Tlr7 drives sex differences in age- and Alzheimer's disease-related demyelination

Abstract

Alzheimer's disease (AD) and other age-related disorders associated with demyelination exhibit sex differences. In this work, we used single-nuclei transcriptomics to dissect the contributions of sex chromosomes and gonads in demyelination and AD. In a mouse model of demyelination, we identified the roles of sex chromosomes and gonads in modifying microglia and oligodendrocyte responses before and after myelin loss. In an AD-related mouse model expressing APOE4, XY sex chromosomes heightened interferon (IFN) response and tau-induced demyelination. The X-linked gene, Toll-like receptor 7 (Tlr7), regulated sex-specific IFN response to myelin. Deletion of Tlr7 dampened sex differences while protecting against demyelination. Administering TLR7 inhibitor mitigated tau-induced motor impairment and demyelination in male mice, indicating that Tlr7 plays a role in the male-biased type I Interferon IFN response in aging- and AD-related demyelination.

Fig. 1.. Sex chromosomes and gonads underlie sex-biased, non- autoimmune demyelination in aging mice.

(A) Average latency to fall during rotarod test following control or 5-week CPZ treatment of 3- to 4-month-old WT mice split by sex; n = 6 to 8 mice in each condition. (B) Average latency to fall during rotarod test following control or 5-week CPZ treatment of 11- to 12-month-old WT mice split by sex; n = 7 to 10 in each condition. (C) Representative images of MBP immunofluorescence in whole hippocampus of control and 5-week CPZ-treated WT mice, 3 to 4 months old, split by sex. (D) Quantification of MBP signal intensity. (E) Representative images of MBP immunofluorescence in whole hippocampus of control and 5-week CPZ-treated WT mice, 11 to 12 months old, split by sex. (F) Quantification of MBP signal intensity. Fluorescence normalized to area. (G) Schematic illustrating the FCG comparison paradigm highlighting sex chromosome versus gonad comparisons. [Schematic created using BioRender.com] (H) Cell ratios of oligodendrocytes within each subcluster colored by genotype and patterned by treatment. (I) Correlation between oligodendrocyte subcluster 4 (OL4) markers and DAO genes from (29). Dashed lines represent the log2FC threshold of 0.1. Red circles represent genes up-in OL4 and DAO, and blue circles represent genes down-regulated in OL4 and DAO. Pearson’s correlation test (two-sided) was used. (J) Cell ratio of 12,473 microglia within each subcluster, colored by genotype and patterned by treatment. (K) Correlation between MG3 markers and DAM genes from (30). Dashed lines represent the log2FC threshold of 0.1. Red circles represent genes up-regulated in MG3 and DAM, and blue circles represent genes down-regulated in MG3 and DAM. Pearson’s correlation test (two-sided) was used. (L) Dot plot of Lpl expression across microglia split by genotype and treatment. (M) Representative images of LPL and IBA1 immunofluorescence taken in the CA3 region of XXO and XYT CPZ-treated samples. (N) Quantification of LPL and IBA1 colocalization. Fluorescence normalized to area. (O) Volcano plot of DEGs in MG4 compared with all other microglia clusters. Dashed lines represent log2FC threshold of 0.1 and adjusted P value (adjpval) threshold of 0.05. All data are represented as mean ± SEM. For data in (A) and (B), significance was determined with a mixed model with Sidak’s post hoc multiple comparisons test. For data in (D), (F), (H), and (J), significance was determined with a two-way analysis of variance (ANOVA) with Tukey’s post hoc multiple comparisons test. In (M), significance was determined with an unpaired, nonparametric student’s t test (two-tailed). Ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. F, female; M, male.

Fig. 2.. Sex chromosome– and gonad-specific effects on microglia and oligodendrocytes during demyelination.

