Cell-free DNA 5-hydroxymethylcytosine is an emerging marker of acute myeloid leukemia

Abstract

Aberrant changes in 5-hydroxymethylcytosine (5hmC) are a unique epigenetic feature in many cancers including acute myeloid leukemia (AML). However, genome-wide analysis of 5hmC in plasma cell-free DNA (cfDNA) remains unexploited in AML patients. We used a highly sensitive and robust nano-5hmC-Seal technology and profiled genome-wide 5hmC distribution in 239 plasma cfDNA samples from 103 AML patients and 81 non-cancer controls. We developed a 5hmC diagnostic model that precisely differentiates AML patients from controls with high sensitivity and specificity. We also developed a 5hmC prognostic model that accurately predicts prognosis in AML patients. High weighted prognostic scores (wp-scores) in AML patients were significantly associated with adverse overall survival (OS) in both training (P = 3.31e-05) and validation (P = 0.000464) sets. The wp-score was also significantly associated with genetic risk stratification and displayed dynamic changes with varied disease burden. Moreover, we found that high wp-scores in a single gene, BMS1 and GEMIN5 predicted OS in AML patients in both the training set (P = 0.023 and 0.031, respectively) and validation set (P = 9.66e-05 and 0.011, respectively). Lastly, our study demonstrated the genome-wide landscape of DNA hydroxymethylation in AML and revealed critical genes and pathways related to AML diagnosis and prognosis. Our data reveal plasma cfDNA 5hmC signatures as sensitive and accurate markers for AML diagnosis and prognosis. Plasma cfDNA 5hmC analysis will be an effective and minimally invasive tool for AML management.

Figure 1

Differentially hydroxymethylated genes (DhMGs) involved in AML signaling pathways. (a) AML signaling pathways generated by IPA. (b) 5hmC levels in DhMGs associated with AML signaling. The DhMGs in AML signaling pathways are all enriched with 5hmC compared to controls (FDR < 0.01). Center line represents median, bounds of box represent 25th and 75th percentiles, and whiskers are Tukey whiskers. (c) Genome browser view of 5hmC distribution of FLT3 in AML and control samples. Peak height represents 5hmC enrichment.

Figure 2

A 5hmC signature differentiates AML patients from controls. (a) Principal component analysis (PCA) in AML patient samples at registration time points and controls using normalized read counts from the 5hmC signature of 70 genes. (b) Unsupervised hierarchical clustering of the 5hmC signature of 70 genes in AML patient samples at registration time points and controls. (c) Boxplot of weighted-diagnostic scores (wd-scores) in controls and AML samples in both training (n = 62) and validation sets (n = 42). Black dashed line represents a cutoff score of 0.263 developed from the training set. P < 0.001, comparison of any of the AML groups to the control group in both training and validation sets. Center line represents median, bounds of box represent 25th and 75th percentiles, and whiskers are Tukey whiskers. (d) Boxplot of wd-scores in controls and other AML samples that were not included in the training, validation, and test sets. Non-HSCT, did not receive hematopoietic stem cell transplantation (n = 26); HSCT Non-CR, received hematopoietic stem cell transplantation and not in complete remission (n = 13); HSCT CR, received hematopoietic stem cell transplantation and in complete remission (n = 21). Center line represents median, bounds of box represent 25th and 75th percentiles, and whiskers are Tukey whiskers. (e) Receiver Operating Characteristics (ROC) analysis of wd-score calculated from the 5hmC signature of 70 genes in validation and test sets. AUC area under curve. CI 95% confidence interval. (f) Boxplot of wd-scores in additional AML samples and controls in the test set. Controls, n = 39. Non-CR did not achieve complete remission, n = 14. CR complete remission, n = 22. Center line represents median, bounds of box represent 25th and 75th percentiles, and whiskers are Tukey whiskers.

Figure 3

A 5hmC signature predicts overall survival in AML patients. (a) Kaplan–Meier analysis of overall survival of AML patients in training set based on wp-scores (n = 50). (b) Kaplan–Meier analysis of overall survival of AML patients in validation set based on wp-scores (n = 26). (c) Overall survival analysis of 61 AML patients with cytogenetic and/or molecular genetic information following the European LeukemiaNet (ELN) risk classification. (d) Association of wp-score with risk stratification of the 61 AML patients categorized using ELN classification. Favorable, n = 7; Intermediate, n = 33; Adverse, n = 21. Black dashed line represents a cutoff wp-score of 18.0 developed from training set. Center line represents median, bounds of box represent 25th and 75th percentiles, and whiskers are Tukey whiskers. (e) Overall survival analysis of the 61 AML patients with cytogenetic and/or molecular information available based on wp-scores. (f) Wp-score further stratifies patients in the intermediate group (n = 33) categorized by ELN recommendations.

Figure 4

The survival-related 5hmC signature is associated with leukemia burden. (a) Boxplot of wp-scores in controls (n = 42) and patient samples at initial diagnosis (n = 11), with chemotherapy (n = 35), after receiving hematopoietic stem cell transplant not in complete remission (HSCT non-CR; n = 13) or HSCT in CR (HSCT CR; n = 11). Red dashed line represents a cutoff wp-score of 18.0 developed from training set. Center line represents median, bounds of box represent 25th and 75th percentiles, and whiskers are Tukey whiskers. (b) Plots of wp-scores in individual AML patients at initial diagnosis and after chemotherapy. Dx, days after initiation of chemotherapy. Black dashed line represents the cutoff wp-score.

Figure 5

A 5hmC levels in a single gene for survival prediction. (a) Kaplan–Meier analysis of overall survival of AML patients in training set based on wp-scores of BMS1 (n = 50). (b) Kaplan–Meier analysis of overall survival of AML patients in validation set based on wp-scores of BMS1 (n = 26). (c) Kaplan–Meier analysis of overall survival of AML patients in training set based on wp-scores of GEMIN5 (n = 50). (d) Kaplan–Meier analysis of overall survival of AML patients in validation set based on wp-scores of GEMIN5 (n = 26).