AUTHORS
Mostafa Mohamed, MS, Saeed Zakakhosravi, MSc2, Leah Gehrs3, Amr Elkholy4, Erin Ahn, PhD5
AFFILIATIONS
1 University of Alabama at Birmingham, Birmingham, Alabama, United States
2 University of Alabama at Birmingham, Birmingham, Alabama, United States
3 University of Alabama at Birmingham, Birmingham, Alabama, United States
4 University of Alabama at Birmingham, Birmingham, Alabama, United States
5 University of Alabama at Birmingham, Birmingham, Alabama, United States, University of Alabama at Birmingham, Birmingham, Alabama, United States
ABSTRACT
Background:
The prognosis and response to treatment in AML are significantly influenced by cytogenetic and molecular abnormalities. KMT2A-rearranged (KMT2Ar) AML is caused by chromosome 11q23.3 translocations that produce KMT2A fusion proteins, impair myeloid development, and enforce aberrant epigenetic programming. KMT2Ar depends on menin-KMT2A-mediated transcriptional activation of leukemogenesis genes, a pathway now therapeutically targeted by menin inhibitors. SON is a DNA/RNA-binding protein that functions as a core scaffold of nuclear speckles. We previously showed that full-length SON (SON-F) interacts with menin and restrains KMT2A/MLL-complex assembly and H3K4me3 deposition at leukemia-associated loci. In contrast, the short N-terminal SON-E isoform lacks the C-terminal domains present in SON-F and may alter SON-mediated transcriptional and RNA-processing functions. However, the distribution and functional significance of SON-E and SON-F isoform states in AML remain unknown.
Methods:
To identify known and novel SON isoforms enriched in specific AML subtypes, we isolated CD34+ cells from bone marrow (BM) samples from 12 patients with primary AML and 2 normal controls. Targeted long-read RNA-seq (TEQUILA-seq) was used to identify and quantify the annotated and novel SON isoforms. qPCR and TaqMan assays validated the findings in the same BM samples. To assess clinical relevance, we analyzed SON isoform expression and clinical outcomes using TCGA-LAML and TCGA Splicing Variants Database (TSVdb) data. The TCGA SON-E fraction was calculated from TSVdb isoform measurements mapped to the N-terminal SON-E structural class defined by long-read sequencing. To determine whether SON status affects leukemogenesis in vivo, we isolated BM cells from C57BL/6-background donor mice of 4 genotypes: WT, Son-F overexpression, Son-E overexpression, and Son haploinsufficiency (Son+/-). Following MLL-AF9 transduction, equal numbers of GFP-positive cells from each genotype were injected retro-orbitally into myeloablatively irradiated recipient mice (n = 10/group), and recipient survival was monitored.
Results:
TEQUILA-seq identified two major structural classes of SON transcripts: short N-terminal-only isoforms, including SON-E, and C-terminal-retaining isoforms, including full-length SON. KMT2Ar samples (n = 5) had a higher median N-terminal/SON-E fraction than normal BM controls (84.40% vs. 65.04%), while C-terminal-retaining isoforms were enriched in normal controls. TaqMan and qPCR results showed that KMT2Ar AML samples with enhanced SON-E signal expressed 12.74-fold higher HOXA9 and 4.72-fold higher MEIS1 levels than normal BM controls (p = 0.0002 and p = 0.0004, respectively).
To assess the clinical significance of SON isoform balance, we analyzed TCGA-LAML isoform measurements and clinical annotations. SON-E fraction significantly interacted with TCGA molecular-risk group for disease-free survival (interaction LRT p = 0.0023; n = 168, 81 events), indicating that the prognostic effect of SON-E varies by clinical context. In models adjusted for age and white blood cell count, each 10-percentage-point increase in SON-E fraction was associated with longer DFS in good-risk AML (HR = 0.52, 95% CI = 0.32–0.85, p = 0.009; n = 31, 10 events), but shorter DFS in intermediate-risk AML (HR = 1.60, 95% CI = 1.15–2.22, p = 0.005; n = 92, 54 events). No significant association was observed in poor-risk AML.
Finally, to determine whether SON state functionally influences leukemia progression, we compared survival among recipients transplanted with MLL-AF9-transduced donor cells from the 4 Son conditions. Recipient survival differed significantly among the groups (global log-rank p = 0.0033). Median survival was shortest in recipients of Son-E-overexpressing cells (41.5 days), followed by recipients of Son+/- cells (48 days), whereas recipients of WT and Son-F-overexpressing cells each had a median survival of 57 days.
Conclusions:
These findings identify SON isoform state as a previously unrecognized regulatory layer in AML progression. Targeted long-read sequencing, orthogonal validation, clinical cohort analysis, and functional transplantation collectively demonstrate that SON-E influences leukemia progression in vivo and has context-dependent prognostic correlations in human AML. In the era of menin inhibitors, these results provide a foundation for evaluating SON isoform status in AML classification and therapeutic investigation.
DISCLOSURES
Mostafa Mohamed, MS: No relevant conflicts of interest to declare.