Monocytes regulate the efficiency of CNS remyelination

To study the specific contribution of monocytes to remyelination, we used a focal lesion model whereby demyelination is induced by stereotaxic injection of the myelin toxin lysophosphatidyl choline (LPC) into the brain white matter tract, the corpus callosum, of young adult mice [4,18,23]. This model is optimal to investigate remyelination specifically, as the well-characterized location and timing of regenerative responses occur without concomitant myelin damage or complicating autoimmunity. We assessed the density of monocytes in lesions at key time points when remyelination is initiated, at 10 days post-lesion (dpl), and complete, at 21 dpl (Fig 1A). This was achieved using a reporter mouse in which selective expression of the chemokine receptor CCR2 by the subtype of monocytes attracted to injured tissue (“classical”) drives RFP expression (Ccr2^RFP/+) (Fig 1B), whereas microglia are RFP negative [24,25]. We detected RFP+ cells in lesions at both time points, yet numbers significantly decreased as remyelination proceeded (Fig 1C and 1D). RFP+ cells were not detected in unlesioned contralateral corpus callosum, indicating recruitment is specific to injured tissue (Fig 1D). We confirmed the presence of monocytes at their peak abundance (10 dpl) via flow cytometry of lesions by selecting for CD11b⁺ myeloid cells, excluding Ly6G⁺ neutrophils and CD3⁺ T cells, then using differential expression levels of CD45 to distinguish from microglia – where it has been established previously in models of injured white matter and bone marrow chimeras that monocytes are CD45^hi and microglia are CD45^lo on a logarithmic scale (Fig 1E; gating strategy in S1A Fig) [24,26–30]. Monocytes were present in lesions, albeit at lower proportions relative to microglia (Fig 1F), but in higher proportions versus sham PBS-injected corpus callosum (Figs 1H and S1A). We confirmed the accuracy of our gating strategy in three ways. First, we demonstrated that the gates distinguish the two cell types clearly in a model where monocyte influx into white matter is robust, following lipopolysaccharide (LPS) injection into the striatum [31] (S1B Fig). Second, we assessed Ccr2^RFP/+ mice for co-expression of RFP and CD45 [24,25] in LPC-induced lesions and found all CD45^hi cells to be RFP⁺ (S1C Fig). Third, we detected expression of a classical monocyte marker, Ly6C, by CD11b⁺CD45^hi lesion cells (S1D and S1E Fig). We next queried the maturation of the cells in lesions. Surprisingly, assessment of macrophage markers by flow cytometry of wild-type mice (F4/80 expression in CD11b⁺CD45^hi Ly6C⁺ cells) (S1F and S1G Fig), or immunofluorescence analysis of Ccr2^RFP/+ mice (IBA1 expression in RFP⁺ cells) (S1H and S1I Fig) indicated that approximately 90% of the cells do not express typical macrophage maturation markers in remyelinating lesions. Considering the absence of expression of the monocyte marker Ly6C in a subset of the cells (S1D and S1E Fig), these findings could suggest that monocytes may be in the process of differentiating into macrophages or that the population represents a mixture of undifferentiated and differentiating cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Classical monocytes regulate central nervous system remyelination.

