AUTHORS
Amy McLemore, MS, Surendra Neupane, PhD2, Maria Balasis3, Zena Komrokji4, Traci Kruer, PhD5, Miguel Martinez6, Adriana Del Pino Herrera7, Angely Pinales8, chaomei Zhang9, Tania Mesa10, Sean Yoder11, Aline Renneville, PhD, PharmD12, Nathalie Droin13, Dorothée Selimoglu-Buet, PhD14, Agnieszka E Gorska, BS15, Kian Rahbari16, Michael Zargari, MD17, Christy Finke, BS18, David Sallman, MD19, Andrew T Kuykendall, MD20, Christopher Ryder, MD PhD21, Jeffrey Lancet, MD22, Rami Komrokji, MD23, Zhuoer Xie, MD,MS24, Rupa Narayan, MD25, Timothy Graubert, MD26, Shannon M. Buckley, PhD27, Ami B. Patel, MD28, Nathan Salomonis, PhD29, Guillermo Montalban-Bravo, MD30, H. Leighton Grimes, PhD31, Mrinal M. Patnaik, MD, MBBS32, Meghan Ferrallfairbanks, PhD33, Michael R. Savona, MD34, Eric Solary, MD35, Eric Padron, MD36
AFFILIATIONS
1 Moffitt Cancer Center, Tampa, Florida, United States
2 Moffitt Cancer Center, Tampa, Florida, United States
3 Moffitt Cancer Center, Tampa, Florida, United States
4 Moffitt Cancer Center, Tampa, Florida, United States
5 Moffitt Cancer Center, Tampa, Florida, United States
6 University of Florida, Gainesville, Florida, United States
7 University of Florida, Gainesville, Florida, United States
8 Moffitt Cancer Center, Tampa, Florida, United States
9 Moffitt Cancer Center, Tampa, Florida, United States
10 Moffitt Cancer Center, Tampa, Florida, United States
11 Moffitt Cancer Center, Tampa, Florida, United States
12 Gustave Roussy, Villejuif, France
13 Gustave Roussy, Villejuif, France
14 Gustave Roussy, Villejuif, France
15 Vanderbilt University Medical Center, Nashville, Tennessee, United States
16 Vanderbilt University Medical Center, Nashville, Tennessee, United States
17 Vanderbilt University Medical Center, Nashville, Tennessee, United States
18 Mayo Clinic, Rochester, Minnesota, United States
19 Moffitt Cancer Center, Tampa, Florida, United States
20 Moffitt Cancer Center, Tampa, Florida, United States
21 Moffitt Cancer Center, Tampa, Florida, United States
22 Moffitt Cancer Center, Tampa, Florida, United States
23 Moffitt Cancer Center, Tampa, Florida, United States
24 Moffitt Cancer Center, Tampa, Florida, United States
25 Massachusetts General Hospital, Boston, Massachusetts, United States
26 Massachusetts General Hospital, Boston, Massachusetts, United States
27 University of Utah, Salt Lake City, Utah, United States
28 University of Utah, Salt Lake City, Utah, United States
29 Cincinnati Children's Hospital, Cincinnati, Ohio, United States
30 University of Texas MD Anderson Cancer Center, Houston, Texas, United States
31 Cincinnati Children's Hospital, Cincinnati, Ohio, United States
32 Mayo Clinic, Rochester, Minnesota, United States
33 University of Florida, Gainesville, Florida, United States
34 Vanderbilt University Medical Center, Nashville, Tennessee, United States
35 Gustave Roussy, Villejuif, France
36 Moffitt Cancer Center, Tampa, Florida, United States
ABSTRACT
CMML is characterized by myeloid-skewed hematopoiesis and expansion of hematopoietic stem and progenitor cells (HSPCs). However, the extent to which malignant HSPC composition and developmental trajectories are altered and clinically relevant remains incompletely defined. To address this, we established an international CITE-seq consortium and generated a large single-cell atlas comprising 2.9 million cells from 307 CMML patients and 80,000 normal CD34-enriched cells. To normalize variability in CD34+ enrichment across samples, we used antibody data to computationally exclude residual CD34- cells. Next, we performed ambient RNA correction and cell annotation using a 300,000 BM-MNC normal reference (Zhang et al., Nat Imm 2024). Using surface antibody data, we manually annotated six major HSPC populations