Microglial heterogeneity: influence of human 2D, 3D, and co-culture models on gene expression and immune function

Microglia, the resident immune cells of the central nervous system, exhibit substantial phenotypic and functional diversity shaped by local microenvironmental cues. While advanced in vitro human microglial models exist, the influence of culture dimensionality and cellular context on microglial state composition remains poorly defined. Here, we analyzed single-cell RNA sequencing datasets from human monocyte-derived microglia (MDMi) cultured under two-dimensional (2D) and three-dimensional (3D) monoculture, as well as 3D neural–glial co-culture conditions. Across platforms, four microglial states were identified, including interferon (IFN)-responsive, chemokine-enriched, metabolically active, and proliferative states, with pronounced environment-dependent transcriptional shifts. 2D cultures were dominated by IFN-responsive microglia characterized by elevated IFITM2 and IFITM3 expression, whereas 3D systems supported greater cellular diversity, including expanded metabolic programs and chemokine remodeling. Co-culture further increased proliferative microglia and induced immune-communication signatures involving CCL2/CCL5/CCL7, CSF1, and VEGF/FLT1 pathways. Pseudotime analysis revealed a largely linear trajectory in 2D cultures, but branching differentiation paths in 3D and co-culture systems, consistent with enhanced microglial heterogeneity. Benchmarking against human microglial reference signatures demonstrated broader and stronger overlap in 3D-based models, with homeostatic and disease-associated modules engaged in a context-specific manner. These findings demonstrate that culture architecture is a major determinant of microglial identity and immune responsiveness; and highlight the value of single-cell datasets to uncover previously underappreciated microglial states with relevance to human neuroimmune biology.

1 Introduction

Chronic neuroinflammation, driven largely by microglia, plays a pivotal role in the progression of neurodegenerative diseases such as Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS). As the resident immune cells of the central nervous system (CNS), microglia exhibit significant phenotypic and functional diversity (Hammond et al., 2019; Masuda et al., 2019; Ochocka et al., 2021; Paolicelli et al., 2022), adopting context-dependent states that can exert neuroprotective or neurotoxic effects depending on disease stage and microenvironmental cues (Michell-Robinson et al., 2015, Colonna and Butovsky, 2017; Friedman et al., 2018; Lloyd et al., 2024). Despite increasing knowledge of microglia’s role in these conditions, accurately modeling this diversity in vitro remains a challenge due to the limitations of current culture systems.

Conventional two-dimensional (2D) culture systems are widely used due to their simplicity and scalability; yet they impose physical constraints that alter metabolic state, functional behavior and give rise to transcriptionally distinct sub-populations (Cadiz et al., 2022; Cuní-López et al., 2023). The lack of extracellular matrix architecture and spatial organization limits cell–cell communication and limits the development of microglial transcription states associated with tissue context. In contrast, three-dimensional (3D) cultures better mimic native tissue environments, supporting spatial organization, improved cellular interactions and in vivo-like tissue architecture (Shamir and Ewald, 2014; Moysidou et al., 2021; Sharaf et al., 2023). Despite their growing use, the impact of dimensionality alone on microglial identity, subtype composition and differentiation trajectories remain unclear. Recent advances in human microglial modeling have highlighted the importance of environmental context in shaping transcriptional and functional states (Haenseler et al., 2017; Depp et al., 2025; Fumagalli et al., 2025). Microglia exposed to defined cytokine milieus, extracellular matrix properties and cellular neighbors adopt gene expression programs and metabolic profiles that more closely resemble ex vivo human microglia (Gosselin et al., 2017; Mancuso et al., 2019; Dolan et al., 2023). 3D culture and co-culture with neural or glial cells promote ramified morphologies, sustained viability and engagement of immune and homeostatic pathways that are poorly captured in 2D monocultures (Schafer et al., 2023; Zhang et al., 2023). Nevertheless, variability in differentiation protocols, matrix composition and co-culture strategies continues to limit cross-study comparability, and systematic evaluations of how dimensionality and cellular context interact to shape microglial differentiation are scarce.

Human primary microglia are difficult to obtain at scale and while induced pluripotent stem cell-derived microglia (iMG) have expanded modeling capacity, differentiation efficiency and maturation state can vary across protocols and donor lines. Monocyte-derived microglia-like cells (MDMi) provide a complementary, accessible human model that retains donor-specific genetic and environmental information, enables longitudinal sampling and is compatible with 3D matrices and multicellular co-culture systems (Ohgidani et al., 2014; Ryan et al., 2017; Cuní-López et al., 2024). Building on prior work establishing patient-derived MDMi platforms (Cuní-López et al., 2024), we reasoned that both culture dimensionality (2D vs. 3D) and cellular context (monoculture versus neural–glial co-culture) would act as key determinants of microglial state composition, maturation dynamics and pathway engagement, even in healthy donor-derived cells.

Here, we leveraged previously characterized MDMi cultured under defined 2D, 3D, and 3D neural-glial co-culture conditions and applied further in-depth analyses to interrogate how microglia adapt to distinct in vitro environments. By mapping transcriptional states and differentiation trajectories across culture platforms, we aimed to define how dimensionality and cellular context shape microglial identity at baseline. This framework informs the selection of human microglial models for mechanistic studies, disease modeling and therapeutic discovery.

2 Materials and methods

2.1 Monocyte-derived microglia-like cell generation and culture conditions

Human monocyte-derived microglia-like cells (MDMi) were generated from peripheral blood monocytes as previously described (Quek et al., 2022; Cuní-López et al., 2024). Briefly, monocytes isolated from healthy donor peripheral blood were differentiated in RPMI-1640 medium supplemented 10 ng/mL GM-CSF and 100 ng/mL IL-34 and maintained under two-dimensional (2D) monoculture, three-dimensional (3D) monoculture or 3D neural–glial co-culture conditions. In 3D cultures, cells were embedded within a Matrigel matrix to provide spatial context, while 3D co-cultures incorporated ReNcell VM–derived neural components that differentiate into mixed neuronal and glial populations.

Prior to single-cell RNA sequencing, cells were dissociated using a non-enzymatic detachment method using a cell recovery solution (Corning, 354253) and fluorescence-activated cell sorting (FACS) was performed to isolate CD11b⁺ cells, thereby excluding leukocytes, neural and glial components. Sorted cells were immediately fixed and processed using the Chromium Next GEM Single Cell Fixed RNA Kit (10x Genomics) according to the manufacturer’s instructions. For scRNA-seq, MDMi were harvested at day 14 in 2D monoculture and at approximately day 30–35 in 3D monoculture and 3D neural–glial co-culture conditions, consistent with the experimental timelines used to generate the GSE255718 dataset (Cuní-López et al., 2024). Detailed experimental protocols are provided in the referenced publication, and no modifications were introduced for the present analysis.

2.2 Single-cell RNA sequencing data resource, processing, QC metrics, and clustering

Single-cell RNA sequencing libraries were generated as described previously (Cuní-López et al., 2024). Briefly, libraries were prepared using the Chromium Next GEM Single Cell Fixed RNA Kit (10x Genomics) and sequenced on an Illumina NextSeq 2000 platform. Raw sequencing data were processed using Cell Ranger v7.0.0 (10x Genomics) (Zheng et al., 2017), including demultiplexing, alignment to the human reference genome, and generation of gene–cell count matrices. For the present study, Cell Ranger–generated count matrices were used as the starting point for all downstream analyses. Filtered feature–barcode matrices were imported into R (v4.0.5) (R Core Team, 2019) using the Read10X function and assembled into Seurat objects using CreateSeuratObject (Hao et al., 2021).

Raw gene expression matrices were processed using Seurat. Cells were filtered based on standard quality control metrics, including the number of detected genes per cell, total UMI counts, and the proportion of mitochondrial transcripts. Cells with fewer than 200 detected genes or more than 6,000 detected genes were excluded to remove low-quality cells and potential doublets. Quality control thresholds were applied uniformly across all samples to minimize systematic bias between experimental groups. Following quality control, 12,244 monocytes (from 12,248), 6,472 2D MDMi (from 7,003), 7,736 3D MDMi (from 7,838), and 6,211 3D co-culture MDMi (from 6,426) were retained for downstream analyses (Supplementary Table 1).

Gene expression values were log-normalized and highly variable genes were identified using Seurat’s standard workflow. Data were scaled to minimize the influence of highly expressed genes prior to dimensionality reduction. Principal component analysis (PCA) was performed, followed by non-linear embedding using Uniform Manifold Approximation and Projection (UMAP) or t-distributed stochastic neighbor embedding (t-SNE) for visualization. Graph-based clustering was conducted using Seurat’s FindClusters function, with optimal resolution determined using the Clustree package (Zappia and Oshlack, 2018) to avoid over- or under-clustering. Individual samples were processed independently through quality control and normalization prior to aggregation for downstream analyses. Samples were subsequently aggregated using Seurat’s merge() function while retaining sample identity metadata (orig.ident) to enable assessment of potential sample-driven effects.