(A) Schematic of in vivo LPC-induced focal demyelinated lesions in corpus callosum of adult mice. (B) Schematic of use of lesioned Ccr2^RFP/+ reporter mice to track classical monocyte entry into brain lesions. (C) Mean numbers of Ccr2-RFP+ cells per lesion image at 10 and 21 days post-lesion (dpl) ± S.E.M. ***P = 0.0003; unpaired 2-tailed Student t test. n = 3 mice/group. (D) Representative immunofluorescent images of Ccr2-RFP+ cells in a corpus callosum lesion (left, outlined), *indicates needle tract. Non-lesion corpus callosum (right) stained with IBA1 (white) and counterstained with Hoechst (blue), **indicates choroid plexus monocyte-derived macrophages. Scale bar, 100 µm. (E) Flow cytometry plot demonstrating differential CD45 expression in CD11b⁺ Ly6G⁻ CD3⁻ cells, used to distinguish between lesion microglia (CD45^lo) and monocytes (CD45^hi) at 10 dpl. SSC: side scatter. Two mice pooled. (F) Mean numbers of lesion CD45^lo microglia and CD45^hi monocytes ± S.E.M. at 10 dpl; **P = 0.0080; unpaired 2-tailed Student t test. n = 6 mice/group, 2 mice pooled per replicate. (G) Schematic of use of Ccr2^−/− mice to impair the egress of classical monocytes from the bone marrow into the blood and subsequently into brain lesions. (H) Mean proportion of monocytes (CD11b+ cells which are CD45^hi) ± S.E.M. in sham PBS-injected corpus callosum, and lesions from wild type (WT) and Ccr2^−/− mice. n = 6 mice/group, 2 mice pooled per replicate. One-way analysis of variance (ANOVA) with Tukey’s multiple comparisons; *P = 0.0445 (sham vs. WT LPC), *P = 0.0356 (WT vs. Ccr2^−/− LPC), I) Mean number of monocytes (CD11b⁺ CD45^hi Ly6C^hi F4/80⁻ cells) ±  S.E.M. in sham PBS-injected corpus callosum, and lesions from WT and Ccr2^−/− mice. n = 6 mice/group, 2 mice pooled per replicate. One-way ANOVA with Tukey’s multiple comparisons; **P  = 0.0040 (WT LPC vs. Ccr2^−/− LPC), *P = 0.0439 (sham PBS vs. Ccr2^−/− LPC). (J) Representative immunofluorescent labeling of oligodendrocyte lineage cells (Olig2+, red) which are mature (CC1+, green) in WT and Ccr2^−/− mice at 10 dpl, counterstained with Hoechst. Scale bar, 20 μm. (K) Mean percentage of Olig2⁺ cells which are CC1⁺ per field ± S.E.M. in WT and Ccr2^−/− mice at 10 dpl; *P = 0.0156; unpaired 2-tailed Student t test. n = 3–4 mice/group. (L) Representative immunofluorescent labeling of myelin-associated glycoprotein (MAG, red) in WT and Ccr2^−/− mice at 10 dpl, counterstained with Hoechst. Scale bar, 20 μm. (M) Quantification of MAG intensity/mm² within lesioned corpus callosum at 10 dpl ± S.E.M. in WT and Ccr2^−/− mice, *P = 0.0135; unpaired 2-tailed Student t test. n = 4–5 mice/group. (N) Mean number of myelinated axons/mm² at 10 dpl ± S.E.M. in WT and Ccr2^−/− lesions quantified from electron micrographs. P = 0.6847; unpaired 2-tailed Student t test. n = 3–6 mice/group. (O) Quantification of MAG intensity/mm² within lesioned corpus callosum at 21 dpl ± S.E.M. in WT and Ccr2^−/− mice, *P = 0.0317; Mann–Whitney test. n = 4–5 mice/group. (P) Mean percentage of Olig2+ cells which are CC1+ per field ± S.E.M. in WT and Ccr2^−/− lesions at 21 dpl; P = 0.2879; unpaired 2-tailed Student t test. n = 5–7 mice/group. (Q) Mean number of myelinated axons in lesions at 21 dpl ± S.E.M. in WT and Ccr2^−/− lesions; **P = 0.0054; unpaired 2-tailed Student t test. n = 3–4 mice/group. (R) Average g-ratio at 21 dpl ± S.E.M. in WT and Ccr2^−/− lesions; P = 0.8466; unpaired 2-tailed Student t test. n = 3 mice/group. (S) Representative electron micrographs of lesioned corpus callosum at 21 dpl in WT and Ccr2^−/− mice. Scale bar, 2 µm. Mice were 8–12 weeks of age in each group. Source data may be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003073.g001

