The microglial response to inhibition of Colony-stimulating-factor-1 receptor by PLX3397 differs by sex in adult mice

SUMMARY

Microglia, the resident macrophages of the brain, are derived from the yolk sac and colonize the brain before the blood-brain barrier forms. Once established, they expand locally and require Colony-stimulating-factor-1 receptor (CSF1R) signaling for their development and maintenance. CSF1R inhibitors have been used extensively to deplete microglia in the healthy and diseased brain. In this study, we demonstrated sex-dependent differences in the microglial response to the CSF1R inhibitor PLX3397. Male mice exhibited greater microglial depletion compared to females. Transcriptomic and flow cytometry analysis revealed sex-specific differences in the remaining microglia population, with female microglia upregulating autophagy and proteostasis pathways while male microglia increased mitobiogenesis. Furthermore, manipulating key microglial receptors by using different transgenic mouse lines resulted in changes in depletion efficacies that were also sex dependent. These findings suggest sex-dependent microglial survival mechanisms, which might contribute to the well-documented sex differences in various neurological disorders.

In brief

Le et al. examined the sex-dependent effects of a drug that removes microglia, called PLX3397. They revealed that male microglia are more susceptible to being depleted by PLX3397 than female microglia and that this difference likely arises from distinct survival strategies employed by male and female microglia.

Graphical Abstract

INTRODUCTION

Microglia are the resident innate immune cells of the central nervous system (CNS), distinguished by their ramified morphology with processes that constantly surveil the surrounding environment. Their constant patrolling activity allows them to serve as the first responders of the CNS, taking central stage as key players in the pathogenesis and progression of numerous neurodegenerative diseases, such as Alzheimer’s disease (AD), Parkinson’s disease (PD), tauopathies, and multiple sclerosis (MS).¹ Aside from classical immune functions, microglia are also crucial for CNS development and plasticity through their dynamic interactions with other cell types that help sculpt the synaptic environment.^2,3

Unlike other cells in the CNS, microglia arise from early, uncommitted embryonic myeloid progenitors between embryonic day (E) 7 and E7.5 in a PU.1/RUNX1/IRF8-dependent manner and migrate to colonize the CNS (E8.5–E9.5) prior to the formation of the blood-brain barrier (BBB) (E13–E14.5).^4,5 Once established, microglia expand their population via local clonal expansion and mature following a stepwise program until adulthood with minimal exchange with circulating monocytes throughout life in the absence of pathology.^6–9 Postnatal maturation of microglia requires transforming growth factor β (TGF-β) signaling,¹⁰ and adult homeostatic microglia are characterized by a distinct transcriptomic signature including microglia-enriched genes like P2ry12, Tmem119, and Sall1.^10–13

Colony-stimulating-factor-1 receptor (CSF1R) signaling is essential for the homing of microglia to the CNS during early development^4,5 and is also crucial for the maintenance of a stable microglial population throughout the animal’s lifespan.^14–18 Null mutation of the Csf1r gene¹⁴ or deletion of the super-enhancer FIRE at the Csf1r locus¹⁵ almost completely ablates microglia with a >99% removal rate. However, these transgenic models cannot distinguish the role of CSF1R signaling during development vs. adulthood. To study the importance of CSF1R in the adult mouse CNS, the small-molecule CSF1R inhibitors PLX3397 and PLX5622 have been widely utilized to deplete microglia.^16–22 While the depletion rate varies with dosage and duration of administration, complete microglia removal with acute PLX3397 or PLX5622 treatment has not been reported.^16–22 A high dose of PLX3397 (600 mg/kg) was shown to completely ablate brain microglia, but only after chronic treatment of 3.5 months.¹⁸ Taken together, accumulating evidence on CSF1R inhibitor treatment suggests the presence of an adult microglial pool that is partially resistant to CSF1R inhibition. This pool is likely to be diverse and require different thresholds for cell death. With the emerging knowledge of the importance of sex-dependent differences in microglia,^23–25 it is important to study the potential divergences in survival strategies utilized by male vs. female microglia.

In the current study, we describe sex-specific differences in the microglial response to CSF1R inhibition with PLX3397 (hereafter referred to as PLX). We show that microglia in male mice were significantly more sensitive to disrupted CSF1R signaling compared to microglia in female mice, as evidenced by a higher depletion efficacy in males. Microglia remaining after PLX treatment in female mice upregulated autophagy and proteostasis pathways, while those remaining in male mice upregulated the expression of electron transport chain complexes and mitobiogenesis, suggesting potential sex divergences in CSF1R downstream signaling. Taken together, our findings uncover sex differences in microglial CSF1R signaling in the adult mouse brain, which warrant further investigation, as the differential survival mechanisms utilized by male and female microglia might contribute to the high degree of sex-related differences in the prevalence of many neurological disorders.

RESULTS

Sex-dependent differences in microglia density and morphology at baseline and under CSF1R inhibition

To determine whether sex influenced microglial responsiveness to CSF1R inhibition, we counted the number of microglia immunolabeled with Iba1 remaining in different brain regions of adult male and female C57BL/6J mice that were fed a diet containing the CSF1R inhibitor PLX (at 290 mg/kg) for 2 weeks and compared them to males and females fed a control diet (Figure 1). To establish a baseline before depletion, we first sought to characterize microglia density and soma size in male and female mice in the absence of PLX treatment in the cortex, hippocampus (including CA1, dentate gyrus, and subiculum), thalamus, and midbrain (including pretectal nuclei, superior colliculus, periaqueductal gray, and midbrain reticular nuclei). We revealed that the cortex had the highest microglia density, followed by the hippocampus, and both these regions had significantly higher microglia density compared to the midbrain and thalamus (Figure S1A). In agreement with previous reports, we observed comparable microglia density between sexes²⁶ and a trend toward smaller soma size in females compared to males in the hippocampus,²⁷ which did not reach statistical significance (Figure S1B). Interestingly, initial observations from a small-scale experiment involving two male and three female mice suggested that male microglia appeared to have higher total mTOR protein levels compared to female microglia at baseline (Figure S1C), suggesting that male microglia might have a slightly higher rate of proliferation.^28,29 While we did not observe significant differences in microglia density between sexes in this study, upregulation of the microglial mTOR pathway in males might help explain previous findings showing a slightly higher microglia density²⁷ at baseline and after acute stab wound³⁰ in males compared to females.

Figure 1. Sex-dependent differences in microglia density and morphology under CSF1R inhibition.

(A) Representative 10× epifluorescence images (cropped from a stitched image representing the whole brain section) showing microglia immunolabeled with Iba1 in hippocampal CA1, white matter (WM), and primary visual cortex V1 in control and PLX-treated male and female mice.

(B) Quantification of depletion efficacies across different brain regions as depletion/control microglia density ratios. Briefly, all control animals of each sex were averaged, and the depletion/control ratios were calculated as each PLX animal divided by the control animal average of the same sex.

(C) Microglial soma size was unaltered after PLX treatment.

(D) Representative 20× confocal image of microglia in V1. Outline shows how individual microglia were cropped out of the image for analysis.

(E) Representative image of a thresholded microglia with overlaid concentric circles for Sholl analysis.

(F–I) Quantification of microglial ramification as total Sholl intersections depending on treatmentand sex in V1 (F); CA1, including SO, SP, and SR (G); SLM (H); and WM (I).

n = 6–7 mice/sex/treatment (B and C), n = 3–5 mice/sex/treatment (F–I); two-way ANOVA with Bonferroni correction, *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars: 200 μm. All data are expressed as the mean ± SEM. See also Figures S1–S3.