2.3 Identifying differentially expressed genes, cell subtypes and pathway enrichment

Differentially expressed genes (DEGs) between clusters within each sample were identified using Seurat’s ‘FindMarkers’ function with the default Wilcoxon rank-sum test for pairwise comparisons. Resulting p-values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) correction. Significantly differentially expressed genes were defined as having an adjusted p < 0.05) and log FC > 0.5. Each transcriptionally distinct cluster was subsequently examined to define microglial subtypes based on the expression of canonical marker genes and functional gene signatures. All quantitative differential gene expression statistics underlying dot plots, heat maps and volcano plots are provided in Supplementary Tables 2–5 to enable direct numerical comparison independent of visual scaling.

To gain biological insight into the molecular/biological functions of each subtype, DEGs were analyzed for functional enrichment in pathways and Gene Ontology (GO) terms. Enrichment analysis was performed using an overrepresentation test implemented in the package clusterProfiler (Wu et al., 2021), with significance determined by adjusted p < 0.05. Annotation and mapping of gene identifiers (e.g., Entrez ID, Ensembl ID, gene description) were carried out using the AnnotationHub package (Morgan and Shepherd, 2017).

2.4 Sample aggregation and pseudotime trajectory mapping of microglial states

Individual samples were processed independently through quality control and normalization and subsequently aggregated using Seurat’s merge() function while retaining sample identity metadata. Because all samples were generated using the same platform, chemistry, and processing pipeline and were processed in parallel, no strong technical batch effects were detected. Data integration methods were therefore not applied, as they may remove genuine biological variability, particularly in highly plastic immune cell populations (Luecken and Theis, 2019; Stuart et al., 2019; Marsh et al., 2022). To examine transition between microglial states and influence of 2D and 3D culture conditions, aggregated scRNA-seq data were used for comparative analysis. Filtered Seurat objects were merged using the merge function, ensuring each cell retained its corresponding sample identifier. Data were normalized using NormalizeData, variable genes were identified via FindVariableFeatures, and expression values were scaled with ScaleData to reduce technical variation across datasets. Dimensionality reduction was performed using PCA (RunPCA), followed by visualization through UMAP and t-SNE). All dimensionality reductions were based on the variable features identified during preprocessing. Cluster visualization and sample-specific comparisons were generated using DimPlot, and expression of key markers was examined using FeaturePlot.

To infer dynamic differentiation processes, pseudotime trajectory analysis was conducted using Monocle3 (Trapnell et al., 2014). This analysis enabled reconstruction of developmental trajectories and identification of transcriptional transitions among microglial populations. The integrated Seurat object was converted to a cell_data_set format, and cell clustering and trajectory learning were performed using the cluster_cells and learn_graph functions. The resulting trajectory was visualized with plot_cells, overlaid onto the UMAP representation. A principal node within the monocyte population was designated as the root for pseudotime ordering using order_cells, enabling visualization of temporal progression and branching of microglial states. Cells were color-coded by pseudotime to illustrate differentiation trajectories and state transitions across 2D, 3D and co-culture conditions.

3 Results

3.1 Four major microglial subtypes identified in 2D and 3D models highlight culture-driven pathway changes

Single-cell RNA sequencing was performed on CD11b⁺ monocyte-derived microglia generated from healthy donors and cultured under 2D, 3D and 3D neural–glial co-culture conditions, as previously established (Cuní-López et al., 2024). To isolate the effects of culture conditions on microglial transcriptional states, analyses were performed independently of disease context. Using an in-depth, unified clustering and annotation framework, we systematically interrogated microglial heterogeneity across culture environments. Clusters were identified independently for each culture condition using unsupervised clustering (Figure 1a) and annotated post hoc based on established microglial marker gene expression (Figures 1b–d). This analysis revealed transcriptional diversity that had not been fully captured previously, identifying four recurrent microglial states including chemokine-enriched (CC), interferon (IFN)-responsive, metabolically active and proliferative, that together comprised the dominant cellular populations across all models (Figures 1a–d). A chemokine-enriched population, characterized by CCL4 and CCL3 expression, was consistently observed across all systems, though its abundance and transcriptional composition differed, comprising 16.6% of cells in 2D cultures (Figure 1b), 20.7% in 3D cultures (Figure 1c), and 5.4% in 3D co-cultures (Figure 1d). This cluster displayed environment-specific transcriptional rewiring—in 3D monoculture, cells upregulated CCL7 and CCL2, while the 3D co-culture environment induced expression of CCL7, CCL5, FLT1, and CSF1, genes associated with immune cell recruitment, angiogenic signaling and microenvironmental crosstalk. Consistent with prior reports that microglial transcriptional states are highly sensitive to cellular context, we observed environment-dependent changes in chemokine and cell-cell interaction pathways, with co-culture conditions preferentially engaging immune recruitment–associated signaling (e.g., CCL2/CCL5/CCL7) (Xue et al., 2021; Gullotta et al., 2023) and microenvironmental/vascular crosstalk (e.g., CSF1 and VEGF/FLT1-associated signaling) (Cakir et al., 2019; Sun et al., 2022). Together, these context-dependent changes illustrate how culture architecture and cellular interaction selectively shape microglial chemokine pathways, revealing transcriptional mechanisms that can confer functional heterogeneity that was obscured in earlier analyses.

FIGURE 1

Figure 1. Comparative analysis of microglial subtypes in 2D, 3D, and 3D co-culture models. (a) The t-SNE plots depict the clustering of microglial cells based on transcriptomic profiles across the 2D (n = 6,472 cells), 3D (n = 7,736 cells), and 3D co-culture (n = 6,211 cells). Numeric cluster IDs reflect unsupervised clustering performed independently for each culture condition and do not imply correspondence across models; biological cluster identities are defined based on marker gene expression shown in the accompanying dot plots. (b–d) Dot plots illustrate gene expression patterns within each identified microglial cluster in 2D, 3D, and 3D co-culture model, with dot size indicating the percentage of cells expressing a given gene and color intensity reflecting the average expression level. Pie charts on the left represent the proportional distribution of the microglial subtypes including proliferative, energy metabolism, chemokine-enriched (CC), and interferon-responsive (IFN) within each culture model.

Immediate-early genes (IEGs), such as FOS and JUN, encode AP-1 transcription factors that are rapidly induced via MAPK signaling in response to environmental and immune cues (Shaulian and Karin, 2002). In our dataset, modest enrichment of AP-1/IEG expression in 2D cultures, particularly within cluster 1, is consistent with rigid substrates and limited 3D cell–cell and extracellular matrix interactions, which promote integrin- and MAPK-dependent transcriptional programs (Discher et al., 2005; Glass et al., 2010; Baker and Chen, 2012). Cluster 1 was primarily defined by an interferon-responsive program (IFITM2, IFITM3, C1QA–C), with FOS and JUN present at low levels and not driving cluster identity, consistent with an immune-alert microglial state rather than dissociation-induced stress (Supplementary Figure 1; Marsh et al., 2022).

The IFN-responsive subtype dominated the 2D model (53.8%), characterized by elevated IFITM2 and IFITM3 expression and declined markedly in 3D (14.7%) and 3D co-culture (28%) conditions. Conversely, the 3D model exhibited a more balanced composition of IFN-responsive (14.7%), chemokine-enriched (20.7%) and metabolically active (14.5%) microglia, reflecting greater transcriptional diversity. A metabolically active microglial population was identified across all models, characterized by elevated mitochondrial gene expression, including MT-ND1, MT-ND2 and MT-ND3, together with immune-regulatory markers FOLR2 and CD163, indicating increased metabolic activity and energy demand. This population accounted for 2.55% of cells (cluster 5) in 2D culture, 14.46% (cluster 4) and 1.22% (cluster 6) in 3D culture and 7.84% (cluster 5) in 3D co-culture (Figures 1b–d). Within the 3D model, cluster 6 also expressed CCL22, CCL17 and IL2RA, consistent with immune-metabolic coupling within the transcriptional profile (Figure 1c).

Interestingly, the 3D co-culture system contained the largest proportion of proliferative microglia (4.56% vs. 1.31% in 2D and 0.93% in 3D), defined by MKI67, TYMS and KIFC1 expression, suggesting that the neural environment supports expansion of a cell-cycle-active compartment (Figure 1d). A distinct 2D-specific cluster enriched for CSF1, LPL and APOC1 exhibited a lipid-metabolic and phagocytic gene signature, consistent with a transcriptionally activated or disease-associated microglial state (Figure 1b); as previously described in neurodegenerative and inflammatory contexts (Keren-Shaul et al., 2017; Krasemann et al., 2017; Shimizu and Prinz, 2025). Comprehensive expression profiles for all clusters are provided in Supplementary Figures 2–4. Together, these analyses show that microglia acquire distinct transcriptional states across culture environments.

