Monocytes differentiate along two alternative pathways during sterile inflammation

Abstract During inflammation, monocytes differentiate within tissues into macrophages (mo‐Mac) or dendritic cells (mo‐DC). Whether these two populations derive from alternative differentiation pathways or represent different stages along a continuum remains unclear. Here, we address this question using temporal single‐cell RNA sequencing in an in vitro model, allowing the simultaneous differentiation of human mo‐Mac and mo‐DC. We find divergent differentiation paths, with a fate decision occurring within the first 24 h and confirm this result in vivo using a mouse model of sterile peritonitis. Using a computational approach, we identify candidate transcription factors potentially involved in monocyte fate commitment. We demonstrate that IRF1 is necessary for mo‐Mac differentiation, independently of its role in regulating transcription of interferon‐stimulated genes. In addition, we describe the transcription factors ZNF366 and MAFF as regulators of mo‐DC development. Our results indicate that mo‐Macs and mo‐DCs represent two alternative cell fates requiring distinct transcription factors for their differentiation.


Introduction
Monocytes are key actors in the maintenance of tissue homeostasis and in inflammatory responses. After exiting the bone marrow, they circulate in the blood and can migrate to peripheral tissues where they rapidly differentiate into macrophages or dendritic cells (DC) (Jakubzick et al, 2017;Guilliams et al, 2018;Coillard & Segura, 2019). The factors orchestrating monocyte fate decision remain poorly understood. In particular, whether these two cellular identities represent alternative differentiation pathways or different stages along a continuum is controversial.
Monocyte-derived macrophages (mo-Mac) and DC (mo-DC) have been evidenced in multiple tissues during acute and chronic inflammation, in both mouse and human (Zigmond et al, 2012;Guilliams et al, 2018;Coillard & Segura, 2019). Mo-Mac and mo-DC are also present in the steady-state in skin, peritoneum, and Peyer's Patches (Tamoutounour et al, 2013;Bonnardel et al, 2015;Goudot et al, 2017). In other tissues at homeostasis, monocytes replenish the pool of macrophages over time, in particular in heart (Epelman et al, 2014;Molawi et al, 2014), intestine (Bain et al, 2014;Shaw et al, 2018) and pancreas (Calderon et al, 2015), but do not seem to differentiate into mo-DC in this context (Liu et al, 2019). Monocytes were initially proposed to be precommitted to become mo-Mac or mo-DC (Menezes et al, 2016). Subsequently, we and others have shown that monocyte differentiation is not transcriptionally imprinted but can be oriented by external signals including cytokines, retinoic acid, and pathogenderived products (Vento-Tormo et al, 2016;Goudot et al, 2017;Mildner et al, 2017;Sander et al, 2017;Devalaraja et al, 2020;Coillard et al, 2021).
Several molecular regulators involved in mo-DC development have been identified using in vitro models and include IRF4, aryl hydrocarbon receptor, BLIMP-1, NCOR2, miR-155, ETV3, and ETV6 (Briseño et al, 2016;Goudot et al, 2017;Sander et al, 2017;Coillard et al, 2021;Mendes et al, 2021;Villar et al, 2022). The role of IRF4, aryl hydrocarbon receptor was confirmed in vivo in genetically deficient mice in homeostatic conditions (Kim et al, 2016;Goudot et al, 2017). The orphan nuclear receptor NR4A3 was also shown to participate in mo-DC differentiation, acting downstream of IRF4, but only during inflammation in vivo (Salix et al, 2019). Similarly, ETV6 was found to control mo-DC differentiation in vivo during inflammation but not in steady-state (Villar et al, 2022). While the transcriptional regulation of macrophage identity in tissues is well characterized (Bl eriot et al, 2020;Guilliams et al, 2020), transcription factors driving the monocyteto-macrophage program remain poorly characterized, besides our finding that MAFB is required for human mo-Mac differentiation (Goudot et al, 2017).
Based on the requirement for distinct transcription factors for their differentiation, we have proposed that mo-Mac and mo-DC constitute two distinct cell lineages stemming from monocytes (Goudot et al, 2017). Recently, single-cell analyses and trajectory reconstruction in a model of neuro-inflammation have suggested that mo-DC could represent an intermediate state of monocyte differentiation, with mo-Mac being the end stage (Amorim et al, 2022). The transcriptional control of these transitions was not explored. More work is therefore needed to better understand the molecular regulation of the monocyte differentiation program.
Here, we used temporal single-cell RNA sequencing (scRNAseq) to analyze human monocyte differentiation trajectories in an in vitro system. We identify divergent differentiation paths toward mo-Mac versus mo-DC, with a fate decision occurring within the first 24 h. Using a fate-mapping approach in a model of sterile peritonitis in mouse, we confirm that monocytes also differentiate along two divergent trajectories in vivo. Using computational approaches, we identify candidate transcription factors involved in human monocyte differentiation. We further show that IRF1 is required for mo-Mac differentiation and that ZNF366 and MAFF are involved in mo-DC differentiation. Our results support a model in which mo-Mac and mo-DC derive from alternative differentiation pathways, controlled by distinct regulatory networks.