We then determined whether differences in morphology existed between male and female microglia in our mice. Microglial ramification was comparable between males and females in V1, CA1, and the corpus callosum (Figures 1F–1I), in agreement with previous findings demonstrating limited differences in microglia morphology between male and female microglia in adulthood.²⁶ Interestingly, we observed heterogeneity in microglia morphology within CA1, between the stratum lacunosum-moleculare (SLM) and layers containing pyramidal neurons and their proximal dendrites (SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum) (Figure S1D).

After 2 weeks of PLX treatment, brains of male mice showed significantly more microglial depletion than those of female mice (approximately 80% compared to 60%, respectively) across all sampled brain regions (Figure 1B). Although microglia density was heterogeneous across the brain at baseline, there was no significant difference in depletion efficacy among assayed brain regions (Figure 1B). The ratio between microglia soma size in PLX-treated and control mice was greater than 1 across all brain areas (Figure 1C), but this soma enlargement measure did not reach statistical significance. It should be noted that the bimodal distribution seen on the graph reflected different immunostaining quality and the resultant segmentation difference between two separate experiments. Furthermore, while microglia remaining after PLX treatment in both males and females exhibited reactive morphologies with decreased ramification (Figures 1F–1I), this effect appeared to be larger in females in the V1 (Figures 1F and S6C), suggesting a heightened reactive state in a region-specific manner.

To determine whether the lower depletion efficacy in females after PLX treatment resulted from lower brain concentrations of the drug itself, we administered PLX at two different doses, the standard 290 mg/kg (1×PLX) and a higher dose of 400 mg/kg (1.4×PLX), and correlated brain PLX levels with the degree of depletion. As expected, the higher dose of PLX enhanced microglia depletion in both male and female mice (Figure S2), with brain-wide microglia in male mice being almost completely ablated. While microglial depletion in females increased at the higher dose, we did not find a strong correlation between brain levels of PLX and the degree of depletion (Figure S2). To examine whether the amount of PLX consumption correlated with depletion efficacy, we housed mice singly and tracked their PLX chow consumption for 2 weeks. We found that female and male mice consumed, on average, similar amounts of PLX (Figure S3A), which were not statistically significantly different between sexes (Figure S3B). Importantly, while PLX intake was similar between the two sexes, microglia were depleted more in males compared to females (Figures S3C–S3F). Thus, our findings suggested that the observed difference in microglia retention at both doses of PLX between the sexes might be due to an inherent sex-dependent difference in vulnerability to CSF1R inhibition.

Microglia in both sexes downregulate homeostatic and upregulate immunoreactivity signatures under CSF1R inhibition

To further investigate the characteristics of microglia that survived the PLX treatment, we performed fluorescence-activated cell sorting (FACS) and sorted microglia after 3 and 7 days of PLX treatment (Figure 2A). Microglia from age-matched non-treated mice were used as controls. CD11b⁺CD45^int expression was used to gate microglia (Figure S4A). FACS analysis validated the sex-dependent depletion efficacy by PLX that we had observed after 2 weeks of treatment using immunohistochemistry (Figure 1). While 7 days of PLX treatment significantly depleted cortical and hippocampal microglia to approximately 30% of their initial numbers in males, depletion in females was highly variable, with around 50% and 70% of microglia remaining in the hippocampus and cortex, respectively (Figures 2B, 2C, S4B, and S4C). Similar trends were observed with 3 days of PLX treatment. Compared to control microglia, PLX treatment for 3 or 7 days significantly reduced the expression of microglial homeostatic proteins P2RY12 and TMEM119 (Figures 2D–2G and S4D–S4G), while the fluorescence intensity of CD68, a lysosomal marker, was significantly increased in microglia after 7 but not 3 days of PLX treatment compared to controls (Figures 2H, 2I, S4H, and S4I), consistent with the observed change in microglia morphology to a potentially reactive phenotype. Altogether, while depletion efficacies were sex specific, the shifts in expression of homeostatic and phagocytotic markers away from control levels were comparable in male and female microglia remaining after PLX treatment.

Figure 2. Cortical microglia remaining after PLX treatment upregulate phagocytosis and inflammatory responses while downregulating homeostatic signatures.

(A) Experimental paradigm depicting duration of PLX treatment and subsequent microglia isolation and analysis.

(B and C) Cortical microglia numbers in female (B) mice were not significantly different at either time point, while those in males (C) decreased at both time points. (D–G) Homeostatic markers P2RY12 and TMEM119 were significantly lower at both time points in both female (D and F) and male (E and G) mice.

(H and I) Phagocytic marker CD68 increased after 7 days of PLX treatment in both females (H) and males (I).

(J and L) Volcano plots showing DEGs between microglia isolated from PLX-treated and control female (J) and male (L) mice after 7 days of PLX treatment.

(K and M) Gene ontology-based analyses revealed upregulation of pathways pertinent to the immune response in both female (K) and male (M) mice after PLX treatment.

n = 5 mice/sex/treatment (B–I); one-way ANOVA with Bonferroni correction, *p < 0.05, **p < 0.01, and ***p < 0.001. All data are expressed as the mean ± SEM.

n = 4 mice/sex/treatment (J–M), adjusted p cutoff <0.05. See also Figures S4 and S5 and Table S1.

We next examined the transcriptomic profiles of male and female microglia after PLX treatment to better elucidate their functional states. Bulk RNA sequencing (RNA-seq) was performed on isolated cortical microglia from control mice and from mice treated with PLX for 7 days (Figure 2A). To avoid pooling samples, we chose to carry out this analysis in the cortex, which yielded enough remaining microglia to use individual mice as biological samples. Principal-component analysis showed clear clustering of samples based on both treatment (principal component 1 [PC1]) and sex (PC2) (Figure S4J). We first investigated the changes in microglial transcriptomic landscape after CSF1R inhibition in each sex. Differential expression analysis of microglia treated with PLX for 7 days vs. control microglia identified 1,596 differentially expressed genes (DEGs) in females and 3,080 DEGs in males (Figures 2J and 2L; Table S1). Although the number of DEGs in males was almost twice that in females, a large subset of DEGs (1,037 genes) were similar between sexes, and almost all these common DEGs changed their expression in the same direction in both sexes (555 downregulated and 473 upregulated genes). Among the shared DEGs, early microglial development genes, such as Lyz2,⁶ and genes linked to disease-associated reactive microglial state, such as Cd68, Clec7a, and Lgals3,^31–33 were significantly upregulated, while mature homeostatic microglia signature genes, such as Tmem119, P2ry12, and Cx3cr1,^6,10,12,13 were downregulated (Figures 2J and 2L). Gene ontology analysis revealed an upregulation of processes associated with cytokine-mediated signaling and immune response to external stimuli and downregulation of pathways related to cell adhesion and chemotaxis in microglia after PLX treatment in both sexes (Figures 2K and 2M). Taken together, these results suggest that the microglia that remain after CSF1R inhibition display a reactive phenotype in both sexes.