3.2 Microglial pathway enrichment patterns differ across culture environments

To determine how subtype-specific differences translate into broader transcriptional programs, we performed pathway enrichment analysis across 2D, 3D and 3D co-culture microglial models. In 2D culture, chemokine-enriched microglia were associated with pathways governing immune surveillance, including viral responses, phagocytosis, mitochondrial organization and apoptotic regulation. Transitioning to a 3D matrix increased the representation of genes involved in cytokine production, immune activation, immune activation and cellular motility, indicating a transcriptional shift toward immune-signaling pathway. Notably, the 3D co-culture model showed further enrichment of pathways linked to ERK1/ERK2 signaling, bacterial pattern recognition and type II interferon responses. Additional gene set enrichment indicated involvement of pathways regulating cell proliferation, adhesion and differentiation, highlighting transcriptional complexity within the 3D co-culture environment (Figure 2a).

FIGURE 2

Figure 2. Pathway enrichment networks of chemokine-enriched and interferon-responsive microglial subtypes across culture models. (a) The network diagrams illustrate enriched top 30 biological pathways activated in chemokine enriched subtype across models. (b) The network diagrams illustrate enriched 30 top biological pathways activated in Interferon responsive subtype across models. Node size reflects the number of genes contributing to each enriched GO term, while edges indicate shared genes between pathways.

IFN-responsive microglia in 2D culture displayed enrichment for immune defense, phagocytosis and viral response pathways. In 3D conditions, these cells showed increased representation of genes programs linked to chemotaxis, cytoskeletal reorganization, NF-κB signaling and broader inflammatory responses. Within 3D co-cultures, further enrichment of pathways related to cell cycle progression, chromatid segregation and migration reflected activation of proliferative and surveillance-associated gene networks (Figure 2b). Together, these results show that increasing culture complexity (from 2D to 3D and then to co-culture), broadens and strengthens transcriptional programs. We next asked whether these changes alter the core molecular identity of microglia by examining conserved and model-specific gene expression signatures across environments.

3.3 Culture environment preserves core microglial identity while shaping transcriptional diversity

Differentially expressed gene (DEG) analysis revealed that key microglial genes—including SELENOP, NUPR1, FOLR2 and C1QA—were consistently expressed across 2D, 3D and 3D co-culture systems compared to monocytes, highlighting their central roles in immune regulation and maintenance of microglial identity (Figures 3a–c). This pattern is consistent with prior studies defining these genes as core components of human microglial transcriptional and immune-regulatory programs (Gosselin et al., 2017; Hammond et al., 2019; Masuda et al., 2019). Complement- and chemokine-associated genes such as C1QB and CXCL12 were likewise conserved across 2D and 3D models, indicating preservation of a stable immune-associated transcriptional signature.

FIGURE 3

Figure 3. Differential gene expression and pseudotime trajectory of monocyte-to-microglia differentiation across in vitro culture models. (a–c) Volcano plots illustrating the top upregulated (red) and downregulated (blue) genes in microglia from each model compared to monocytes. (d) The UMAP plot illustrates the transcriptomic aggregation of monocytes and microglial cells across the 2D, 3D, and 3D co-culture models. (e) UMAP visualization of the pseudotime trajectory depicting monocyte-to-microglia differentiation across the three in vitro models. Pseudotime values mapped onto cells, colored from early (dark blue) to late (yellow) differentiation stages. (f) Left: UMAP plot with annotated trajectory clusters (1–6), corresponding to distinct phases of differentiation or cellular states. Right: Trajectory structure with directional arrows indicating the progression along pseudotime. (g) The UMAPs plots (right) show expression patterns of key genes, including AIF1, CCL24, C1QA, GPR34, MEF2C, OLFML3, MKI67, and TYMS.

The culture environment additionally influenced other transcriptional programs, with 3D conditions promoting increased expression of FCGBP and IGSF21, genes previously associated with cell–cell communication and extracellular matrix-related processes (Hedegaard et al., 2020). These findings indicate that while the core molecular identity of microglia is largely preserved across models, the magnitude and composition of transcriptional variation are strongly shaped by culture environment.

To explore transcriptional heterogeneity across environments, we applied pseudotime trajectory analysis to integrated single-cell datasets from monocytes, 2D, 3D and 3D co-culture systems (Figure 3d). This approach reconstructed lineage trajectories and captured transitional states across environments. In 2D cultures, cells followed a relatively linear trajectory, consistent with a more uniform transcriptional state. By contrast, 3D and 3D co-cultures exhibited multiple branches, reflecting enhanced transcriptional heterogeneity and the presence of diverse microglial states (Figures 3e,f).

Lineage markers were consistent across trajectories, with sustained expression of the myeloid marker AIF1 and canonical microglial markers such as C1QA, confirming microglial lineage identity in line with prior single-cell studies defining these genes as core components of the microglial transcriptional program (Masuda et al., 2019). In contrast, CCL24 was enriched at early pseudotime, consistent with progenitor-like or transitional myeloid states as previously reported (Van Hove et al., 2019).

Genes including GPR34 and OLFML3 were enriched in distinct trajectory branches, highlighting environment-dependent microglial transcriptional programs (Grabert et al., 2016). Notably, proliferation-associated genes (TYMS, MKI67) were enriched in terminal branches of the 3D co-culture trajectory, consistent with the emergence of proliferative microglial states in complex, tissue-like conditions (Hammond et al., 2019; Masuda et al., 2019; Figure 3g).

3.4 Environmental context drives distinct transcriptional programs in human microglial models

To determine how culture architecture and cellular context shape microglial transcriptional identity, we compared gene expression and pathway activity across MDMi cultured in 2D, 3D monoculture and 3D neural–glial co-culture systems. By integrating differential gene expression and pathway enrichment analyses, we aim to resolve how dimensionality alone, as well as the presence of neural and glial components, influences microglial activation, metabolic and intercellular communication. This unified framework enabled direct comparison of microglial transcriptional programs across progressively complex in vitro environments, providing mechanistic insight into how microenvironmental cues drive microglial heterogeneity and putative functional specialization.

In 2D cultures, upregulated genes such as THBS1, RSAD2, MX2, and PERM1 (Figure 4a, blue) were associated with activation of interferon, cytokine and NF-κB–related transcriptional programs, consistent with interferon-responsive microglial signatures observed in single-cell studies of immune-activated states previously reported (Lopez-Atalaya and Bhojwani-Cabrera, 2025; Tang et al., 2025). Pathway enrichment further revealed processes related to cell differentiation and adhesion, suggesting microglial adaptation to 2D culture environment (Figure 4b). In contrast, 3D cultures display pronounced up-regulation of CXADR, ROBO4, CCL17, CXCL5, and MMP1 (Figure 4a, red), indicative of engagement of extracellular-matrix remodeling, immune-cell recruitment and metabolic activation—a pattern aligned with context-dependent microglial transcriptional diversity described in recent reviews of microglia heterogeneity (Fumagalli et al., 2025). Enriched pathways in 3D included mitochondrial respiration, vesicle organization and protein transport, reflecting enhanced transcriptional signatures linked to metabolic and secretory capacity. Collectively, these analyses highlight that 2D cultures preferentially support stress- and immune-primed states, whereas 3D models promote metabolically active and transcriptionally diverse programs, providing a molecular context for the microglial heterogeneity observed in previous analyses (Figure 4c).

FIGURE 4

Figure 4. Comparative gene expression and functional pathway analysis between microglial culture models. (a) Volcano plot showing differentially expressed genes between 3D and 2D cultures. Genes significantly upregulated in 3D cultures are shown in red, while those upregulated in 2D cultures are shown in blue. (b) Network diagram displaying the top 100 enriched biological pathways activated in 2D cultures compared to 3D cultures. (c) Network diagram displaying the top 100 enriched biological pathways activated in 3D cultures compared to 2D cultures. (d) Volcano plot showing differentially expressed genes between 3D co-culture and 3D monoculture. Genes significantly upregulated in the 3D co-culture are shown in red, while those upregulated in 3D monoculture are shown in blue. (e) Network diagram of the top 100 enriched pathways activated in 3D monoculture relative to 3D co-culture. (f) Network diagram of the top 100 enriched pathways activated in 3D co-culture relative to 3D monoculture. Genes shown in volcano plots meet differential expression criteria (adjusted p < 0.05 and |log2FC| ≥ 0.3); complete DEG statistics and top regulated genes are provided in Supplementary Data. Node size in Network diagrams reflects the number of genes contributing to each enriched GO term, while edges indicate shared genes between pathways.

Building on the environment-dependent effects, we next compared microglial transcriptional profiles between 3D monoculture and a 3D neural–glial system incorporating ReNcell VM, a human neural progenitor line that spontaneously differentiates into a mixed population of neurons and glial (predominantly astrocytes) (Cuní-López et al., 2024), thereby providing a more complex neural-glial context.

In the 3D monoculture condition, upregulated genes including FABP4, COL1A1, CAMP, and SDC2 were associated with pathways involved in lipid metabolism and extracellular matrix remodeling (Figure 4d, blue), consistent with previously reported metabolically active or matrix-interacting myeloid and microglial populations (Gosselin et al., 2017; Chia et al., 2025; He et al., 2025). Gene set enrichment further indicated increased representation of transcriptional programs related to cellular respiration, phospholipid biosynthesis and protein folding, consistent with previously described stress-adaptive metabolic states in microglia (Gosselin et al., 2017; Ulland et al., 2017; Figure 4e).

