.

Table 1.

Primers for RT-qPCR used in this study

Gene	Forward (5’-3’)	Reverse (5’-3’)
BYSL	CTCTCCAACTGGGAGCAAATCC	TGCGTTCCTTCAGGTTAGAGGC
GAPDH	AATGAAGGGGTCATTGATGG	AAGGTGAAGGTCGGAGTCAA
BYSL promoter	GGACCTGGCACCTTACAACT	GGACCCGTTTCCCACTATGT

Western blotting

Cells were collected and lysed on ice for 30 min with lysis buffer containing a protease inhibitor cocktail (Roche, USA). Supernatants were collected after centrifugation, and protein concentrations were measured using a BCA protein quantification assay (Beyotime, Shanghai, China). After heat denaturation, the proteins were separated on 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a 0.22 μm polyvinylidene fluoride (PVDF) membrane (BioRad, USA). The membranes were blocked with 5% fat-free milk in TBST for 1 h at room temperature, and then incubated with specific primary antibodies at 4 °C overnight. After washing with TBST buffer, membranes were incubated for 2 h with HRP-conjugated secondary antibodies IgG-HRP (1:5000, Cell Signaling, Germany) at room temperature. Finally, the enhanced chemiluminescence reagent was used to visualize protein signals, which were recorded by the chemiluminescence image analysis system (Tanon5200, Shanghai, China). Densitometric quantification of Western blot bands was performed using ImageJ software (NIH, USA). Phosphorylated protein levels were expressed as the ratio of phosphorylated to total protein (e.g., p-AKT/AKT), while total protein levels were normalized to GAPDH. Quantitative analyses were based on three independent biological replicates.

Cell transfection for BYSL knockdown

To transiently knock down BYSL, BYSL-targeting siRNAs and a negative control (si-NC) were synthesized by GenePharma (Shanghai, China), and the procedure followed the manufacturer’s instructions as well as our previous work [29]. To establish stable BYSL knockdown AML cell lines, we used lentiviral infection. Briefly, shRNA sequences targeting BYSL were designed and cloned into the pLKO.1-Puro vector to construct BYSL interference lentivirus. The modified plasmid or control plasmid, along with packaging plasmids, was co-transfected into 293T cells using Lipofectamine 2000 (Invitrogen, USA). After 48 h of transfection, the supernatants containing lentivirus particles were collected and filtered. AML cells Molm-13 and U937 were seeded into 6-well plates at 4 × 10⁵ cells per well and incubated overnight, then infected with lentivirus. Eight hours after infection, the culture medium was replaced. Cells were cultured for another 72 h before puromycin selection was performed to isolate stable BYSL knockdown cells. The sequences of siRNA and shRNA are listed in Table 2, and transfection efficiency was evaluated by qPCR assay.

Table 2.

Sequences of specific siRNAs and shRNAs for knockdown

Name	Sense (5’-3’)	Antisense (5’-3’)
siNC	UUCUCCGAACGUGUCACGUTT	ACGUGACACGUUCGGAGAATT
siBYSL-1	GUGCCAUAGAGAUGUUCAUTT	AUGAACAUCUCUAUGGCACTT
siBYSL-2	GCUGCUGGAUAAGAAGUAUTT	AUACUUCUUAUCCAGCAGCTT
shBYSL	CCGGGGGTTTGGAAGGACACCAAGACTCGAGTCTTGGTGTCCTTCCAAACCCTTTTTG	AATTCAAAAAGGGTTTGGAAGGACACCAAGACTCGAGTCTTGGTGTCCTTCCAAACCC
shNC	CCGGCCTAAGGTTAAGTCGCCCTCGTTCAAGAGACGTAGGGTGACTTAACCTTAGTTTTTTG	AATTCAAAAACCTAAGGTTAAGTCACCCTACGCTCTCTTGACGAGGGCTGACTTAACCTTAGG

NC: Negative Control

Cell proliferation and soft-agar colony formation assay

Cell counting kit-8 (CCK-8) assay was used to assess the proliferation of AML cells. 1 × 10⁴ cells/well in 100 µl volume were seeded in 96-well culture plates with triplicate wells. After being cultured for 24, 48 and 72 h, 10 ul CCK-8 reagent (Meilunbio, China, MA0218) was added to each well and then incubated for an additional four hours at 37 °C. The optical density (OD) at 450 nm was measured using a microplate reader (Bio-Rad Laboratories, USA). To observe the cell colony formation ability, AML cells were suspended in 0.3% agar in complete medium and seeded into a 24-well plate containing a base layer of 0.5% agar in complete medium. The cells were cultured for 3 weeks, after which the colonies were stained and counted using AlphaView software.

