Additional genetic targets are present in the majority of AML patients with NPM1, KMT2A or NUP98 aberrations potentially impacting combination therapy with menin inhibitors Free

Introduction:

The exploration of menin inhibitors has opened new ways for targeted treatment – however, it is still unclear exactly which AML cases would benefit from menin inhibition. Recent studies have highlighted the clinical potential of menin inhibitors for AML with NPM1 mutations (NMP1m) or rearrangements of KMT2A (KMT2Ar) or NUP98 (NUP98r), all of which are characterized by activated HOX/MEIS genes, while the efficacy is unclear for patients with KMT2A-PTD (partial tandem duplication). In addition, the combination of menin inhibitors with FLT3 inhibitors or other targeted drugs has shown first promising results.

Aim:

We aimed to identify - based on HOX/MEIS gene expression signatures - AML cases, which may benefit from menin inhibition therapy, and to comprehensively evaluate the frequency, pattern, and clonal hierarchy of targetable co-mutations.

Methods:

We enrolled 13,458 adult patients diagnosed with de novo AML. Bone marrow and/or peripheral blood samples underwent examination through cytomorphology, cytogenetics, FISH, panel sequencing, fragment length analysis for FLT3-ITD detection, and RT-PCR for KMT2A-PTD detection. For a subset of 769 patients, whole genome and transcriptome sequencing was performed, and cases were clustered according to the expression of 21 HOX/MEIS-signature genes.

Results:

Hierarchical clustering of AML patients according to expression of HOX/MEIS-signature genes revealed four different clusters. Cases with KMT2Ar, NPM1m, or NUP98r, which have all been connected to menin sensitivity, were exclusively present (NPM1m, NUP98r) or enriched (KMT2Ar) in the two clusters characterized by high HOX/MEIS expression (p<0.001). In addition, cases with KMT2A-PTD were also highly enriched in these HOX/MEIS-high clusters (19-fold, p<0.001); we therefore included KMT2A-PTD as a distinct subset for further analysis. Within the entire cohort, 3,837 patients (29%) showed a genetic aberration potentially eligible for menin inhibitor therapy: 367 KMT2Ar (3%), 3,403 NPM1m (25%) and 67 NUP98r (3% of cases specifically tested). In addition, we identified KMT2A-PTD in 544 patients (7% of cases specifically tested). Targetable co-mutations in FLT3, IDH1 and IDH2 were detected in 63% of menin inhibitor eligible patients indicating a high potential for combination therapies. However, these targetable co-mutations were much less frequent in KMT2Ar cases (17%) compared to the others (NPM1m: 68%; NUP98r: 69%; KMT2A-PTD: 64%). In the KMT2Ar cohort, only FLT3 tyrosine kinase domain (TKD) mutations were detected at relevant frequency (11%), while FLT3-ITD or IDH1/2 mutations were very rare (≤2%). The NUP98r cohort showed a unique mutational landscape, with FLT3-ITD mutations being notably prevalent (57%), while all others were rare (≤5%). The NPM1m cohort was more heterogeneous with FLT3-ITD (39%), IDH2 (19%) and IDH1 (15%) mutations, and FLT3-TKD (13%) recurrently present. Interestingly, the co-mutation pattern of cases with KMT2A-PTD (FLT3-ITD 27%, IDH2 24%, IDH1 18%, and FLT3-TKD 7%) was very similar to the one of NPM1m but differed significantly from the KMT2Ar cohort. Analysis of clonal hierarchy revealed KMT2Ar and NUP98r as mainly ancestral (median clone size 85% and 84%), while NPM1m were often secondary (62%). In particular, IDH1 (77%) and IDH2 (79%) mainly showed larger clone sizes, while the clone size was comparable for FLT3 (54%) and NPM1.

Conclusions:

The HOX/MEIS-high gene expression signature of AML with KMT2A-PTD warrants further investigation of sensitivity to menin inhibitor treatment. The frequent presence of targetable FLT3 or IDH1/2 co-mutations in cases with NPM1m, KMT2Ar and NUP98r AML offers possibilities for combination therapies in the majority of patients. While the combination of menin and FLT3 inhibition is already being clinically tested in NPM1m and KMT2Ar patients, our findings on prevalence, mutational patterns, and clonal hierarchies in 13,458 cases provide evidence for additional combinations. The targetable co-mutational pattern of KMT2A-PTD cases was similar to NPM1m but completely different to KMT2Ar patients where mainly FLT3-TKD mutations were found. Within the small subset of NUP98r cases, FLT3-ITD inhibition could be particularly interesting for combination therapy. As we continue to unravel the genetic intricacies of AML, the prospect of personalized, mutation-specific treatments becomes increasingly tangible, promising progress of precision medicine in hematology.