Results and Discussion
Temporal scRNA-seq analysis reveals early divergent paths of human monocyte differentiation To dissect monocyte fate decision mechanisms, we used our previously established in vitro differentiation model, in which human monocytes are cultured with a cocktail of M-CSF, IL4, and TNF-a and differentiate in the same culture into mo-DC and mo-Mac resembling the ones found in clinical samples (Goudot et al, 2017). Of note, we have previously shown using scRNA-seq that monocytes used in these cultures do not contain contaminating DC precursors (Villar et al, 2022). To address whether monocytes followed a binary fate decision or differentiated along a continuum of states, we sought to reconstruct monocyte differentiation trajectories using single-cell transcriptomes. We profiled monocyte cultures from two individual donors by scRNA-seq using a droplet-based method and focused on early time points: 3, 9, 24, and 48 h after the start of the culture ( Fig 1A). We integrated the different datasets using STACAS (Fig EV1A-C) (Andreatta & Carmona, 2021) and performed unsupervised clustering using a graph-based approach with the Seurat package (Hao et al, 2021) (Fig EV1D). We identified 3 clusters of contaminating B cells, NK cells, and T cells, that we excluded from subsequent analysis (Fig EV1E and F). Cells from both donors were distributed homogenously (Fig EV1F). Unsupervised analysis grouped cells according to time points (Figs 1A and EV1G). Analysis of differential gene expression between time points showed a temporal regulation of gene expression (Fig EV1H), validating the experimental set-up.
We identified 14 different clusters in the dataset, most of them with mixed sample origins ( Fig 1B). Clusters 1 and 2 contained almost exclusively cells captured at 3 h, cluster 5 contained cells captured at 9 h, while clusters 11, 12, 13, and 14 displayed a majority of cells from the 48 h time point (Fig 1B). To decipher cluster identity, we analyzed the expression of the top differentially expressed genes between clusters (Fig 1C; Dataset EV1) and of canonical genes of monocytes (S100A9, S100A8, VCAN, CD14), mo-Mac (MAFB, CXCL8, THBS1, APOE) and mo-DC (IRF4, PRDM1, CD1B, CD1A; Fig 1D). Clusters 1, 2, 3 and 4 had the highest expression of monocyte genes. Consistent with this, clusters 1, 2, 3, 4, 5, and 6 were enriched for the monocyte signature compared to other clusters ( Fig 1E). Macrophage genes were predominantly expressed in clusters 4, 5, 6, 13, and 14, including APOC1, sterol 27-hydroxylase (CYP27A1), fatty acid binding proteins FABP4 and FABP5, and chemokine CCL22. DC genes were found in clusters 3, 7, 8, 11, and 12, including CD1E and CST3. Of note, cluster 10 contained cells from all time points and expressed genes consistent with a contaminating population of intermediate CD14 + CD16 + monocytes, including FCGR3A (encoding CD16) and AIF1 (Villani et al, 2017). We have previously shown in this model that IRF4 and MAFB are necessary for mo-DC and mo-Mac differentiation respectively (Goudot et al, 2017). Their expression appeared largely exclusive, with IRF4 expression highest in cluster 3 while MAFB was mainly expressed in clusters 1, 4, 5, and 6 ( Fig 1E). These results suggest that clusters 1 and 2 correspond to undifferentiated monocytes, while cells from other clusters express either a DC-or a macrophage-oriented transcriptional program starting after 3-9 h.
To study differentiation trajectories, we analyzed RNA velocity using scVelo (Bergen et al, 2020). This method estimates the kinetics of mRNA splicing in single-cell transcriptomes, in order to reconstruct cellular dynamics. Inference of the velocity pseudotime using scVelo was able to reconstruct independently the temporal sequence of the dataset (Fig EV1I), confirming the robustness of this approach. RNA velocity projection indicated two main dynamics ( Fig 1F). Clusters 1 and 2 showed a strong activity directed towards clusters 3 and 4, suggesting early transcriptional changes consecutive to fate decision. At later time points, two parallel differentiation trajectories were evidenced, with velocities from clusters 7 and 11 directed towards clusters 8 and 12 on one hand (DC path), and velocities from cluster 13 directed towards cluster 14 on the other hand (macrophage path). Collectively, these results suggest two divergent and alternative differentiation pathways ( Fig 1G), with fate commitment occurring in the first 24 h after monocytes are exposed to differentiation cues.