Sex-dependent differences in microglial transcriptomic landscape under CSF1R inhibition

We then sought to understand the underlying mechanisms that gave rise to the sex-dependent disparity in PLX-induced microglial depletion. To this end, we focused on examining sex differences in the transcriptome between the microglia remaining in females and males after PLX treatment. Similar to a previous study using cortical microglia from young adult mice,²⁷ we observed limited differences between male and female microglia at baseline (44 DEGs, 24 upregulated and 20 downregulated genes), with the most differentially regulated genes being sex-linked genes such as Ddx3y, Eif2s3, Kdm5d, Uty, and Ddx3x (Figure 3A; Table S2). Fisher’s exact test also showed significant overlap between our DEG gene set and that reported by Guneykaya et al.²⁷ Enrichment analysis suggested that in control brains, female cortical microglia downregulated processes related to the inflammatory response compared to male microglia (Figure 3B), in agreement with previous findings looking at whole brain and hippocampal microglia.^34,35

Figure 3. Sex-dependent differences in microglial transcriptomic landscape during CSF1R inhibition.

(A) Volcano plot showing DEGs between female and male microglia in control animals.

(B) Enrichment analysis revealed upregulation of immune-related pathways in control male microglia compared to female. The small number of upregulated DEGs in females compared to males precluded an enrichment analysis.

(C) Volcano plot showing DEGs between female and male microglia after 7 days of PLX treatment.

(D) After PLX treatment, upregulated DEGs in females were enriched in pathways related to autophagy and protein homeostasis.

(E) After PLX treatment, upregulated DEGs in males were enriched in mitochondrial oxidative phosphorylation and mitotic processes.

n = 4 mice/sex/treatment; adjusted p cutoff <0.05. See also Figures S4–S6 and Tables S2 and S5.

We next compared transcriptomic profiles of remaining female vs. male microglia after PLX treatment and identified 2,158 DEGs with 1,504 upregulated and 654 downregulated genes in female compared to male microglia (Figure 3C; Table S2). This large divergence suggests that the PLX-induced phenotypic switch to a reactive morphology and expression pattern common to both male and female microglia may represent different signaling states in the two sexes. Since there were significantly fewer surviving microglia in male compared to female brains after PLX treatment at D7, it is possible that the sex differences in transcriptomic profiles were influenced by the amount of apoptosis occurring in male vs. female mice. We indeed observed a higher proportion of apoptotic microglia in male compared to female at D7 PLX (Figures S5C–S5F). However, the degrees to which apoptotic microglia increased between control and PLX treatment were similar between the two sexes, as male mice had more apoptotic events in the microglia population at baseline compared to females (Figure S5). Thus, it is unlikely that the difference in the apoptotic milieu led to differences in the transcriptomic signatures of surviving male and female microglia.

Gene ontology analyses revealed enrichment of genes associated with autophagy and proteostasis pathways in upregulated DEGs in female microglia (Figure 3D). On the other hand, signaling pathways associated with mitotic processes and mitobiogenesis appeared to be upregulated in male microglia (Figure 3E). Taken together, these findings suggest that PLX induces sex-specific metabolic changes that might affect microglial survival. This prompted us to focus on the role of mitochondria in regulating metabolic states in response to PLX in the two sexes. We performed further pathway enrichment analyses using Mito-Carta3.0³⁶ pathways and found that female microglia upregulated mitochondrial pathways associated with protein homeostasis, mitochondrial dynamics and surveillance, and mtDNA maintenance in response to PLX (Table S3). On the other hand, genes coding for various subunits of electron transport chain complexes, especially complex I (6/14 DEGs belonged to complex I subunits), were upregulated in remaining male microglia (Table S3).

Sex-specific differences in microglial Galectin-3 expression at baseline and under CSF1R inhibition

Because microglia expressing Galectin-3 have been reported to adopt a progenitor-like state and be resistant to CSF1R inhibition in the young adult brain,²² we asked whether there were any sex differences in the population of Galectin-3⁺ microglia in control animals that could confer higher resistance of female microglia to PLX-induced depletion. To answer this question, we immunolabeled brains from PLX-treated and non-treated young adult mice with Iba1 and Galectin-3 and quantified the number and morphology of Iba1⁺Galectin 3⁺ cells (Figures 4 and 6A–6D). In control conditions, very few cortical and hippocampal microglia expressed Galectin-3, and while the percentage of Galectin-3⁺ microglia tended to be higher in females, this effect was not significant due to high biological variability (Figure 4C). After PLX treatment, the proportion of the remaining microglia population that expressed Galectin-3 was dramatically increased with more modest increases in Galectin-3⁺ microglial cell density (Figures 4D and 4E; Tables S4 and S5), suggesting that this microglial population survives CSF1R inhibition and that CSF1R induces Galectin-3⁺ upregulation in a proportion of surviving microglia. Indeed, none of the Galectin-3⁺ microglia after 7 days or 2 weeks of PLX treatment expressed the proliferation marker pH3 (Figure S6A), indicating that the increase in number of Galectin-3⁺ microglia was not from proliferation of an existing Galectin-3⁺ population but the result of Galectin-3 upregulation. Furthermore, we noticed that expression of Galectin-3 in microglia was not homogeneous, with some cells expressing much higher levels compared to others (Figure S6B). Thus, we divided the Galectin-3⁺ population into low (Galectin-3^lo) and high (Galectin-3^hi) expressing and found that 2 weeks of PLX treatment increased the density and proportion of Galectin-3^hi microglia more so than Galectin-3^lo microglia (Figures 4D and 4E). After PLX treatment, Galectin-3⁺ and Galectin-3⁻ microglia in both males and females had similar morphologies, as assayed using Sholl analysis (Figures S6C and S6D), suggesting that the two populations are not morphologically distinct. Due to their very small numbers, we could not quantify the morphology of Galectin-3⁺ microglia before PLX treatment. All together, these observations align with previously published data,²² further strengthening the notion that the Galectin-3-expressing subpopulation of microglia can survive when CSF1R signaling is inhibited.

Figure 4. Sex-dependent differences in microglial Galectin-3 expression at baseline and under CSF1R inhibition.

(A and B) Representative 20× confocal images showing microglia immunolabeled with Iba1 (green) and Galectin-3 (magenta) in control (A) and PLX-treated mice (B). For ease of visualization, representative images were subjected to altered brightness/contrast that masks the different Galectin-3 fluorescence intensity.

(C) The percentage of Galectin-3⁺ microglia was higher in females compared to males in both cortex (left) and hippocampus (right), although the effect did not reach statistical significance.

(D and E) PLX treatment increased the proportion (D) and density (E) of Galectin-3⁺ microglia in both males and females, also increasing the level of expression in individual microglia.

n = 6–8 mice/sex/treatment (C), n = 3–8 mice/sex/dose (D and E), Student’s t test (C), two-way ANOVA with Bonferroni correction (D and E), p values are reported in Tables S4 and S5. Scale bar: 200 μm. All data are expressed as the mean ± SEM.

See also Figure S6.

Figure 6. Sex-dependent depletion efficacies among commonly used transgenic mouse models.

Depletion/control density ratios are shown for the whole brain (A and B), cortex (C and D), and hippocampus (E and F) in female (top) and male (bottom) mice treated with PLX for 2 weeks. CX3CR1^KO females showed an increased depletion efficacy, while TREM2^KO CX3CR1^Het of both sexes had increased PLX-induced removal of microglia. The transgenic models used were CX3CR1 haplodeficiency, CX3^Het, and knockout, CX3^KO; TREM2 knockout, TREM2^KO; and the double transgenic line TREM2^KO CX3^Het.

n = 3–7 mice/sex/genotype/treatment; one-way ANOVA with Bonferroni correction, *p < 0.05, and **p < 0.01. All data are expressed as the mean ± SEM.