Having demonstrated that monocytes are present during remyelination, we then investigated how they influence this process. We assessed the requirement for classical monocytes in remyelination by lesioning transgenic mice lacking expression of CCR2 (Ccr2^−/−), in which their egress from bone marrow is impaired [25,32] (Fig 1G). Microglia do not express CCR2 under homeostasis nor following myelin injury [25], and CCR2-expressing cells do not contribute to CNS-resident macrophages in the corpus callosum [33]. We confirmed that Ccr2^−/− lesions had reduced recruitment of classical monocytes compared to wild-type lesions, with significantly reduced CD11b⁺CD45^hi cells (Fig 1H) and, of those, reduced Ly6C⁺ F4/80⁻ cells (Fig 1I). We observed no difference in lesioning between groups using the myelin protein, myelin basic protein (MBP) (S2A Fig). Next, we investigated how reduced monocyte presence in lesions impacts the regenerative processes regulated by myeloid cells during remyelination: myelin debris clearance and OPC differentiation. Myelin debris had been mostly cleared in both Ccr2^−/− and wild-type mice at the peak of demyelination, 3 dpl, as indicated by MBP and Oil Red O staining (S2A and S2B Fig). However, differentiation of OPCs into mature oligodendrocytes was significantly impaired in Ccr2^−/− lesions at the onset of remyelination (10 dpl), with a reduced proportion of oligodendrocyte lineage cells (Olig2+) which were mature (CC1+) (Fig 1J and 1K). We next asked how this affected remyelination; whereas remyelination was initiated normally at 10 dpl in wild-type mice as indicated by expression of the early remyelination marker and myelin protein myelin-associated glycoprotein (MAG), MAG expression was significantly reduced in lesions from Ccr2^−/− mice at this time (Fig 1L and 1M). To complement this, we assessed myelin by electron microscopy-based ultrastructural analysis, first confirming that unlesioned adult Ccr2^−/− mice have unaffected myelinated axon density and myelin thickness (“g-ratio”; S2C-S2E Fig). Remyelinating lesions surprisingly had no significant difference in the density of myelinated axons between genotypes at 10 dpl (Fig 1N), despite the decrease in mature oligodendrocytes and myelin protein expression in Ccr2^−/− lesions at this time; we confirmed lesions were formed in these mice via toluidine blue-stained semi-thin sections (S2F Fig). These findings indicate that when classical monocytes are reduced in remyelinating lesions, oligodendrocyte differentiation is impaired yet more myelin is generated, suggesting that oligodendrocytes may compensate by producing more myelin. Overall, this points to monocytes restricting myelin production during remyelination. To further test this postulate, we investigated lesions when myelin production is normally complete in wild-type controls, at 21 dpl. At this time, MAG remained downregulated in Ccr2^−/− lesions versus wild-type control (Fig 1O). However, the proportion of lesion oligodendrocyte lineage cells which had matured in Ccr2^−/− mice was now equivalent to that in wild-type controls (Fig 1P). By electron microscopy, we found that Ccr2^−/− mice had ~25% increase in number of myelinated axons compared to wild-type controls at 21 dpl (Fig 1Q and 1S). We assessed the structural integrity of this remyelination and found no significant differences from wild-type control in myelin thickness (g-ratio; Figs 1R and S2G), nor in frequency of structural abnormalities (S2H and S1I Figs). These findings reveal that, although classical monocytes are required for the timely onset of oligodendrocyte differentiation and myelin protein expression during remyelination, they limit myelin production. The latter suggests that monocytes reduce the efficiency of remyelination itself.

