Transcriptional and Genomic Characterization of Measurable Residual Disease (MRD) Cells in Acute Myeloid Leukemia (AML)

Background: Failure to achieve complete response (CR) and undetectable MRD is associated with inferior survival in AML. However, there is virtually no data on the molecular traits of treatment resistant cells to understand the mechanisms of MRD resistance. This would be particularly relevant in elderly AML, for which outcomes remain dismal despite increasing CR rates with novel combinations.

Aim: Uncover mechanisms of MRD resistance in elderly AML by comparing the transcriptional and genomic profile of patient-matched leukemic cells at diagnosis and after treatment with semi-intensive chemotherapy or hypomethylating agents.

Methods: 285 AML patients with a median age of 75 were included in the phase 3 PETHEMA-FLUGAZA trial and were randomized to receive induction and consolidation with fludarabine and cytarabine (FLUGA) vs 5-azacitidine (AZA). After consolidation, patients continued with the same treatment if MRD ≥ 0.01% or stopped if MRD < 0.01%. MRD was assessed using multidimensional flow cytometry, and highly-purified MRD cells were isolated according to patient-specific aberrant phenotypes using FACS. Leukemic blasts and T cells (germline control) were also sorted at diagnosis.

A 3’ end RNAseq method optimized for generating libraries from low-input starting material (MARSeq) was used. RNAseq was performed at diagnosis in 220 patients, and in paired diagnostic-MRD leukemic cells from 22 patients. Whole-exome sequencing (WES) was performed using molecular barcoding in paired diagnostic-MRD leukemic cells from 14 patients.

Results: After induction, 35 patients were in partial remission (PR), 51 attained CR/CRi but persistent MRD (CR/MRD+), and 13 achieved CR/CRi and undetectable MRD (CR/MRD-). There were no differences in median overall survival (OS) between patients in PR vs CR/MRD+ (median of 21 vs 20 months, p = .603), and both were numerically inferior to that observed in CR/MRD- patients (median of 27 months, p = .097).

There was only one gene (PIEZO2) differentially expressed between leukemic cells at diagnosis vs after treatment in patients achieving PR (n = 10). By contrast, there were 117 differentially expressed genes between leukemic cells at diagnosis vs after treatment in CR/MRD+ patients (n = 12). To confirm that gene deregulation was intrinsically associated with treatment resistance, the prognostic value of gene (over or under) expression in accordance with the differential observed between diagnostic and MRD leukemic cells, was subsequently analyzed in the 220 patients with RNAseq data at diagnosis. Over expression of 3 genes (IGFBP2, LINC00475 and PLEKHG3) and under expression of 12 genes (eg, CASC15, KDM7A and MEX3B) among the 117 identified above, was significantly associated with inferior OS. Interestingly, none of the genes was prognostic in the BeatAML cohort, suggesting that the transcriptome of MRD resistance may be specific of treatment.

A total of 6,054 mutations were detected after WES of matched leukemic cells at diagnosis and after treatment, in 14 patients achieving CR/MRD+. Amongst these, 4,708 (78%) were detected at both time points, 354 (6%) were present at diagnosis while absent in MRD blasts, and 992 (16%) emerged during MRD resistance (p ≤ 0.008). There were no differences between treatment arms in the number of shared and private mutations in diagnostic vs MRD cells. Amongst the 1,346 mutations that either became undetectable or present at MRD, recurrence (ie, ≥ 3 patients) was observed in 47 genes. Acquired mutations in AVP and UBXN11 were observed in MRD cells from 6 of the 14 patients.

Conclusions: This study showed that elderly AML patients treated with semi-intensive chemotherapy or hypomethylating agents, have dire survival regardless of achieving PR vs CR/MRD+. However, RNAseq analyses of resistant cells uncovered that whereas PR is characterized by primary resistance and transcriptionally stability from diagnosis, CR/MRD+ is associated with the emergence of molecular traits of acquired resistance. These findings could help explaining why additional treatment with the same schema is unable to overcome the poor prognosis of persistent MRD in elderly AML patients achieving CR/CRi. First-ever simultaneous RNAseq and WES of patient-matched diagnostic and MRD blasts, suggests that a few deregulated genes and recurrent mutations with limited or previously unknown role in AML pathobiology, are associated with MRD resistance and inferior OS.