(A) Venn diagram showing number of microglia DEGs between females (XXO) and males (XYT) in 3- and 5-week CPZ treatments as well as overlap. (B) Heatmap of microglia DEGs between females (XXO) and males (XYT) treated with CPZ occurring in both 3- and 5-week CPZ treatments. Heat indicates average log2FC expression of a given gene between XXO CPZ-treated and XYT CPZ-treated samples. (C) Cell ratios within each microglia subcluster by genotype and treatment; n =3 mice per condition. (D) Cell ratios of MG2 in control and 5-week CPZ-treated samples between XX genotypes (XXO, XXT) and XY genotypes (XYO, XYT). The adjacent graph shows cell ratios of MG2 between genotypes with ovaries (XXO, XYO) and genotypes with testes (XXT, XYT). (E) Correlation between MG2 markers and published DAM genes from (30). Dashed lines represent the log2FC threshold of 0.1. Red circles represent genes up-regulated by MG2 and DAM, and blue circles represent genes down-regulated by MG2 and DAM. Pearson’s correlation test (two-sided) was used. (F) Venn diagram showing number of oligodendrocyte DEGs between females (XXO) and males (XYT) in 3-week CPZ-treated mice, 5-week CPZ-treated mice, and the overlap. (G) Heatmap showing log2FC values of DEGs between females (XXO) and males (XYT) overlapping between 3- and 5-week CPZ treatment in oligodendrocytes. (H) Venn diagram of number of DEGs between female and male mice (bottom) and DEG contribution by sex chromosome (green), gonad (pink), or both (purple). (I) CellChat analysis showing ligand-receptor pairs with sex-biased expression (purple) or sex-biased and gonad-driven expression (pink). (J) Chord diagram showing CDH signaling split by sex or gonad. (K) Representative images of CDH2 and OLIG2 immunofluorescence in dentate gyrus of control and 5-week CPZ-treated mice. (L) Quantification of CDH2⁺ OLIG2⁺ signal intensity; n = 3 to 6 mice per condition. All data are represented as mean ± SEM. For data in (C) and (D), significance was determined by two-way ANOVA with Tukey’s post hoc multiple comparisons test. For data in (L), significance was determined by unpaired, nonparametric student’s t test (two-tailed). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 3.. Sex chromosome– and gonad-specific effects on CPZ-induced myelin loss and neuroinflammation.

(A) Average latency to fall during rotarod test for control and 5-week CPZ-treated FCG mice, split by gonad; n = 6 to 10 mice per genotype per treatment. (B) Average latency to fall during rotarod test for control and 5-week CPZ-treated FCG mice, split by sex chromosome. (C) Representative lanes of western blot for MBP (18 and 22 kDa) and GAPDH (36 kDa) performed on frontal cortex lysate of mice treated with control or 5-week CPZ diet; n = 3 to 4 mice per genotype per treatment. (D) Quantification of percent myelin loss in 5-week CPZ-treated samples between mice with ovaries versus testes and (E) XX versus XY genotypes. (F) Quantification of estradiol amounts in plasma of FCG mice split by gonad; n = 4 to 8 mice per genotype. (G and H) Quantification of cytokine expression average log2FC in the frontal cortex between genotypes with ovaries (XXO, XYO) and genotypes with testes (XXT, XYT) for (G) 3- and (H) 5-week CPZ treatments; n = 3 to 5 mice per genotype. (I) Quantification of hippocampal cytokine expression average log2FC between genotypes with ovaries (XXO, XYO) and testes (XXT, XYT) for 3- and (J) 5-week CPZ treatments. (K) Heatmap of multiplex cytokine panel of frontal cortex lysate between CPZ-treated XYO and XYT mice and (L) XXT and XXO mice. All data are represented as mean ± SEM. For data in (A) and (B), significance was determined by mixed model with Sidak’s post hoc multiple comparisons test. In (D) to (F), significance was determined by unpaired, nonparametric student’s t test (two-tailed). Significance for cytokine expression was determined using linear regression package Limma in R (materials and methods). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 4.. Sex chromosomes and gonads affect tau-induced microglial responses and demyelination on humanized APOE4 background.

(A) Breeding scheme for generating FCG mice homozygous for human APOE4 on P301S background. [Schematic created using BioRender.com] (B) Representative images of MC-1 immunofluorescence in CA3 of P301S⁺ FCG mice homozygous for E4 (E4/E4); n = 4 to 9 mice per genotype. (C) Quantification of MC-1 signal intensity between gonadal genotypes and (D) sex chromosomal genotypes. (E) Microglia cell ratios of each subcluster by genotype; n = 3 mice per genotype. (F) Correlation between MG3 markers and published DAM genes from (30). Dashed lines represent the log2FC threshold of 0.1. Red circles represent genes up-regulated by MG3 and DAM, and blue circles represent genes down-regulated by MG3 and DAM. Pearson’s correlation test (two-sided) was used. (G) Cell ratios of MG3 in P301S^– and P301S⁺ samples between XX genotypes (XXO, XXT) and XY genotypes (XYO, XYT). (H) Cell ratios of MG3 between genotypes with ovaries (XXO, XYO) and genotypes with testes (XXT, XYT). (I) Ingenuity Pathway Analysis–predicted increased upstream regulators in XYT; E4/E4; P301S⁺ microglia compared with XXO; E4/E4; P301S⁺ microglia. (J) Western blot for STING, NF-κB p65, MBP, and GAPDH; n = 3 mice per genotype except XYO; P301S^–, n = 2 mice. (K and L) Quantification of STING (K) and NF-κB p65 (L) between XX and XY genotypes with or without P301S. (M) Quantification of percent MBP loss within each genotype. (N) Dot plot of IFN gene expression for each genotype split by sex in P301S mice with E4/E4; n = 3 mice per genotype. All data are represented as mean ± SEM. For data in (E), (G), (H), (K), (L), and (M), significance was determined by two-way ANOVA with Tukey’s post hoc multiple comparisons test. For data in (C) and (D), significance was determined by unpaired, nonparametric student’s t test (two-tailed). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Fig. 5.. Tlr7 drives male-biased type I IFN response in mouse brains and cultured microglia stimulated with myelin.

