Infiltrating blood-derived macrophages are vital cells playing an anti-inflammatory role in recovery from spinal cord injury in mice

Abstract

Background: Although macrophages (MPhi) are known as essential players in wound healing, their contribution to recovery from spinal cord injury (SCI) is a subject of debate. The difficulties in distinguishing between different MPhi subpopulations at the lesion site have further contributed to the controversy and led to the common view of MPhi as functionally homogenous. Given the massive accumulation in the injured spinal cord of activated resident microglia, which are the native immune occupants of the central nervous system (CNS), the recruitment of additional infiltrating monocytes from the peripheral blood seems puzzling. A key question that remains is whether the infiltrating monocyte-derived MPhi contribute to repair, or represent an unavoidable detrimental response. The hypothesis of the current study is that a specific population of infiltrating monocyte-derived MPhi is functionally distinct from the inflammatory resident microglia and is essential for recovery from SCI.

Methods and findings: We inflicted SCI in adult mice, and tested the effect of infiltrating monocyte-derived MPhi on the recovery process. Adoptive transfer experiments and bone marrow chimeras were used to functionally distinguish between the resident microglia and the infiltrating monocyte-derived MPhi. We followed the infiltration of the monocyte-derived MPhi to the injured site and characterized their spatial distribution and phenotype. Increasing the naïve monocyte pool by either adoptive transfer or CNS-specific vaccination resulted in a higher number of spontaneously recruited cells and improved recovery. Selective ablation of infiltrating monocyte-derived MPhi following SCI while sparing the resident microglia, using either antibody-mediated depletion or conditional ablation by diphtheria toxin, impaired recovery. Reconstitution of the peripheral blood with monocytes resistant to ablation restored the lost motor functions. Importantly, the infiltrating monocyte-derived MPhi displayed a local anti-inflammatory beneficial role, which was critically dependent upon their expression of interleukin 10.

Conclusions: The results of this study attribute a novel anti-inflammatory role to a unique subset of infiltrating monocyte-derived MPhi in SCI recovery, which cannot be provided by the activated resident microglia. According to our results, limited recovery following SCI can be attributed in part to the inadequate, untimely, spontaneous recruitment of monocytes. This process is amenable to boosting either by active vaccination with a myelin-derived altered peptide ligand, which indicates involvement of adaptive immunity in monocyte recruitment, or by augmenting the naïve monocyte pool in the peripheral blood. Thus, our study sheds new light on the long-held debate regarding the contribution of MPhi to recovery from CNS injuries, and has potentially far-reaching therapeutic implications.

Figure 1. Monocyte-derived macrophages, spontaneously recruited to the injured spinal cord following the injury, promote functional recovery.

Wild-type mice were subjected to SCI and received passive transfer (injected intravenously) of monocytes (CD45.1⁺ or Cx₃cr1 ^GFP) during the first week of recovery. (A–C) Spinal cords were excised 7 d after the injury and analyzed for the presence of infiltrating monocyte-derived MΦ. Flow cytometric analysis of (A) lesion area (4 mm segment) of injured spinal cord from mice treated with and without (w/o) adoptive transfer of monocytes (CD45.1⁺/CD11b⁺), indicating the arrival of graft-derived MΦ to the lesion area. (B) Flow cytometric analysis of lesion and distal areas (4 mm segment each) from injured spinal cords of adoptively transferred mice indicating the accumulation of the graft-derived MΦ (CD45.1⁺/CD11b⁺) mainly at the lesion and not at the distal areas (2,259±431 engrafted cells per gram of tissue taken from lesion area [mean±SE]). (C) Immunohistochemical analysis showing the adoptively transferred cells (Cx₃cr1 ^GFP/+; green) restricted to the margins of the lesion site, delineated by GFAP expression (red, right frame) (scale bar = 100 µm). (D) Similarly treated animals were followed for locomotor activity assessed according to the BMS (repeated measures ANOVA; F[between groups]_1,18 = 16.7; p = 0.0007). y-Axis error bar represents SE. (E) Mean locomotor score (BMS) of individual mice on d28 after spinal cord injury (Student's t-test; t = −5.09; df = 15; p = 0.0001), suggesting that increasing the pool of naïve monocytes by IV injection of wt mice following SCI enhanced recovery beyond spontaneous levels. The assessment of the functional outcome presented here is from one experiment representative of three independent experiments performed.