In contrast, microglia within the 3D co-culture system exhibited broader transcriptional diversity, with pronounced enrichment of immune and cell communication–related pathways. Upregulation of LYVE1, SERPINB2, HIC1, and FCGR3A pointed to elevated immune responsiveness, phagocytic activity and inflammatory signaling (Figure 4d, red). Expression of GFAP and PLA2G2C further suggested astrocyte activation and lipid-mediated intercellular signaling, reflecting ongoing neuron-astrocyte-microglia cross-talk. Together, these findings indicate that incorporating ReNcell VM–derived neural components create a more complex transcriptional environment, enhancing immune-related gene expression and intercellular communication compared with 3D monocultures (Figure 4f). Consistent with these environment-dependent transcriptional differences, we also observed model-specific expression of genes that have been previously implicated in neurodegenerative disease pathways, suggesting that distinct culture contexts may differentially engage transcriptional programs that are relevant to, but not specific for, disease-associated states (Supplementary Figure 5).

3.5 Benchmarking MDMi clusters against human and mouse microglial reference datasets

We next benchmarked cluster-specific gene signatures from 2D, 3D, and 3D co-culture MDMi against ten independent human microglial reference datasets compiled by Oldham et al. (2008), Miller et al. (2010), Hawrylycz et al. (2012), Darmanis et al. (2015), Wehrspaun et al. (2015), Zhang et al. (2016), Galatro et al. (2017), Gosselin et al. (2017), Patir et al. (2019), and Olah et al. (2020). These reference lists comprise genes annotated to core human microglial identity and to canonical microglial states, including interferon-responsive, chemokine-enriched, metabolic, proliferative and border-associated–like programs, as defined in prior single-cell studies (Grabert et al., 2016; Hammond et al., 2019; Masuda et al., 2019). Across all culture conditions, the largest overlap was consistently observed with the primary human microglia dataset from Olah et al., with cluster-level overlaps ranging from 139–483 genes in 2D, 133–597 genes in 3D and 252–557 genes in 3D co-culture (Supplementary Table 6).

Notably, 3D and 3D co-culture models showed both higher overlap and a broader distribution of overlapping genes across clusters compared with 2D cultures. In 3D MDMi, clusters 2 and 3 each overlapped with more than 500 microglial genes from Olah et al., whereas in 3D co-culture four clusters (1, 3, 4, and 5) overlapped with more than 430 genes. In contrast, 2D clusters showed more restricted overlap, with only clusters 2 and 3 exceeding 450 overlapping genes (Supplementary Tables 6, 7 and Supplementary Figure 6a). These results indicate that 3D-based models capture a larger fraction of established human microglial gene expression.

Given the high overlap with human microglial reference signatures, we next examined the expression of homeostatic and disease-associated microglia (DAM) genes across MDMi clusters (Supplementary Figures 6b–d), using the mouse DAM signature (Keren-Shaul et al., 2017) and recently reviewed (Fumagalli et al., 2025). Genes including HEXB, CST3, and C1QA/C1QB were broadly expressed across clusters in all three culture conditions, indicating stable microglial gene expression across models. By contrast, DAM-associated genes showed heterogeneous, cluster-restricted expression rather than uniform activation, with representative markers such as TREM2 and LPL varying between clusters and culture conditions. Overall, these data show that MDMi preserve key microglial features, and that more complex culture environments drive them into a wider range of microglia-like states that activate established human microglial gene programs.

4 Discussion

Microglia display remarkable heterogeneity shaped by developmental origin, regional cues, and environmental context (Yang et al., 2013; Dando et al., 2016, Grabert et al., 2016; De Biase et al., 2017; Ayata et al., 2018; Dubbelaar et al., 2018; Hammond et al., 2019; Marsh et al., 2022; Stogsdill et al., 2022; Yaqubi et al., 2023). To capture this complexity, in vitro models have evolved from 2D cultures to iPSC-derived microglia, 3D systems and 3D co-cultures that better reflect human microglial transcriptional signatures and behavior (Park et al., 2018; Liu et al., 2022; Stöberl et al., 2023; Tujula et al., 2025). Building on prior work and our development of patient-derived MDMi systems, this study shows that culture dimensionality and neural context profoundly influence microglial transcriptional profiles, even under homeostatic conditions (Ohgidani et al., 2014; Ryan et al., 2017; Sellgren et al., 2017; Park et al., 2023; Cuní-López et al., 2024; Gonul et al., 2025; Llaves-López et al., 2025; Risby-Jones et al., 2025).

In 2D systems, microglia were dominated by an IFN-responsive state (53.77%), characterized by elevated IFITM2 and IFITM3, markers of antiviral surveillance and innate immune priming (Diamond and Farzan, 2013; Zhou et al., 2013; Fensterl and Sen, 2015). The upregulation of IFITM3, a member of the microglial “sensome,” highlights its specialized role in sensing pathogen- and danger-associated molecular patterns (Hickman et al., 2013), and has also been shown to be upregulated in glial cells under AD-related neuroinflammation, supporting its role in mediating immune responses in disease contexts (Hur et al., 2020). The predominance of this state in 2D indicates that the cultures maintain microglia in a more uniform, immune-alert condition, likely reflecting the absence of native extracellular matrix (ECM) cues. In contrast, 3D and co-culture systems diversified the microglial landscape, generating metabolic and proliferative subtypes alongside chemokine-enriched populations. These findings suggest that 3D environments provide spatial and biochemical complexity required to support microglial transcriptional state diversity.

Chemokine-enriched microglia were consistently identified across models but exhibited distinct transcriptional features depending on culture context. In 2D cultures, these microglia expressed antiviral and immune-surveillance genes, whereas in 3D and co-culture systems they upregulated CCL4, CCL3, CCL7, and CSF1, consistent with enhanced inflammatory signaling, immune-cell recruitment and matrix interaction. These chemokines have been implicated in neuroinflammation through modulation of BBB integrity and immune cell trafficking (Estevao et al., 2021), highlighting the influence of the microenvironment on immune signaling pathways (Cherry et al., 2020).

Notably, 3D systems contained a higher proportion of metabolically active and proliferative microglia, characterized by increased mitochondrial and biosynthetic gene programs. Although metabolic variation among microglial subtypes remains underexplored, several studies support the influence of local cues on microglial energy-related transcriptional states. For instance, studies have shown that hippocampal and cerebellar microglia exhibit greater expression of metabolic genes than those in cortex and striatum, reflecting region-specific energy demands (Grabert et al., 2016). Similarly, microglia co-culture within human dorsal or ventral forebrain spheroids displays distinct metabolic pathway activity (Song et al., 2019; Zhang et al., 2023). Our findings therefore align with the broader literature indicating that 3D or multicellular contexts promote energy-demanding, adaptive transcriptional programs associated with increased metabolic activity.

Across all models, conserved microglial markers such as SELENOP, NUPR1, FOLR2, C1QA, F13A1, ALK, CCL13, CXCL12, C1QB, HS3ST2, and TMEM37 were consistently expressed, reaffirming their fundamental role in maintaining microglial identity. However, only the 3D and 3D co-culture conditions uniquely induce expression of cell-cell interactions and extracellular matrix components, including FCGBP and IGSF21, consistent with enhanced cell attachment, growth and differentiation in spatially structured systems (Baker and Chen, 2012; Hedegaard et al., 2020). The 3D co-culture model further distinguished itself by upregulating genes such as LYVE1, SERPINB2, and HIC1, indicative of active cross-talk with neighboring neural and glial cells. These results align with previous findings that 3D co-culture systems recapitulate key tissue-relevant features, including multicellularity and neuronal-glial interactions (Moysidou et al., 2021; Wareham and Calkins, 2025), reinforcing the physiological relevance of this model. Expression of genes such as LYVE1, CD163, and CD38 likely reflects the intrinsic plasticity of monocyte-derived microglia and the influence of complex 3D and co-culture environments, which are known to promote macrophage-like, reactive, or inflammatory transcriptional programs. Similar transcriptional states have been reported in in vitro microglial and macrophage models exposed to extracellular matrices or inflammatory cues (Wen et al., 2023).

Pseudotime trajectory analysis provided complementary evidence that culture dimensionality influences microglial transcriptional progression and state complexity (Cheng et al., 2023). Microglia in 2D cultures followed a largely linear trajectory consistent with a more constrained and uniform transcriptional profile, whereas 3D and co-culture conditions exhibited branching trajectories, indicating increased transcriptional diversity and alternative state transitions. These patterns are consistent with prior observations that microglial transcriptional states can progress through multiple transitional programs. In 3D, enrichment of genes associated with cytokine signaling, immune pathways and cellular motility was observed, while 3D co-cultures further showed increased representation of ERK1/2 signaling and bacterial recognition-related pathways. Collectively, these findings suggest that 3D monoculture and co-culture systems support more dynamic transcriptional state transitions, providing a framework for modeling context-dependent microglial adaptations observed in neurodegeneration or injury.