Cell apoptosis and cycle assays

The Annexin V-FITC/PI kit (Invitrogen, USA) was used for flow cytometry to detect and quantify apoptotic cells. Cells were collected and washed with PBS, then resuspended in staining buffer at a concentration of 1 × 10⁶ cells/ml. Subsequently, the cells were incubated with 5 µl of Annexin V-FITC and 10 µl of PI at room temperature in the dark for 10 min. Finally, the stained cells were analyzed using a flow cytometer (Beckman, USA) and FlowJo software (LLC, USA). For cell cycle analysis, cells were collected, washed twice with cold PBS, fixed in 70% ethanol at 4 °C overnight, then washed twice with cold PBS and incubated with 500 µl of a mixture containing PI and RNase A for 30 min at 37 °C in the dark. Flow cytometric analysis was then performed using a flow cytometer (Beckman, USA).

Gene set enrichment analysis (GSEA)

Functional enrichment analysis comparing the low- and high-BYSL expression groups was conducted using GSEA 4.4.0 software, and significance in enrichment was established with a normal p-value (NOR) < 0.05 and a false discovery rate (FDR) q-value < 0.25.

Chromatin immunoprecipitation (ChIP)-qPCR assay

We performed the chromatin immunoprecipitation (ChIP) assay according to the protocol provided with the ChIP kit (Millipore, USA). Cells were treated with formaldehyde to cross-link proteins and DNA at the BYSL promoter, and sonicated to shear the DNA. Chromatin was immunoprecipitated using antibodies against c-MYC or negative control IgG. We designed primers (Table 1) to amplify the BYSL promoter region, covering predicted c-MYC binding site identified by JASPAR. The immunoprecipitated DNA was quantified by qPCR, and normalized to the input.

Statistical analysis

All data are presented as the mean ± standard deviation (SD). The relationship between BYSL expression and clinicopathological characteristics was assessed using Pearson’s χ² test. Survival curves were generated by the Kaplan-Meier method and compared by the log-rank test. For univariate survival analyses, patients were stratified into high- and low-BYSL expression groups using the median BYSL expression value as the cut-off threshold. Hazard ratio (HR) and 95% confidence interval (CI) were calculated using Cox proportional hazards regression models, with low-BYSL expression as the reference group. Assumptions of proportional hazards were verified using Schoenfeld residuals. For comparisons between two groups, statistical significance was determined using the Student’s t-test. For experiments involving more than two groups, including rescue assays, one-way analysis of variance (ANOVA) followed by post hoc testing was performed. A two-tailed P value of less than 0.05 was considered statistically significant.

Results

BYSL is upregulated in AML patients based on database analysis and clinical validation

To investigate the potential role of BYSL in AML, we analyzed its expression in AML patients and normal controls using datasets from the GEO database. Analysis of the GSE114868 and GSE7186 datasets revealed that BYSL is significantly upregulated in AML patients compared to healthy controls (P < 0.0001) (Fig. 1A). To validate these findings, we measured BYSL expression by RT-qPCR in peripheral blood mononuclear cells (PBMCs) from 93 newly diagnosed AML patients and 10 healthy donors. BYSL expression was significantly higher in AML patients compared to healthy controls (P = 0.0086) (Fig. 1B).

Fig. 1.

BYSL is aberrantly overexpressed in AML patients. (A) Transcriptional expression of BYSL in normal controls and AML patients from two GSE datasets. (B) BYSL expression in PBMC of normal controls and AML patients analyzed by RT-qPCR in the validation cohort. ** P < 0.01, **** P < 0.0001. PBMC: peripheral blood mononuclear cell

Analysis between BYSL expression and clinicopathological characteristics of AML patients

A total of 80 patients with complete clinical data were enrolled, with a median age of 45 years (range: 15–84 years), comprising 42 males and 38 females. Of these, 69 patients received intensive chemotherapy based on an anthracycline and cytarabine regimen, while 11 patients received hypomethylating agent (HMA)-based therapy. Patients were stratified into two groups based on BYSL expression levels, using the median value (1.7173, calculated with GAPDH as the reference) as the cutoff. Patients with BYSL expression above this value were categorized into the high-expression group, while those below were classified into the low-expression group. We analyzed the correlation between BYSL expression and various clinical factors. As shown in Table 3, patients with low BYSL expression had a mean hemoglobin concentration of 86.23 g/L, compared to 74.24 g/L in those with high BYSL expression. Patients with high BYSL expression had more severe anemia (P = 0.044). No statistically significant differences were observed in white blood cell (WBC) or platelet (PLT) counts between the two groups. Additionally, no significant differences were found in bone marrow blasts proportion, risk stratifications distribution, or the presence of gene mutations including FLT3-ITD/NPM1, DNMT3A, and TET2.

