Steven M. Foltz

The bone marrow immune microenvironment (BMME) is a critical factor in multiple myeloma (MM) pathogenesis. Despite higher incidence of MM among self-identified African Americans, the relationship between African genetic similarity and the BMME remains unclear. We present the largest analysis to date assessing the impact of Ancestry on the BMME in myeloma (N=320, including 64 with ≥50% African genetic similarity, AFR-High) leveraging our recently generated Immune Atlas. AFR-High was associated with enrichment of late T-effector cells, driven by rare cytotoxic CD4⁺ T cells, suggesting a distinct adaptive BMME. AFR-High exhibited expansion of late-differentiated adaptive NK cells with high cytotoxic GNLY and GZMH expression. AFR-High depicted reduced CD14⁺ monocytes, increased CD16⁺ non-classical monocytes, and expansion of macrophage-like monocyte populations with enrichment of metabolic adaptation, efferocytosis and phagocytic clearance programs, and immunoregulatory pathways. These findings reveal a shift from an innate inflammatory immune landscape toward adaptive cytotoxic and immunoregulatory remodeling associated with African genetic similarity in MM.

Compared with the general population, individuals with neurofibromatosis type 1 have a 50-fold higher risk of developing a high-grade glioma (HGG) in their lifetime. Despite improved understanding of the molecular and cellular drivers of these neoplasms, we have yet to translate this knowledge into therapies that improve overall survival. One limitation has been the paucity of in vivo models for drug testing within this population. We generated three distinct glioma stem cell lines from HGGs arising in mice with Nf1 and Trp53 mutations in cis (NPcis) and characterized the allografts resulting from one Nf1-glioma stem cell line (Nf1-HGG17) by single-cell and single-nucleus RNA sequencing. Because our cell lines are grown in stem cell media, there is an inherent reduction of the differentiation states present in primary HGG, with only oligodendrocyte precursor-like and neural-like cells identified. However, orthotopic allografts of Nf1-HGG17 regained the other differentiation states typically observed in HGG and glioblastoma (neuronal progenitor cell-like and astrocyte-like). About half of the neoplastic cells cluster separately from these classical groups and highly express genes of the antigen presentation machinery, including Cd74, which we also observed in patient samples. Notably, heterozygosity of Nf1 within the tumor microenvironment does not cause marked changes in the immune tumor microenvironment, gene expression, or differentiation states of neoplastic cells. These data indicate that allografted HGG lines from NPcis mice are an effective model of NF1-HGG, mimicking the complexity observed in human HGG that can be used for larger-scale in vivo drug screening and evaluation.

Supplementary Figure S2. Metrics of quality control and variability in the snRNA-seq experiments.

Supplementary Figure S1. Nf1-HGG5746 cells readily engraft in wild-type mice and form tumors with 100% penetrance that histopathologically resemble human disease.

Supplementary Figure S3. Neoplastic cells and cell states.

Supplementary Figure S4. Characterization of antigen presenting tumor cells.

Supplementary Figure S5. Characterization of immune cluster.

Supplementary Table S1. Immunofluorescence markers used to identify cellular differentiation states.

Supplementary Table S2. Pseudobulk comparison of the Antigen Presenting Tumor cells to all other neoplastic cells.

Supplementary Table S3. Marker genes used for immune cluster merging and annotation.