Benzene is a widely recognized environmental pollutant that chronic exposure can induce hematotoxicity, augmenting the risk of aplastic anemia and hematological neoplasms, especially in occupationally exposed populations^[1,2]. After entering the body, benzene is metabolized in the liver, lung, and bone marrow, transforming into various metabolites including phenol and hydroquinone (HQ), while an increasing number of studies have demonstrated the pivotal role of these phenolic metabolites in the benzene-induced hematotoxicity^[3-9].

Among these metabolites, HQ is a major active metabolite and often serves as a substitute for benzene in vitro experiments. Our previous study identified that, during hemin-induced erythroid differentiation, HQ distinctly inhibited the expression of globin genes and critical transcription factor genes (GATA-1 and NF-E2)^[7], while transcription of some erythroid-related genes was down-regulated by HQ-induced DNA methylation^[10,11]. Other epigenetic changes, including histone modifications, aberrant expression of ncRNAs, and chromatin remodeling, also involve in HQ-induced hematotoxicity^[12,13]. It has been reported that HQ induced the generation of reactive oxygen species (ROS), the phosphorylation of histone γ-H2AX, and the production of the DNA damage-responsive enzyme PARP-1, consequently leading to cell apoptosis via a mitochondria-mediated apoptotic pathway in TK6 cells^[14]. Consistently, we have proved that exposure to HQ induced increases in ROS, the activity of caspase-8, the expression of Fas and FasL on the cell surface, and the decrease in the cell surface sialic acid level, which not only affected cell cycles but induced the inhibition of erythroid differentiation and apoptosis as well in K562 cells^[15,16]. Recent studies focused on the effects of microRNAs (miRNAs) expression changes in HQ-induced hemotoxicity and leukemogenesis^[9,17,18]. Some researchers have proposed that the down-regulation of miRNA-451a and miRNA-486-5p might involve in HQ-induced inhibition of erythroid cell differentiation in CD34⁺ hematopoietic progenitor cells^[9], besides miR-1246 and miR-224 were the potential major regulators in HQ-exposed K562 cells^[18].

Proteomics aims to identify and quantify the compositions, expression levels, post-translational modification, interactions, and functions in tissues and cells of all proteins in a given sample^[19,20]. Proteins are the final product of gene expression and directly regulate the life activities of organisms. Applying proteomics to environmental toxicology research is conducive to identifying and screening new high-sensitivity protein markers, which will contribute to revealing the toxicological molecular mechanism of exogenous substances more intuitively and deeply^[21-23]. Reports on the toxicological mechanism of hematotoxicity of benzene and HQ mostly focus on genomics and transcriptomic analysis^{[9,17,18,24-26]}, especially the analysis of miRNA expression^[9,17,18,25]. However, there are few relevant reports on proteomic analysis. Although these studies have clarified some important potential mechanisms of benzene and HQ-induced hematotoxicity, they are not as intuitive as proteomic analysis of the final product protein of gene expression and may omit information leading to incomplete analysis.

Based on previous studies, we hypothesized that HQ mainly induces hematotoxicity by inhibiting hematopoietic differentiation, especially erythroid differentiation. Human leukemia K562 cells were derived from a patient with chronic myeloid leukemia and can be induced to red cell differentiation in vitro, which can be served as a good model system for investigating erythroid differentiation^[27,28]. Therefore, through label-free proteomic analysis and bioinformatics methods, this study aims to screen HQ-induced differentially expressed proteins (DEPs) in K562 cells and identify their functions and signaling pathways involved in regulation and the networks of protein interactions, especially in erythroid differentiation, so that we can have a deeper understanding of the mechanism of HQ-induced hemotoxicity.

Human leukemia K562 cells were purchased from Cell Resource Center, Peking Union Medical College (CRC/PUMC, China), and were cultured in RPMI-1640 culture medium (Invitrogen) supplemented with 10% (V/V) fetal bovine serum (FBS) (HyClone), 100 U/mL penicillin (Sigma-Aldrich), and 100 μg/mL streptomycin (Sigma-Aldrich) in a humidified 5% CO₂ incubator at 37 °C. Exponentially growing K562 cells at passages 4–8 after recovery were collected and re-suspended in a fresh culture medium. According to our previous study^[7], after K562 cells were treated with 40 μmol/L HQ (Sigma-Aldrich) for 72 h, the cells were harvested for further study.