While the density of Iba1⁺Galectin-3^hi cells after PLX treatment was comparable between males and females (Figure 4E), males had a significantly higher proportion of Galectin-3^hi cells in the remaining microglia population (Figure 4D), due to the higher depletion rate in males. To compensate for the sex-specific differences in depletion, we compared males treated with the standard PLX dose with females treated with 1.4× the PLX dose, equalizing depletion efficacies between the sexes (Figure S2). The 1.4× PLX dose increased the percentage of Iba1⁺Galectin-3^hi in the remaining microglia population in females to an intermediate phenotype between the 1× PLX-treated males and females (Figure 4D). Furthermore, the density of PLX-induced Iba1⁺Galectin-3^hi cells in females treated with 1.4× PLX was significantly lower than that in males treated with 1× PLX (Figure 4E), suggesting sex-dependent differences in the response of male and female microglia to CSF1R inhibition.

Galectin-3⁺ microglia are distinct from Galectin-3⁻ microglia at baseline and under CSF1R inhibition

We next performed flow cytometry to examine whether Galectin-3⁺ microglia differ from Galectin-3⁻ microglia in their survival pathways and metabolism. We constructed a flow cytometry panel consisting of Galectin-3, MitoTracker green, CSF1R, and TREM2, in addition to microglia gating markers, and examined cortical microglia from males and females after 7 days of PLX treatment compared to controls (Figures 5A and S6E). This approach validated the results of Figure 4, showing that the percentage of Galectin-3⁺ microglia increases after PLX treatment in both males and females, with males showing more pronounced changes (Figure 5B). In addition, the expression of Galectin-3 was increased in both sexes after PLX treatment, and this increase was larger in males than in females (Figure 5C).

Figure 5. Expression patterns of Galectin-3⁺ microglia at baseline and under CSF1R inhibition.

(A) Experimental paradigm depicting duration of PLX treatment and subsequent microglia gating and analysis.

(B) PLX treatment increases the proportion of microglia expressing Galectin-3 to a greater extent in males compared to females.

(C) PLX treatment increases the expression levels of Galectin-3 in Galectin-3⁺ microglia to a greater extent in males compared to females.

(D–F) PLX treatment led to increased mitochondrial mass in female and male microglia, as estimated by expression of MitoTracker green. This effect was larger in males than in females (D). In both control and PLX groups, Galectin-3⁺ microglia appear to have a higher mitochondrial mass compared to Galectin-3⁻ microglia in females (E) and males (F).

(G–I) PLX treatment led to a significant reduction in CSF1R expression in both female and male microglia. Male microglia downregulated CSF1R levels to a greater extent compared to females (G). PLX treatment reduced CSF1R expression in both Galectin-3⁺ and Galectin-3⁻ microglia in both females (H) and males (I), and Galectin 3⁺ microglia generally had higher expression of CSF1R than Galectin-3⁻ microglia.

(J–L) PLX treatment led to increased TREM2 expression in both male and female microglia (J). In both control and PLX groups, Galectin 3⁺ microglia expressed higher levels of TREM2 compared to Galectin-3⁻ microglia in females (K) and males (L).

n = 3–4 mice/sex/treatment; two-way ANOVA with Bonferroni correction, *p < 0.05, **p < 0.01, and ***p < 0.001. All data are expressed as the mean ± SEM. See also Figure S6.

Because we had found a difference in the regulation of mitochondrial pathways after PLX treatment between the two sexes (Figure 3), we examined the incorporation of MitoTracker green as a proxy for mitochondrial mass.^37,38 While microglia from both male and female mice had significantly higher mitochondrial content after PLX treatment, the magnitude of the increase was much higher in males (Figure 5D), supporting the idea that PLX induces mitogenesis in males more than in females. We also found that Galectin-3⁺ microglia in both sexes at baseline had higher mitochondrial content compared to Galectin-3⁻ cells (Figures 5E and 5F), a difference that persisted under PLX treatment (Figures 5E and 5F).

We then assayed the expression of CSF1R on male and female microglia to determine if sex-specific expression could underlie the differential responses to PLX treatment. Although male microglia were more reactive at baseline, showing an upregulation of inflammatory pathways (Figure 3B),^27,34,35 female microglia expressed slightly lower levels of CSF1R, which is considered to be a homeostatic gene, compared to male microglia (Figure 5G). However, after PLX treatment, CSF1R levels were higher in females compared to males (Figure 5G), which aligned with our transcriptomic analysis showing significant downregulation of Csf1r in male but not in female microglia (Table S1). Looking at Galectin-3⁺ vs. Galectin-3⁻ cells, we observed significantly higher CSF1R expression in the Galectin-3⁺ population in both male and female control mice (Figures 5H and 5I). After PLX treatment, CSF1R expression decreased in both Galectin-3⁺ and Galectin-3⁻ microglia in both sexes (Figures 5H and 5I), and Galectin-3⁺ microglia had higher CSF1R expression than Galectin-3⁻ microglia in males (Figure 5I), while expression normalized between the two populations in females (Figure 5H).

We next examined TREM2, as it signals via its interaction with DAP12, an adapter protein that is also essential for the downstream effects of CSF1R signaling.³⁹ Furthermore, the TREM2/β-Catenin pathway has been shown to be important for microglial survival and proliferation.^40–42 We found that after PLX treatment, TREM2 protein expression was upregulated in both female and male microglia (Figure 5J), while at the transcript level, TREM2 was upregulated in female but not in male microglia (Table S1). Upregulation of TREM2 signaling could compensate for loss of CSF1R signaling and play an important role in maintaining a subset of microglia during PLX treatment. Interestingly, we observed significantly higher levels of TREM2 in Galectin-3⁺ microglia compared to Galectin-3⁻ microglia in both control and PLX groups of both sexes and an increase in TREM2 expression in Galectin-3⁻ microglia after PLX treatment, which approached significance in males (Figures 5K and 5L). These results suggest an upregulation of multiple survival pathways in the Galectin-3⁺ microglia population at baseline, which may explain their resistance to depletion.

Loss of CX3CR1, but not TREM2, confers a higher vulnerability to PLX-induced microglia depletion specifically in female mice

Since microglia upregulated TREM2 signaling after PLX treatment, we wondered if TREM2 could act as an alternative survival pathway in the presence of CSF1R inhibitor (Figure 5). To address this question, we assayed whether PLX-mediated microglial depletion would differ in male and female TREM2^KO mice compared to wild-type mice. While we expected that the loss of TREM2 signaling would facilitate microglia removal, we found that microglial depletion in TREM2^KO mice was similar to that in wild-type C57BL/6J mice in both sexes, with trends toward a reduction in depletion efficacy (Figure 6). Since loss of TREM2 has been suggested to “lock” microglia in a homeostatic state,⁴³ we then asked whether altering other pathways known to contribute to microglial reactivity could alter PLX-mediated depletion. We used mice deficient in CX3CR1, which are known to exhibit an inflammatory premature aging phenotype, and mice haplodeficient in CX3CR1, with an intermediate phenotype between wild type and that of CX3CR1^KO.⁴⁴ While loss of one copy of CX3CR1 did not alter depletion efficacy of PLX, the loss of both copies of CX3CR1 increased depletion in females, but not males (Figure 6). Interestingly, CX3CR1^Het in combination with TREM2^KO conferred higher sensitivity to PLX-induced microglia depletion in both sexes to comparable extents, although this did not reach statistical significance in males (Figure 6).