Monocytes differentiate along two divergent pathways during inflammation in mice
To validate the physiological relevance of our findings, we sought to track in vivo monocyte differentiation in a tissue where mo-DC and mo-Mac coexist. To this end, we used a model of sterile peritonitis, induced by thioglycolate injection into the peritoneum. We have previously shown that mo-DC and macrophages can be distinguished in steady-state peritoneum by the markers CD226 and ICAM2, respectively (Goudot et al, 2017). To confirm that mo-DC and mo-Mac had a stable phenotype during peritonitis, we profiled the expression of DC and macrophage markers after thioglycolate injection. While both populations expressed CD64, only mo-DC expressed CD11c and mo-Mac displayed higher levels of F4/80 and MerTK ( Fig EV2A). We adoptively transferred CD45.2 + Ly6C + monocytes into the inflamed peritoneum of CD45.1 + mice and analyzed by flow cytometry the CD45.2 + monocyte-derived cells at different time points (Fig 2A), using a panel of surface markers for monocytes, mo-DC, and macrophages. We concatenated 11.000 downsampled CD45.2 + cells for each time point and we performed dimension reduction into a UMAP projection (Figs 2B and EV2B). Importantly, injected monocytes contained negligible quantities of DC precursors or Ly6C À CD43 + monocytes ( Fig EV2C). Temporal analysis showed that the phenotype of CD45.2 + monocyte-derived cells changed over time (Fig 2C-E), with monocyte markers Ly6C and CCR2 being down-regulated in the first 24 h and MHC II and CD115 being up-regulated within 24-48 h ( Fig 2D). CD45.2 + mo-DC (expressing CD11c, MHC II, and CD226) and mo-Mac (expressing ICAM2 and F4/80) emerged from 48 h (Fig 2D-G). To assess differentiation trajectories, we performed unsupervised clustering using This analysis identified several paths, with two distinct branching trajectories leading to mo-DC (clusters 8, 13, 18) or mo-mac (cluster 11) as terminal points ( Fig 2J). These results suggest that in vivo, during inflammation, monocytes give rise to mo-DC or mo-Mac via two divergent differentiation pathways.

Identification of candidate transcription factors involved in monocyte differentiation
To identify transcriptional regulators of the early steps of monocyte differentiation, we used our human scRNA-seq dataset and analyzed differentially expressed genes between the clusters showing a commitment towards mo-DC versus mo-Mac, and extracted genes annotated as transcription factors. We focused on the comparison between clusters 3 and 4 ( Fig 3A), clusters 6 and 7 (Fig 3B), clusters 6 and 8 (Fig 3C), and clusters 12 and 14 ( Fig 3D). To infer transcriptional activity, we also performed network analysis using DoRoThEa (Garcia-Alonso et al, 2019; Holland et al, 2020). Based on these analyses (Dataset EV2), we identified as candidate IRF1, which is more expressed in macrophage-engaged clusters 4 ( Fig 3A) and 6 (Fig 3B and C) and shows high transcriptional activity in the clusters 5 and 6 ( Fig 3E). Among differentially expressed transcription factors, we also selected as candidate ZNF366, which is more expressed in DCcommitted clusters 3 ( Fig 3A) and 8 ( Fig 3C) and whose mouse ortholog is involved in classical DC1 (cDC1) terminal differentiation (Chopin et al, 2019;Shengbo et al, 2021). Finally, we selected MAFF, which is more expressed in macrophage-committed clusters 6 ( Fig 3B) and 14 ( Fig 3D) and heterodimerizes with the transcription factor NFE2L2 (Katsuoka & Yamamoto, 2016) which showed high predicted activity in clusters 5 and 6 ( Fig 3E).