Table 3.

Correlation of BYSL expression with clinicopathological characteristics of AML patients in the validation cohort

Characteristics	BYSLLow (n = 40)	BYSLHigh (n = 40)	P-value
Age (Years, mean ± SD)	43.75 ± 1.83	46.65 ± 2.92	0.403
Sex (Male/Female)	20/20	22/18	0.654
WBC (× 109/L)	39.95 ± 6.52	55.49 ± 12.9	0.271
Hb (g/L)	86.23 ± 3.95	74.24 ± 4.356	0.044
PLT (× 109/L)	48.61 ± 6.58	60.98 ± 11.9	0.367
Blasts in bone marrow (%)	60.61 ± 3.909	66.77 ± 4.604	0.309
Treatment regimen			0.518
Intensive chemotherapy	36 (90.0%)	33 (82.5%)
HMA-based therapy	4 (10.0%)	7 (17.5%)
Treatment response (CR1, %)	22 (55.0%)	21 (52.5%)	0.823
Risk stratification (n, %)			0.893
Favorable	6 (15.0%)	7 (17.5%)
Intermediate	28 (70.0%)	26 (65.0%)
Adverse	6 (14.0%)	7 (17.5%)
Gene mutation (n, %)
FLT3-ITD/NPM1			0.584
ITD-/NPM1-	25 (62.5%)	29 (72.5%)
ITD+/NPM1	9 (22.5%)	6 (15.0%)
ITD-/NPM1+	3 (7.5%)	1 (2.5%)
ITD+/NPM1+	3 (7.5%)	4 (10.0%)
FLT3-TKD	2 (5.0%)	1 (2.5%)	1.000
DNMT3A	3 (7.5%)	6 (15.0%)	0.479
IDH1/IDH2	6 (15.0%)	9 (22.5%)	0.390
KRAS/NRAS	2 (5.0%)	3 (7.5%)	1.000
TP53	2 (5.0%)	3 (7.5%)	1.000
TET2	3 (7.5%)	9 (22.5%)	0.117
RUNX1	5 (12.5%)	2 (5.0%)	0.429
CEBPA	8 (20.0%)	10 (25.0%)	0.592
WT1	4 (10.0%)	4 (10.0%)	1.000
KIT	2 (5.0%)	4 (10.0%)	0.675

WBC: White blood cell; Hb: Hemoglobin; HMA: hypomethylating agent; Cytogenetic risk was classified as favorable, intermediate, or adverse according to the European LeukemiaNet (ELN) 2017 guidelines [30]

High BYSL is correlated with poor prognosis in AML patients

To determine whether BYSL expression impacts the prognosis of AML patients, we analyzed BYSL levels and survival data using the TCGA-LAML and TARGET-AML datasets. Patients in each cohort were divided into high and low BYSL expression groups based on the median BYSL expression level. In the TCGA-LAML dataset, analysis revealed that both OS and event-free survival (EFS) were significantly shorter in the high BYSL group compared to the low BYSL group, with HR = 1.55 (95% CI: 1.04–2.32), P = 0.03 for OS and HR = 1.53 (95% CI: 1.06–2.23), P = 0.02 for EFS (Fig. 2A). Similarly, in the TARGET-AML dataset, patients with high BYSL expression exhibited shorter OS (HR = 1.94 (95% CI: 1.24–3.05), P = 0.004), although no significant difference was observed in EFS between the two groups (Fig. 2B). These results suggest that BYSL may serve as a valuable prognostic marker for AML patients.

Fig. 2.

Impact of BYSL gene expression on survival for AML patients. (A, B) Kaplan-Meier survival analyses of patients from the TCGA-LAML (A) and TARGET-AML (B) cohorts, stratified by high- and low-BYSL expression groups (median cut-off). Log-rank test P-values and Cox HR (95% CI) are indicated. (C, D, E) Association between BYSL expression levels and OS and EFS in de novo AML patients at our center (median cut-off). Subgroup analyses in chemotherapy-only (D) and normal karyotype (E) patients include log-rank P-values and Cox HR (95% CI). OS: Overall survival; EFS: Event-free survival; HR: Hazard ratio; CI: confidence interval

Next, we analyzed the prognostic data of AML patients at our center. As in the previous cohorts, patients were divided into high and low BYSL expression groups using the median BYSL expression as the cut-off. No statistically significant differences were found in EFS and OS between the two groups (Fig. 2C); however, the OS curve showed a trend toward separation, prompting a subgroup analysis. Among the 80 patients, 9 underwent hematopoietic stem cell transplantation. Analysis of the 71 patients who received chemotherapy only revealed that those with high BYSL expression had a shorter OS (median OS 13.6 months vs. 18.8 months), with a statistically significant difference between the two groups (HR = 2.58 (95% CI: 1.03–6.46), P = 0.037) (Fig. 2D). In a subsequent analysis of 36 patients with a normal karyotype, those with high BYSL expression showed a worse OS (HR = 3.15 (95% CI: 1.01–9.88), P = 0.048). Together, these findings indicate that BYSL overexpression predicts poor prognosis in AML patients.