in accordance with immunophenotyping conventions and asked whether RNA- or protein-defined HSPC architecture would reflect clinically meaningful biology. Higher LMPP abundance, defined by RNA or protein annotation, was associated with inferior OS and PFS in univariable analyses. After multivariable adjustment for age, CPSS-Mol, marrow blasts, blood counts, and mutation burden, RNA-defined but not protein-defined LMPP abundance remained independently associated with inferior PFS (HR 1.39 per SD, p=0.016). In contrast, those cases with a more balanced progenitor composition, as defined by higher normalized RNA HSPC Shannon entropy, were associated with improved OS and remained significant after adjustment for age and CPSS-Mol (HR 0.69, p<0.001). Unsupervised RNA compositional clustering identified LMPP-high, MPP-high, ERP-high, and normal-like progenitor-bias states. Analysis of PCA-informed superclustering showed that LMPP/MPP-skewed patients had significantly higher molecular and clinical risks (IPSS-M q<0.001, CPSS-Mol q=0.001, IPSS-R q=0.001), were enriched for myeloproliferative CMML (p<0.001), HMA exposure prior to sampling (p=0.007), and mutations in ASXL1 (q<0.001), EZH2 (q=0.007), SETBP1 (q=0.027), and NRAS (q=0.043), and had inferior OS (p=0.0015) and PFS (p=0.018) compared to the ERP-high/normal-like supercluster. Interestingly, pseudotime analysis of CMML CD34+ cells showed lower HSC-rooted pseudotime compared with normal controls (q<0.001), indicating retention of a more immature transcriptional state. When pseudotime was performed in a patient specific manner, patients with more apically rooted CD34+ cells had inferior OS and PFS. Collectively, these data suggest that static RNA-defined HSPC composition and trajectory are clinically meaningful.
We next evaluated dynamic cell-state features in CMML using scVelo (Bergen et al, Nature Biotech, 2020), which models the kinetics of spliced and unspliced transcripts to infer RNA velocity at single-cell resolution and observed an increased RNA velocity magnitude across 9/15 cell types (q<0.001 to 0.048), even after exclusion of CMML patients with splicing mutations. We assessed the clinical relevance of cell-type specific RNA velocity and observed that higher GMP and LMPP RNA velocity length was associated with inferior OS (p<0.001 and p=0.046) and PFS (p=0.002 and p=0.06), which remained significant after Cox adjustment. Higher LMPP velocity was also associated with RBC transfusion dependence (TD) (p<0.001). Consistent with a more immature transcriptional state in CD34+ cells, CMML RNA velocity was not driven by upregulation of differentiation programs or correlated with downstream cell-type abundance. Instead, RNA velocity and conventional differential gene expression analysis demonstrated that TNF/NFkB and other inflammatory programs were the primary drivers of transcriptional kinetics, particularly in LMPP and GMP subsets. Indeed GMP and LMPP velocity length was enriched in MPN-CMML (q=0.018, q=0.068), linking inflammatory progenitor transcriptional remodeling to proliferative disease.
Our findings demonstrate that progenitor composition and inflammatory cell-state dynamics are clinically relevant features of CMML. RNA-defined HSPC states, differentiation trajectories, and RNA velocity driven by TNF/NFkB gene sets were associated with molecular risk, disease phenotype, TD, and survival, supporting the concept that perturbations in malignant hematopoietic trajectories contribute to CMML progression. These data establish one of the largest single-cell malignancy atlases and nominates inflammatory LMPPs and GMPs as clinically relevant disease-driving HSPC subsets.
DISCLOSURES
Amy McLemore, MS: No relevant conflicts of interest to declare.