IRF1 controls mo-Mac differentiation independently of its role in regulating interferon-stimulated genes
To analyze the expression of IRF1 at the protein level, we performed Western Blot on freshly isolated monocytes and at different time points of culture ( Fig 4A). IRF1 was expressed in monocytes and its expression was upregulated after 1 day of culture. To address the role of IRF1 in monocyte differentiation, we silenced its expression at the start of the culture using shRNA delivered through a lentivirus. We tested the impact of IRF1 deficiency on mo-Mac differentiation in a linear model (monocytes cultured with M-CSF alone in which only mo-Mac are produced), as well as our differentiation model with M-CSF, IL4, and TNF-a (Fig 4B-D). We used two different shRNA resulting in efficient silencing of IRF1 ( Fig 4B). IRF1 silencing significantly impaired macrophage differentiation in both culture systems (Fig 4C and D), without affecting mo-DC ◀ Figure 2. Monocytes differentiate along two parallel pathways during peritonitis in vivo.
A CD45.2 + monocytes were injected in CD45.1 + recipient mice previously injected with thioglycolate. Peritoneal lavage was analyzed at 3, 9, 24, 48, and 72 h post transfer (n = 4 biological replicates). B UMAP displaying CD45.2 + Lineage À Ly6G À Tim4 À peritoneal cells analyzed by flow cytometry and concatenated from all time points. C UMAP of cells from each time point. D Expression of phenotypic markers used for UMAP projection. E Supervised gating strategy for moDC and moMac at each time point (gated on live CD11b + CD115 + cells). F Supervised gating strategy to identify monocyte-derived cells. G Distribution of manual gates in the UMAP projection. H Unsupervised clustering using Phenograph. I Heatmap of median marker expression values for each Phenograph cluster. J Trajectory and pseudotime analysis using Slingshot. Cluster 12 (monocytes) was chosen as starting point. Trajectories from monocytes towards moDC or moMac are highlighted. Cells are colored by pseudotime.  6 of 13 EMBO reports 24: e56308 | 2023 Ó 2023 The Authors differentiation ( Fig 4D). These results validate a role for IRF1 in mo-Mac differentiation. IRF1 is known to regulate the expression of interferon-stimulated genes (ISG) in macrophages (Kamijo et al, 1994;Langlais et al, 2016;Song et al, 2021;Rosain et al, 2023). Of note, cluster 5 showed high expression of ISG such as MX1 and ISG15 and was enriched for the ISG signature ( Fig 4E). In addition, network analysis suggested that IRF1 controlled the expression of several ISG expressed in our dataset ( Fig 4F). Therefore, we sought to address whether IRF1 may impact mo-Mac differentiation through the regulation of ISG expression. To confirm the expression of ISG during monocyte differentiation in vitro, we measured by RT-qPCR the expression of the ISG MX1, IFIT3, and CXCL10 after 9 h of culture with different cytokine combinations ( Fig 4G). We found that ISG expression were significantly induced upon exposure to M-CSF, with TNF-a decreasing ISG expression. To directly address the effect of ISG on monocyte differentiation, we used a soluble type I interferon receptor (B18R) which acts as an inhibitor of interferon signaling (Colamonici et al, 1995). We validated the inhibitory effect of B18R on the expression of MX1, IFIT3, and CXCL10 ( Fig 4G). However, we did not observe any impact of B18R on mo-Mac differentiation, whether using M-CSF alone, the combination of M-CSF and IL4, or the cocktail M-CSF, IL4, and TNF-a ( Fig 4H). Taken together, these results suggest that IRF1 controls mo-Mac differentiation independently of its role in ISG expression.
IRF1 is known to be involved in myelopoiesis. Irf1-deficient mice display reduced granulocyte numbers, but have normal monocyte numbers in circulation with reduced expression of PU.1 (Testa et al, 2004). A role for IRF1 in the function and differentiation of osteoclasts, a bone-resident population of the monocyte lineage (Arai et al, 1999), was also proposed (Salem et al, 2014;Place et al, 2021). IRF1 was shown to modulate in human macrophages, but not in monocytes, the expression of ISG by controlling chromatin accessibility upon stimulation (Song et al, 2021). This finding together with our results suggests that the role of IRF1 in monocytes may be different from that observed in macrophages.