Mass spectrometry (MS) analyses were performed at the Peking University medical and health analytical center using 200 μg protein samples. The protein samples were denatured using 8 mol/L urea, diluted using 10 mmol/L dithiothreitol (DTT), alkylated using 1 mol/L iodoacetamide, washed three times with 50 mmol/L ammonium bicarbonate and digested with trypsin at 37 °C overnight. The digestion was quenched with 50 mmol/L ammonium bicarbonate and dried in a vacuum concentrator.

The desalted peptides were separated by high-pH reversed-phase high-performance liquid chromatography (RP-HPLC). The samples were dissolved in mobile phase A (20 mmol/L ammonium formate, pH = 10), mixed well by vortex shock, and centrifuged at 12,000 ×g for 20 min. The supernatants were desalted on the C18 precolumn (3.5 μm, 4.6 mm × 20 mm, water) and eluted in mobile phase B (20 mmol/L ammonium formate (pH = 10) in 80% (v/v) acetonitrile) on BEH130C18 column (3.5 μm, 2.1 mm × 150 mm column, water). The flow rate of mobile phase B was 230 μL/min with a linear gradient as followed: (1) 0–9 min, 5%–15% B; (2) 9–30 min, 15%–50% B; (3) 30–41 min, 50%–90% B; (4) 41–60 min, 90% B. After elution, 27 fractions were collected, combined into 8 fractions with UV absorption at 214 nm, and then dried in a vacuum concentrator.

The dried fractions were dissolved in mobile phase C [0.1% (v/v) formic acid in water] and analyzed by nano-liquid chromatography-tandem mass spectrometry (LC-MS/MS) consisting of ultimate 3,000 coupled to Q-Exactive HF. The mobile phase D [0.1% (v/v) formic acid in acetonitrile] increased from 5% to 90% in 110 mins. The mass spectrometer was operated in the positive-ion mode at an ion transfer tube temperature of 300 °C with 2.2 kV positive-ion spray voltage. Orbitrap Velos was set up in Data-Dependent Acquisition (DDA) mode. The resolution of full-scan MS (350–2,000 m/z) in the Orbitrap analyzer was 60,000; the 20 most intense ions from the preceding MS were sequentially collected; the Normalized collision energy (NCE) was 27%; dynamic exclusion was set to an exclusion duration of 30 s and a repetition count of 1. The tandem mass spectra were searched against the human UniProt Non-redundant Reference database (2018.05) using Proteome Discoverer 2.2 software. As search parameters, a tolerance of 10 ppm was considered for precursor ions (MS search) and 0.02 Da for fragment ions (MS/MS search); the maximum number of missed cleavages was 2; carbamidomethylation of cysteine was set as a fixed modification; protein N-terminal oxidation of methionine was set as a variable modification. The search results are imported into skyline for data analysis of Data-Independent Acquisition (DIA).

The DIA was performed using the same liquid phase separation gradient as DDA; full scan MS spectra (350–1,200 m/z) is divided into 40 Da precursor isolation windows, and each parent window is sequentially selected, fragmented, and collected. All sub-ion information is used for quantification. The search results are imported into Skyline for data analysis. Protein abundances were calculated using the label-free quantitation algorithm (LFQ). LFQ Intensity was used for protein quantitation. The minimum ratio count of LFQ was set to 2. The false discovery rate (FDR) was set to 0.05.

The samples were collected and pretreated in a similar way as Proteomics. The mobile phase D increased from 4% to 90% in 72 mins. The mass range of full-scan MS was 350–1,500 m/z. The resolution, ACG target, and Maximum IT were 30,000 at 200 m/z, 3e6, and 20 ms for the first ion MS, respectively. The ACG target was 1e5 and the maximum IT was 45 ms for the secondary ion MS. Isolation Windows is 1.0 m/z. Skyline software was used to analyze PRM data and differential expression protein data was analyzed with MSstats.

Proteins with a P-value < 0.05 were considered differentially expressed. Differentially expressed proteins (DEPs) analysis of two groups (three biological replicates per group) was indicated with the absolute log₂ (fold change of HQ/C) values. Gene Ontology (GO) enrichment analysis of DEPs was conducted using PANTHER (http://pantherdb.org/). The pathway analysis of DEPs was carried out using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.kegg.jp/). The protein-protein-interaction network was constructed using STRING (http://string-db.org/).

