Write a comment

This preprint investigates the adaptations that brain-infiltrating peripheral monocytes undergo to phenotypically resemble resident microglia in the central nervous system (CNS). Using single-cell RNA sequencing, ATACseq, flow cytometry, immunohistochemistry, and qPCR in a busulfan-conditioned bone marrow chimera mouse model, the authors report that once infiltrated, monocytes progressively upregulate canonical “microglia-specific” markers such as P2RY12, TMEM119, FCRLS, and IBA1. Moreover, they undergo extensive chromatin remodeling at microglial gene loci, accompanied by enrichment of transcription factor motifs associated with microglial identity. The authors ultimately conclude that infiltrating monocytes acquire transcriptional and epigenetic features that make them indistinguishable from embryonic microglia by conventional phenotyping, while maintaining potentially distinct inflammatory potential. This study is highly relevant to ongoing efforts to rigorously define both microglial identity and the roles of infiltrating monocytes in the CNS, particularly at a time when interest in their contribution to neuroinflammatory and neurodegenerative processes is rapidly expanding.

Strengths: The study is technically sophisticated, integrating complementary molecular, cellular, and epigenetic approaches that together provide a comprehensive view of monocyte-to-microglia reprogramming. The dataset provides valuable multiomic insight into the dynamic plasticity of CNS myeloid populations and meaningfully contributes to a field faced with the limitations of canonical markers.

Weaknesses: Some interpretative claims—particularly regarding the extent of “indistinguishability by conventional phenotyping” and the generalization of these findings to physiological contexts—would benefit either from additional supporting data or from language that more explicitly acknowledges the limitations of the experimental model. Clearer figure labeling and methodological detail would also improve transparency and strengthen alignment between data and conclusions.

Major Points

1.     The evidence presented shows that infiltrating monocytes acquire many microglia-like molecular and epigenetic features, but the conclusion that they become fully “indistinguishable” from embryonic microglia by conventional phenotyping requires additional evidence. The current data suggest substantial convergence toward a microglial profile, yet also indicate retained specificity such as residual separation between day 0 and day 7 clusters (Fig. 1A), partial overlap rather than merging of CD45 expression (Fig. 1E), lower P2RY12 and TMEM119 levels in engrafted cells compared to resident microglia (Figs. 2–3, S2), persistence of CCR2⁺Ly6C⁺ subsets (Fig. 5), and distinct transcription factor motif enrichment (Fig. 6). Together, these examples suggest that infiltrating monocytes adopt many canonical microglial features while maintaining discrete molecular signatures, which influences how the results can be interpreted. It would be helpful for the authors to comment on this possibility and clarify whether they view these differences as biologically meaningful or within the expected range of phenotypic convergence. Quantifying the fraction of marker-positive cells across time points and assessing functional microglial behaviors (e.g., cytokine release) would further clarify the extent of convergence.

2.     The authors suggest that monocyte infiltration and reprogramming occur under physiological conditions. However, the study design relies on busulfan-mediated myeloablation to label infiltrating monocytes, an approach that alters hematopoiesis, cytokine signaling, and blood–brain barrier integrity, potentially facilitating both monocyte entry and reprogramming. Without controls in unconditioned or lineage-traced animals, it remains uncertain whether the observed engraftment represents a physiological process or an artifact of conditioning. The authors should clarify the limitations of this approach and suggest follow-up studies using complementary models such as genetic lineage tracing in order to determine whether similar reprogramming occurs under physiological conditions.

3.     ATAC-seq results clearly demonstrate increased accessibility at microglial gene loci (e.g., P2ry12, Tmem119, Ms4a7, C1qa) among infiltrating monocytes. However, without direct manipulation of key transcriptional regulators (e.g., PU.1, IRF8, MITF), the observed motif enrichment indicates only a correlation between the loci accessibility during monocyte reprogramming and transcription factor sequences associated with microglial identity. Without demonstrating TF binding or necessity, motif enrichment alone cannot establish mechanistic causation. Integrating ATAC-seq and scRNA-seq data, such as by linking accessibility peaks to differentially expressed genes or motif expression correlations, would connect chromatin remodeling to transcriptional outcomes more convincingly, and incorporating CUT&RUN, ChIP-seq, or perturbation experiments would further strengthen mechanistic inference. At minimum, the authors’ claim that chromatin remodeling “drives” phenotypic convergence should be rephrased to emphasize association, and the discussion could be further strengthened by offering alternative interpretations such as chromatin remodeling functioning as a permissive or downstream adaptation to microenvironmental cues.

4. The discussion effectively highlights the implications of marker ambiguity. However, while the authors persuasively challenge the field’s reliance on canonical markers, they stop short of leveraging their own data to propose distinguishing criteria or mechanistic frameworks for future studies. For example, sustained enrichment of MITF and NF-κB1 motifs and persistent accessibility of inflammatory gene loci could represent actionable distinguishing features. Emphasizing these differences would shift the discussion from identifying what existing approaches get wrong to demonstrating what the field can do right.

Minor Points

1.     The manuscript alternates between “monocyte,” “macrophage,” “microglia-like,” and “monocyte-derived microglia” when describing the same population. Consistent nomenclature would prevent confusion between lineage and phenotype.

2.     Exact n values, statistical tests, and effect sizes are not consistently reported in figure legends. Including these for each panel would improve transparency and reproducibility.

3.     In Figure 1, the caption does not define what the t-SNE plots in panels C–D represent or how they relate to the UMAP in 1A. Clarifying whether these reflect transcriptional clusters, marker expression, or sample time points would improve interpretability.

4.     In Figure 4, motif analysis compares naïve microglia to engrafted macrophages but omits injury conditioned microglia as a control, making it difficult to separate monocyte-specific injury responses from those of the broader myeloid population. Including injury-conditioned microglia in the comparison, or explicitly clarifying why such a control was not incorporated, would help contextualize the motif differences.

5.     In Figure 6, the caption could briefly clarify that this panel summarizes preceding findings rather than presenting new quantitative data, to ensure readers interpret it as a schematic overview.

6.     The limitations section could be expanded beyond busulfan conditioning to address restricted sample sizes, potential phenotypic reversal at later timepoints, lack of functional validation, and the conflation of injury-related and physiological contexts.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.