ZNF366/DC-SCRIPT is involved in mo-DC differentiation
We then examined the role of ZNF366 (also known as DC-SCRIPT) in monocyte differentiation. DC-SCRIPT was shown to be highly expressed in human DC, either derived in vitro or isolated from tissues (Triantis et al, 2006), and in murine cDC (Shengbo et al, 2021). ZNF366 was highly expressed in DC-committed clusters 7, 8, and 12 (Fig 5A and B). The velocity of ZNF366 was highest in the cluster 3, suggesting that the transcription of ZNF366 was initiated in the first few hours of mo-DC fate decision (Fig 5A). To confirm this, we analyzed the protein expression at different time points using Western Blot (Fig 5C). ZNF366 was not detected in freshly isolated monocytes and its expression was induced upon culture. To directly address the role of ZNF366, we silenced its expression at the start of the culture using shRNA. We used two different shRNA with efficient down-regulation of ZNF366 (Fig 5D). Silencing of ZNF366 significantly decreased mo-DC differentiation without affecting mo-Mac ( Fig 5E). Instead, we observed an increased proportion of CD16 À CD1a À cells, a population of undifferentiated cells (Goudot et al, 2017). These results indicate that ZNF366 is involved in mo-DC differentiation.
In mouse, DC-SCRIPT controls in cDC1 the production of IL-12, a cytokine expressed specifically by human mo-DC from clinical samples compared to mo-Mac (Tang-Huau et al, 2018). Understanding what gene programs are controlled by DC-SCRIPT during monocyte differentiation remains open for future investigation.

MAFF regulates mo-DC differentiation
Our transcriptomic data indicated that MAFF is more expressed in macrophage-committed clusters (Fig 5F). To analyze its expression at the protein level, we performed Western Blot at different time points (Fig 5G). MAFF was not detected in monocytes and was expressed upon culture. To test its role in monocyte differentiation, we silenced its expression as above and assessed the outcome of the monocyte culture. We used two different shRNA resulting in efficient decrease of MAFF expression (Fig 5H). Silencing of MAFF decreased mo-DC differentiation, without affecting mo-Mac proportions ( Fig 5I). We also observed an increased proportion of undifferentiated CD16 À CD1a À cells.
MAFF belongs to the family of 'small' MAF proteins, which heterodimerize with other transcription factors including NF-E2related factors, Bach proteins and c-Maf (Katsuoka & Yamamoto, 2016). MAFF has been reported to be a transcriptional repressor, but its function remains poorly understood (Blank, 2008;Katsuoka & Yamamoto, 2016). Our results show that MAFF is involved in mo-DC differentiation, although being enriched at the transcript level in ◀ Figure 4. IRF1 controls mo-Mac differentiation independently of its regulation of interferon-stimulated genes.

A
Monocytes were cultured with M-CSF, IL-4, and TNFa for 5 days. Protein quantification by Western Blot. Actin was used as loading control. Representative results are shown (n = 6 biological replicates). Quantification was performed by densitometry. Each symbol represents an individual donor. Paired one-way ANOVA. B-D Monocytes were cultured with M-CSF (B, C) or M-CSF, IL-4, and TNFa (D) for 5 days. IRF1 expression was silenced using a lentivirus containing shRNA. (B) Protein quantification by Western Blot after 5 days. Actin was used as loading control. Representative results are shown (n = 8 biological replicates). Quantification was performed by densitometry. Each symbol represents an individual donor. Paired one-way ANOVA. (C, D) Macrophage differentiation from monocytes was assessed by flow cytometry. One representative donor is shown (n = 8 for C and n = 9 for D). DN = double negative. Median is shown (n = 8-9 biological replicates in 3 independent experiments). Paired one-way ANOVA. E Normalized expression of ISG signature in the different clusters. Pattern of expression for MX1 and ISG15. F Gene network from IRF1 regulon in cluster 5. G, H Monocytes were cultured with RPMI medium (R10), or combinations of M-CSF, IL-4, and TNFa in the presence or absence of B18R (type I interferon inhibitor). (G) ISG expression after 9 h was analyzed by RT-qPCR. Each symbol represents an individual donor (n = 5 biological replicates). (H) Monocyte differentiation after 5 days. One representative donor is shown (n = 5 biological replicates). Proportion of mo-Mac and mo-DC after 5 days. Each symbol represents an individual donor (n = 5 biological replicates). Paired t-test.
Data information: For all panels: *P < 0.05, **P < 0.01, ***P < 0.001. Absence of star indicates 'not significant'. Source data are available online for this figure.
Ó 2023 The Authors EMBO reports 24: e56308 | 2023 In this work, we used temporal scRNA-seq analysis to reveal differentiation trajectories of human monocytes. In addition, we confirmed the existence of two parallel differentiation pathways in vivo in a mouse model of peritonitis. We identified IRF1, ZNF366/DC-SCRIPT and MAFF as molecular regulators of the early stages of monocyte differentiation. Our results contribute to the evidence supporting a model of divergent pathways leading to mo-DC versus mo-Mac differentiation. Recently, it was proposed in a model of neuro-inflammation that monocytes differentiate along a single path, with mo-DC being an intermediate and mo-Mac the end stage (Amorim et al, 2022). The discrepancy with our results may reflect context-or tissue-dependent phenomena. Because mo-DC and mo-Mac possess distinct functional properties, a better understanding of monocyte differentiation trajectories during pathogenic inflammation, and of their transcriptional control, will be essential for manipulating their fate for therapeutic purposes.