A novel risk model based on the BYSL expression exhibits good predictive power for OS in AML patients

We developed a novel risk model for predicting OS in AML patients (Fig. 3). This model was constructed using multivariable Cox regression analysis using data from the TCGA database (Table 4). The factors included in the model were: age (< 60 years or ≥ 60 years), BYSL expression level (high or low), cytogenetic risk status (favorable, intermediate, or adverse), transplantation status (yes or no), and ITD/NPM1 mutation status. (Transplantation was modeled as a binary variable; its prognostic association should be interpreted with consideration of potential temporal bias, as discussed in the Limitations section). The concordance index (c-index) of our model was 0.754, indicating high predictive accuracy. Risk score was calculated based on the weighted factors in the nomogram. Using the median risk score as the cutoff, patients were stratified into low-risk and high-risk groups. Kaplan-Meier survival analysis revealed that patients in the high-risk group had significantly worse OS compared to those in the low-risk group (P < 0.001). ROC curve analysis of the internal validation cohort demonstrated that the model retained good predictive accuracy for 1-year and 2-year OS, with AUC values of 0.80 and 0.82, respectively (Fig. 4A). Consistent with this, survival analysis based on the risk score grouping showed that patients in the high-risk group had significantly poorer OS (HR = 3.58 (95% CI: 2.35–5.44), P < 0.0001) (Fig. 4B).

Fig. 3.

Construction of a novel risk model for the OS of AML patients

Table 4.

Univariate and multivariate Cox regression analyses of OS in AML patients from the TCGA database

	Univariate Cox analysis			Multivariate Cox analysis
	HR (95%CI)	P-value		HR (95%CI)	P-value
BYSL (High vs. Low)	1.718(1.063–2.778)	0.027		1.665(1.001–2.771)	0.040
Age (≥ 60y vs. <60y)	2.607(1.721–3.947)	< 0.001		1.588(1.168–2.603)	0.046
WBC (× 109/L) (≥ 100 vs. <100)	1.874(0.939–3.740)	0.174
Blasts in bone marrow (%) (≥ 70 vs. <70)	1.032(0.662–1.608)	0.889
Hb (g/L) (≥ 90 vs. <90)	1.359(0.903–2.045)	0.140
PLT (× 109/L)  50–99 vs. <50	1.179(0.740–1.878)	0.488
≥100 vs. <50	1.020(0.595–1.748)	0.943
Transplantation(Yes vs. No)	0.401(0.264–0.608)	< 0.001		0.329(0.198–0.547)	< 0.001
Cytogenetic risk status
intermediate vs. favorable	2.964(1.273–6.899)	0.011		2.951(1.076–8.088)	0.035
adverse vs. favorable	4.732(1.976–11.329)	< 0.001		6.397(2.352–17.406)	< 0.001
Gene mutation (Yes vs. No)
FLT3-ITD/NPM1
ITD+/NPM1- vs. ITD-/NPM1-	1.455(0.991–3.063)	0.073		2.466(1.102–5.518)	0.028
ITD-/NPM1 + vs. ITD-/NPM1-	0.899(0.503–1.610)	0.722		1.003(0.479–2.098)	0.993
ITD+/NPM1 + vs. ITD-/NPM1-	1.073(0.548–2.101)	0.836		1.246(0.595–2.610)	0.558
FLT3-TKD	1.781(1.022–3.554)	0.043		2.113(0.978–4.565)	0.056
DNMT3A	1.465(1.034–2.323)	0.043		1.332(0.808–2.197)	0.260
CEBPA	0.969(0.487–1.931)	0.930
IDH1/IDH2	0.791(0.472–1.325)	0.374
KRAS/NRAS	0.777(0.389–1.549)	0.474
TP53	4.331(2.185–8.585)	< 0.001		2.523(1.152–5.526)	0.020
TET2	1.052(0.509–2.173)	0.891
RUNX1	1.652(0.994–3.052)	0.059		1.873(0.943–3.719)	0.072
WT1	0.868(0.450–1.674)	0.673
KIT	0.457(0.144–1.446)	0.183

WBC: White blood cell; Hb: Hemoglobin; CR1: Complete remission 1; Cytogenetic risk was classified as favorable, intermediate, or adverse according to the European LeukemiaNet (ELN) 2017 guidelines [30], which were applied to ensure consistency with the TCGA-LAML dataset used for model development and the timing of cohort enrollment

Fig. 4.