Our previous studies have demonstrated that hemin-induced erythroid differentiation is concentration-dependently and time-dependently inhibited by 0–80 μmol/L hydroquinone in K562 cells^[7]. In the present study, the concentration of 40 μmol/L HQ was selected to correspond to no obvious cytotoxicity but markedly inhibiting erythroid differentiation in K562 cells when exposed for 72 h^[7,11,16]. Label-free proteomic analysis was applied to analyze the differentially expressed proteins in HQ-induced K562 cells. The DEPs were assessed using a P-value less than 0.05. In Figure 1, the volcano plots summarized the fold change and significance for the protein levels in HQ-induced K562 cells. The results showed that 4,082 proteins were identified and 466 proteins were differentially expressed (P < 0.05) including 187 upregulated proteins and 279 downregulated proteins after HQ exposure. Moreover, there were 121 upregulated DEPs and 154 downregulated DEPs over 1.3-fold change after HQ exposure. Table 1 illustrated the top 10 upregulated DEPs and top 10 downregulated DEPs after HQ exposure for 72 h. ERCC1 increased to 9.62-fold of the control group after HQ exposure, which was the most upregulated DEP. IFITM1 decreased to 0.07-fold of the control group after HQ exposure, which was the most downregulated DEP.

Figure 1.  DEPs in HQ-induced K562 cells. K562 cells were induced with 0 and 40 μmol/L HQ for 72 h. The label-free proteomic assay was applied to analyze the DEPs in HQ-induced K562 cells. The volcano plot displays the differences in protein abundance detected between HQ and C groups. Each point represents one of the detected genes. Green points are downregulated DEPs, red points are upregulated DEPs and blue points are proteins without significant changes in HQ-induced K562 cells. Horizontal dotted lines indicate statistical thresholds for a P-value < 0.05 calculated using a two-tailed t-test. Data are representative of three independent experiments. HQ, hydroquinone-induced K562 cells; C, the control group; DEPs, differentially expressed proteins.

PRM is a targeted proteomics strategy that enables simultaneous monitoring of all products of a target peptide under conditions that offer high-resolution and high-mass accuracy^[29]. Therefore, it has been widely applied to the validation of proteomics^[30,31]. In our research, PRM quantitative analysis was performed. Based on the possible mechanisms suggested in the previous studies of HQ-induced hematotoxicity, including inhibition of erythroid differentiation, intervention of cell cycle and cell apoptosis, epigenetic changes (DNA methylation and histone modification) and signaling pathways, 16 upregulated DEPs and 17 downregulated DEPs were validated with PRM (Table 2). In terms of erythroid differentiation, CD44, ECM1, HBAT, ITA2B, CWC25, HBE, GATA1, HBAZ, and HBA were considerably upregulated, while HBG2 and CD11B were downregulated. In terms of cell cycle and cell apoptosis, CDK4 was upregulated, while IKKB and FADD were downregulated. In terms of DNA methylation and histone modification, TRAM1, NPL4, DNMT1 and HAT1 were significantly upregulated, while H2AZ, KDM1A, HDAC2, H2AY, H12, H31, and H32 were downregulated. In terms of signaling pathways, TGFB1 and BRAT1 were considerably up-regulated, while MTOR, SMAD3, STAT3, BCAT1, STA5B, and GSK3B were down-regulated. The changing trends of these DEPs were consistent with the analysis results of proteomic methods, confirming the reliability of the data.

The HQ-upregulated DEPs were annotated with 820 GO annotations according to their biological processes, cellular components, and molecular functions. The top 30 enriched GO terms were shown in Figure 2. GO analysis of HQ-upregulated DEPs showed that the biological processes were enriched in neutrophil mediated immunity, granulocyte activation, proteasomal protein catabolic process, vacuolar transport, and aging; the cellular components were enriched in blood microparticle, vacuolar lumen, vesicle lumen, pigment granule and primary lysosome; the molecular functions were enriched in oxidoreductase activity acting on CH-OH group of donors, misfolded protein binding, ATPase binding, cell adhesion molecule binding and structural constituent of muscle (Figure 2A).

Figure 2.  GO analysis of DEPs in HQ-induced K562 cells. (A) GO enrichment histogram of upregulated DEPs. (B) GO enrichment histogram of downregulated DEPs. GO, Gene Ontology; DEPs, differentially expressed proteins; HQ, hydroquinone.

