AUTHORS
Caroline C Ennis, BS, Alana M Franceski, BS, MSc2, Daniel Ward Phillips, N/A3, Paran Goel, PhD4, Ashley N Connelly, PhD5, Arthur W Meehan6, Christina R Larson7, Francesca Luca, PhD8, Nicholas Stevanov9, Pritam Apitkatla10, Manasa Vadapalli11, Jamie Lo12, Esha K Patel13, Rob S Welner, PhD14
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
6 University of Alabama at Birmingham, Birmingham, Alabama, United States
7 University of Alabama at Birmingham, Birmingham, Alabama, United States
8 University of Alabama at Birmingham, Birmingham, Alabama, United States
9 University of Alabama at Birmingham, Birmingham, Alabama, United States
10 University of Alabama at Birmingham, Birmingham, Alabama, United States
11 University of Alabama at Birmingham, Birmingham, Alabama, United States
12 University of Alabama at Birmingham, Birmingham, Alabama, United States
13 University of Alabama at Birmingham, Birmingham, Alabama, United States
14 University of Alabama at Birmingham, Birmingham, Alabama, United States
ABSTRACT
Introduction
Acute myeloid leukemia (AML) arising from antecedent myelodysplastic syndrome (secondary AML, sAML) demonstrates poor response to induction and to hypomethylating agent regimens. sAML is also significantly more relapse-prone following transplantation compared to de novo AML (deAML), resulting in a disproportionate share of AML-related mortality among adults over age 60, whose five-year survival remains below 10%. Prognostic frameworks (ELN 2022, IPSS-M) and nearly all drug development approaches conceptualize AML as a cell-autonomous genetic disease. However, clonal hematopoiesis (CH) and MDS frequently share founder mutations with the sAML that arises from them, and sAML exhibits worse clinical outcomes than genotype-matched deAML. This observation suggests that factors beyond the genome contribute to disease aggression. We hypothesized that the CH/MDS-remodeled marrow environment is a key determinant. Testing this hypothesis has been challenging, as no preclinical system previously introduced a fixed AML genotype into CH/MDS-conditioned versus healthy marrow.
Methods
A sequential transplantation system was developed in age-matched recipient mice. Tet2 single knockout cells were engrafted to establish CH, followed 6 months later by Tet2/Asxl1 double-knockout cells to establish MDS. Following confirmation of disease (dysplasia and cytopenias, 3 months), Tet2-knockout/Flt3-ITD AML was transplanted into CH-, MDS-, CH/MDS-bearing, or age-matched healthy hosts (n=5/group). The resulting sAML and deAML populations were genetically identical and differed only in the disease history of the recipient. Experimental readouts included serial peripheral blood, spleen, and marrow chimerism, complete blood counts (CBCs), survival, immunophenotyping of HSPCs, and single-cell transcriptomics of marrow populations.
In parallel, paired marrow specimens collected at MDS diagnosis and at subsequent sAML transformation from the same individuals (n=9 pairs) were obtained from the UAB biorepository under IRB and informed consent. Specimens were analyzed by spectral flow cytometry and single-cell transcriptomics of CD34+ and CD45+ populations, with malignant cell states annotated using established AML hierarchy classifiers.
Results
CH/MDS-bearing hosts developed anemia (RBC 11.2 M/uL to 10.3 M/uL; p=0.0002; n=21/group) and neutropenia (Neus 4.1 k/uL to 2.5 k/uL; p=0.0048; n=21/group), accompanied by reduced survival compared to single-disease-bearing counterparts, confirming the presence of a remodeled marrow environment prior to AML addition. Contrary to expectations, the identical AML clone expanded significantly more slowly in CH, MDS, or CH/MDS-conditioned host marrows than in healthy marrows, in peripheral blood and marrows (~3-fold lower burden at 2.5wks, p=0.0004), with correspondingly prolonged disease latency. As the clone genotype remained constant, this restraint is attributed to host-extrinsic factors. Blasts from marrow with prior MDS also exhibited elevated NFκB signaling relative to those with AML alone (p=0.01, FDR<0.001), indicating the MDS environment influences the AML clone outgrowth. sAML displayed a more primitive stem cell-like phenotype relative to deAML, including elevated SLAM (3.5% of LSK to 37.4% of LSK; p=0.0066; n=4/group), and c-KIT expression (5.4% of Lin- to 15.6% of Lin-; p=0.006; n=4/group), consistent with leukemic stem cell (LSC) enrichment. These blast features suggest antecedent disease states may exert selective pressure favoring more primitive, stem-like clones. We generated single-cell transcriptomic libraries and spectral immune phenotyping of the paired bone marrow aspirates from MDS diagnosis to sAML transformation. sAML marrows indicate enrichment for more primitive blast populations distinct from MDS or deAML marrows. Key signatures derived from our murine sAML, such as NFκB pathway activation and niche-driven stem-like enrichment, are being integrated with our human malignant cell populations.
Conclusions
The CH/MDS marrow microenvironment restrains leukemic proliferation while selecting for and imprinting a primitive, NFκB-activated, LSC-enriched state on a genetically fixed emergent clone. This finding resolves a longstanding clinical paradox: sAML frequently presents with a lower proliferative burden yet demonstrates poorer response and higher relapse rates. The premalignant bone marrow niche thus constitutes a critical target for preventative and therapeutic strategies in sAML.
DISCLOSURES
Caroline C Ennis, BS: No relevant conflicts of interest to declare.