Analysis of CNS autoimmunity in genetically diverse mice reveals unique phenotypes and mechanisms

Abstract

Multiple sclerosis (MS) is a complex disease with significant heterogeneity in disease course and progression. Genetic studies have identified numerous loci associated with MS risk, but the genetic basis of disease progression remains elusive. To address this, we leveraged the Collaborative Cross (CC), a genetically diverse mouse strain panel, and experimental autoimmune encephalomyelitis (EAE). The 32 CC strains studied captured a wide spectrum of EAE severity, trajectory, and presentation, including severe-progressive, monophasic, relapsing remitting, and axial rotary-EAE (AR-EAE), accompanied by distinct immunopathology. Sex differences in EAE severity were observed in 6 strains. Quantitative trait locus analysis revealed distinct genetic linkage patterns for different EAE phenotypes, including EAE severity and incidence of AR-EAE. Machine learning-based approaches prioritized candidate genes for loci underlying EAE severity (Abcc4 and Gpc6) and AR-EAE (Yap1 and Dync2h1). This work expands the EAE phenotypic repertoire and identifies potentially novel loci controlling unique EAE phenotypes, supporting the hypothesis that heterogeneity in MS disease course is driven by genetic variation.

Keywords: Autoimmunity; Genetic variation; Genetics; Mouse models; Multiple sclerosis.

Figure 1. MOG_35–55 induced EAE in CC strains results in heterogeneous disease profiles.

EAE was induced via 200 μg MOG_35–55 in CFA (s.c.) and 200 ng PTX (i.p.) in 8- to 14-week-old male and female B6 (18 male [M], 18 female [F]) and H2^b or H2^g7 CC mice (32 strains, ~5M, ~5F; Table 1). (A) Schematic illustrating the study design. (B) Percent EAE incidence per CC strain with B6 shown for reference control. Bar color denotes EAE subtype (classic,gay; AR, orange). (C) Percent incidence of RR (green), monophasic (light green), and chronic (gray) EAE in CC strains, with B6 shown for reference control. (D) Comparison of EAE disease severity in CC strains, as calculated by CDS, versus B6 reference controls. Significance of differences of each CC strain from B6 reference control was determined via 1-way ANOVA with Dunnett’s multiple-comparison test and indicated by asterisks where significant. Corresponding colors indicate directionality as compared with B6 (blue, less severe; red, more severe). (E–K) Distribution of strain CDS and incidence of EAE, classic-EAE, AR-EAE, chronic-EAE, RR-EAE, and monophasic-EAE, grouped by H2^b and H2^g7 homozygous haplotypes. Each data point in E–K represents a strain average. Significance of differences between haplotypes was determined by 2-tailed unpaired t test.

Figure 2. EAE in CC strains captures clinically relevant disease courses.

EAE was induced and evaluated in CC and B6 reference control mice as described in Figure 1. (A–E) Daily strain disease course profiles for strains of interest are shown, including severe-progressive EAE in CC028 (5M, 5F) (red) and EAE resistance in CC011 (5M, 5F) (blue), compared with B6 (18M, 18F) reference controls (gray) (sexes pooled) (A); AR-EAE in CC004 (7M, 7F) (orange) (sexes pooled) (B); RR-EAE in CC002 (6M, 5F) (sexes pooled) (C); secondary progressive EAE in CC043 (5M, 5F) (sexes pooled) (D); and monophasic-EAE in CC068 (4M, 5F) (sexes shown separately due to timing of disease onset) (E). All panels show classic-EAE scores, except B, which shows AR-EAE scores, as indicated on the y axes.

Figure 3. EAE in CC strains demonstrates bidirectional effects of sex on disease course.

EAE was induced and observed in CC mice as described in Figure 1. (A–C) Disease severity was assessed for effects of sex within strain using CDS (A), classic-CDS (B), and AR-CDS (C). Significance of differences between sexes was determined by 2-way ANOVA with Fisher’s LSD multiple-comparison test (Supplemental Table 3). Comparisons are indicated by asterisks where significant. (D–G) Disease course profiles of sex differences in classic-EAE in CC046 (D) and CC042 (E), and AR-EAE in CC038 (F) and CC072 (G).

Figure 4. Severe progressive EAE in CC028 mice and AR-EAE in CC004 mice is associated with distinct pathology in the spinal cord and brain, respectively.

EAE was induced and evaluated as described in Figure 1. On D50, or at humane endpoint, spinal cord and brains were collected and processed for staining with H&E with or without LFB. Histopathologic evaluation of B6 reference control (9M, 9F), CC002 (6M, 5M), CC004 (7M, 7F), and CC028 (5M, 5F) (sexes pooled) was performed as described in Methods. (A and B) Spinal cord inflammation scores by strain and corresponding representative images. (C and D) Spinal cord demyelination scores by strain and corresponding representative images. Spinal cord images (B and D) were captured at 10× objective. Scale bar: 100 μm. (E and F) Brain inflammation scores by strain and corresponding representative images. (G and H) Brain demyelination scores by strain and corresponding representative images. Brain images (F and H) were captured at 5× objective. Scale bar: 200 μm. For all images, the arrows mark regions of inflammatory infiltrates or demyelination. Significance of differences of each CC strain from B6 reference control was determined by ordinary 1-way ANOVA, with Fishers LSD multiple-comparison test (A, C, and G), or by Brown-Forsythe and Welch ANOVA, with unpaired t test with Welch’s correction for multiple comparison testing when appropriate (E). *P ≤ 0.05, **P ≤ 0.01.

Figure 5. Severe EAE in CC028 and CC004 mice is associated with unique CNS immune profiles.