Together with prior work, our findings support the use of advanced human in vitro systems, particularly 3D and co-culture models, to capture key aspects of microglial biology that are poorly represented in simpler systems. Comparative studies have shown that human primary and iPSC-derived microglia exhibit more robust inflammatory and phagocytic responses than immortalized or murine models (Maguire et al., 2022; Woolf et al., 2025). Consistent with this, MDMi expressed canonical microglial markers absent in monocytes, supporting their utility as an accessible human platform for studying microglia-associated pathways.

4.1 Current gaps and outlook

The conclusions of this study apply specifically to monocyte-derived microglia-like cells, and differences in microglial developmental origin are likely to further shape transcriptional identity. While transcriptomic analyses are critical for defining cellular heterogeneity, complementary functional validation is required to link transcriptional diversity to biological outcomes. Functional assays, including phagocytosis, cytokine and chemokine secretion, oxidative stress responses and synaptic pruning, will be necessary to determine whether the microglial states identified here correspond to functionally distinct phenotypes. Integrating such phenotypic readouts with transcriptomic profiling will provide a more comprehensive understanding of microglial behavior across different culture environments. Consistent with this, prior studies have shown that gene expression profiles alone do not fully predict cellular function, underscoring the importance of correlating molecular signatures with functional outputs (Cakir et al., 2022).

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

Ethics statement

The studies involving humans were approved by the National Health and Medical Research Council (NHMRC) guidelines for human research ethics and were approved by the QIMR Berghofer Medical Research Institute Human Research Ethics Committee. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

FE: Conceptualization, Validation, Data curation, Writing – review & editing, Methodology, Writing – original draft, Investigation, Formal analysis, Visualization. AW: Project administration, Supervision, Funding acquisition, Conceptualization, Resources, Writing – review & editing. HQ: Visualization, Resources, Data curation, Conceptualization, Project administration, Investigation, Funding acquisition, Methodology, Writing – review & editing, Supervision.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research was funded by the National Health and Medical Research Council of Australia (NHMRC) (APP1125796), a National Foundation for Medical Research and Innovation (NFMRI) grant, and a US Dept Defense CDMRP ALS grant (W81XWH-22-1-0714) to AW. AW was supported by an NHMRC Senior Research Fellowship (APP1118452), and HQ was supported by an NHMRC Ideas grant (APP2029183) and MND Research Australia Fat Rabbit Innovator grant (IG_2439).

Acknowledgments

We would like to thank all authors who contributed to the work in Cuní-López et al. (2024), whose dataset were used in this manuscript. We also sincerely thank the QIMR Berghofer HPC and Bioinformatics team—Ross Koufariotis, Scott Wood, David Burrows, Rebecca Johnston and Sharon Hoyte for their valuable assistance and support throughout this work.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fncel.2026.1770518/full#supplementary-material

Supplementary Figure 1 | Immediate-early gene expression across culture conditions. (a) Dot plot showing expression of immediate-early genes (IEGs) across monocytes, 2D MDMi, 3D MDMi, and 3D MDMi co-cultures. Dot size represents the percentage of cells expressing each gene and color intensity indicates average normalized expression. (b) Condition-specific dot plots showing immediate-early gene expression across clusters in 2D MDMi, 3D MDMi, and 3D co-culture conditions.

Supplementary Figure 2 | Heatmap of normalized expression for selected differentially expressed genes in each cluster in 2D MDMi model. Range of log2 normalized counts are shown at right hand of heatmap. Blue indicates low expression; and orange indicates high expression.

Supplementary Figure 3 | Heatmap of normalized expression for selected differentially expressed genes in each cluster in 3D MDMi model. Range of log2 normalized counts are shown at right hand of heatmap. Blue indicates low expression; and orange indicates high expression.

Supplementary Figure 4 | Heatmap of normalized expression for selected differentially expressed genes in each cluster in 3D co-culture. Range of log2 normalized counts are shown at right hand of heatmap. Blue indicates low expression; and orange indicates high expression.

Supplementary Figure 5 | Comparative analysis of microglial disease related genes in healthy 2D, 3D, and 3D co-culture models. (a) Dot plots illustrate AD related gene expression patterns in microglial models and monocytes, (b) dot plots illustrate PD related gene expression patterns in microglial models and monocytes, (c) dot plots illustrate ALS related gene expression patterns in microglial models and monocytes. Dot size indicating the percentage of cells expressing a given gene and color intensity reflecting the average expression level.

Supplementary Figure 6 | Human microglial signature overlap and marker gene expression across MDMi culture conditions. (a) Bar plots showing the number of genes overlapping the human microglial transcriptional signature defined by Olah et al. (2020) across clusters in 2D MDMi, 3D MDMi, and 3D MDMi co-culture conditions. (b–d) Violin plots showing expression of selected microglial marker genes across clusters in 2D MDMi (b), 3D MDMi (c), and 3D co-culture (d) conditions. Genes are ordered from top to bottom according to canonical microglial identity, homeostatic, and activation-associated DAM genes. Each violin represents the distribution of normalized expression levels within a cluster.

Supplementary Table 1 | scRNAseq quality control metrics.

Supplementary Table 2 | Shows cluster-specific gene expression in 2D MDMi.

Supplementary Table 3 | Shows cluster-specific gene expression in 3D MDMi.

Supplementary Table 4 | Shows cluster-specific gene expression in 3D co-culture MDMi.

Supplementary Table 5 | Shows DE analysis of all cultures.

Supplementary Table 6 | Comparison of MDMi with 10 published human microglia signatures.

Supplementary Table 7 | Overlap of cluster-specific MDMi genes with the Olah human microglia gene set.

References

Ayata, P., Badimon, A., Strasburger, H., Duff, M., Montgomery, S., Loh, Y., et al. (2018). Epigenetic regulation of brain region-specific microglia clearance activity. Nat. Neurosci. 21, 1049–1060. doi: 10.1038/s41593-018-0192-3

PubMed Abstract | Crossref Full Text | Google Scholar

Baker, B., and Chen, C. (2012). Deconstructing the third dimension: How 3D culture microenvironments alter cellular cues. J. Cell. Sci. 125, 3015–3024. doi: 10.1242/jcs.079509

PubMed Abstract | Crossref Full Text | Google Scholar

Cadiz, M., Jensen, T., Sens, J., Zhu, K., Song, W., Zhang, B., et al. (2022). Culture shock: Microglial heterogeneity, activation, and disrupted single-cell microglial networks in vitro. Mol. Neurodegener. 17:26. doi: 10.1186/s13024-022-00531-1

PubMed Abstract | Crossref Full Text | Google Scholar

Cakir, B., Kiral, F. R., and Park, I.-H. (2022). Advanced in vitro models: Microglia in action. Neuron 110, 3444–3457. doi: 10.1016/j.neuron.2022.10.004

PubMed Abstract | Crossref Full Text | Google Scholar

Cakir, B., Xiang, Y., Tanaka, Y., Kural, M., Parent, M., Kang, Y., et al. (2019). Engineering of human brain organoids with a functional vascular-like system. Nat. Methods 16, 1169–1175. doi: 10.1038/s41592-019-0586-5

PubMed Abstract | Crossref Full Text | Google Scholar

Cheng, S., Brenière-Letuffe, D., Ahola, V., Wong, A., Keung, H., Gurung, B., et al. (2023). Single-cell RNA sequencing reveals maturation trajectory in human pluripotent stem cell-derived cardiomyocytes in engineered tissues. iScience 26:106302. doi: 10.1016/j.isci.2023.106302

PubMed Abstract | Crossref Full Text | Google Scholar

Cherry, J., Meng, G., Daley, S., Xia, W., Svirsky, S., Alvarez, V., et al. (2020). CCL2 is associated with microglia and macrophage recruitment in chronic traumatic encephalopathy. J. Neuroinflammation 17:370. doi: 10.1186/s12974-020-02036-4

PubMed Abstract | Crossref Full Text | Google Scholar

Chia, S., Li, M., Li, Z., Tu, H., Lee, J., Qiu, L., et al. (2025). Single-nucleus transcriptomics reveals a distinct microglial state and increased MSR1-mediated phagocytosis as common features across dementia subtypes. Genome Med. 17:92. doi: 10.1186/s13073-025-01519-4

PubMed Abstract | Crossref Full Text | Google Scholar

Colonna, M., and Butovsky, O. (2017). Microglia function in the central nervous system during health and neurodegeneration. Annu. Rev. Immunol. 35, 441–468. doi: 10.1146/annurev-immunol-051116-052358

PubMed Abstract | Crossref Full Text | Google Scholar

Cuní-López, C., Stewart, R., Oikari, L., Nguyen, T., Roberts, T., Sun, Y., et al. (2024). Advanced patient-specific microglia cell models for pre-clinical studies in Alzheimer’s disease. J. Neuroinflammation 21:50. doi: 10.1186/s12974-024-03037-3

PubMed Abstract | Crossref Full Text | Google Scholar

Cuní-López, C., Stewart, R., White, A., and Quek, H. (2023). 3D in vitro modelling of human patient microglia: A focus on clinical translation and drug development in neurodegenerative diseases. J. Neuroimmunol. 375:578017. doi: 10.1016/j.jneuroim.2023.578017

