Stem Cells and Stealth Immunosurveillance

The hematopoietic compartment is highly susceptible to genomic perturbations, as evidenced by the frequency of mutation-driven clonal hematopoiesis in later decades of life. Several questions remain regarding the protection mechanisms that prevent emergent and potentially leukemogenic clones from becoming dominant — for example, how cells of the adaptive immune system are able to survey hematopoietic stem cells (HSCs) for aberrancy, while also protecting the hematopoietic microenvironment from excessive inflammatory insults.

The acquisition of mutations in protein coding genes can lead to the presentation of neoantigens — peptide sequences that have not previously been encountered by the immune system. All cells present endogenous antigens to a certain degree and are subject to immunosurveillance. Neoantigens are presented by major histocompatibility complex (MHC) class I or class II (MHC-II) molecules that activate CD8+ cytotoxic T cells or CD4+ helper T cells, respectively. While MHC class I proteins are ubiquitously expressed, high constitutive expression of MHC-II and the ability to activate CD4+ T cells are restricted to professional antigen-presenting cells (APCs), such as dendritic cells, macrophages, and B cells.¹  Almost all myeloid cells, including neutrophils, megakaryocytes, mast cells, basophils, and eosinophils; innate lymphoid cells; and certain epithelial, endothelial, and stromal cell types can express MHC-II following stimulation.² 

In a recent report in Cell Stem Cell, Dr. Pablo Hernández-Malmierca and colleagues show that HSCs from mice and humans continuously express high levels of MHC-II at levels only slightly lower than professional APCs, and that constitutive MHC-II expression gradually decreases during differentiation. Indeed, functional stem cell activity correlated with high MHC-II activity. Similar to professional APCs, HSCs efficiently present peptides via MHC-II at homeostasis and further upregulate expression with inflammatory stimulation, and activate antigen-specific CD4+ T cells both in vitro and in vivo. Immunoprecipitation of MHC-II from HSCs of healthy mice and identification of the peptides using mass spectrometry revealed that the majority of peptides were “self” antigens, indicating that this feature may be important for anticancer immunosurveillance.

In line with this theory, interaction with CD4+ T cells led to pronounced HSC proliferation, with differentiation to the myeloid lineage and loss of stem cell potential. To test the consequences of sustained antigen presentation in vivo, the authors generated chimeric mice by co-transplanting equal numbers of wild-type hematopoietic stem and progenitor cells (HSPCs) and HSPCs that constitutively presented ovalbumin peptide. In the absence of antigen specific CD4+ T cells, this resulted in stable 50:50 chimeras. But, when OVA-specific CD4+ T cells were co-transplanted, OVA-expressing HSPCs were rapidly eliminated, resulting in complete engraftment with non–OVA-presenting cells. This confirmed that antigen presentation by HSPCs leads to their elimination.

They next examined mechanisms that prevented interactions between HSPCs and T cells from causing excessive inflammation. The CD4+ T cells interacting with HSPCs adopted immunosuppressive and anti-inflammatory phenotypes, with upregulation of PD-L1 and the immunoregulatory cytokines IL10 and IL27.

To validate the importance of this for anticancer immunosurveillance in humans, they examined the expression of MHC-II in a dataset of 532 acute myeloid leukemia (AML) samples.³  The cellular origin of the leukemia was found to determine the antigen-presenting capacity, with high expression of MHC-II found to be associated with “stem-cell”–like leukemic transcriptome signatures. AMLs arising from genomic perturbations that transform HSCs (e.g., FLT3-ITD) displayed high MHC-II, while AMLs that occur in downstream progenitors (e.g., in association with a NPM1 mutation) had lower expression of MHC-II.

Finally, they showed that the IDH1(R132H) mutation, which generates an MHC-II restricted immunogenic neoantigen, occurred almost exclusively in differentiated NPM1-mutated AML but not in more immature NPM1–wild-type AMLs, whereas AMLs with a non-immunogenic IDH1(R132C) mutation occurred in a comparable proportion of the general AML cohort. This suggested that mutations that generate MHC-II–restricted neoantigens in humans are extinguished if they arise in stem cell–like AMLs with substantial APC capacity, but not in differentiated leukemias. To test this experimentally, they transformed OVA-expressing HSPCs with the MLL-AF9 oncogenic translocation and transplanted these cells into mice with or without OVA-specific OT-II CD4+ T cells. The transformed HSPCs were efficiently removed, and CD4+ T cells expanded and acquired a PD-L1 high phenotype.