Figure 2. Monocyte-derived macrophages acquire a unique phenotype in close proximity to the lesion site.

Chimeric mice were subjected to SCI and analyzed a week later for homing of cells. (A) Cells labeled for IB-4 (red), and GFP (green) at the lesion site of [Cx₃cr1 ^GFP/+>wt] BM chimeric mice. (B) Cells labeled for GFAP (red) and GFP (green), demonstrating that the infiltrating myeloid cells barely penetrate the lesion epicenter. (C) Representative flow cytometric analysis showing the extent of expression of various markers by the infiltrating myeloid cells (CD11b⁺/ Cx₃cr1 ^GFP/+/CD45.1⁺) in the injured spinal cords of [Cx₃cr1 ^GFP/+ (CD45.1)>wt (CD45.2)] BM chimeras. The numbers above the bars refer to the percentage of the cells positive for the indicated marker out of the R1×R2 population (representing the infiltrating monocytes). The bars point to cells positive for the indicated marker (isotype control, gray line). (D) Spatial distribution map of monocyte-derived MΦ (GFP⁺) at specific distances relative to the epicenter of the lesion site in [Cx₃cr1 ^GFP/+>wt] BM chimeras based on immunohistochemical analysis. (E) High magnification of GFP⁺ cells (green) from distal and marginal areas of the lesion, demonstrating morphological differences. (F) Representative confocal micrograph of longitudinal sections from injured spinal cord of [Cx₃cr1 ^GFP/+>wt] BM chimeras, labeled for Ly6C (red), and infiltrating monocyte-derived MΦ by GFP (green). Lower panel: z-axis projection of a single cell. (G) Representative confocal micrograph of longitudinal sections of injured spinal cord, labeled for monocyte-derived MΦ by GFP (green) and CD11c (red). Lower panel: z-axis projection of single cell. (H) Flow cytometric analysis of distal and lesion spinal cord samples for CD11c expression on MΦ (CD11b+ cells) in [Cx₃cr1 ^GFP/+(CD45.1)>wt(CD45.2)] BM chimeras. Note the higher incidence of CD11c^high cells in the lesion sample. The histogram to the right is gated on CD11b⁺/CD11c^high cells at the lesion area, showing that both resident (CD45.1⁻) and infiltrating (CD45.1⁺) cells express CD11c. The dashed line demarcates the lesion site in (A), (B), (F), and (G), as determined by GFAP immunoreactivity. Scale bar in (B) and (G), represents 250 µm; in (A) and (F), 100 µm; in (E), and lower panels of (F) and (G), 10 µm. Five mice were analyzed in each group.

Figure 3. Infiltrating monocyte-derived macrophages are involved in the spontaneous process of functional recovery from spinal cord injury.