PubMed Abstract | Crossref Full Text | Google Scholar

Dando, S., Naranjo Golborne, C., Chinnery, H., Ruitenberg, M., and McMenamin, P. G. A. (2016). case of mistaken identity: cd11c-eyfp(+) cells in the normal mouse brain parenchyma and neural retina display the phenotype of microglia, not dendritic cells. Glia 64, 1331–1349. doi: 10.1002/glia.23005

PubMed Abstract | Crossref Full Text | Google Scholar

Darmanis, S., Sloan, S., Zhang, Y., Enge, M., Caneda, C., Shuer, L., et al. (2015). A survey of human brain transcriptome diversity at the single cell level. Proc. Natl. Acad. Sci. U S A. 112, 7285–7290. doi: 10.1073/pnas.1507125112

PubMed Abstract | Crossref Full Text | Google Scholar

De Biase, L., Schuebel, K., Fusfeld, Z., Jair, K., Hawes, I., Cimbro, R., et al. (2017). Local cues establish and maintain region-specific phenotypes of basal ganglia microglia. Neuron 95, 341–356.e6. doi: 10.1016/j.neuron.2017.06.020

PubMed Abstract | Crossref Full Text | Google Scholar

Depp, C., Doman, J., Hingerl, M., Xia, J., and Stevens, B. (2025). Microglia transcriptional states and their functional significance: Context drives diversity. Immunity 58, 1052–1067. doi: 10.1016/j.immuni.2025.04.009

PubMed Abstract | Crossref Full Text | Google Scholar

Diamond, M., and Farzan, M. (2013). The broad-spectrum antiviral functions of IFIT and IFITM proteins. Nat. Rev. Immunol. 13, 46–57. doi: 10.1038/nri3344

PubMed Abstract | Crossref Full Text | Google Scholar

Discher, D., Janmey, P., and Wang, Y. (2005). Tissue cells feel and respond to the stiffness of their substrate. Science 310, 1139–1143. doi: 10.1126/science.1116995

PubMed Abstract | Crossref Full Text | Google Scholar

Dolan, M., Therrien, M., Jereb, S., Kamath, T., Gazestani, V., Atkeson, T., et al. (2023). Exposure of iPSC-derived human microglia to brain substrates enables the generation and manipulation of diverse transcriptional states in vitro. Nat. Immunol. 24, 1382–1390. doi: 10.1038/s41590-023-01558-2

PubMed Abstract | Crossref Full Text | Google Scholar

Dubbelaar, M., Kracht, L., Eggen, B., and Boddeke, E. (2018). The kaleidoscope of microglial phenotypes. Front. Immunol. 9:1753. doi: 10.3389/fimmu.2018.01753

PubMed Abstract | Crossref Full Text | Google Scholar

Estevao, C., Bowers, C., Luo, D., Sarker, M., Hoeh, A., Frudd, K., et al. (2021). CCL4 induces inflammatory signalling and barrier disruption in the neurovascular endothelium. Brain Behav. Immun. Health 18:100370. doi: 10.1016/j.bbih.2021.100370

PubMed Abstract | Crossref Full Text | Google Scholar

Fensterl, V., and Sen, G. (2015). Interferon-induced Ifit proteins: Their role in viral pathogenesis. J. Virol. 89, 2462–2468. doi: 10.1128/JVI.02744-14

PubMed Abstract | Crossref Full Text | Google Scholar

Friedman, B., Srinivasan, K., Ayalon, G., Meilandt, W., Lin, H., Huntley, M., et al. (2018). Diverse brain myeloid expression profiles reveal distinct microglial activation states and aspects of Alzheimer’s disease not evident in mouse models. Cell. Rep. 22, 832–847. doi: 10.1016/j.celrep.2017.12.066

PubMed Abstract | Crossref Full Text | Google Scholar

Fumagalli, L., Nazlie Mohebiany, A., Premereur, J., Polanco Miquel, P., Bijnens, B., Van de Walle, P., et al. (2025). Microglia heterogeneity, modeling and cell-state annotation in development and neurodegeneration. Nat. Neurosci. 28, 1381–1392. doi: 10.1038/s41593-025-01931-4

PubMed Abstract | Crossref Full Text | Google Scholar

Galatro, T., Holtman, I., Lerario, A., Vainchtein, I., Brouwer, N., Sola, P., et al. (2017). Transcriptomic analysis of purified human cortical microglia reveals age-associated changes. Nat. Neurosci. 20, 1162–1171. doi: 10.1038/nn.4597

PubMed Abstract | Crossref Full Text | Google Scholar

Glass, C., Saijo, K., Winner, B., Marchetto, M., and Gage, F. (2010). Mechanisms underlying inflammation in neurodegeneration. Cell 140, 918–934. doi: 10.1016/j.cell.2010.02.016

PubMed Abstract | Crossref Full Text | Google Scholar

Gonul, C., Kiser, C., Yaka, E., Oz, D., Hunerli, D., Yerlikaya, D., et al. (2025). Microglia-like cells from patient monocytes demonstrate increased phagocytic activity in probable Alzheimer’s disease. Mol. Cell. Neurosci. 132:103990. doi: 10.1016/j.mcn.2024.103990

PubMed Abstract | Crossref Full Text | Google Scholar

Gosselin, D., Skola, D., Coufal, N., Holtman, I., Schlachetzki, J., Sajti, E., et al. (2017). An environment-dependent transcriptional network specifies human microglia identity. Science 356:eaal3222. doi: 10.1126/science.aal3222

PubMed Abstract | Crossref Full Text | Google Scholar

Grabert, K., Michoel, T., Karavolos, M., Clohisey, S., Baillie, J., Stevens, M., et al. (2016). Microglial brain region-dependent diversity and selective regional sensitivities to aging. Nat. Neurosci. 19, 504–516. doi: 10.1038/nn.4222

PubMed Abstract | Crossref Full Text | Google Scholar

Gullotta, G., Costantino, G., Sortino, M., and Spampinato, S. (2023). Microglia and the blood-brain barrier: An external player in acute and chronic neuroinflammatory conditions. Int. J. Mol. Sci. 24:9144. doi: 10.3390/ijms24119144

PubMed Abstract | Crossref Full Text | Google Scholar

Haenseler, W., Sansom, S., Buchrieser, J., Newey, S., Moore, C., Nicholls, F., et al. (2017). A highly efficient human pluripotent stem cell microglia model displays a neuronal-co-culture-specific expression profile and inflammatory response. Stem. Cell. Reports 8, 1727–1742. doi: 10.1016/j.stemcr.2017.05.017

PubMed Abstract | Crossref Full Text | Google Scholar

Hammond, T., Dufort, C., Dissing-Olesen, L., Giera, S., Young, A., Wysoker, A., et al. (2019). Single-Cell RNA sequencing of microglia throughout the mouse lifespan and in the injured brain reveals complex cell-state changes. Immunity 50, 253–271.e6. doi: 10.1016/j.immuni.2018.11.004

PubMed Abstract | Crossref Full Text | Google Scholar

Hao, Y., Hao, S., Andersen-Nissen, E., Mauck, W., Zheng, S., Butler, A., et al. (2021). Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e29. doi: 10.1016/j.cell.2021.04.048

PubMed Abstract | Crossref Full Text | Google Scholar

Hawrylycz, M., Lein, E., Guillozet-Bongaarts, A., Shen, E., Ng, L., Miller, J., et al. (2012). An anatomically comprehensive atlas of the adult human brain transcriptome. Nature 489, 391–399. doi: 10.1038/nature11405

PubMed Abstract | Crossref Full Text | Google Scholar

He, J., Qin, Z., Liu, H., Cai, Y., Wu, H., Wang, T., et al. (2025). Fatty acid-binding protein 4 drives microglia-mediated neuroinflammation through promoting S100A9 expression and lipid droplet accumulation after intracerebral hemorrhage. J. Neuroinflammation 22:263. doi: 10.1186/s12974-025-03573-6

PubMed Abstract | Crossref Full Text | Google Scholar

Hedegaard, A., Stodolak, S., James, W., and Cowley, S. (2020). Honing the double-edged sword: Improving human iPSC-Microglia models. Front. Immunol. 11:614972. doi: 10.3389/fimmu.2020.614972

PubMed Abstract | Crossref Full Text | Google Scholar

Hickman, S., Kingery, N., Ohsumi, T., Borowsky, M., Wang, L., Means, T., et al. (2013). The microglial sensome revealed by direct RNA sequencing. Nat. Neurosci. 16, 1896–1905. doi: 10.1038/nn.3554

PubMed Abstract | Crossref Full Text | Google Scholar

Hur, J., Frost, G., Wu, X., Crump, C., Pan, S., Wong, E., et al. (2020). The innate immunity protein IFITM3 modulates γ-secretase in Alzheimer’s disease. Nature 586, 735–740. doi: 10.1038/s41586-020-2681-2