DISCUSSION

In the current study, we explored sex-dependent differences in microglial responses to CSF1R inhibition and demonstrated that female microglia were more resistant to PLX-induced depletion compared to male microglia. Comparing microglial transcriptomic profiles across both treatment and sex, we identified shared pathways induced by PLX treatment in both sexes as well as sex differences in the transcriptome and surface protein expression of remaining microglia after PLX treatment. Interestingly, while both male and female microglia upregulated TREM2 expression in response to CSF1R inhibitor, the loss of TREM2 did not promote microglia ablation by PLX. CX3CR1 deficiency, in contrast, facilitated microglial removal by PLX in females, resulting in a comparable depletion rate compared to males. Altogether, our findings demonstrate sex-dependent differences in the microglial response to CSF1R inhibitor, suggesting sex-dependent differential survival mechanisms that warrant further investigation.

We first confirmed that in control young adult mice, males and females had comparable densities of microglia in different brain areas and similar morphologies apart from a small increase in the size of microglial somas in the hippocampus of male vs. female mice (Figure S1). Our results aligned with a recent study showing equal distribution of microglia with no differences in several scalar morphometric descriptors between sexes.²⁶ However, a previous report did demonstrate mild but significantly higher microglia density in the cortex and hippocampus in adult male compared to female mice.²⁷ Furthermore, a recent study developed and applied a topological data analysis approach to uncover sex differences in microglial structure in all examined brain regions.⁴⁵ Hence, we cannot rule out the possibility that sex-dependent disparities in microglia density exist, but further systemic characterization is required. In fact, we found that male microglia appeared to express higher total mTOR levels at baseline compared to female microglia (Figure S1C). Since mTOR signaling has been well known to regulate growth and proliferation,^28,29 higher mTOR levels in male microglia might suggest higher cell density compared to females at baseline that we did not detect with our methods or a higher rate of microglial turnover.⁸ Similarly, we did not observe any significant differences in microglia morphology at baseline between sexes (Figures 1F–1I), but we did notice the well-documented differences in degree of microglial ramification among brain regions,^26,45,46 with white matter microglia being less ramified compared to both hippocampal and cortical microglia (Figures 1F–1I). While we report limited sex differences in the transcriptomic signature of adult cortical microglia, we did identify an upregulation of processes related to the inflammatory response in male compared to female microglia (Figure 3B). These results also support previous work showing that adult male microglia exist in a slightly heightened inflammatory state compared to female microglia with larger soma size²⁷; higher expression of proteins linked to a reactive phenotype such as MHCI, MHCII, or CD68^27,35; and an upregulation of NF-κB-dependent inflammatory process.³⁴

Although it is well known that myeloid cells require CSF1R signaling for survival during development,^4,5 previous studies have suggested the presence of a CSF1R-independent microglial subpopulation in adulthood, with the existence of a small microglial population observed regardless of dose or duration of acute PLX treatment or inducible deletion of CSF1R.^{11,16–22,42} However, to the best of our knowledge, no studies have extensively categorized possible sex differences in the microglial response to CSF1R inhibition. Sex-dependent differences in microglial signatures have received increasing focus recently^23–25 after extensive evidence highlighting the crucial role of these cells in brain development and plasticity, as well as their contributions to pathogenesis of diseases with high sex-dependent prevalence and severity. Here, we demonstrated that female microglia were more resistant to PLX treatment (Figure 1B), with around a 1.5-fold larger population of remaining cells in female compared to male brains after 2 weeks of treatment. The morphology of the remaining cortical female microglia appeared to undergo larger changes that typically reflect higher reactivity compared to male microglia after PLX treatment (Figure S6C), supporting previous evidence showing that male microglia mount a smaller immune response compared to female microglia under long-lasting pathological insults and as a result of aging despite their higher inflammatory state at baseline.^35,47,48 In response to maternal immune activation, female microglia increase their phagolysosomal activity, while male microglia do not.³⁵ In addition, compared to female microglia, male microglia manifest lower inflammatory signaling with age in wild-type mice⁴⁷ and upregulate the disease-associated-microglia (DAM) transcriptional signature to a lesser extent in AD mice.⁴⁸

Gene ontology enrichment analyses with DEGs between female and male PLX-resistant microglia identified upregulation in autophagy and proteostasis pathways in female microglia (Figure 3D). Microglial autophagy is increasingly regarded as an important contributor to various neurodevelopmental⁴⁹ and neurodegenerative disorders,^50–53 with recent evidence suggesting that dysregulation of autophagy in microglia induces a senescent microglial phenotype that worsens AD pathology in a mouse model.⁵¹ On the other hand, PLX-resistant male microglia upregulated their mitobiogenesis compared to females, evidenced by both higher expression of genes encoding the electron transport chain complexes, especially complex I, and higher mitochondrial mass (Figure 5D; Table S3). Recent research has highlighted the role of complex I upregulation in sustaining inflammation by driving a reverse electron transport (RET) process that generates heightened levels of reactive oxygen species (ROS).⁵⁴ RET has been shown to be regulated by infection and inflammatory conditions, a hypoxic environment such as in ischemic stroke, and aging.^55–57 Thus, it is not unreasonable to suspect that the immune-activated phenotype of PLX-treated microglia in males leads to upregulation of complex I and mitogenesis in the direction of increasing RET and ROS production, which could lead to faster cell death compared to females. In contrast, PLX-treated female microglia, which also increased their mitochondrial mass compared to controls, upregulated pathways related to maintaining mitochondria homeostasis (Table S3), which might partially explain their higher resistance to CSF1R inhibition.

In agreement with a previous study using a similar but more specific CSF1R inhibitor (PLX5622),²² we also showed a significant increase in the density of Iba1⁺Galectin-3⁺ population after PLX treatment, which was accompanied by a marked depletion of Iba1⁺Galectin3⁻ microglia (Figures 4 and 5B). We revealed that Galectin-3⁺ microglia, in both control and PLX-treated brains, expressed higher mitochondrial content as well as elevated levels of both CSF1R and TREM2 compared to Galectin-3⁻ cells. While TREM2 signaling is not essential for maintaining the homeostatic microglia population, it is important for microgliosis in response to acute and chronic brain insults^41,58 and is upregulated in lieu of CSF1R signaling.⁴² Thus, this finding suggests that Galectin-3⁺ microglia might utilize a unique set of survival mechanisms and differentially regulate their metabolism.

Although the density of Iba1⁺Galectin-3⁺ cells after PLX treatment was comparable in male and female brains, the population of these cells may be higher in females at baseline (Figure 4D). Sex-specific Galectin-3 expression could contribute to sex differences in PLX-mediated microglial depletion, as previous studies have suggested a link between Galectin-3 expression and microglial survival after PLX.²² This link was also supported by work in AD mouse models, which showed that plaque-associated microglia were much more resistant to being depleted by PLX treatment compared to plaque-distal and wild-type microglia.^59,60 Plaque-associated microglia are characterized by high expression of the DAM signature, in which Galectin-3 and TREM2 are among the most important defining genes.^31,32 Furthermore, at baseline, female microglia expressed slightly lower levels of CSF1R (Figure 5G), suggesting they might not be as dependent on CSF1R signaling as male microglia. While 7 days of PLX treatment decreased CSF1R expression and increased TREM2 expression in both female and male microglia, female microglia expressed less Galectin-3 and TREM2 and more CSF1R compared to males (Figure 5). It is interesting to note that TREM2 has been suggested to act as a receptor for Galectin-3,⁶¹ which may lead to co-regulation of the levels of these two proteins. All together, these findings further support the notion that female and male microglia show differential responses to CSF1R inhibitor, on top of a generalized response that occurs in both sexes.