Materials and Methods
Mice C57BL/6 mice (CD45.2 + ) were purchased from Charles River (France). CD45.1 + C57BL/6 mice were produced in-house. Mice were maintained under specific pathogen-free conditions at the animal facility of Institut Curie. Female mice were used at age 7-9 weeks. Sample size was not calculated a priori. No animal was excluded from analysis. Blinding was performed during outcome assessment. All animal procedures were in accordance with the guidelines and regulations of the French Veterinary Department and approved by the Institut Curie ethics committee (authorization APAFIS #25217-2020042522586261 v1).

Human samples
Buffy coats from healthy donors (both male and female donors) were obtained from Etablissement Français du Sang (Paris) in accordance with INSERM ethical guidelines. According to French Public Health Law (art L 1121-1-1, art L 1121-1-2), written consent and IRB approval are not required for human non-interventional studies.

Monocyte isolation and culture
Peripheral Blood Mononuclear Cells (PBMC) were prepared by centrifugation on a Ficoll gradient (Lymphoprep, StemCell). Blood CD14+ monocytes were isolated from healthy donors' PBMC by positive selection using magnetic beads (Miltenyi). Monocytes (2 × 10 6 cells/ml) were cultured for indicated times in RPMI-Glutamax medium (GIBCO) supplemented with antibiotics (penicillin and streptomycin) and 10% Fetal Calf Serum in the presence or absence of 100 ng/ml M-CSF (Miltenyi), 5 ng/ml IL-4 (Miltenyi) and 5 ng/ml TNF-a (R&D Biotechne). Cytokines were added only at the start of the culture, and medium was not refreshed during the course of the culture. In some experiments, monocytes were cultured in the presence of 1 lg/ml of recombinant B18R (StemCell).

Single-cell RNA library preparation
For each time point, monocytes were detached from the plate and dead cells were removed with Dead Cell Removal Kit (Miltenyi) according to manufacturer's instructions. Cells were barcoded per donor (Donor A and B) using TotalSeq TM -anti-human Hashtag antibody (A0251, A02052, respectively; Biolegend) according to manufacturer's instruction. Barcoded cells were counted and mixed in a 1:1 ratio. Finally, barcoded single-cell suspension was loaded into 10× Genomics Chromium (CA, USA). Libraries were prepared as per manufacturer's protocol (Chromium Single Cell 3' Reagent Kits v3 protocol) and sequenced on an Illumina NovaSeq sequencer according with 10× Genomics recommendations (paired-end reads) to a depth of approximately 50,000 reads per cell.