Predictive performance of the novel risk model for the OS of AML patients. (A, D) The sensitivity and specificity of the novel risk model for AML patients were assessed using the ROC curves and AUC values. (B, E) Kaplan-Meier survival analysis of OS for AML patients in the low- and high-risk groups. (C, F) Calibration plot of the nomogram model. AUC: Area under the curve; OS: Overall survival; ROC: Receiver operating characteristic

We then validated the model using the independent validation cohort. The ROC curve shows that the AUC for predicting 1-year and 2-year OS is 0.70 and 0.69, respectively. Patients with high-risk had worse OS compared to those with the low-risk (HR = 2.73 (95% CI: 1.01–7.06), P = 0.019) (Fig. 4E). These findings suggest that the newly developed nomogram-based prognostic model effectively distinguishes between high-risk and low-risk patients.

Knockdown of BYSL suppresses AML cell growth and colony formation ability

We measured BYSL expression in several AML cell lines and found that, compared to PBMC, most AML cell lines showed higher BYSL mRNA levels (Fig. 5A). We also investigated the protein expression level of BYSL (Fig. 5B). Given the significantly elevated expression of BYSL in AML cell lines, we used siRNAs to knock down BYSL in Molm-13 and U937 cell lines to investigate its functional role and underlying mechanism. BYSL knockdown efficacy was confirmed 48 h after transfection, with qPCR showing reduced mRNA levels (Fig. 5C) and Western blot analysis showing decreased protein levels (Fig. 5D). We further evaluated the effect of BYSL knockdown on AML cell proliferation using the CCK-8 assay. As shown in Fig. 5E, BYSL knockdown significantly inhibited cell proliferation in a time-dependent manner. The soft-agar colony formation assay using shRNA demonstrated that BYSL silencing reduced AML cell viability and colony numbers (Fig. 5G), confirming that BYSL knockdown suppresses AML cells growth and colony formation.

Fig. 5.

Knockdown of BYSL suppresses AML cells growth and colony formation ability. (A and B) Investigation of BYSL expression in seven AML cell lines. (C) RT-qPCR analysis of BYSL mRNA expression in Molm-13 and U937 cells transfected with specific siRNAs. (D) Western blot analysis of BYSL protein expression in Molm-13 and U937 cells transfected with specific siRNAs. (E) Detection of AML cell proliferation by CCK-8 assay at 24, 48, and 72 h after BYSL knockdown. (F) qPCR and Western blot analysis of BYSL expression after transfected with specific shRNA. (G) Significant reduction in cell colony number following BYSL knockdown. PBMCs: Peripheral blood mononuclear cells; ** P < 0.01; *** P < 0.001; **** P < 0.0001; ^## P < 0.01; ^### P < 0.001; ^#### P < 0.0001

Knockdown of BYSL induces cell cycle arrest at the G0/G1 phase in AML cells

Next, we investigated the effect of BYSL knockdown on cell apoptosis using flow cytometry with Annexin V/PI staining. No significant increase in the proportion of early or late apoptotic cells was observed in si-BYSL AML cells compared to the control group (Fig. 6A), suggesting that the inhibition of cell proliferation caused by BYSL knockdown is not primarily due to increased apoptosis. We also assessed the impact of BYSL knockdown on cell cycle progression in AML cells. Cell cycle analysis revealed a decrease in the S phase population and induction of G0/G1 arrest in BYSL-knockdown cells compared with the control group (Fig. 6B). Consistently, Western blot analysis showed significant downregulation of key cell cycle regulators, including Cyclin D1 and Cyclin A2, following BYSL knockdown (Fig. 6C).

Fig. 6.

Effects of BYSL knockdown on AML cell apoptosis and cell cycle distribution. (A) Cell apoptosis rate of Molm-13 and U937 cells transfected with specific siRNAs or the negative control. (B) Cell cycle distribution of Molm-13 and U937 cells transfected with specific siRNAs or the negative control. (C) Expressions of cyclinD1 and cyclin A2 in Molm-13 and U937 cells after BYSL knockdown. * P < 0.05; ** P < 0.01; ns: no significance

BYSL is associated with activation of the PI3K/AKT/Cyclin D1 pathway in AML cells

To better understand the underlying mechanism of BYSL in AML, we divided 136 AML patients from the TCGA database into two groups based on BYSL expression. Then, we performed GSEA analysis. We found that the PI3K_AKT_SIGNALING (normalized enrichment score (NES) = 2.19, NOM p = 0.006, FDR = 0.008) and MYC_TARGETS (NES = 2.19, NOM p < 0.001, FDR = 0.004) gene sets were significantly enriched in AML patients with high BYSL expression (Fig. 7A and B).

