Single-Cell Multiomic Profiling of TP53 Mutated Myelodysplastic Neoplasms - All Clones Are Not Equal

Background

In patients with myelodysplastic neoplasms (MDS) the acquisition of multi-hit rather than single hit TP53 mutations and chromosomal copy number alterations (CNAs) confers a dismal prognosis with no standard of care therapeutics. Currently available clinical diagnostic tests cannot confidently determine if TP53 mutations are single or multi-hit events or whether CNAs are part of the same clone.

Aims

We set out to apply a multiomic platform incorporating single-cell DNA sequencing and surface protein mapping to uncover the clonal architecture and clinically relevant genetic heterogeneity, with the ambition of utilising this test to determine cell surface phenotype and the landscape of TP53 mutations and CNAs.

Methods

We performed single-cell sequencing on non-enriched bone marrow mononuclear cells (BM MNCs) in a cohort of 16 MDS patients with reported TP53 mutation/s. All patients had either an additional chromosome 5 or 7 alteration detected by routine clinical diagnostic testing. We designed a bespoke 621 amplicon panel including 45 recurrently mutated genes in MDS and common SNPs across regions of significant chromosomal copy loss and gain, additionally staining samples with a 42-plex oligo-antibody cocktail for cell surface phenotyping.

Results

13,480 BM MNCs passed protein expression quality control criteria. Of these, 10,791 cells (~80%) presented with concurrent high genotype quality (GQ>20) variant calling data, confirming the expected TP53 mutations in 14/16 patients. Clustering of protein expression data resulted in 10 distinct cell types. Analysis of TP53 variant distribution across the haematopoietic hierarchy revealed that in all but one case, TP53 mutations were absent in T cells, confirming their somatic nature. Comparison of the fraction of wild-type (WT) to TP53 mutated cell types in the remaining patients (n=13) revealed a significant relative expansion of stem and progenitor cells, most strikingly in HSPCs (0.01 vs 0.15, p=0.0007) and erythroid progenitors (0.05 vs 0.25, p=0.0007), with a relative decrease of lymphoid progenitors (0.05 vs 0.01; p=0.0007), CD8 T cell (0.24 vs 0.01, p=0.0002), CD4 T cell (0.22 vs 0.01, p=0.0002), B cell (0.05 vs 0.004, p=0.033).

Leveraging single-cell allele frequency, we then stratified TP53 mutant cells into single and multi-hit TP53 clones, with and without coexistent CNAs, specifically focusing on chromosome 5 and 7 due to their dominance in our cohort and prognostic importance. Most patients (11/13; 85%) had multi-hit clones with linear clonal evolution from wild type through to multi-hit TP53 clones followed by CNAs, with the remainder presenting with branching evolution dictated by loss or amplification (LOH or CNA) of the TP53 mutant alleles. Two patients (2/13; 15%) presented with only single hit mutations, with one harboring del5q.

In addition, single-cell sequencing revealed that a TP53 mutation in one patient detected at a bulk VAF of 44% was in fact a previously unrecognised germline mutation, detected in all cells including T cells. Inferring clonal evolution in this patient, we observed the linear acquisition of several subclones stemming from the germline mutant clone, defined by the acquisition of multiple mutations in TET2 and ZRSR2. Interestingly, the cellular distribution of these clones was sparse and did not mimic that of somatic TP53 mutant clones.

Importantly, we identified TP53 multi-hit or del5q only clones in two low VAF TP53 mutated patients (<20% VAF) MDS patients, respectively 17% and 13.4%, where cellular distribution mirrored multi-hit clones observed in higher VAF (>20% VAF) MDS. This observation challenges the current clinical diagnostics criteria which commonly stratifies TP53 mutated MDS patients into single vs multi-hit based on a 20% VAF cut-off, supporting the potential added value of single-cell genomics analyses.

Conclusion

Our single-cell multi-modal investigation of TP53 mutant MDS revealed distinct cellular distributions of single versus multi-hit clones with a marked HSPC and erythroid expansion, providing additional information with important clinical implications above that provided by routine diagnostic testing.