EAE was induced in 8- to 14-week-old male B6 (n = 5), CC004 (n = 4), and CC028 (n = 5) mice as described in Figure 1. On D14, spinal cord and brains were collected and processed for flow cytometric staining. (A and B) Disease course profiles for B6, CC004, and CC028 mice displayed as classic-EAE or AR-EAE. (C) Representative gating scheme for flow cytometric analysis. (D–I) Scatter plots demonstrating frequencies of key immune cell subsets in the spinal cord of by strain, including CD11b⁺ cells (CD45⁺CD11b⁺) (D), microglial cells (CD45^intCD11b⁺CX3CR1⁺) (E), myeloid cells (CD45⁺CD11b⁺Cx3CR1^lo/–) (F), neutrophils (CD45⁺CD11b⁺CX3CR1^–Ly6G⁺) (G), B cells (CD45⁺CD11b^–CD19⁺) (H), and T cells (CD45⁺CD11b^–CD19^–TCRβ⁺) (I). (J–O) Frequencies of key immune cell subsets in the brain by strain, including CD11b⁺ cells (J), microglial cells (K), myeloid cells (L), neutrophils (M), B cells (N), and T cells (O). (P and Q) Frequencies of CD4⁺ T cells (CD45⁺CD11b^–CD19^–TCRβ⁺CD4⁺) producing IFN-γ (P) and IL-17 (Q) in the spinal cord by strain. (R and S) Frequencies of CD4⁺ T cells producing IFN-γ (R) and IL-17 (S) in the brain by strain. Significance of differences between each CC strain and B6 reference control was determined via 1-way ANOVA with Dunnett’s multiple-comparison test.*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 6. Peripheral immune and CNS intrinsic factors drive RR-EAE in CC002 mice.

(A) B6 and CC002 mice were subjected to BM ablation and reconstitution to create reciprocal BM chimeric mice, designated as B6→B6 (7M, 4F), B6→02 (7M, 1F), 02→B6 (6M, 4F), 02→02 (6M, 1F) and illustrated in the schematic. Mice were rested for a total of 8 weeks prior to EAE induction as described in Figure 1. Mice were observed for a total of 34 days. On D34, spleen (B6→B6: 7M, 4F; B6→02: 7M, 1F; 02→B6: 6M, 4F; 02→02: 6M, 1F), and spinal cord (n = 4 males/chimera) tissues were collected and processed for flow cytometric staining. (B–G) Percent chimerism was assessed in B6→B6, B6→02, and 02→B6 for splenic CD11b⁺ cells (CD45⁺CD11b⁺CD19^–) (B), CD19⁺ cells (CD45⁺CD11b^–CD19⁺) (C), CD4⁺ T cells (CD45⁺CD11b^–CD19^–TCRβ⁺CD4⁺) (D), and CD8⁺ T cells (CD45⁺CD11b^–CD19^–TCRβ⁺CD8⁺) (E), as well as infiltrating CD4⁺ (F) and CD8⁺ (G) T cells in the spinal cord. Bars in B–G represent average percent chimerism for donor (dark gray) and host (light gray), while corresponding colored dots demonstrate individual samples. (H) Disease course profiles for B6→B6, B6→02, 02→B6, and 02→02 displayed as classic-EAE. (I–K) Comparison of spinal cord infiltrating immune cell populations in B6→B6, B6→02, 02→B6, and 02→02 for CD11b⁺ cells (I), CD19⁺ cells (J), CD4⁺ T cells (K), and CD8⁺ T cells (L). Significance of differences between groups was determined via 1-way ANOVA with Tukey’s multiple-comparison test and indicated by brackets and asterisks where significant. P ≤ 0.05.

Figure 7. QTL analysis reveals distinct genetic linkage patterns for AR-EAE incidence and EAE severity.

EAE was induced and evaluated in CC strains, as described in Figure 1. EAE QTVs were calculated, and QTL mapping was performed as described in Methods. (A) Manhattan plot demonstrating LOD traces for AR-EAE incidence, 15% and 20% genome-wide significance is indicated by the solid and dashed lines, respectively. (B) Corresponding CC founder allele effects plot for lead QTL on Chr9 — Eaecc3. (C) Heatmap demonstrating CC strain distribution based on genotype-by-phenotype analysis for Eaecc3. (D) Manhattan plot demonstrating LOD traces for EAE severity, 15% and 20% genome-wide significance is indicated by the solid and dashed lines, respectively. (E) Corresponding CC founder allele effects plot for lead QTL on Chr14 — Eaecc6. (F) Box and whisker plot (mean ± 95% CI) demonstrating distribution of CC founder alleles within strains.

Figure 8. Machine learning–based functional candidate gene prioritization nominates distinct genes associated with QTL for AR-EAE incidence, EAE severity, and monophasic-EAE incidence.

(A) SVM classifiers were trained using MS GWAS genes and integrated with tissue-specific connectivity networks to rank gene candidates associated with Eaecc QTL in the context of either the CNS or immune system, as illustrated by the schematic. (B and C) Ranked candidate genes for AR-EAE incidence (Eaecc3) in the immune system (B) and CNS (C). Genes are plotted by genomic position on the x axis and –log(FPR) on the y axis, dotted lines demonstrating QTL boundaries. (D–G) Ranked candidate genes, graphed independent of genomic position, for EAE severity (Eaecc6) in the immune system (D) and CNS (E), and monophasic-EAE incidence (Eaecc5) in the immune system (F) and CNS (G). The solid line in B–G corresponds an FPR threshold of 0.05.