Fig. 7.

BYSL is functionally associated with the PI3K/AKT/Cyclin D1 pathway in AML cells. (A and B) PI3K_AKT_SIGNALING and MYC_TARGETS gene sets are enriched in AML patients with high-BYSL expression in the TCGA database. (C) Expression of PI3K, p-PI3K, AKT, and p-AKT in Molm-13 and U937 cells after BYSL knockdown. (D) Expression of PI3K/AKT/Cyclin D1 pathway proteins in Molm-13 cells pre-treated with 20 µmol/L 740Y-P for 24 h, with or without BYSL knockdown. Phosphorylated protein levels were expressed as the ratio of phosphorylated to total protein, while total protein levels were normalized to GAPDH. *** P < 0.001; **** P < 0.0001

We next evaluated the association between BYSL expression and PI3K/AKT signaling in AML cells. Western blot analysis showed that BYSL knockdown reduced the phosphorylation levels of PI3K and AKT compared with control cells (Fig. 7C). Densitometric quantification confirmed significant decreases in p-PI3K/PI3K and p-AKT/AKT ratios following BYSL silencing. To assess whether activation of the PI3K/AKT pathway could mitigate the effects of BYSL knockdown, Molm-13 cells were treated with the PI3K activator 740Y-P (20 µmol/L) after transfection with control or BYSL-targeting siRNA. As shown in Fig. 7D, densitometric analysis across three independent biological replicates demonstrated that 740Y-P treatment significantly increased p-PI3K/PI3K and p-AKT/AKT levels compared with BYSL knockdown alone. Similarly, the reduced expression of Cyclin D1 and Cyclin A2 induced by BYSL silencing was partially restored upon 740Y-P treatment. These findings suggest that BYSL expression is functionally linked to PI3K/AKT/Cyclin D1 pathway activity in AML cells.

c-MYC binds to the promoter of BYSL and regulates its expression in AML cells

To investigate the underlying mechanism of BYSL upregulation in AML, we first analyzed the expression correlation between BYSL and c-MYC in 55 AML cell lines from the DepMap database. The results showed a moderate positive correlation (Pearson r = 0.370) (Fig. 8A). Additionally, we analyzed the correlation between BYSL and c-MYC expression in the TCGA-LAML dataset, which also showed a moderate positive correlation in clinical samples (Pearson r = 0.505) (Fig. 8B). Based on the GSEA analysis and prediction from the JASPAR database we hypothesized that c-MYC could bind to the promoter region of the BYSL gene (Fig. 8C). To verify this, we performed a ChIP-qPCR assay using anti-c-MYC antibody to pull down the binding DNA. As shown in Fig. 8D, BYSL promoter DNA sequences were significantly enriched in the c-MYC antibody pulldown compared to the negative control IgG (P < 0.0001), suggesting that c-MYC binds to the BYSL gene promoter region. Furthermore, we treated Molm-13 and U937 cell lines with the c-MYC inhibitor 10,058-F4 (30 µmol/L, 24 h) (#CSN12467, CNS pharm). The expression of c-MYC and BYSL proteins was significantly decreased (Fig. 8E), suggesting that BYSL overexpression in AML is potentially regulated by c-MYC. To assess reciprocal regulation, c-MYC protein levels were examined following BYSL knockdown in Molm-13 and U937 cells. As shown in Fig. 8F, BYSL knockdown did not alter c-MYC protein expression in either cell line.

Fig. 8.

C-MYC binds to the promoter of BYSL and regulates its expression in AML cells. (A) Pearson correlation analysis of c-MYC and BYSL expression in 55 AML cell lines in the DepMap database, showing a moderate positive correlation (Pearson r = 0.370). (B) Pearson correlation analysis of c-MYC and BYSL expression in TCGA-LAML samples, showing a moderate positive correlation (Pearson r = 0.505). (C) Prediction of binding site for c-MYC to the promoter of BYSL by JASPAR database. (D) ChIP-qPCR verifying the binding of c-MYC to the BYSL promoter. IgG antibody was used as the negative control. (E) Western blot analysis of c-MYC and BYSL protein levels in AML cells pre-treated with 30 µmol/L 10,058-F4 for 24 h. (F) Expression of c-MYC in Molm-13 and U937 cells after BYSL knockdown. TSS: Transcription start site; NC: Negative control; *** P < 0.001

Discussion

Previous studies have shown that BYSL overexpression is strongly associated with higher recurrence and mortality in breast cancer patients [31]. Additionally, BYSL promotes glioblastoma cell migration, invasion, and mesenchymal transformation via the GSK-3β/β-Catenin signaling pathway [32], and enhances cell proliferation by activating the AKT/mTOR pathway through complex formation with RIOK2 [24]. BYSL overexpression has been confirmed in various solid tumors, implicating its role in tumorigenesis and progression. Furthermore, studies using different tumor cell lines have shown that BYSL knockdown leads to changes in cellular proliferation and clonogenicity, suggesting its regulatory function in these processes. However, limited research exists on its role in AML.