PubMed Abstract | Crossref Full Text | Google Scholar

Keren-Shaul, H., Spinrad, A., Weiner, A., Matcovitch-Natan, O., Dvir-Szternfeld, R., Ulland, T., et al. (2017). A unique microglia type associated with restricting development of Alzheimer’s disease. Cell 169, 1276–1290.e17. doi: 10.1016/j.cell.2017.05.018

PubMed Abstract | Crossref Full Text | Google Scholar

Krasemann, S., Madore, C., Cialic, R., Baufeld, C., Calcagno, N., El Fatimy, R., et al. (2017). The TREM2-APOE pathway drives the transcriptional phenotype of dysfunctional microglia in neurodegenerative diseases. Immunity 47, 566–581.e9. doi: 10.1016/j.immuni.2017.08.008

PubMed Abstract | Crossref Full Text | Google Scholar

Liu, R., Meng, X., Yu, X., Wang, G., Dong, Z., Zhou, Z., et al. (2022). From 2D to 3D Co-culture systems: A review of co-culture models to study the neural cells interaction. Int. J. Mol. Sci. 23:13116. doi: 10.3390/ijms232113116

PubMed Abstract | Crossref Full Text | Google Scholar

Llaves-López, A., Micoli, E., Belmonte-Mateos, C., Aguilar, G., Alba, C., Marsal, A., et al. (2025). Human microglia-like cells differentiated from monocytes with GM-CSF and IL-34 show phagocytosis of α-Synuclein aggregates and C/EBPβ-Dependent proinflammatory activation. Mol. Neurobiol. 62, 756–772. doi: 10.1007/s12035-024-04289-z

PubMed Abstract | Crossref Full Text | Google Scholar

Lloyd, A., Martinez-Muriana, A., Davis, E., Daniels, M., Hou, P., Mancuso, R., et al. (2024). Deep proteomic analysis of microglia reveals fundamental biological differences between model systems. Cell. Rep. 43:114908. doi: 10.1016/j.celrep.2024.114908

PubMed Abstract | Crossref Full Text | Google Scholar

Lopez-Atalaya, J., and Bhojwani-Cabrera, A. (2025). Type I interferon signalling and interferon-responsive microglia in health and disease. FEBS J. 292, 5921–5940. doi: 10.1111/febs.70126

PubMed Abstract | Crossref Full Text | Google Scholar

Luecken, M., and Theis, F. (2019). Current best practices in single-cell RNA-seq analysis: A tutorial. Mol. Syst. Biol. 15:e8746. doi: 10.15252/msb.20188746

PubMed Abstract | Crossref Full Text | Google Scholar

Maguire, E., Connor-Robson, N., Shaw, B., O’Donoghue, R., Stöberl, N., and Hall-Roberts, H. (2022). Assaying microglia functions in vitro. Cells 11:3414. doi: 10.3390/cells11213414

PubMed Abstract | Crossref Full Text | Google Scholar

Mancuso, R., Van Den Daele, J., Fattorelli, N., Wolfs, L., Balusu, S., Burton, O., et al. (2019). Stem-cell-derived human microglia transplanted in mouse brain to study human disease. Nat. Neurosci. 22, 2111–2116. doi: 10.1038/s41593-019-0525-x

PubMed Abstract | Crossref Full Text | Google Scholar

Marsh, S., Walker, A., Kamath, T., Dissing-Olesen, L., Hammond, T., de Soysa, T., et al. (2022). Dissection of artifactual and confounding glial signatures by single-cell sequencing of mouse and human brain. Nat. Neurosci. 25, 306–316. doi: 10.1038/s41593-022-01022-8

PubMed Abstract | Crossref Full Text | Google Scholar

Masuda, T., Sankowski, R., Staszewski, O., Böttcher, C., Amann, L., and Sagar, et al. (2019). Spatial and temporal heterogeneity of mouse and human microglia at single-cell resolution. Nature 566, 388–392. doi: 10.1038/s41586-019-0924-x

PubMed Abstract | Crossref Full Text | Google Scholar

Michell-Robinson, M., Touil, H., Healy, L., Owen, D., Durafourt, B., Bar-Or, A., et al. (2015). Roles of microglia in brain development, tissue maintenance and repair. Brain 138, 1138–1159. doi: 10.1093/brain/awv066

PubMed Abstract | Crossref Full Text | Google Scholar

Miller, J., Horvath, S., and Geschwind, D. (2010). Divergence of human and mouse brain transcriptome highlights Alzheimer disease pathways. Proc. Natl. Acad. Sci. U S A. 107, 12698–12703. doi: 10.1073/pnas.0914257107

PubMed Abstract | Crossref Full Text | Google Scholar

Morgan, M., and Shepherd, L. (2017). AnnotationHub: Client to Access AnnotationHub resources. R package version 2, no. 1.

Google Scholar

Moysidou, C., Barberio, C., and Owens, R. (2021). Advances in engineering human tissue models. Front. Bioeng. Biotechnol. 8:620962. doi: 10.3389/fbioe.2020.620962

PubMed Abstract | Crossref Full Text | Google Scholar

Ochocka, N., Segit, P., Walentynowicz, K., Wojnicki, K., Cyranowski, S., Swatler, J., et al. (2021). Single-cell RNA sequencing reveals functional heterogeneity of glioma-associated brain macrophages. Nat. Commun. 12:1151. doi: 10.1038/s41467-021-21407-w

PubMed Abstract | Crossref Full Text | Google Scholar

Ohgidani, M., Kato, T., Setoyama, D., Sagata, N., Hashimoto, R., Shigenobu, K., et al. (2014). Direct induction of ramified microglia-like cells from human monocytes: Dynamic microglial dysfunction in Nasu-Hakola disease. Sci. Rep. 4:4957. doi: 10.1038/srep04957

PubMed Abstract | Crossref Full Text | Google Scholar

Olah, M., Menon, V., Habib, N., Taga, M., Ma, Y., Yung, C., et al. (2020). Single cell RNA sequencing of human microglia uncovers a subset associated with Alzheimer’s disease. Nat. Commun. 11:6129. doi: 10.1038/s41467-020-19737-2

PubMed Abstract | Crossref Full Text | Google Scholar

Oldham, M., Konopka, G., Iwamoto, K., Langfelder, P., Kato, T., Horvath, S., et al. (2008). Functional organization of the transcriptome in human brain. Nat. Neurosci. 11, 1271–1282. doi: 10.1038/nn.2207

PubMed Abstract | Crossref Full Text | Google Scholar

Paolicelli, R. C., Sierra, A., Stevens, B., Tremblay, M.-E., Aguzzi, A., Ajami, B., et al. (2022). Microglia states and nomenclature: A field at its crossroads. Neuron 110, 3458–3483. doi: 10.1016/j.neuron.2022.10.020

PubMed Abstract | Crossref Full Text | Google Scholar

Park, D., Kozaki, T., Tiwari, S., Moreira, M., Khalilnezhad, A., Torta, F., et al. (2023). iPS-cell-derived microglia promote brain organoid maturation via cholesterol transfer. Nature 623, 397–405. doi: 10.1038/s41586-023-06713-1

PubMed Abstract | Crossref Full Text | Google Scholar

Park, J., Wetzel, I., Dréau, D., D’Avanzo, C., Kim, D., et al. (2018). A 3D human triculture system modeling neurodegeneration and neuroinflammation in Alzheimer’s disease. Nat. Neurosci. 21, 941–951. doi: 10.1038/s41593-018-0175-4

PubMed Abstract | Crossref Full Text | Google Scholar

Patir, A., Shih, B., McColl, B., and Freeman, T. C. A. (2019). core transcriptional signature of human microglia: Derivation and utility in describing region-dependent alterations associated with Alzheimer’s disease. Glia 67, 1240–1253. doi: 10.1002/glia.23572

PubMed Abstract | Crossref Full Text | Google Scholar

Quek, H., Cuní-López, C., Stewart, R., Lim, Y., Roberts, T., and White, A. R. (2022). A robust approach to differentiate human monocyte-derived microglia from peripheral blood mononuclear cells. STAR Protoc. 3:101747. doi: 10.1016/j.xpro.2022.101747

PubMed Abstract | Crossref Full Text | Google Scholar

R Core Team. (2019). R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.

