Work

Exploring the genetic and epigenetic origins of juvenile myelomonocytic leukemia using newborn screening samples

Year: 2021

Type: article

Abstract: We identified 35 patients that were born in the state of California from 1990 to 2017 who were confirmed to have JMML per World Health Organization criteria [ 1 ] and were previously consented to participate in a JMML tissue bank study. None of the patients had Noonan syndrome. Diagnostic JMML material was available for all 35 of these patients. Guthrie cards from the 35 patients were obtained from the California Department of Public Health California Biobank Program. In addition, Guthrie cards were obtained from the California Biobank Program for 12 healthy control subjects born in California who did not develop cancer during childhood. DNA was extracted using standard methods and targeted deep sequencing (Supplemental Table 1 ) as well as methylation profiling were performed. For additional details, see Supplemental Methods . We analyzed a total of 35 newborn blood screening (NBS) cards from children who developed JMML later in childhood and from 12 healthy controls. Clinical characteristics at disease onset showed a median age at diagnosis of 1.5 years, elevated white blood cell count (WBC), monocytes and hemoglobin F (Supplementary Tables 2 and 3 ). At diagnosis of JMML, somatic mutations were identified in 34 of 35 patients. The most common mutations were in PTPN11 (12), NRAS (7) and KRAS (7) (Supplementary Fig. 1A ). Ten of the 34 patients had a secondary mutation at diagnosis. The most common secondary mutations were in NF1 (2) and SETBP1 (2). Of the 34 patients who had a somatic mutation present upon diagnosis of JMML, those mutations were found in newborn cards in 13 (38%) patients using a VAF cut-off of 0.01. Clonal mutations (VAF > 15%) were found in 9 of the 34 (26%) patients. Somatic mutations in NRAS (4) and PTPN11 (3) were the most common detected at birth (Supplementary Fig. 1B ). In three patients with germline CBL mutations, loss of heterozygosity (LOH) was detected at diagnosis but not at birth. Of the 10 patients who had a secondary mutation at diagnosis, none of those were found in newborn blood samples. Patients who had a somatic mutation detected at birth were significantly younger at diagnosis with a median age of 7.1 months compared to 19.8 months in patients who had no mutations at birth (p = 0.03) (Supplemental Table 4 ). However, no difference was observed in EFS or OS for patients with or without somatic mutations at birth (Supplementary Fig. 2 ). Targeted deep sequencing did not identify any mutations in the 12 healthy controls included in our study. To better understand when DNA methylation changes occur in patients, we profiled all 35 patient and 12 control NBS cards with a custom-capture targeted MethylSeq assay. Minimum distance to the nearest centroid classified all NBS cards as having “low” DNA methylation using the international, consensus definition [ 12 ] (Fig. 1A ). While all NBS cards were categorized as displaying a low methylation (LM) signature, three patients clustered separately from all other samples (Supplementary Fig. 3 ). Notably, all three of these patients had a clonal NRAS mutation detected at birth. A JMML NBS or control NBS samples profiled by MethylSeq classified according to the international JMML methylation consensus signature [ 12 ]. B Methylation status at diagnosis of patients ( n = 16) reported in this study. Panel A and B Heatmaps show the beta values of 1386 CpG loci used for methylation classification. Of the 35 patients, 16 had sufficient DNA available from diagnosis for DNA methylation analysis (Fig. 2 ). These samples were classified as LM, intermediate (IM) or high methylation (HM). Nine patients had a LM signature at diagnosis, 4 were IM and 3 were HM (Figs. 1B and 2 ). One patient, UPN3153, had serial samples permitting longitudinal profiling at birth, diagnosis, and post-chemotherapy. The allelic frequency of the patient’s KRAS mutation was 0% at birth, 39% at diagnosis and 27% post-chemotherapy. The NBS card was designated LM while both the diagnostic and post-chemotherapy timepoints were IM (Supplemental Figure 4 ). Despite detection of mutations in nine of the corresponding NBS cards, of which seven were present at a VAF ≥ 19%, all patients had lower methylation at birth compared to diagnosis of JMML (Fig. 2 ). Genetic and methylation profiling of paired NBS and diagnostic samples from JMML patients are displayed. The star indicates the NBS sample of the pair.

Source: Leukemia

DOI: 10.1038/s41375-021-01331-0