MicroRNA signature of central nervous system-infiltrating dendritic cells in an animal model of multiple sclerosis

Abstract

Innate immune cells are integral to the pathogenesis of several diseases of the central nervous system (CNS), including multiple sclerosis (MS). Dendritic cells (DCs) are potent CD11c⁺ antigen-presenting cells that are critical regulators of adaptive immune responses, particularly in autoimmune diseases such as MS. The regulation of DC function in both the periphery and CNS compartment has not been fully elucidated. One limitation to studying the role of CD11c⁺ DCs in the CNS is that microglia can upregulate CD11c during inflammation, making it challenging to distinguish bone marrow-derived DCs (BMDCs) from microglia. Selective expression of microRNAs (miRNAs) has been shown to distinguish populations of innate cells and regulate their function within the CNS during neuro-inflammation. Using the experimental autoimmune encephalomyelitis (EAE) murine model of MS, we characterized the expression of miRNAs in CD11c⁺ cells using a non-biased murine array. Several miRNAs, including miR-31, were enriched in CD11c⁺ cells within the CNS during EAE, but not LysM⁺ microglia. Moreover, to distinguish CD11c⁺ DCs from microglia that upregulate CD11c, we generated bone marrow chimeras and found that miR-31 expression was specific to BMDCs. Interestingly, miR-31-binding sites were enriched in mRNAs downregulated in BMDCs that migrated into the CNS, and a subset was confirmed to be regulated by miR-31. Finally, miR-31 was elevated in DCs migrating through an in vitro blood-brain barrier. Our findings suggest miRNAs, including miR-31, may regulate entry of DCs into the CNS during EAE, and could potentially represent therapeutic targets for CNS autoimmune diseases such as MS.

Keywords: dendritic cells; microRNAs; multiple sclerosis; neuroinflammation.

Figure 1

The miRNA signature of migratory CD11c⁺ cells in experimental autoimmune encephalomyelitis (EAE) spinal cord tissue is distinct from LysM⁺ microglia. (a) Schematic of miRNA affinity and tagging purification (miRAP) approach where CD11c‐Cre mice are crossed with mice expressing LSL‐tAgo2. miRAP from spleen and spinal cord of CD11c Cre, LSL‐tAgo2 mice isolates miRNAs only from CD11c⁺ DCs. (b) miR‐31, 339‐5p, 301a and 301b levels are enriched in CD11c⁺ cells of EAE spinal cord tissue as compared to spleen; n = 6/tissue. (c, d) miRAP of CD11c⁺ cells as compared with LysM⁺ microglia in the spinal cord indicates elevated miR‐339‐5p, miR‐301a and miR‐31 is specific to CD11c⁺ cells, whereas miR‐155 is more enriched in LysM⁺ microglia; n = 6/line. Values represented as mean ± SEM. Relative expression normalized to a geometric mean of miR‐24 and miR‐191. Student's two‐tailed, unpaired t‐tests with Bonferroni correction for multiple (4) comparisons (b, c–d). Adjusted P‐values: *P ≤ 0·05, **P ≤ 0·01, ***P ≤ 0·001, ****P ≤ 0·0001.

Figure 2

miR‐31 enrichment in central nervous system (CNS) localized CD11c⁺ cells is driven by bone marrow‐derived dendritic cells (BMDCs), not resident CD11c⁺ cells during experimental autoimmune encephalomyelitis (EAE). (a) miRNA affinity and tagging purification (miRAP) of all CD11c⁺ cells or specifically CD11c⁺ BMDCs from EAE tissue confirms enrichment of miR‐31 in the spinal cord as compared with spleen, particularly in CD11c⁺ BMDCs; n = 6/condition/tissue. CD45^intCD11b^int microglia (b) and CD11b^hiCD11c^int/hi BMDCs (c) were isolated by FACS from the spinal cord (n = 4–8) of mice with EAE. (d) Expression of miR‐31 by CD11b^hiCD11c^int/hi BMDCs and CD45^intCD11b^int microglia isolated by FACS from the spinal cord of mice with EAE; n = 4–8. (e) miR‐34b‐3p is selectively depleted from CNS‐infiltrating BMDCs as compared with resident CNS CD11c⁺ cells. (f) Expression of miR‐34b‐3p by CD11c^int/hi BMDCs and CD45^intCD11b^int microglia isolated by FACS from the spinal cord of mice with EAE; n = 4–8. Values represented as mean ± SEM. Relative expression normalized to a geometric mean of miR‐24 and miR‐191 (a, e) or U6 snRNA (d, f). One‐way anova with multiple comparisons (Dunnett's) (a, e) or Student's two‐tailed, unpaired t‐tests (d, f). P‐values: *P ≤ 0·05, **P ≤ 0·01, ***P ≤ 0·001.

Figure 3

miR‐31 sites are enriched in mRNAs downregulated in central nervous system (CNS) bone marrow‐derived dendritic cells (BMDCs) compared with spleen BMDCs. (a) We found that of the 4480 mRNAs downregulated by ≥ 1·5‐fold in CNS BMDCs compared with spleen BMDCs, 101 transcripts were putative miR‐31 targets (Targetscan.org). This is a significant enrichment when compared with the mouse genome of ~23 000 genes. (b–e) Putative miR‐31 targets, Hiat1 (b), Srp54b (d) and Tspan31 (e) were confirmed using luciferase assays; the luminescence activity of the reporter bearing the corresponding 3′‐untranslated region (UTR) was downregulated in the presence of miR‐31 as compared with scrambled miRNA control. Putative miR‐31 target, Ndrg3, (c) was not regulated by miR‐31. n = 10/condition (three biological replicates with three–four technical replicates each). Values represented as mean ± SEM. Student's two‐tailed, unpaired t‐tests (b–e). P‐values: *P ≤ 0·05, **P ≤ 0·01, ***P ≤ 0·001.

Figure 4

MiR‐31 is increased in dendritic cells (DCs) that migrate through an in vitro blood–brain barrier (BBB). (a, b) miR‐31, but not miR‐155, is elevated in cultured bone marrow‐derived dendritic cells (BMDCs) that migrate through (bottom DCs) an in vitro BBB assay, n = 6–9/condition. (c) Lipopolysaccharide (LPS) stimulation of BMDCs results in increased expression of miR‐155, but not miR‐31; n = 2/condition. (d, e) LPS‐activated DCs also have elevated miR‐31, but not miR‐155, upon migrating through (bottom DCs) an in vitro BBB, n = 6/condition. Values represented as mean ± SEM. Relative expression normalized to U6 snRNA. Student's unpaired, two‐tailed t‐test (a, b, d, e). P ‐values: *P ≤ 0·05, **P ≤ 0·01.