Since microglia depletion with CSF1R can be titrated,^17–19 we explored whether the sex-related differences observed could solely reflect a pharmacokinetic discrepancy between females and males. We treated both sexes with two doses of PLX, a standard dose (290 mg/kg, 1× PLX) and a 140% increased dose (400 mg/kg, 1.4× PLX) and indeed did observe that depletion efficacy in females treated with 1.4× PLX was similar to that of males treated with the standard dose, while almost all microglia were ablated in males treated with 1.4× PLX (Figure S2). However, we were unable to detect a strong correlation between the microglial depletion rate and the concentration of PLX in the brain (Figure S2). A previous study using 290–300 mg/kg PLX did report slightly lower concentrations of plasma PLX levels in females vs. males after 1 month of treatment in young adult wild-type mice.⁶² Thus, we cannot rule out the possibility that some pharmacokinetic differences can account for the disparate depletion efficacy in the female vs. male brains. Further experiments with different doses and durations of PLX treatment and larger sample size are needed for clarification. However, the fact that males lost more microglia despite similar PLX chow intake compared to females (Figure S3) suggests that there might be inherent sex-dependent vulnerabilities to CSF1R inhibition. This is further supported by our observation that even with similar microglial depletion ratios (1.4× PLX-treated females vs. 1× PLX-treated males), males and females differed in the proportions of Iba1⁺Galectin-3^hi in their remaining microglia populations (Figures 4D and 4E; Tables S1 and S2). In addition, we found that at baseline, female microglia expressed lower levels of CSF1R than males, although this effect was not statistically significant (Figure 5G), while PLX treatment induced a larger reduction in CSF1R expression in males but a similar (~2-fold) upregulation of Csf1 in male and female microglia, potentially to counteract CSF1R inhibition in an autocrine manner. Together, these results may suggest that female microglia utilize different survival signaling pathways and do not rely as much on CSF1R signaling as their male counterparts but do compensate for loss of CSF1R signaling by upregulating CSF1 itself, similar to male microglia. Future experiments looking in-depth at the sex differences in CSF1R receptor signaling will further elucidate the basal survival mechanisms engaged by female and male microglia.

It is also important to note that CX3CR1^Het mice had similar depletion efficacies compared to WT mice, which has ramifications for the many studies that are currently being done using this mouse line. However, pairing CX3CR1 haploinsufficiency with loss of TREM2 increased depletion efficacy in both sexes, suggesting that studies that pair multiple manipulations to microglia may need to evaluate the efficiency of PLX-mediated depletion and take care when comparing to data obtained using different genotypes. Intriguingly, complete CX3CR1 deficiency, which could shift microglia to a more reactive state at baseline,⁴⁴ led to higher PLX depletion efficacy in females, resulting in comparable percentages of microglial removal compared to males (Figure 6). On the other hand, while microglia upregulated TREM2 in response to CSF1R inhibition, possibly as an alternative survival signal^39–42 (Figure 5J), loss of TREM2, which is thought to lock microglia in a homeostatic state,⁴³ did not increase depletion efficacy and even led to slightly lower removal rate of microglia (Figure 6). These findings further support the notion that the basal reactivity of microglia governs microglial susceptibility to CSF1R inhibition. This may underlie the sex-dependent differences we report in microglial depletion efficacy as male microglia are more inflammatory at baseline,^27,34,35 and this may make them more vulnerable to PLX-induced depletion. To consolidate these notions, future experiments looking at PLX pharmacokinetics in these transgenic animals are needed for both sexes.

Limitations of the study

We explored the effects of only one CSF1R inhibitor, PLX3397, an older PLX compound whose off-target effects have been described⁶³ and whose contribution to sex-specific effects is unknown. Hence, studies with additional, and more specific, inhibitors should be undertaken. In addition, while we confirmed that the intake of PLX was similar in males and females, studies on its pharmacokinetics in the brain should be performed to determine whether there are sex-specific alterations in PLX entry into the brain and its metabolism. Last, our results do not address the question of whether sex-specific microglial responses to CSF1R inhibition are cell autonomous or whether they are shaped by the different environments in the adult male and female brain, a question that should be explored.

RESOURCE AVAILABILITY

Lead contact

Please direct all requests to the lead contact, Ania K. Majewska (ania_majewska@urmc.rochester.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

Raw sequencing data were deposited in the NCBI GEO database with accession ID GEO: GSE278897. All analysis scripts were deposited in GitHub and released at https://doi.org/10.5281/Zenodo.14252261. The raw imaging data and training models for image analysis will be made available by the lead contact upon reasonable request.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

All animal procedures were reviewed and approved by the University Committee on Animal Resources of the University of Rochester Medical Center and carried out according to the Institutional Animal Care and Use Committee and guidelines from the National Institute of Health (NIH). Animals were housed in a 12-hour light/12-hour dark cycle with ad libitum access to food and water. All mice were derived from and maintained on a C57BL/6J background. Transgenic strains used in this study included CX3CR1-GFP reporter mice (JAX stock no. 005582⁶⁵), CX3CR1^GFP/+ generated in-house from crossing CX3CR1-GFP with C57/Bl6, TREM2KO (JAX stock no. 027197⁶⁶, courtesy of Dr. Ukpong Eyo at University of Virginia), and TREM2KO CX3CR1^GFP/+ obtained from crossing TREM2KO with TREM2KO CX3CR1-GFP. To minimize potential age-related variability in microglial response to CSF1R inhibition, all animals used in this study were between 2–3 months of age at the start of treatment.

On the day of sacrifice, mice were deeply sedated with sodium pentobarbital overdose (Euthasol 1:10; Virbac) and perfused intracardially with 0.1M phosphate buffer saline (PBS). After perfusion, hemispheres were separated: one was immediately submerged in fixative solution (4% paraformaldehyde (PFA), pH 7.2 in PB, 4°C) to be used for immunofluorescence experiments while the cortex and hippocampus were dissected from the other hemisphere for fluorescence activated cell sorting (FACS).

METHOD DETAILS

Microglia depletion with PLX3397

Mice received a chow diet (AIN-76A-D1001, Research Diets) containing 290 mg/kg (standard dose, 1x) or 400 mg/kg PLX3397 (high dose, 1.4x) (Chemgood) ad libitum to deplete microglia for different durations (3, 7 or 14 days) as detailed in each experiment appropriately. Control chow with the same base formula without PLX3397 was given to the control group.

PLX3397 pharmacokinetics analysis

Male and female animals fed with PLX3397 at two different concentrations (290 mg/kg and 400 mg/kg) and control chow were sacrificed after 2 weeks as described above. After perfusion, hemispheres were separated, one was immediately submerged in 4% PFA for use in immunofluorescence experiments (described below), while the other was flash frozen with liquid nitrogen. All frozen samples were shipped on dry ice to Cytoscient (Berkeley, CA) for pharmacokinetic (PK) analysis.