(A) Experimental schematic for treatment of 10- to 12-month-old WT mice with saline or TLR7 agonist, split by sex. [Schematic created using BioRender.com] (B) Venn diagram of significant number of DEGs between agonist versus saline-treated within females and males; n = 7 mice per sex per treatment. Numbers or genes in red are up-regulated, whereas those in blue are down-regulated. The pink-shaded region indicates the portion of genes specifically altered by agonist in males only. (C) Gene ontology pathway analysis of DEGs up-regulated by agonist in males only. (D) Schematic illustrating the treatment of WT (Tlr7^+/+) and TLR7 knockout (Tlr7^−/−) in vitro microglia with myelin, split by sex; n = 3 replicates per condition. (E) Venn diagram of DEGs between TLR7 knockout and WT microglia (MG) treated with myelin within males and females. The pink-shaded region indicates portion of genes specifically altered by agonist in males only. (F) Volcano plot of specific DEGs in TLR7 knockout male microglia versus WT male microglia, both treated with myelin. Dashed lines represent the log2FC threshold of 0.1 and P value threshold of 0.05. (G) Gene ontology pathway analysis of DEGs down-regulated by TLR7KO in males only. (H) Volcano plot of specific DEGs in TLR7 knockout female microglia versus WT female microglia, both treated with myelin. Dashed lines represent the log2FC threshold of 0.1 and P value threshold of 0.05. (I) Gene ontology pathway analysis of DEGs down-regulated by TLR7 knockout in females only.

Fig. 6.. TLR7 deficiency abolishes sex differences and protects against nonautoimmune and tau-induced demyelination.

(A) Experimental design for CPZ treatment of TLR7 knockout mice. [Schematic created using BioRender.com] (B) Violin plot showing pseudobulk Tlr7 expression in microglia. (C and E) Subcluster cell ratios of microglia (C) or oligodendrocytes (E) in XXO or XYT mice treated with control or 3 weeks of CPZ. (D and F) Subcluster cell ratios of microglia (D) or oligodendrocytes (F) from TLR7 knockout female (KO-XXO) or male (KO-XYT) mice treated with control or 3 weeks of CPZ; n = 3 mice per condition. (G) Representative images of pSTAT1 and IBA1 immunofluorescence in hippocampal CA1 of CPZ-treated FCG male, FCG female, TLR7 knockout female, and TLR7 knockout male mice. (H) Quantification of pSTAT1 and IBA1–positive signal between male and female, FCG and TLR7 knockout CPZ-treated samples. n = 6 to 9 mice per condition. (I) Representative images of MBP immunofluorescence in whole hippocampus of control and CPZ-treated FCG (Tlr7^+/+) and TLR7 knockout mice. (J) Quantification of MBP signal intensity. (K) Average latency to fall during rotarod test for WT and TLR7 knockout control and 5-week CPZ-treated female mice; n = 11 to 16 per condition. (L) Experimental schematic of TLR7 inhibitor (TLR7i) treatment of PSUK mice. [Schematic created using BioRender.com] (M) Venn diagram of DEGs between inhibitor- and control-treated hippocampi from PSUK males and females; n = 3 to 6 mice per condition. (N) Gene ontology pathway analysis of DEGs down-regulated by the inhibitor treatment specifically in males. (O) Correlation between inhibitor versus control overlapping DEGs in males and females. Dashed lines represent the log2FC threshold of 0.1. Red circles represent genes up-regulated by males, and blue circles represent genes down-regulated by males. Pearson’s correlation test (two-sided) was used. (P) Representative images of MBP immunofluorescence in dentate gyrus of inhibitor and control-treated PSUK or WT mice. (Q) Quantification of MBP intensity. (R) Hindlimb paralysis scores of PSUK mice treated with TLR7 inhibitor or control diet and WT mice on control chow split by sex. n = 5 to 8 mice per condition. All data are represented as mean ± SEM. For data in (C) to (F), (H), (J), (Q), and (R) significance was determined by two-way ANOVA with Tukey’s post hoc multiple comparisons test. For data in (K), significance was determined by mixed model with Sidak’s post hoc multiple comparisons test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.