Monocytes have a Wnt signature during remyelination

Having identified that monocytes regulate remyelination efficiency, we next investigated the molecular mechanisms influencing this function by transcriptomic analysis of lesion monocytes in comparison with microglia. Given the low number of lesion monocytes, we achieved this by performing ultra-low input bulk ribonucleic acid (RNA) sequencing of these cells subsequent to fluorescence-activated cell sorting (FACS)-sorting of CD11b⁺CD3⁻ Ly6G⁻ CD45^hi cells at 10 dpl (S1A Fig). We confirmed enrichment of microglia-associated genes in the CD45^lo gate relative to the CD45^hi gate (e.g., P2ry12, Tgfbr1, Sall1, Tmem119) by RNA sequencing (Fig 2A) and quantitative polymerase chain reaction (qPCR) (S3A Fig); our findings are in line with some of these genes having previously been shown to be expressed in monocyte-derived macrophages albeit at a relatively lower level than in microglia (e.g., P2ry12, Tmem119), whereas Sall1 is never expressed by monocytes [34–37]. Furthermore, we found enrichment of monocyte-enriched genes in the CD45^hi cells (S3B Fig). Our gating strategy successfully excluded other neural and immune cell populations (S3C Fig). A low level of transcript for the myelin/oligodendrocyte gene Mbp was detected in microglia but not monocytes (S3C Fig), consistent with previous observations of myelin transcripts normally located in oligodendrocyte processes being found in microglia during CNS injury, likely reflecting phagocytosis [38]. However, no significant difference in expression of genes associated with phagocytosis was detected between microglia and monocytes (S3D Fig). Lesion monocytes and microglia showed significantly different transcriptomes (S1 Sheet). Analysis of the top 10% of all expressed genes revealed that monocytes and microglia share expression of 1,566 genes (Fig 2B), including myeloid cell identity genes such as Csf1r, Itgam, Cx3cr1, and Ptprc, as expected. However, of these top expressed genes, monocytes, and microglia both express 934 genes that are not highly expressed by the other population (Fig 2B; top 50 indicated in Fig 2C). We interrogated the differentially expressed genes (DEGs) between the two cell types using a stringent significance threshold (P < 0.01) and observed that whereas only 4 genes were upregulated in microglia versus monocytes, monocytes upregulated 457 genes relative to microglia (Fig 2D). This indicated that these myeloid cell populations have differing transcriptomes during remyelination, with monocytes primarily upregulating genes relative to microglia.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Monocytes have a Wnt signature during remyelination.

(A) Mean expression in Fragments per Kilobase of Transcript per Million mapped reads (FPKM) of microglia-enriched genes in lesion monocytes (red) and microglia (blue) at 10 dpl. **P = 0.0026; 2-tailed paired Student t test. n = 2–3 mice/group. (B) Venn diagram of top 10% of all expressed genes in lesion microglia and monocytes at 10 dpl. (C) Log2 fold change (FC) in top 50 enriched genes in microglia (left) and monocytes (right). Values were generated by comparing microglia transcriptomes to monocyte transcriptomes; the positive values indicate an upregulation (enrichment) in microglia, whereas negative values indicate a downregulation in microglia, therefore, an enrichment in monocytes. Red indicates highest enrichment and purple indicates the lowest enrichment. n = 2–3 mice/group. (D) Volcano plot showing significance as –Log10 transformed P values against Log2FC. Black indicates no significant change, red indicates genes upregulated in monocytes vs. microglia, and blue indicates genes upregulated in microglia vs. monocytes. P value cutoff 0.01. n = 2–3 mice per group. (E) Pathway analysis of genes upregulated in monocytes vs. microglia at 10 dpl. (F) Average FPKM of Wnt ligand genes in monocytes (red) and microglia (blue) in lesions at 10 dpl. ****P < 0.0001; one sample t test against microglia value of 0. (G) Average FPKM of Wnt signaling genes in monocytes (red) and microglia (blue) in lesions at 10 dpl. * P = 0.0152; one sample t test against microglia value of 0. (H) RFP+ lesion monocytes (red) expressing Axin2 (green) at 10 dpl, counterstained with myeloid cell nuclear marker PU.1 (purple). Inset, Axin2 isotype control. Scale bar, 100 µm. (I) Mean number of RFP+ cells expressing Axin2 ± S.E.M. at 10 and 21 dpl in Ccr2^RFP/+ reporter mouse lesions. *P = 0.00111; unpaired 2-tailed Student t test. n = 3 mice/time point. (J) Mean percentage of Olig2+ cells expressing Axin2 ±  S.E.M. at 21 dpl in WT and Ccr2^−/− mice; * P = 0.0162; unpaired 2-tailed Student t test. n = 3 mice/group. (K) Oligodendrocyte lineage cells (Olig2+; purple) expressing Axin2 (green) at 21 dpl. Scale bar, 50 µm. Mice were 8–12 weeks of age in each group. Source data may be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003073.g002