[CD11c-DTR: Cx₃cr1 ^GFP/+>wt] BM chimeras were generated by reconstitution of irradiated C57BL/6J mice with CD11c-DTR: Cx₃cr1 ^GFP/+ BM. All mice were subjected to SCI, and half of them received IP injections of DTx starting immediately after the injury, then every other day. (A) GFP staining of the injury site to detect the presence of monocyte-derived MΦ, without (w/o; left panel) or with (right panel) DTx ablation (scale bar = 250 µm). (B) Quantitative analysis of monocyte-derived MΦ (number of GFP⁺ cells) at the injured site. Treatment with DTx significantly reduced the number of monocyte-derived MΦ (Student's t-test; t = −7.39; df = 5; p = 0.0007). (C) Representative micrograph of injured spinal cord sections, labeled for monocyte-derived MΦ by GFP (green) and for CD11c (red). The sections were taken from DTx-treated animals, showing that the remaining GFP⁺ cells following ablation are CD11c negative, whereas the remaining CD11c⁺ cells are GFP⁻, representing the resident microglia (scale bar = 100 µm). (D) Hind-limb locomotor performance was assessed according to the BMS (repeated measures ANOVA; F[between groups]_1,15 = 4.88; p = 0.04). (E) Mean locomotor score (BMS) of individual mice on d28 after SCI (Student's t-test; t = −2.9; df = 5; p = 0.01). (F) Staining for myelin by Luxol (blue), and nuclei by Nissl (pink) both in the absence (w/o; upper panel) or in the presence (lower panel) of DTx (scale bar = 250 µm). (G) Quantitative analysis of the size of the injury site as a function of treatment with DTx, determined by Luxol and Nissl staining. Ablation of infiltrating CD11c⁺ monocyte-derived MΦ resulted in increased lesion size following SCI (Student's t-test; t = 3.6; df = 3.82; p = 0.02). y-Axis error bar represents SE. The assessment of the functional outcome is from one experiment, representative of two independent experiments performed.

Figure 4. Vaccination results in enhanced infiltrat? ion of monocyte-derived macrophages to the injured spinal cord.

[Cx₃cr1 ^GFP/+>wt] BM chimeric mice were generated by reconstitution of irradiated C57BL/6J mice with bone marrow cells from Cx₃cr1 ^GFP/+-transgenic mice. Half of these chimeras were vaccinated with myelin-derived altered peptide ligand (45D) 7 d before the SCI. Spinal cord sections were prepared from mice killed 1, 4, 7, 14, and 21 d after injury. (A) Quantitative analysis of cells labeled with IB-4, to identify both resident microglia and monocyte-derived MΦ, at different time points after the contusion, in either vaccinated (gray bar) or unvaccinated (black bar) mice (ANOVA; F_9,20 = 16.7, p = 0.0001). (B) Kinetics of monocyte-derived MΦ infiltration to the injured site, measured by the number of GFP⁺ cells in both vaccinated (gray bar) and unvaccinated (black bar) mice. Vaccination significantly enhanced the monocyte-derived MΦ infiltration to the injured site (ANOVA; F_9,37 = 52.12; p = 0.0001). (C) Sections from [Cx₃cr1 ^GFP/+>wt] BM chimeras labeled for GFP (representing infiltrating monocyte-derived MΦ; green) in unvaccinated (upper panel) and vaccinated mice (lower panel) (scale bar 250 µm). (D) Quantitative analysis of GFP⁺/CD11c⁺ cells at the injury site revealing increased numbers of infiltrating CD11c⁺ monocyte-derived MΦ following vaccination (Student's t-test; p<0.001). In the kinetic analysis, asterisks indicate significant differences between vaccinated and unvaccinated groups. Significant differences were also found between different time points. y-Axis error bar represents SE. The dashed line in (C) that demarcates the lesion site was determined according to GFAP immunoreactivity.

Figure 5. Conditional ablation of monocyte-derived macrophages at the injured site or antibody-mediated monocyte-depletion in the peripheral blood abolishes the augmented recovery in a paradigm of vaccination.