In the present study, we found that BYSL was markedly upregulated in AML across public datasets and our validation cohort. High BYSL expression was significantly associated with more severe anemia. This association may be related to BYSL’s role in ribosome biogenesis, although direct evidence linking BYSL overexpression to nucleolar stress or impaired erythropoiesis in AML remains to be established. In particular, excessive ribosome production can lead to inefficient processing of rRNA precursors, resulting in the accumulation of free ribosomal proteins that activate p53-dependent apoptosis in erythroid progenitors, thereby hindering terminal erythroid maturation and contributing to anemia [33, 34]. Such mechanisms are well-documented in ribosomopathies, like Diamond-Blackfan anemia, where ribosomal defects selectively block erythroid differentiation while sparing other lineages [35]. In AML, high BYSL expression may further promote leukemic dominance in the bone marrow niche, suppressing normal erythropoiesis through PI3K/AKT hyperactivation and fostering ineffective hematopoiesis [36, 37]. These insights suggest that BYSL not only serves as a prognostic marker but also as a potential mediator of AML-associated cytopenias, warranting further investigation in erythroid-specific models. Moreover, BYSL’s role in ribosome biogenesis underscores its oncogenic potential in AML. As a key component in 40 S subunit maturation with PNO1 and NOB1 [17, 18], dysregulated BYSL may trigger ribosomal stress, characterized by nucleolar disruption and p53/MDM2 activation [38, 39]. In AML, this stress paradoxically promotes leukemic survival by rewiring metabolism and influencing hematopoietic stem cell (HSC) regulation, disrupting quiescence and leading to HSC exhaustion and leukemic progenitor expansion. From a therapeutic perspective, targeting BYSL or ribosome biogenesis holds promise. Inhibiting RNA polymerase I (e.g., CX-5461) or inducing nucleolar stress sensitizes AML cells to chemotherapy or venetoclax by exacerbating ribosomal stress [40, 41]. BYSL knockdown may synergize, overcoming resistance in high-BYSL subsets. Future studies should explore BYSL inhibitors or CRISPR targeting in xenografts.

The PI3K/AKT pathway regulates crucial cellular processes such as proliferation, apoptosis, and the cell cycle [42–45], with constitutive activation observed in AML [46–51]. In the present study, we observed that BYSL knockdown resulted in a consistent reduction in PI3K and AKT phosphorylation, accompanied by altered expression of Cyclin D1 and Cyclin A2. These findings provide functional evidence that BYSL is involved in the modulation of PI3K/AKT signaling activity in AML cells. Previous studies in glioma have reported that BYSL can interact with RIOK2 and the mTORC2 complex to enhance AKT phosphorylation and downstream signaling [24]. However, this protein-protein interaction has not been experimentally validated in AML, and no direct evidence currently supports the existence of a BYSL-RIOK2-mTOR complex in hematologic malignancies. Therefore, whether BYSL influences PI3K/AKT signaling through a similar molecular mechanism in AML remains speculative. BYSL functions as a critical regulator of 40 S ribosomal subunit maturation and its dysregulation is expected to trigger the canonical ribosomal stress response observed across multiple ribosome biogenesis factors. GSEA analysis links high BYSL expression to MYC targets and PI3K/AKT activation, potentially linking ribosome biogenesis and oncogenic signaling programs [38, 52]. Further studies incorporating co-immunoprecipitation and in vivo models will be required to determine whether the BYSL-RIOK2-mTOR axis described in solid tumors is conserved in AML and whether such interactions represent exploitable therapeutic vulnerabilities [48].