GO analysis of HQ-downregulated DEPs with 806 GO annotations showed that the biological processes were enriched in ribonucleotide metabolic process, cellular amino acid metabolic process, nucleoside monophosphate metabolic process, generation of precursor metabolites and energy and organic acid biosynthetic process; the cellular components were enriched in oxidoreductase complex, respiratory chain, mitochondrial protein complex, mitochondrial membrane part, and mitochondrial matrix; the molecular functions were enriched in electron transfer activity, cofactor binding, metal cluster binding, ligase activity and magnesium ion binding (Figure 2B).

Previous research has shown that HQ considerably inhibits hemin-induced erythroid differentiation in K562 cells^[7,15,32,33]. Transcriptomic analysis and proteomic analysis contributed to high-throughput data for HQ-induced DEGs and DEPs in K562 cells. According to GO analysis results, there were a total of 23 DEGs and 5 DEPs in erythroid differentiation-related pathways (Table 3). In the positive regulation of erythrocyte differentiation term, HSPA1A, HSPA1B, PRMT1, and STAT1 were considerably upregulated, meanwhile, ARNT, STAT3, NCKAP1L, STAT5B, and TRIM58 were downregulated. In the primitive erythrocyte differentiation term, VEGFA was significantly downregulated. In the enucleate erythrocyte development term, MAEA and LYAR were considerably upregulated, while DMTN, CITED2, BCL6, and LYAR were considerably downregulated. In the erythrocyte differentiation term, KIT, EPAS1, CASP3, and HSPA9 were significantly upregulated, while ALAS2, THRA, DNASE2, INHA, and SPI1 were significantly downregulated. In the erythrocyte maturation, HBAZ and G6PD were considerably upregulated, while ERCC2 was considerably downregulated. These results confirmed that HQ exposure might contribute to these genes and protein expression changes, affecting erythroid differentiation and maturation.

A Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway-based functional enrichment analysis was performed to determine the pathways associated with HQ-induced DEPs. The HQ-upregulated DEPs were annotated with 162 KEGG pathway map identifiers and the downregulated DEPs included 169 KEGG pathway identifiers. The top 14 enriched pathways of upregulated DEPs and the top 19 enriched pathways of downregulated DEPs were shown in Figure 3. KEGG pathway analysis demonstrated that HQ-upregulated DEPs were the most significantly enriched in the lysosome (hsa04142), followed by metabolic pathways (hsa01100), cell cycle (hsa04110), and cellular senescence (hsa04218) (Figure 3A). In addition, the HQ-downregulated DEPs were the most significantly enriched in metabolic pathways, followed by biosynthesis of amino acids (hsa01230), alzheimer disease (hsa05010), and thermogenesis (hsa04714) (Figure 3B).

Figure 3.  KEGG analysis for DEPs in HQ-induced K562 cells. (A) KEGG pathway enrichment analysis of upregulated DEPs. (B) KEGG pathway enrichment analysis of downregulated DEPs. Rich factor indicates the annotated DEPs number and whole background genes number ratio in the corresponding pathway. The size of each bubble corresponds to the number of annotated DEPs and the color gradient depicts the P-value of enrichment significance. KEGG, Kyoto Encyclopedia of Genes and Genomes; DEPs, differentially expressed proteins; HQ, hydroquinone.

Focusing on potential mechanisms of HQ-induced hematotoxicity revealed in previous research^[14-16], there were 3, 6, 5, 4, and 4 DEPs annotated as the KEGG terms of cell cycle, cellular senescence, apoptosis, autophagy–animal, and necroptosis (Table 4). KEGG pathway analysis also showed the potentially related signaling pathways such as mTOR, p53, PI3K-Akt, MAPK, and AMPK signaling pathways (Supplementary Table S1 available in www.besjournal.com). Meanwhile, biological processes like ECM-receptor interaction, protein processing in the endoplasmic reticulum, hematopoietic cell lineage, and nucleotide excision repair were revealed (Supplementary Table S2, available in www.besjournal.com). Input DEPs of these pathways were illustrated in Supplementary Table S3, available in www.besjournal.com. These results suggested that HQ-induced hematotoxicity might be achieved by modulating the expression of these DEPs to affect related pathways.