(A–E) [CD11c-DTR: Cx₃cr1 ^GFP/+>wt] BM chimeras were vaccinated 7 d before SCI and divided into two groups; one group was left untreated and one group was injected IP with DTx every other day starting immediately after the injury. (A) Flow cytometric analysis of lesion sites in injured DTx-treated and non-treated [CD11c-DTR>wt] BM chimeras. Note specific ablation of CD11c⁺ infiltrating monocyte-derived MΦ in DTx-treated mice, but persistence of CD11c⁺ resident microglia (GFP⁻). The cells in the dot plots to the right are gated according to CD11b expression, as indicated. (B) Hind-limb locomotor performance was assessed according to the BMS (repeated measures ANOVA; F[between groups]_1,23 = 144; p = 0.0001). (C) Mean locomotor score (BMS) of individual mice on d28 after spinal cord injury (Student's t-test; t = 8.6; df = 12.21; p<0.0001). (D) Staining for myelin by Luxol (blue), and nuclei by Nissl (pink) in the absence (w/o; upper panel) or in the presence (lower panel) of DTx (scale bar = 250 µm). (E) Quantitative analysis of the size of the injury site as a function of treatment with DTx, determined by Luxol and Nissl staining. Ablation of infiltrating CD11c⁺ monocyte-derived MΦ resulted in significantly increased lesion size following SCI (Student's t-test; t = −6.5; df = 8; p = 0.0002). (F–I) Depletion of Ly6C(Gr1⁺)CCR2⁺ monocytes in the blood, using MC-21 antibody (F), resulted in decreased recruitment of infiltrating monocytes to the injured cord (G), worsening of recovery (H; repeated measures ANOVA; F[between groups]_1,11 = 73.623; p = 0.0001), and larger lesion size compared to control group (I; Student's t-test; t = −3.77; df = 4.2; p = 0.017). Asterisks denote statistically significant differences between the indicated groups in (C), (E), and (I), and compared to the relevant controls in (B) and (H). y-Axis error bar represents SE. The assessment of the functional outcome presented is from one experiment representative of four independent experiments performed.

Figure 6. Replenishment of the monocyte pool restores functional recovery from spinal cord injury.

Vaccinated [CD11c-DTR>wt] BM chimeras were subjected to SCI and subsequently divided into three groups; one group remained untreated and served as a control. The two others were treated with DTx during the first week after injury with or without replenishment of their monocyte pool with Cx₃cr1 ^GFP/+ monocytes (CD45.1). (A) Flow cytometric analysis of the naïve monocyte graft, isolated from the BM, prior to its injection to the injured mice. Note that the vast majority of cells represent Cx₃cr1 ^GFP/+/CD115⁺ monocytes. (B) Flow cytometric analysis of spinal cord samples to identify descendents of grafted monocytes (CD45.1⁺). Note specific homing of transferred monocytes to the lesion area. The dot plot to the right is gated on CD11b⁺ MΦ indicating CD11c expression on both grafted and endogenous cells. (C) Immunohistological analysis of engrafted cells (GFP⁺; green), demonstrating their preferential localization to the margins of the lesion site demarcated by GFAP (red). (D) Morphological differences between marginal (right) and distal (left) engrafted cells (GFP⁺; green). (E) Mean locomotor score (BMS) for each group as a function of time postinjury. Adoptive transfer of monocytes resistant to ablation administered in parallel to DTx treatment during the first week restored recovery (repeated measures ANOVA; F[between groups]_2,29 = 15.97; p = 0.0001). (F) Staining for myelin by Luxol (blue), and nuclei by Nissl (pink) of sections taken from DTx-treated mice with or without adoptive transfer of monocytes. (G) Quantitative analysis of the size of the injury site as a function of monocyte transfer, determined by Luxol and Nissl staining. Adoptive transfer of monocytes significantly reduced the size of the lesion site compared to the group that was DTx-ablated and did not receive monocytes (ANOVA; F_2,9 = 42.17, p = 0.0001). Asterisks indicate significant differences compared to the relevant controls in (E) and between the indicated groups in (G). y-Axis error bar represents SE. Scale bar representation: (C) 100 µm; (D) 10 µm; and (F) 200 µm. The dashed line that demarcates the lesion site was determined according to GFAP immunoreactivity. The assessment of the functional outcome presented is from one experiment representative of two independent experiments performed.

Figure 7. Infiltrating monocyte-derived macrophages exhibit an immunoregulatory phenotype.