Single-cell RNA-sequencing analysis
Initial processing was done using CellRanger (v3.1.0). Starting from the filtered gene-cell count matrix produced by CellRanger, we proceeded with the Seurat v4.0 workflow (Hao et al, 2021) for demultiplex cells based on hashtag barcode and quality control (QC) processing. Briefly, we filtered out cells with mitochondrial genes < 20%, as well as genes expressed in less than 3 cells. Representative results are shown (n = 4 biological replicates). Quantification was performed by densitometry. Each symbol represents an individual donor. Paired one-way ANOVA. (I) Mo-mac and mo-DC differentiation after 5 days was assessed by flow cytometry. One representative donor is shown (n = 7 biological replicates). DN = double negative. Proportion of mo-DC, mo-Mac and DN cells after 5 days. Median is shown (n = 7 biological replicates). Paired one-way ANOVA.
Data information: *P < 0.05, **P < 0.01. Absence of star indicates 'not significant'. Source data are available online for this figure.
Ó 2023 The Authors EMBO reports 24: e56308 | 2023 Hashtag demultiplexing was performed with the function HTODemux() and positive.quantile = 0.99. Integration of the 8 samples (2 donors, 4 time points) was subsequently performed with the STACAS workflow (Andreatta & Carmona, 2021). Briefly, normalization and identification of variable features was performed with standard Seurat pipelines and then anchors were identified with FindAnchors.STACAS (. . ., dims = 1:30, anchor.features = 5,000) and filter with FilterAnchors.STACAS (. . ., dist.thr = 0.8). The order for integration was calculated with SampleTree.STACAS(). Finally, integration was performed with the filtered anchors using the function IntegrateData(). Downstream analysis, graph-based clustering, visualization and differential gene expression analyses of the scRNA-seq data were performed using Seurat v4.0. For clustering analysis, FindNeighbors() and FindClusters() functions of the Seurat package were used with the first 50 significant PCs and a resolution of 1.3, respectively. For signatures enrichment, we used the function AddModuleScore(). For identification of DEGs, we used the Find-Markers or FindAllMarkers function (test.use = 't', logfc. threshold = log[0.25]) based on normalized data. DEGs with adjusted P values of > 0.05 were filtered out. Data has been deposited in GEO (accession number GSE218483).

RNA velocity
Annotations of unspliced/spliced reads were obtained using velocyto CLI with default parameters (La Manno et al, 2018). Next, unspliced counts were merged with the preprocessed, normalized, integrated and annotated spliced count matrix (Seurat v4.0) and the RNA velocity analysis was performed using scVelo (v0.2.4) workflow (Bergen et al, 2020). Briefly, we computed moments for velocity estimation and then the stochastic model to learn the transcriptional dynamics of splicing kinetics, transcriptional state, and velocity pseudo-time across the complete dataset.

Monocyte adoptive transfer
Monocytes were isolated from the pooled bone marrows of 3 individual mice using the EasySep Mouse Monocyte Isolation Kit (Stemcell) according to manufacturer's instructions. 1 million monocytes were injected intra-peritonally into CD45.1 C57BL/6 mice which had been injected 18 h before with 1 ml of 3.8% brewer's thioglycolate medium (Sigma). Peritoneal lavage was analyzed by flow cytometry at indicated time points after monocyte injection. Peritoneal lavage was recovered by intraperitoneal injection of 5 ml of PBS.

Flow cytometry
Human cells were stained in PBS containing 0.5% human AB serum and 2 mM EDTA for 30-45 min on ice with APC anti-CD1a (Biolegend, clone HI149) and FITC anti-CD16 (Biolegend, clone 3G8). For experiments involving shRNA, cells were stained with fixable Aqua Live/Dead (ThermoFisher) for 10 min at 4°C, and fixed prior to acquisition with fixation buffer from eBioscience Intracellular Fixation and Permeabilization kit (ThermoFisher). Otherwise, DAPI (Fischer Scientific, 100 ng/ml) was added immediately prior to acquisition. Cells were acquired on a FacsVerse instrument (BD Biosciences).
qPCR Cells were harvested and lysed in RLT buffer (QIAGEN). RNA extraction was carried out using the RNAeasy micro kit (QIAGEN) according to manufacturer's instructions. Total RNA was retro-transcribed using the superscript II polymerase (Invitrogen), in combination with random hexamers, oligo dT, and dNTPs (Promega). Transcripts were quantified by real-time PCR on a 480 LightCycler instrument (Roche). Reactions were carried out in 10 ll, using a master mix (Eurogentec), with the following Taqman Assays primers (Merk), for human samples: B2M (Hs99999907_m1), RPL34 (Hs00241560_m1), HPRT1 (Hs02800695_m1), MX1 (Hs00895608_m1), IFIT3 (Hs00155468_m1), CXCL10 (Hs00895608_m1). The second derivative method was used to determine each Cp and the expression of genes of interest relative to the housekeeping genes (B2M, HPRT, RPL34) was quantified.

Statistical analysis
Statistical tests were performed using Prism v9 (GraphPad Software). Statistical details for each experiment can be found in the corresponding figure legend. N corresponds to the number of biological replicates.
The flow cytometry data from this publication have been deposited to the Flow repository database (http://flowrepository.org/id/ FR-FCM-Z678 and assigned the identifier FR-FCM-Z678).