To further interrogate the molecular mechanisms regulating monocyte function during remyelination, we performed pathway analysis on DEGs. The Wnt signaling regulator β-catenin (CTNNB1) was predicted to be a top upstream regulator of monocyte DEGs, and gene ontology (GO) term and Ingenuity Pathway Analysis (IPA) suggested significant regulation of the Wnt signaling pathway (Fig 2E). This is of interest given the previous association of Wnt signaling with monocyte responses to injury in other organs [16,39–41] and with the regulation of remyelination [42–44]. In monocytes, we detected expression of 13 of the 19 Wnt ligands (Fig 2F), 6 of the Wnt receptors (Fzd1, 2, 3, 6, 9, 10), and a major downstream Wnt target and signaling component (Axin2) (Fig 2G), whereas these gene products were not detectable in microglia. These findings suggest active autocrine Wnt signaling within monocytes during remyelination. To validate this postulate, we assessed Axin2 protein expression in lesion monocytes using Ccr2^RFP/+ reporter mice. Axin2+ RFP+ cells were detected at both 10 and 21 dpl, significantly decreasing over time as remyelination proceeds (Fig 2H and 2I). We also detected Wnt-responsive oligodendrocyte lineage cells in remyelinating lesions (Axin2+ Olig2+), which were significantly decreased in Ccr2^−/− lesions (Fig 2J and 2K). Altogether, these findings suggest that monocytes have a Wnt signature during remyelination and that Wnt responsiveness is both autocrine and paracrine within lesions.

Wnt signature in monocytes in multiple sclerosis

We queried the translational relevance of our findings by investigating changes in monocyte Wnt activity in MS white matter lesions. As it is challenging to confidently distinguish between monocytes and microglia by immunostaining of human post-mortem tissue, we took advantage of a recently published single-nuclei RNA sequencing dataset of the rim of mixed MS lesions, which are sites of transition from efficient to poor remyelination [16,45]. Data-mining the genes significantly enriched in the 2 classical monocyte clusters (renamed clusters 1 and 2) relative to other immune cell clusters revealed that ~50% of genes are known to either be regulated by Wnts or regulate Wnt signaling (Fig 3A and 3B). We then asked whether such changes could also be detected in MS blood monocytes in association with clinical progression, as poor remyelination is observed in individuals with high disability [46], by assessing a highly-expressed subset of these genes using qPCR (Figs 3C and S3E). Of the 18 genes showing a ≥2-fold increase in expression in progressive MS monocytes over the average in healthy controls, 14 genes were significantly increased in cases with high clinical disability (secondary progressive multiple sclerosis [MS]; EDSS ≥ 6) versus those with low clinical disability (relapse-remitting MS; EDSS ≤ 2) (Fig 3C). This was not a result of higher age in the progressive MS group (S1 Table). Our results indicate that a Wnt signature is already partially induced in MS monocytes prior to CNS entry, which is increased in association with clinical disability progression.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Wnt signature in MS monocytes.

(A) Expression of Wnt-related genes in monocyte cluster 1, sequenced in the rim of mixed active MS lesions, represented as percentage of total cluster genes (left) and Log2 fold change (FC) (right). (B) Expression of Wnt-related genes in monocyte cluster 2, sequenced in the rim of mixed active MS lesions, represented as percentage of total cluster genes (left) and Log2FC (right). (C) Mean expression of Wnt-related genes in blood monocytes from individuals with relapse-remitting MS (RRMS; EDSS ≤ 2; n = 5) and secondary progressive MS (SPMS; EDSS ≥ 6; n = 6), which showed a ≥ 2-fold increase over healthy control (n = 4), represented as fold change over control. Brown-Forsythe and Welch ANOVA with Dunnet’s T3 multiple comparisons test; P = 0.0202 (BCAT1), 0.006 (HIF1A), 0.0058 (HSPH1), 0.0097 (NAMPT), <0.0001 (FXR1), 0.0014(BCAS2), 0.003 (CLK1), 0.0002 (ERO1A), 0.0050 (SNAP23), 0.0012 (FMNL2), 0.0002 (P4HA1), 0.0002 (CACYBP), 0.0001 (TCP1), 0.0084 (ZNF331). Source data may be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003073.g003