[CD11c-DTR: Cx₃cr1 ^GFP/+>wt] BM chimeras were treated and analyzed as follows. (A) Labeling of spinal cord sections with IB-4 (red) to detect the presence of local inflammation, in DTx untreated (w/o), and treated mice. (B, C) Quantitation of IB-4⁺ cells representing local inflammation, in mice treated with DTx along the entire recovery process, only during the second week, or untreated (B; ANOVA; F_2,11 = 35.69; p = 0.0001); or with and without treatment with MC-21 (C; Student's t-test; t = −4.9; df = 8; p = 0.001). (D) Quantification of the number of IB-4⁺ cells in injured mice, with and without adoptive transfer of DTx-resistant monocytes given in parallel with the DTx treatment (ANOVA; F_2,8 = 6.1; p = 0.025). (E) Arginase I–expressing cells (ARG; red) are restricted to the margins of the lesion site demarcated by GFAP (green). (F) In [Cx₃cr1 ^GFP/+>wt] BM chimeras, infiltrating monocyte-derived MΦ (GFP⁺; green) expressed arginase I (ARG; red), arrows indicate double-labeled cells. (G) Interleukin 10 (IL-10; red) expression is confined to the margins of the lesion site demarcated by GFAP (green). (H) In [Cx₃cr1 ^GFP/+>wt] BM chimeras, monocyte-derived MΦ (GFP⁺; green) expressed IL-10 (red) only in close proximity to the lesion site. Right panel represents z-axis projection of individual cell. (I) In [CD11c-GFP>wt] BM chimeras, monocyte-derived MΦ that expressed CD11c (GFP⁺; green) coexpressed IL-10 (red). Right panel represents z-axis projection of individual cell. (J) Graft-derived MΦ (GFP⁺) expressing interleukin 10 (IL-10; red) at the margins of the lesion site following adoptive transfer to DTx-treated mice. Asterisks denote statistically significant differences between the indicated groups. y-Axis error bar represents SE. Scale bar in (A) represents 250 µm; in (E), (G), (H), and (I), 100 µm; in (J), 50 µm; and in (F) and right panels of (H) and (I), 10 µm. The dashed line that demarcates the lesion site was determined according to GFAP immunoreactivity.

Figure 8. The contribution of monocyte-derived macrophages to the healing process is crucially dependent upon their expression of IL-10.

[CD11c-DTR (CD45.1)>CD45.1] BM mice were vaccinated with 45D, a week later subjected to SCI, and then divided into four groups. One group remained untreated and served as a control. The three other groups were treated during the first week, with DTx and with or without administration of DTx-resistant monocytes isolated from either wt or IL-10–deficient mice. (A) Flow cytometric analysis of engrafted wt or IL-10– deficient (IL-10D) monocytes (CD45.2) found at the lesion area of vaccinated [CD11c-DTR (CD45.1)>CD45.1] BM chimeras treated with DTx. The cells in the dot plots to the right were gated according to CD45.2 expression, as indicated. MΦ derived from wt and IL-10–deficient monocytes were located at the lesion area, and expressed CD11c to a similar extent (284.4±45.9 versus 301.4±10.8 CD11b⁺/CD45.2⁺ [mean±SE] engrafted IL-10 deficient and wt MΦ, respectively; Student's t-test; t = 0.36; df = 4.4; p = 0.73). (B) Mean locomotor score (BMS) for each group as a function of time post injury (repeated measures ANOVA; F[between groups]_3,33 = 23.4; p = 0.0001). (C) Quantitative analysis of the size of the injury site, determined according to Luxol-Nissl staining, as a function of the IL-10 expression by the transferred monocytes (wt; IL-10D) (ANOVA; F_3,29 = 8.32, p = 0.0004). (D) Quantification of the number of IB-4⁺ cells in the lesion area of the different groups (ANOVA; F_3,24 = 21.05; p = 0.0001). Asterisks denote statistically significant differences between the indicated groups in (C) and (D), and compared to the relevant controls in (B). y-Axis error bar represents SE. The assessment of the functional outcome presented is from one experiment representative of two independent experiments performed.