Focusing on the erythroid differentiation-related pathways in the GO analysis results, the network of the 27 differently expressed proteins and differently expressed genes-related proteins were analyzed using STRING with a medium confidence score threshold of 0.4. After kmeans-clustering these proteins into four categories, an interactome network was built to find out protein-protein interaction and predict functional associations among these significantly altered proteins (Figure 4). By analyzing the predicted protein networks, we determined six differentially expressed genes with the most interactions and they were STAT1, STAT3, CASP3, KIT, STAT5B, and VEGFA. STATs, named signal transducer and transcription activator, were identified as critical transcription factors in mediating virtually all cytokine-driven signaling and are thought to involve in various aspects of hematopoiesis, affecting cell proliferation, differentiation, and cell survival, meaning the dysregulation^[34-36]. As shown in the networks, the differential expressions of STAT1, STAT3, and STAT5B significantly impact the entire protein interaction network in HQ-induced K562 cells, which means the dysregulation of these proteins highly demands our attention.

Figure 4.  Predicted protein networks associated with the proteins upregulated or downregulated in HQ-exposed K562 cells. Cytoscape networks were constructed using 27 DEPs in erythroid differentiation-related pathways. Four bubbles in different colors indicate four clusters. DEPs, differentially expressed proteins; HQ, hydroquinone.

Proteomics is the study of information about all the proteins expressed in an organism at a specific time and place, including the structure, location, and quantities^[19,21,23]. Based on the bioinformatics database and bioinformatics analysis, we can further analyze the interaction relationship between DEPs and the signaling pathways and biological processes they affect. Compared with traditional genomics and transcriptomics analysis, proteomics can help us elucidate environmental pollutants’ toxicological mechanisms more clearly and find the biomarkers more directly and precisely.

In this study, on comparing the HQ-induced and control K562 cells, 466 differentially expressed proteins were screened out, of which 187 proteins were upregulated and 279 were downregulated. In our study, Excision repair cross-complementation group 1 (ERCC1), which forms heterodimers with the XPF endonuclease participating in nucleotide excision repair (NER) and genomic instability^[37,38], was the most upregulated DEP. Myelodysplastic syndromes (MDS) are characterized by ineffective hemopoiesis leading to blood cytopenias^[39]. Previous research implied that ERCC1 promotes the aging process of erythroid cells and the induced ERCC1 expression might reflect the premature aging of the MDS-derived erythroid precursors^[40]. Our result enhanced this discovery and showed that the role of ERCC1 and DNA damage repair in HQ-induced hematotoxicity needs further study.

Following the GO classification of DEPs, the HQ upregulated and downregulated DEPs were annotated with GO annotations including neutrophil-mediated immunity, blood microparticle, ribonucleotide metabolic process, cellular amino acid metabolic process, electron transfer activity, and so on, which were associated with the occurrence of hematotoxicity. Since HQ-induced increase in intracellular ROS levels has been shown to be associated with erythroid differentiation and apoptosis in HL-60 promyelocytic leukemia cells, Jurkat T-lymphoblastic leukemia cells, JHP lymphoblastoid cells and K562 cells^{[12,15,16,41,42]}, we focused on the GO term of oxidative stress and found there were 9 upregulated DEPs (ERCC1, AGAP3, SIRT1, G6PD, HSPB1, TOR1A, SOD2, CAPN2, NUDT1) and 7 downregulated DEPs (NDUFS8, PRDX2, NDUFS2, OXSR1, ADPRHL2, LDHA, PYCR1) in the response to oxidative stress term (GO:0006979), further indicating the essential roles of HQ-induced oxidative stress in K562 cells. Focusing on cell apoptotic, there were 5 upregulated DEPs (CD44, TAF9, SIRT1, HSPB1, SOD2) in the intrinsic apoptotic signaling pathway term (GO:0097193), 6 upregulated DEPs (CD44, TAF9, SIRT1, GSN, HSPB1, SOD2) in the regulation of apoptotic signaling pathway term (GO:2001233) and 4 downregulated DEPs (HK2, NDUFS1, BID, TIMM50) in the apoptotic mitochondrial changes term (GO:0008637). These results confirmed the HQ-induced apoptotic in K562 cells, human TK6 lymphoblastoid cells, human ARPE-19 retinal pigment epithelial cells, and so on described in previous literature^{[14-16,43,44]}, and revealed several potentially related downstream proteins.

Besides HQ-induced oxidative stress and apoptotic, we used KEGG pathway annotations and analysis to find that HQ-upregulated DEPs were the most significantly enriched in the lysosome, metabolic pathways, cell cycle, and cellular senescence. Previously, we found that HQ could induce a significant increase in cell population in the G₀/G₁ phase and a significant decrease in the S phase^[16]. In recent years, it was reported that HQ increased the expressions of E4 transcription factor 1 (E4F1) mRNA and protein in HQ-induced malignant transformed TK6 cells (TK6-HT), and silencing E4F1 could block the G₂/M phase of the cell cycle, inhibiting the progress of cell cycle and cell growth^[8]. The results of our research and analysis were consistent with these previous findings and elucidated the relevant pathways and specific mechanisms.