Immunofluorescence

Half-brains were fixed overnight in 4% PFA at 4°C, dehydrated in 30% sucrose overnight, and sectioned on a freezing stage microtome into 50 μm thick coronal slices stored in cryoprotectant solution. For immunofluorescence, sections were washed extensively in PBS and blocked with 10% bovine serum albumin (BSA, A2153, Sigma) for 1 h at room temperature (RT). Sections were immunolabeled for microglia (rabbit anti-Iba1, 019–19741, 1:2000, Wako) alone, with Galectin-3 (goat anti-Galectin3, AF1197, 1:1500, R&D Systems), or a combination of pH3 (rabbit anti-pH3, 06–570, 1:500, Millipore) and Galectin-3. Sections were incubated in primary antibodies for 48 h at 4°C. The sections were washed with PBS and incubated in appropriate fluorescently labeled secondary antibodies (Alexa Fluor 488, Alexa Fluor 594, and Alexa Fluor 647; all at 1:1000; Invitrogen) for 4 h at RT, then washed, mounted and coverslipped (Prolong Diamond, ThermoFisher Scientific).

Figures 1B, 1C, and 4 contained combined data from 2 experiments using C57BL/6J animals.

Image acquisition

For all experiments, 3–4 coronal tissue sections were imaged per animal, evenly spaced between Bregma −2 mm to −3.75 mm. For microglia quantification across different brain regions, slides were imaged on an Axioplan II (Carl Zeiss) epifluorescence microscope with 10x objective (0.4 NA). Image acquisition and montage generation were performed in Slidebook software (Intelligent Imaging Innovations). Brain regions of interest (ROIs) were manually selected in ImageJ FIJI (NIH) and consisted of the cortex, hippocampal formation (containing the CA1, dentate gyrus and subiculum), thalamus, and midbrain. For quantification of Galectin-3 expressing microglia, Z-stacks of the ROIs were captured with a Stellaris 5 (Leica) confocal microscope using a 20x (0.75 NA) objective lens. Imaging parameters were kept constant across all sections for each set of immunofluorescent labels. All image analysis was performed using Ilastik⁶⁴ with custom trained models and ImageJ FIJI (NIH) with semi-automated custom macros (available at https://github.com/majewska-lab).

Microglia isolation, RNA-sequencing and western blotting

Fluorescence activated cell sorting (FACS) / Flow cytometry

The cortex and hippocampus were dissected out and homogenized in 3 mL FACS buffer (1X PBS + 0.5% BSA). Homogenates were filtered through a 70 μm cell strainer into a 15 ml tube containing 3 ml FACS buffer. The strainer was washed with an additional 3 ml of FACS buffer, and the cell suspensions were centrifuged at 400 × g for 5 min at 4°C. The supernatants were discarded, and the remaining pellets were resuspended in 40% Percoll (Cytiva) (diluted in PBS), then centrifuged at 400 x g for 30 min with no braking. After removing the supernatants, the pellets were resuspended in 90 μL FACS buffer with Fc block (2.4G2, 553141, 1:90, BioLegend) and transferred to a 96 well-plate.

For microglia isolation, after a 15 min incubation with Fc block at 4°C, the following antibodies were added in a 10 μl master mix: CD11b-FITC (M1/70, 101206, 1:400, BioLegend), CD45-APC/Cy7 (30F11, 103116, 1:400, BioLegend), CD68-PerCP/Cy5.5 (FA-11, 137010, 1:400, BioLegend), P2Ry12-APC (S16007D, 848006, 1:50, BioLegend) & TMEM119-PE (106–6, ab225496, 1:500, Abcam). The latter two cell surface molecules are considered homeostatic microglial markers.^31,32 The plate was then incubated for 30 mins at 4°C in the dark.

For flow cytometry experiment in Figure 5, after a 15 min incubation with Fc block at 4°C, the following antibodies were added in a 10 μl master mix: CD11b-PerCP/Cy5.5 (M1/70, 561114, 1:400, BD Biosciences), CD45-APC/Cy7 (30F11, 103116, 1:400, BioLegend), MitoTracker Green (M7514, 100nM, Invitrogen), Galectin-3-PE (M3/38, 12–5301-82, 1:250, Invitrogen), TREM2-APC (FAB17291A, 1:100, R&D Systems) & CD115-BV421 (AFS98, 135513, 1:200, BioLegend). The plate was then incubated for 1 h at 4°C in the dark.

For detection of apoptotic microglia in Figure S7, Annexin V-FITC Apoptosis Detection Kit (eBioscience) was used as per manufacturer’s recommendations in tandem with CD11b-PerCP/Cy5.5 (M1/70, 561114, 1:400, BD Biosciences) and CD45-BUV395 (30F11, 564279, 1:400, BD Biosciences) and P2Ry12-APC (S16007D, 848006, 1:50, BioLegend).

The samples were washed once with FACS buffer and transferred to 5 ml tubes containing 7AAD (Invitrogen) such that its final dilution was 1:80. Appropriate fluorescent-minus-one (FMO) and single-stained bead controls (Ultracomp eBeads, Invitrogen) were prepared in tandem with samples. After excluding debris, doublets, and dead cells, CD11b⁺/CD45int was used to gate for microglia on a FACSAria II (BD) for sorting, or on a BD Fortessa (BD) without sorting. To maximize amount of RNA for sequencing, microglia from the entire sample volume were sorted. All events were recorded, and data were analyzed with FCS Express 7 (DeNovo Software). Galectin-3 FMO was used to gate Galectin-3⁺ and Galectin-3⁻ cells. Data were recorded as median fluorescence intensity (FMI) ± SEM.

RNA-sequencing

Sorted cells were collected in 300 μL RLT Buffer (Qiagen) containing 2-mercaptoethanol (1μL/100μL RLT) and total RNA was isolated using the RNeasy Plus Micro Kit (Qiagen). RNA concentration was determined with the NanoDrop One spectrophotometer (NanoDrop) and RNA quality assessed with the Agilent Bioanalyzer 2100 (Agilent). 500 pg of total RNA was pre-amplified with the SMARTer Ultra Low Input kit v4 (Clontech) per manufacturer’s recommendations. The quantity and quality of the subsequent cDNA was determined using the Qubit Fluorometer 3.0 (Life Technnologies) and the Agilent Bioanalyzer 2100 (Agilent). 150pg of cDNA was used to generate Illumina compatible sequencing libraries with the Nextera XT library preparation kit (Illumina) per manufacturer’s protocols. The amplified libraries were hybridized to the Illumina flow cell and sequenced using the NovaSeq6000 sequencer (Illumina) with target depth of 20–25 million of 50 nt paired-end reads per sample. Four samples, each representing individual animal, were sequenced per sex per treatment.

Western blotting

Cortical microglia were sorted into RIPA buffer (ThermoFisher Scientific) as described above. Due to low protein yield, only 2 control males and 3 control females were included for subsequent western blotting experiments. Protein concentrations were determined using Bradford Protein Quantification reagent (B6916, Sigma) and 15ug of protein per sample were denatured in Laemmli Sample Buffer (1610747, BioRad). Collected fractions were analyzed on 8–16% gradient SDS-PAGE gels (4561105, BioRad). Antibodies used for immunoblotting were β-actin (A2066, 1:5000, Sigma), S6 (2217, 1:1000, CST), and mTOR (2983, 1:1000, CST). Quantification was done using FIJI ImageJ (NIH) software.

