TMT-based comprehensive proteomic profiling identifies serum prognostic signatures of acute myeloid leukemia

Abstract

Acute myeloid leukemia (AML) is classified into favorable-risk, intermediate-risk, and poor-risk subtypes. This study aimed to compare the serum proteomic signatures of the three AML subtypes and identify prognostic biomarkers for AML. Serum samples from patients with favorable-risk (n = 14), intermediate-risk (n = 19), and poor-risk AMLs (n = 18) were used for the analysis of tandem mass tag (TMT) labeling-based quantitative proteomics. Comparative analysis was performed to identify differentially expressed proteins (DEPs) between groups. Prognostic proteins were screened using binary logistics regression analysis. TMT-MS/MS proteomics analysis identified 138 DEPs. Fumarate hydratase (FH), isocitrate dehydrogenase 2 (IDH2), and enolase 1 (ENO1) were significantly upregulated in poor-risk patients compared with favorable-risk patients. ELISA assay confirmed that patients with poor-risk AMLs had higher levels of IDH2, ENO1, and FH compared with intermediate-risk AML patients. Logistics analysis identified that proteins 3-hydroxyacyl-CoA dehydrogenase type-2 (HADH, odds ratio (OR) = 1.035, p = 0.010), glutamine synthetase (GLUL, OR = 1.022, p = 0.039), and lactotransferrin (LTF, OR = 1.1224, p = 0.016) were associated with poor prognosis, and proteins ENO1 (OR = 1.154, p = 0.053), FH (OR = 1.043, p = 0.059), and IDH2 (OR = 3.350, p = 0.055) were associated with AML prognosis. This study showed that AML patients had elevated levels of FH, IDH2, ENO1, LTF, and GLUL proteins and might be at high risk of poor prognosis.

Keywords: TMT labeling-based quantitative proteomics; favorable-risk acute myeloid leukemia; isocitrate dehydrogenase 2; prognostic biomarker; proteomic profiling.

Figure 1

The statistics of DEPs in patients with acute myeloid leukemia. (a and b) The statistics diagram and the Venn diagram of the DEPs by different comparisons, respectively. FR, favorable-risk. IR, intermediate-risk. PR, poor-risk.

Figure 2

The PPI network of DEPs in patients with acute myeloid leukemia. (a and b) The PPI network of DEPs by different comparisons: favorable-risk vs intermediate-risk (Cyan), favorable-risk vs poor-risk (Red), and intermediate-risk vs poor-risk (Green), involving biological processes and pathways, respectively. Black stars indicate six potent proteins: FH, GLUL, LTF, ENO1, HADH, and IDH2. (c and d) The two modules. Upregulated and downregulated proteins are shown by circles and triangles, respectively. The two significant modules were identified using the MCODE plugin in Cytoscape (http://apps.cytoscape.org/apps/mcode) with the threshold score ≥5.0.

Figure 3

The serum contents of six proteins by ELISA assay. ENO1, enolase 1; FH, fumarate hydratase, mitochondrial; GLUL, glutamine synthetase; HADH, hydroxyacyl-CoA dehydrogenase; IDH2, isocitrate dehydrogenase 2; LTF, lactotransferrin. Differences across groups were analyzed using the non-parametric Kruskal–Wallis H test (Tukey post-hoc test). Patients with myelodysplastic syndromes, lupus nephritis, or thrombopenia were enrolled as controls. Data are expressed as scattered plots, and the boxed values indicate median with range (min to max).

Figure 4

Correlation of protein-encoding genes with overall survival in acute myeloid leukemia. Prognosis verification was performed using the GEPIA online tool, based on the TCGA data (n = 106). The cutoff of high and low expression is the median value of gene expression. The HR was calculated based on Cox proportional hazard model. The dotted line indicates 95% CI.