Google Scholar

Risby-Jones, G., Marallag, J., Jagaraj, C., Atkin, J., Walker, A., Woodruff, T., et al. (2025). IL-6 trans-signalling is elevated in ALS models and drives TDP-43 induced inflammatory responses in microglia. Brain Behav. Immun. 129, 296–304. doi: 10.1016/j.bbi.2025.06.021

PubMed Abstract | Crossref Full Text | Google Scholar

Ryan, K., White, C., Patel, K., Xu, J., Olah, M., Replogle, J., et al. (2017). A human microglia-like cellular model for assessing the effects of neurodegenerative disease gene variants. Sci. Transl. Med. 9:eaai7635. doi: 10.1126/scitranslmed.aai7635

PubMed Abstract | Crossref Full Text | Google Scholar

Schafer, S., Mansour, A., Schlachetzki, J., Pena, M., Ghassemzadeh, S., Mitchell, L., et al. (2023). An in vivo neuroimmune organoid model to study human microglia phenotypes. Cell 186, 2111–2126.e20. doi: 10.1016/j.cell.2023.04.022

PubMed Abstract | Crossref Full Text | Google Scholar

Sellgren, C., Sheridan, S., Gracias, J., Xuan, D., Fu, T., and Perlis, R. (2017). Patient-specific models of microglia-mediated engulfment of synapses and neural progenitors. Mol. Psychiatry 22, 170–177. doi: 10.1038/mp.2016.220

PubMed Abstract | Crossref Full Text | Google Scholar

Shamir, E., and Ewald, A. (2014). Three-dimensional organotypic culture: Experimental models of mammalian biology and disease. Nat. Rev. Mol. Cell. Biol. 15, 647–664. doi: 10.1038/nrm3873

PubMed Abstract | Crossref Full Text | Google Scholar

Sharaf, A., Timmerman, R., Bajramovic, J., and Accardo, A. (2023). In vitro microglia models: The era of engineered cell microenvironments. Neural Regen. Res. 18, 1709–1710. doi: 10.4103/1673-5374.363828

PubMed Abstract | Crossref Full Text | Google Scholar

Shaulian, E., and Karin, M. (2002). AP-1 as a regulator of cell life and death. Nat. Cell. Biol. 4, E131–E136. doi: 10.1038/ncb0502-e131

PubMed Abstract | Crossref Full Text | Google Scholar

Shimizu, T., and Prinz, M. (2025). Microglia across evolution: From conserved origins to functional divergence. Cell. Mol. Immunol. 22, 1533–1548. doi: 10.1038/s41423-025-01368-6

PubMed Abstract | Crossref Full Text | Google Scholar

Song, L., Yuan, X., Jones, Z., Vied, C., Miao, Y., Marzano, M., et al. (2019). Functionalization of brain region-specific spheroids with isogenic microglia-like cells. Sci. Rep. 9:11055. doi: 10.1038/s41598-019-47444-6

PubMed Abstract | Crossref Full Text | Google Scholar

Stöberl, N., Maguire, E., Salis, E., Shaw, B., and Hall-Roberts, H. (2023). Human iPSC-derived glia models for the study of neuroinflammation. J. Neuroinflammation 20:231. doi: 10.1186/s12974-023-02919-2

PubMed Abstract | Crossref Full Text | Google Scholar

Stogsdill, J., Kim, K., Binan, L., Farhi, S., Levin, J., and Arlotta, P. (2022). Pyramidal neuron subtype diversity governs microglia states in the neocortex. Nature 608, 750–756. doi: 10.1038/s41586-022-05056-7

PubMed Abstract | Crossref Full Text | Google Scholar

Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W., et al. (2019). Comprehensive integration of single-cell data. Cell 177, 1888–1902.e21. doi: 10.1016/j.cell.2019.05.031

PubMed Abstract | Crossref Full Text | Google Scholar

Sun, X., Ju, X., Li, Y., Zeng, P., Wu, J., Zhou, Y., et al. (2022). Generation of vascularized brain organoids to study neurovascular interactions. Elife 11:e76707. doi: 10.7554/eLife.76707

PubMed Abstract | Crossref Full Text | Google Scholar

Tang, Y., Zhang, Q., Qu, Y., Yi, L., Li, F., Qu, C., et al. (2025). Single-cell RNA sequencing reveals microglial heterogeneity and functional states after cerebral ischemia-reperfusion injury. J. Inflamm. Res. 18, 16931–16956. doi: 10.2147/JIR.S539404

PubMed Abstract | Crossref Full Text | Google Scholar

Trapnell, C., Cacchiarelli, D., Grimsby, J., Pokharel, P., Li, S., Morse, M., et al. (2014). The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386. doi: 10.1038/nbt.2859

PubMed Abstract | Crossref Full Text | Google Scholar

Tujula, I., Hyvärinen, T., Lotila, J., Rogal, J., Voulgaris, D., Sukki, L., et al. (2025). Modeling neuroinflammatory interactions between microglia and astrocytes in a human iPSC-based coculture platform. Cell Commun. Signal. 23:298. doi: 10.1186/s12964-025-02304-x

PubMed Abstract | Crossref Full Text | Google Scholar

Ulland, T., Song, W., Huang, S., Ulrich, J., Sergushichev, A., Beatty, W., et al. (2017). TREM2 maintains microglial metabolic fitness in Alzheimer’s disease. Cell 170, 649–663.e13. doi: 10.1016/j.cell.2017.07.023

PubMed Abstract | Crossref Full Text | Google Scholar

Van Hove, H., Martens, L., Scheyltjens, I., De Vlaminck, K., Pombo Antunes, A., De Prijck, S., et al. (2019). A single-cell atlas of mouse brain macrophages reveals unique transcriptional identities shaped by ontogeny and tissue environment. Nat. Neurosci. 22, 1021–1035. doi: 10.1038/s41593-019-0393-4

PubMed Abstract | Crossref Full Text | Google Scholar

Wareham, L., and Calkins, D. (2025). Making tracks: Microglia and the extracellular matrix. Mol. Neurodegener. 20:101. doi: 10.1186/s13024-025-00898-x

PubMed Abstract | Crossref Full Text | Google Scholar

Wehrspaun, C., Haerty, W., and Ponting, C. (2015). Microglia recapitulate a hematopoietic master regulator network in the aging human frontal cortex. Neurobiol. Aging 36, 2443.e9–2443.e20. doi: 10.1016/j.neurobiolaging.2015.04.008

PubMed Abstract | Crossref Full Text | Google Scholar

Wen, W., Cheng, J., and Tang, Y. (2023). Brain perivascular macrophages: Current understanding and future prospects. Brain 147, 39–55. doi: 10.1093/brain/awad304

PubMed Abstract | Crossref Full Text | Google Scholar

Woolf, Z., Stevenson, T., Lee, K., Highet, B., Macapagal Foliaki, J., Ratiu, R., et al. (2025). In vitro models of microglia: A comparative study. Sci. Rep. 15:15621. doi: 10.1038/s41598-025-99867-z

PubMed Abstract | Crossref Full Text | Google Scholar

Wu, T., Hu, E., Xu, S., Chen, M., Guo, P., Dai, Z., et al. (2021). clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation 2:100141. doi: 10.1016/j.xinn.2021.100141

PubMed Abstract | Crossref Full Text | Google Scholar

Xue, J., Zhang, Y., Zhang, J., Zhu, Z., Lv, Q., and Su, J. (2021). Astrocyte-derived CCL7 promotes microglia-mediated inflammation following traumatic brain injury. Int. Immunopharmacol. 99:107975. doi: 10.1016/j.intimp.2021.107975

PubMed Abstract | Crossref Full Text | Google Scholar

Yang, T., Lin, C., Hsu, C., Wang, T., Ke, F., and Kuo, Y. (2013). Differential distribution and activation of microglia in the brain of male C57BL/6J mice. Brain Struct. Funct. 218, 1051–1060. doi: 10.1007/s00429-012-0446-x

PubMed Abstract | Crossref Full Text | Google Scholar

Yaqubi, M., Groh, A., Dorion, M., Afanasiev, E., Luo, J., Hashemi, H., et al. (2023). Analysis of the microglia transcriptome across the human lifespan using single cell RNA sequencing. J. Neuroinflammation 20:132. doi: 10.1186/s12974-023-02809-7

PubMed Abstract | Crossref Full Text | Google Scholar

Zappia, L., and Oshlack, A. (2018). Clustering trees: A visualization for evaluating clusterings at multiple resolutions. Gigascience 7:giy083. doi: 10.1093/gigascience/giy083

PubMed Abstract | Crossref Full Text | Google Scholar

Zhang, W., Jiang, J., Xu, Z., Yan, H., Tang, B., Liu, C., et al. (2023). Microglia-containing human brain organoids for the study of brain development and pathology. Mol. Psychiatry 28, 96–107. doi: 10.1038/s41380-022-01892-1

PubMed Abstract | Crossref Full Text | Google Scholar

Zhang, Y., Sloan, S., Clarke, L., Caneda, C., Plaza, C., Blumenthal, P., et al. (2016). Purification and characterization of progenitor and mature human astrocytes reveals transcriptional and functional differences with mouse. Neuron 89, 37–53. doi: 10.1016/j.neuron.2015.11.013

PubMed Abstract | Crossref Full Text | Google Scholar

Zheng, G., Terry, J., Belgrader, P., Ryvkin, P., Bent, Z., Wilson, R., et al. (2017). Massively parallel digital transcriptional profiling of single cells. Nat. Commun. 8:14049. doi: 10.1038/ncomms14049

PubMed Abstract | Crossref Full Text | Google Scholar

Zhou, X., Michal, J., Zhang, L., Ding, B., Lunney, J., Liu, B., et al. (2013). Interferon induced IFIT family genes in host antiviral defense. Int. J. Biol. Sci. 9, 200–208. doi: 10.7150/ijbs.5613

PubMed Abstract | Crossref Full Text | Google Scholar