c-MYC is a key transcription factor that regulates 15% to 20% of the genome, playing crucial roles in proliferation, differentiation, and metabolism. It is overexpressed in various malignancies, including AML [53–56]. While the mechanisms behind c-MYC upregulation in AML remain incompletely understood, studies suggest that aberrant fusion genes like RUNX1-RUNX1T1 and PML-RARA contribute to its overexpression [57]. c-MYC upregulation is sufficient to induce significant changes in myeloid precursor differentiation and promote leukemogenesis [56–60]. Additionally, c-MYC has been shown to regulate BYSL expression in Burkitt lymphoma cells, highlighting its broader impact on tumorigenesis [61]. Consistent with these findings, our study indicate that c-MYC transcriptionally regulates BYSL in AML through direct binding to its promoter region, as confirmed by ChIP-qPCR assays and pharmacological inhibition of c-MYC, which resulted in a marked reduction of BYSL protein expression. Assessment of reciprocal regulation showed that knockdown of BYSL did not alter c-MYC protein levels in AML cell lines, suggesting that BYSL does not exert feedback regulation on c-MYC at the protein level under the conditions tested. Together, these results are consistent with a unidirectional regulatory relationship, in which BYSL functions as a downstream transcriptional target embedded within a broader c-MYC-driven transcriptional program. As a c-MYC target gene, BYSL may contribute to leukemic cell proliferation and cell-cycle progression by functionally associating with oncogenic signaling pathways such as PI3K/AKT/mTOR or GSK-3β/β-catenin, as suggested by prior studies in solid tumors [20–24, 32]. While the full c-MYC transcriptional network is likely to carry greater prognostic and biological significance, our findings identify BYSL as a functionally relevant downstream effector and a potentially more accessible therapeutic target within AML.

High BYSL expression is associated with poor prognosis, serving as a potential biomarker for risk stratification and treatment guidance in AML. Although our findings are encouraging, some limitations should be acknowledged. The relatively small size of our institutional validation cohort may limit the generalizability of the results. Additionally, while our prognostic associations are significant, they remain correlative rather than causal. Our single-gene model may have lower predictive power compared to multi-gene signatures [62, 63], such as those based on immune or proptosis-related panels [64–66]. Although our bioinformatics and clinical analyses demonstrate a strong association between BYSL expression and AML prognosis, further studies are needed to validate its etiological role in leukemogenesis. While in vitro knockdown experiments provide mechanistic evidence that BYSL promotes AML cell proliferation and cell-cycle progression via the PI3K/AKT pathway, causal validation at the population level is still required. Future studies should incorporate BYSL into integrative multi-gene or multi-omics prognostic models (e.g., gene-pair or transcriptomic-genomic-clinical integrations) to enhance risk stratification and enable personalized therapeutic strategies. Moreover, Mendelian randomization analyses [67], as recently applied to immune phenotypes [68], ascorbic acid exposure [69], and telomere length [70] in AML, could rigorously assess the causal impact of BYSL genetic variants on disease risk and outcome. Ultimately, in vivo validation using inducible depletion systems, patient-derived xenografts, or BYSL-specific inhibitors will be essential to confirm its therapeutic potential and overcome its essential role in normal hematopoiesis. Finally, transplantation status was coded as a binary variable in the prognostic model rather than as a time-dependent covariate. Because transplantation is a post-diagnosis intervention, this approach may introduce immortal time (temporal) bias. Importantly, the TCGA-LAML dataset used for model development lacks precise transplantation timing information for most patients, which precluded the use of landmark analysis or time-dependent Cox regression. As a result, the predictive performance of the model should be interpreted with caution, particularly when applied to treatment-naïve patients at the time of diagnosis. Nevertheless, given the strong and well-established prognostic impact of transplantation in AML and its central role in clinical decision-making, we retained this variable to enhance the clinical relevance of the nomogram. Future validation in large, multicenter cohorts with standardized treatment timelines will be necessary to further assess model robustness and generalizability.

In summary, our findings support a role for BYSL in AML, as a transcriptional target of c-MYC and potentially through modulation of the PI3K/AKT signaling pathway. These results lay the foundation for future investigations into targeted therapeutic strategies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Jie Gao, Fujue Wang, Yingying Chen and Pengqiang Wu. The first draft of the manuscript was written by Jie Gao and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the Bethune Charitable Foundation (YDTR-046 for Jie Gao) and the National Natural Science Foundation of China (81570148 for Xianmin Song).

Data availability

Due to the inclusion of potentially identifiable clinical and genomic information and restrictions imposed by participants’ consent and applicable regulations, the datasets are not publicly available. De-identified data may be made available from the corresponding author upon reasonable request, subject to prior approval by the Ethics Committee of West China Hospital of Sichuan University and execution of a data use agreement.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

This study was approved by the Ethics Committee of West China Hospital of Sichuan University, in accordance with the Declaration of Helsinki, as well as applicable national regulations on human subjects and genomic data.

Consent to participate

Written informed consent was obtained from all participants or their authorized guardians.

Consent to publish

Not applicable. This manuscript contains no individual person’s data (including images, videos, or detailed case descriptions) that would require consent to publish.