QUANTIFICATION AND STATISTICAL ANALYSIS

Image analysis

For automated detection of microglia expressing either Iba1 or Galectin-3, the image classification and segmentation software, Ilastik⁶⁴, was used. Three labels were used to train for pixel classification including microglial soma, processes, and background. Approximately 5–10% of all images in an experimental set were manually annotated for approximately 25–50% area of the image. After sufficient training, appropriate thresholding and size exclusion criteria were applied for batch object classification of microglial soma. Due to limited resolution of the epifluorescence microscope, microglial processes were only demarcated for training to provide contrast for better soma detection. To separate low vs. high Galectin-3 expression, an arbitrary intensity cut-off for high Galectin-3 based on approximately 3 times brighter than low Galectin-3 expression was used on 8-bit images. For Iba1⁺ microglia density and soma size, depletion/control ratios were reported for ease of visualization and interpretation. For each sex, all controls were averaged, and the depletion/control ratios were calculated as PLX-treated animals divided by the control animal average of the same sex.

For assessment of microglia morphology in confocal images, Sholl analysis was performed on at least 10 microglia per ROI per animal to quantify degree of ramification. Microglia were selected from z-maximum intensity projections, thresholded, binarized and analyzed with an automated ImageJ Sholl Analysis plug-in (available at https://github.com/majewska-lab). The maximum and total number of intersections were used for statistical analyses.

Bioinformatics analysis

Raw reads generated from the Illumina basecalls were demultiplexed using bcl2fastq v2.19.1. Quality filtering and adapter removal was performed using FastP v.0.23.1. Processed reads were then mapped to the human reference genome (GRCm39 + gencode-M31 annotation) using STAR_2.7.9a. Read counts were quantified using both Subread-featureCounts v2.0.1 with “-s 0” indicating unstranded reads, and Salmon v1.5.2. All downstream analysis was carried out within R v4.2.2. The tximport package was used to import and summarize Salmon transcript-level abundance, estimated counts and transcript lengths into gene-level counts and offset matrices for use with downstream gene-level analysis packages. Differential expression analysis was performed using DESeq2. Gene ontology analyses were performed using the clusterProfiler package. Gene list enrichment analyses for mitochondrial pathways were carried out with the GeneOverlap package using the MitoCarta3.0³⁶ database. For all analyses, p_adj < 0.05 was use as a cut-off for statistical significance.

Statistical analysis

Data organization and summaries were carried out in RStudio v4.2.2. All statistical analyses and graphing were performed in GraphPad Prism v10. Comparisons between PLX and control-treated groups were made using Student’s t-test, one-way or two-way ANOVA when suitable with appropriate post-hoc correction, as detailed appropriately in figure legends. Linear regression was used to determine correlation between PLX concentration in the brain and depletion efficacy. Bioinformatics analysis was described above. Alpha was set to 0.05. All data points that represent individual animal averages are presented as mean ± SEM.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
rabbit anti-Iba1	Wako	Cat# 019-19741; RRID:AB_839504
goat anti-Galectin3	R&D Systems	Cat# AF1197; RRID:AB_2234687
rabbit anti-pH3	Millipore	Cat# 06-570; RRID:AB_310177
Donkey anti-rabbit 594nm	Invitrogen	Cat# A21207; RRID:AB_141637
Donkey anti-rabbit 488nm	Invitrogen	Cat# A21206; RRID:AB_2535792
Donkey anti-goat 647nm	Invitrogen	Cat# A21447; RRID:AB_2535864
CD11b- FITC	BioLegend	Clone M1/70, Cat# 101206; RRID:AB_312789
CD45-APC/Cy7	BioLegend	Clone 30F11, Cat# 103116; RRID:AB_312981
P2Ry12-APC	BioLegend	Clone S16007D, Cat# 848006; RRID:AB_2721469
TMEM119-PE	Abcam	Clone 106-6, Cat# ab225496
CD68- PerCP/Cy5.5	BioLegend	Clone FA-11, Cat# 137010; RRID:AB_2260046
CD11b- PerCP/Cy5.5	BD Biosciences	Clone M1/70, Cat# 561114; RRID:AB_2033995
MitoTracker Green	Invitrogen	M7514
Galectin-3-PE	Invitrogen	Clone M3/38, Cat# 12-5301-82; RRID:AB_842792
TREM2-APC	R&D Systems	Cat# FAB17291A; RRID:AB_884527
CD115- BV421	BioLegend	Clone AFS98, Cat# 135513; RRID:AB_2562667
Annexin V-FITC Apoptosis Detection Kit	eBioscience	Cat# 88-8005-72; RRID:AB_2575162
CD45- BUV395	BD Biosciences	Clone 30F11, Cat# 564279
7AAD	Invitrogen	Cat# 50-112-8859
β-actin	Sigma	Cat# A2066
S6	CST	Cat# 2217
mTOR	CST	Cat# 2983
BSA	Sigma	Cat# A2153
Percoll	Cytiva	Cat# 17-0891-01
Fc block	BioLegend	Clone 2.4G2, Cat# 553141
Bradford Protein Quantification reagent	Sigma	Cat# B6916
Laemmli Sample Buffer	BioRad	Cat# 1610747
SDS-PAGE gels	BioRad	Cat# 4561105
Deposited data
RNA-seq of microglia from male and female mice treated with PLX or control chow	This paper	GSE278897
Experimental models: Organisms/strains
CX3CR1-GFP	The Jackson Laboratory	Cat# 005582
TREM2KO	The Jackson Laboratory	Cat# 027197
Software and algorithms
Fiji v2.16	NIH	https://imagej.net/software/fiji/
FCS Express 7	De Novo Software	https://denovosoftware.com/
Ilastik	Berg et al.64	https://www.ilastik.org/download
R v4.2.2	R Core Team, 2019	http://www.r-project.org/
RStudio 2023.03.0 Build 386	R Core Team, 2019	https://cran.rstudio.com/
GraphPad Prism v10	GraphPad	https://www.graphpad.com/
RNA-seq analysis with existing packages	This paper	https://doi.org/10.5281/zenodo.14252261
MATLAB R2020a	MathWorks	https://www.mathworks.com/products/new_products/release2020a.html
Other
PLX3397	Chemgood	Cat# C-1271
BD FACS Aria II	BD Biosciences	https://www.bdbiosciences.com/en-us
BD LSRFortessa	BD Biosciences	https://www.bdbiosciences.com/en-us
RNeasy Plus Micro Kit	Qiagen	Cat# 74034
NanoDrop One spectrophotometer	NanoDrop	Cat# ND-ONE-W
Agilent Bioanalyzer 2100	Agilent	https://www.agilent.com/en/product/automated-electrophoresis/bioanalyzer-systems/bioanalyzer-instrument/2100-bioanalyzer-instrument-228250
SMARTer Ultra Low Input kit v4	Clontech	Cat# 634888
Qubit Fluorometer 3.0	Life Technnologies	Cat# 15387293
Nextera XT library preparation kit	Illumina	https://www.illumina.com/products/by-type/sequencing-kits/library-prep-kits/nextera-xt-dna.html
NovaSeq6000 sequencer	Illumina	https://www.illumina.com/systems/sequencing-platforms/novaseq.html

Highlights.

• Male mice exhibit greater microglial depletion than females after PLX3397 treatment

• Female and male microglia upregulate different signaling pathways during depletion

• Transgenic mouse models show sex-dependent differences in microglial depletion efficacy