top of page

Monocyte-derived microglia with Dnmt3a mutation cause motor pathology in aging mice

Kim et al. (BioRxiv) DOI:10.1101/2023.11.16.567402

Monocyte-derived microglia with Dnmt3a mutation cause motor pathology in aging mice


  • Monocyte-derived macrophages

  • Dnmt3a mutation

  • Clonal hematopoiesis (CH)

Main Findings

Microglia, originating during embryogenesis, establish a self-contained cellular domain resilient against infiltration from adult definitive hematopoiesis, serving as vital guardians of brain homeostasis. However, the authors findings reveal an unforeseen occurrence within the aging mouse brain. Monocyte-derived macrophages (MoMΦ) exhibit a unique spatial preference by favouring areas such as the nigrostriatum and medulla while notably avoiding colonization of the frontal cortex. Remarkably, these parenchymal MoMΦ seamlessly adopt expression profiles akin to bona fide microglia, introducing a novel subset—monocyte-derived microglia. Originating from hematopoietic sources, MoMΦ become targets of clonal hematopoiesis (CH).

In a chimeric transfer model, the study illustrates the pathogenic potential of MoMΦ expressing Dnmt3aR822H, a prevalent mutation in human CH. These pathogenic MoMΦ, when accumulated, actively contribute to motor deficits reminiscent of atypical Parkinsonian disorders. The mouse model establishes a progressive seeding of MoMΦ within the aging healthy mouse brain, selective accumulation in distinct regions, and the capacity to induce brain pathology when harbouring somatic mutations associated with CH.

This integration of developmental insights and the emergent role of MoMΦ in the aging brain, coupled with the ramifications of Dnmt3a mutations, underscores the intricate interplay between microglial origins, clonal hematopoiesis, and neurodegenerative outcomes. Navigating these complexities, the study's implications extend beyond the mouse model, providing a foundation for understanding and intervening in the broader landscape of neurodegenerative diseases in humans.

  • Monocyte-derived macrophages accumulate in selected brain regions during healthy ageing excluding large areas of the frontal cortex.

  • Age-accumulating monocyte-derived macrophages acquire transcriptomic identities similar to YS-derived microglia including expression of the ‘signature microglia identity” gene Sall1.

  • The common clonal haemopoiesis variant Dnmt3aR822H impacts the global epigenome and transcriptome of monocyte-derived macrophages that seed the brain.

  • BM transplants compromising of HSC with Dnmt3aR822H leads to pathogenetic monocyte-derived macrophages that enter the brain and promote motor deficits resembling Parkinsonian disorders.


  1. To seed the CNS with MoMΦ derived from Dnmt3aR822H HSC, the authors engrafted Cx3cr1gfp/+ animals with BM of 10-month-old VavCre: hDNMT3AR822H mice, or age matched Cre-negative littermate BM as controls. However, it would have been interesting and made a stronger statement if the authors investigated if aged VavCre: Dnmt3aR822H also exhibit spontaneous age-associated accumulation of monocyte-derived macrophages in certain brain regions including the substantia nigra pars compacta, and if these mice also developed motor dysfunction. Also, it is not clear why 10-month-old mice were selected as 10 months is not regarded as ‘aged’ – rather age mice to 20-24 months where possible.

  2. Although the authors show that there is broad transcriptional and epigenetic dysregulation in MoMacs that harbour Dnmt3aR822H, this is very descriptive referring to a ‘more activated’ state. More details on what pathways were affected in mutated cells would enhance our understanding of the effect of the mutation on these cells.

  3. It would further contribute to the authors findings if they also showed that these cells react differentially to an age-related challenge for example LPS due to the epigenic reprogramming caused by the mutation.

  4. Moreover, the validation of these findings in patient biopsies is necessary in understanding the relevance of the disease pathology and relation of mutated Dnmt3a in CH affecting brain pathologies through the MoMΦ seeding from clonal hematopoiesis.

It would further be interesting to know:

  • Why do monocyte-derived macrophages only seed selected areas of the brain during healthy ageing? What are the signals that direct this? Why is the cortex largely excluded for example?

  • Dnmt3aR822H is important in DNA methylation – what effect does this mutation have on the epigenetic profile of these monocyte-derived macrophages and their ability to respond to stimuli?


This study first demonstrates that Dnmt3a mutations in clonal hematopoiesis can result in pathogenic macrophages seeding the brain and resulting in Parkinsonian-like deficits in mice. This provides a unique perspective on the relation between hematologic and neurological disorders that if validated in humans would bring insight into possible future treatment strategies.


Reviewed by Austeja Baleviciute as part of a cross-institutional journal club between the Icahn School of Medicine at Mount Sinai, the University of Oxford, the Karolinska Institute and the University of Toronto.

The author declares no conflict of interests in relation to their involvement in the review.

Point by Point Response by the Preprint Authors

  1. Vav-Cre mice do show rearrangements in microglia, and to the best of our knowledge there is no animals models available that strictly targets HSC, but spares other lineages. To restrict the mutation to HSC and thereby mimic the physiological situation of CH, we therefore resorted to the chimera approach. However, more of concern could be the high frequency of mutant cells in our model. In a revision of our study we therefore will include results with head shielded chimeras that are closer to the situation in patients. The chimeras we analyzed develop the motoric defects by 4 months after irradiation, at an age of 6 months. It is probably fair to assume that the irradiation accelerated aging. In the revised manuscript we will however also include data on the frequency and distribution of MoMg in mice older than 18 months.

  2. We do not show epigenetic dysregulation of mutant MoMg, but just transcriptomic changes. The latter are global, as would be expected from hypomethylation. We show a few selected genes, incl. Ccl5, whose expression reaches significance. However, the definition of the molecular pathways that drive the pathology will require a dedicated study with selected mutagenesis of genes in MoMg.

  3. We already reported in a previous study that HSC-derived macrophages, even when not carrying mutations, show distinct responses to a peripheral endotoxin challenge. Please see Shemer et al. Nat Comm. 2018 (PMID 30523248 ) for details. We agree that this issue should be further explored, especially in the context of mutant MoMg, but this is beyond the scope of the present study.

  4. Our experiments in the BM chimeras suggest that the mere presence of hDNMT3a mutant monocytes in BM or blood does not cause pathology, but  only with time, when the progeny of these cells, accumulates in specific locations of brain and impinges on the CNS environment the consequences of the DNMT3aR882H mutations became evident. It has by now been shown  in two studies that MoMF carrying DNMT3a mutations can be found in human  brains of the elderly (pmid: 30323172, 37322115). The next important  question is whether there is an association of CH mutations, and  specifically DNMT3a variants, and PD risk or other neuro-pathologies. To  provide evidence for that one will have to analyze the brains of PD  patients for presence of CH mutation carrying macrophages in relevant  brain regions, and compare them to aged-matched brains of individuals  without PD. To achieve a conclusive result, sufficient numbers of brain  samples will needed to be analyzed but to the best of our knowledge  brain bank depositories currently do not hold sufficient numbers of  paired blood and tissue samples.

bottom of page