The bone marrow (BM) microenvironment (ME) is critical for homeostasis of hematopoiesis. However, in the context of acute myeloid leukemia (AML), how the BM ME is altered by the malignant clone and its contribution to AML progression, treatment responsiveness, and relapse remains to be elucidated. While our AML clone-centric understanding of the oncogenesis is expanding, robust assessment of the AML-impacted stroma is difficult due to limited accessibility of adequate sampling material. In this abstract, we identified a potential set of stromal markers for informing induction chemotherapy response. Previously, we performed bulk RNA-sequencing (RNA-seq) of core bone marrow biopsies (cBMBs) collected at diagnosis and at 14/15 days after the 7+3 (daunorubicin/cytarabine) chemotherapy and detected significant differences in the stromal profiles (GEO GSE213210) between samples obtained from responders vs. non-responders (defined as > 5% AML blasts after induction chemotherapy). Strikingly, cBMBs of responders post treatment compared to diagnosis demonstrated increased expression of mesenchymal stem and progenitor cells (MSPC) markers compared to non-responder cBMBs, which showed enhanced expression of osteo-committed markers (Treaba et al., British Journal of Haematology, 2023). These observations led us to hypothesize that preexisting stromal characteristics are linked to a signature of favorable prognosis in AML.

To first characterize transcriptomic signatures at diagnosis, we analyzed differentially expressed transcripts in cBMBs from responders compared to non-responders. At this time point, responder cBMBs exhibited significantly upregulated presence of transcripts of key osteogenic markers such as osterix (SP7), alkaline phosphatase (ALPL), and biglycan (BGN), with log2 fold changes of 1.95, 2.84, and 2.00, respectively (all adj p-value < .05). Expression of key chondrogenic markers including collagen 2 alpha 1 (COL2a1), collagen 9 alpha 1 (COL9a1), aggrecan (ACAN), and SRY-box transcription factor 6 (SOX6) were also significantly increased, with a log2 fold difference of 2.01, 3.02, 2.55, and 2.12, respectively (all adj p-value < .05). Consistently higher expression of these osteo-chondrogenic transcripts in responders suggests the presence of a well-differentiated stroma in these patients prior to treatment that may better support residual normal hematopoiesis and be linked to a better response to standard cytotoxic chemotherapy.

To then validate the contrasting differentiation states of the BM stroma in responders and non-responders, we collected diagnostic BM aspirates of 6 new patients (3 responders/3 non-responders). Aspirates were sorted into hematopoietic and stromal components based on expression of hematopoietic lineage markers, and bulk RNA-seq was performed on both fractions. In both cBMBs and stromal fraction aspirates, osteo-, chondro-, and mesenchymal-associated transcripts LDL receptor-related protein 1 (LRP1), syndecan 1 (SDC1), transglutaminase 2 (TGM2), and galectin 3 (LGALS3) were upregulated in responders. LRP1 had a log2 fold change in cBMBs and aspirates of 0.74 (adj p-value = .09) and 1.72 (adj p-value = .06), respectively. SDC1 had a log2 fold change in cBMBs and aspirates of 1.52 (adj p-value < .05) and 1.77 (adj p-value = .07), respectively. TGM2 had a log2 fold change in cBMBs and aspirates of 1.48 (adj p-value < .05) and 2.87 (adj p-value < .05), respectively. LGALS3 had a log2 fold change in cBMBs and aspirates of 1.42 (adj p-value < .05) and 1.02 (adj p-value = .15), respectively. Of these targets, SDC1 has previously been shown to be associated with a favorable treatment response in AML (Cheng et al., Trans Cancer Res, 2020). Enhanced expression of a common set of transcripts between two sample types at diagnosis in responders suggests the presence of signature biomarkers predictive of treatment response.

In summary, using two different AML sample types, BM aspirates and cBMBs, we observed that a well-differentiated stroma at diagnosis may be associated with a favorable treatment response to 7+3 induction chemotherapy, and LRP1, SDC1, TGM2, and LGALS3 should be validated as potential positive prognostic biomarkers. These findings call for further in-depth studies into diagnosis-specific stromal markers in AML that could complement standard cytogenetics and somatic mutation testing in predicting therapeutic outcomes.

Disclosures

Reagan:Bristol Myers Squibb: Membership on an entity's Board of Directors or advisory committees; Sanofi: Consultancy; Pfizer: Research Funding. Olszewski:Genmab, Schrodinger, Genentech, Inc., Precision Biosciences, Artiva, Pfizer, Kymera Therapeutics: Research Funding; Genmab, Schrodinger, ADC Therapeutics, BeiGene, Bristol-Myers Squibb: Consultancy.

This content is only available as a PDF.
Sign in via your Institution