In addition, most DEPs were annotated as the metabolic pathways map identifier and played important roles in multiple signaling pathways and various biological processes. Among them, the upregulated CDK4 was annotated as an important participant in the p53 and PI3K-Akt signaling pathway, modulating cellular process and life course including cell cycle and cellular senescence. The uncontrolled proliferation of immature myeloid cells is often associated with cell cycle dysregulation, which might inhibit cell differentiation processes^[45,46]. Cyclin-dependent kinases (CDKs) take an important part in cell cycle control, among which CDK4 mediates progression through the G₁ metaphase in conjunction with D-type cyclins^[47,48]. Therefore, the mutation and abnormal expression of CDK4 have important effects on hematotoxicity and disease progression, meanwhile, researchers are committed to developing CDK4-targeted treatment strategies^[47,49].

To further explore the mechanism of HQ-induced hematotoxicity, we focused on the 23 DEGs and 5 DEPs in erythroid differentiation-related pathways in the GO analysis results and constructed the network of protein interactions in K562 cells. We determined 6 DEPs with the most interactions, which were STAT1, STAT3, CASP3, KIT, STAT5B, and VEGFA, and the differential expressions of STAT1, STAT3, and STAT5B make a central impact. Hematopoietic development is highly dependent on cytokine/receptor-initiated signaling pathways^[50]. The activation of STATs mediates cellular signal transduction initiated by cytokines and growth factors with cognate receptors, especially the Janus kinase (JAK)-STAT pathway plays important roles in hematopoiesis by affecting the production of mature hematopoietic cells via effects on cellular proliferation, survival, and lineage-specific differentiation^[51-53]. STAT1 is a pivotal downstream mediator of interferon (IFN) signaling required for IFN-induced hematopoietic stem cells (HSCs) proliferation^[54]. The phosphorylated STAT1 enhanced secretion of TRAIL to induce cell apoptotic in KMS-20 cells through JAK/STAT pathway activation^[55], while recent research showed that STAT1 is also critical for hematopoietic stem and progenitor cells (HSPCs) homeostasis maintenance by regulating HSCs self-renewal and intrinsic functions^[56]. STAT3 is also an important gene expression regulator in the cell cycle, antiapoptosis, angiogenesis, and invasion/migration^[52]. It not only promotes the development of B and T cell subsets but also regulates neutrophil numbers and hematopoietic stem cell self-renewal^[57]. Thus STAT3 has been identified as a therapeutic target for a wide range of diseases^[52,57,58]. In addition to STAT1 and STAT3, STAT5 also regulates cell proliferation, differentiation, and survival critically^[59]. STAT5 deletion can cause significant development defects in both lymphoid and myeloid lineages, and STAT5 is indispensable at later stages for normal granulopoiesis by mechanisms like regulating proliferation and survival genes to direct Granulocyte macrophage-colony stimulating factor (GM-CSF) signaling^[60,61]. Therefore, HQ regulates the expression of these DEGs and DEPs associated with hematopoietic cell development, maturation, and differentiation to affect the production of mature erythrocytes and other blood cells with normal physiological functions. Further elucidation of the functions, interactions, and signal transduction involved in them will help us understand HQ-induced hematotoxicity.

In summary, this study demonstrated that the benzene metabolite HQ induced various protein expression changes in K562 cells, directly affecting intracellular signaling pathways and metabolic activities and consequently leading to hematotoxicity. Our study showed that STATs may be potential biomarkers of HQ-induced hematotoxicity and deserve attention. The DEP-associated changes shown in this article will shed light on future in-depth studies of the underlying mechanisms of HQ-induced hematotoxicity. Focusing on critical DEPs identified, more research can be carried out to find the intrinsic relationship between DEPs and other HQ-induced changes, such as DNA methylation, ROS increase, miRNAs expression changes, and histone modification. Further research and validation at the molecular level are needed to investigate these DEPs and their associated potential mechanisms. Coupled with cross-omics analysis and animal model studies, the mechanism of HQ-induced hematotoxicity will be presented more clearly and completely.