Monocyte Wnt signaling reduces the efficiency of remyelination

We next explored the role of Wnts in monocyte function during remyelination by inhibiting their Wnt activity. First, we used a mouse model previously employed to assess the requirement of Wnt signaling in monocytes in the repair of other organs [41], where Wnt activity is conditionally knocked out by eliminating expression of the Porcupine enzyme (Porc) involved in the post-translational modifications of Wnts required for their release. In the absence of an existing transgenic mouse in which Cre recombinase is specifically active in monocytes without potential confounding activity in other immune cells [47,48], we used the Csf1r-iCre line which is effective in knocking out Porc in myeloid cells [41] (Fig 4A). Although this approach targets both microglia and monocytes, we did not detect a Wnt signature in microglia during remyelination. We found that in floxed controls, remyelination occurred at a slower pace than expected, with minimal remyelination at 14 dpl likely resulting from the Friend leukemia virus B background (versus the standard C57Bl/6) [4,18]. Despite this finding, at this time, Csf1r-iCre;Porc^fl/fl mice showed enhanced remyelination compared to floxed controls, as indicated by increased expression of the late-stage remyelination marker myelin basic protein (MBP; Fig 4B and 4C). However, there was no significant difference in the proportion of mature oligodendrocytes (CC1+ Olig2+) in Csf1r-iCre;Porc^fl/fl versus floxed control lesions at 14 or 21 dpl, indicating increased remyelination independent of an impact on oligodendrocyte differentiation (Fig 4D and 4E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Wnt signaling in monocytes regulates remyelination efficiency.

(A) Csf1r-iCre;Porc^fl/fl were lesioned to assess the impact on remyelination. (B) Representative images of myelin basic protein (MBP) expression within lesions (outlined) at 14 dpl in Csf1r-iCre;Porc^fl/fl compared to Porc^fl/fl controls. Scale bar, 25 µm. (C) Mean percentage area of MBP staining ± S.E.M. in corpus callosum of Csf1r-iCre;Porc^fl/fl compared to Porc^fl/fl controls at 14 dpl. *P = 0.0130, one-tailed t test against control value of 1. n = 3–5 mice/group. (D) Representative immunofluorescent labeling of oligodendrocyte lineage cells (Olig2⁺, green) which are mature (CC1⁺, magenta) in Csf1r-iCre;Porc^fl/fl compared to Porc^fl/fl control mice at 14 dpl, counterstained with Hoechst. Scale bar, 20 μm. (E) Mean percentage of Olig2⁺ cells which are CC1⁺ per field ± S.E.M. in Csf1r-iCre;Porc^fl/fl compared to Porc^fl/fl controls at 14 and 21 dpl; ^*P = 0.1210 (14 dpl), 0.1741 (21 dpl); unpaired 2-tailed Student t test. n = 3 mice/group. (F) Schematic of transplant of C59- or Vehicle-treated GFP⁺ monocytes into lesioned Ccr2^−/− mouse circulation. (G) GFP⁺ transplanted cells detected in lesions of Ccr2^−/− mice, following ex vivo treatment with vehicle or C59, and stained for Axin2 (red), at 10 dpl. Double positive cells indicated with arrows. Scale bar, 25 µm. (H) Representative images of MAG expression within lesions at 10 dpl in lesioned Ccr2^−/− mice transplanted with C59- or vehicle-treated monocytes. Scale bar, 25 µm. (I) Mean percentage area of MAG staining ± S.E.M. in C59-treated monocyte transplanted Ccr2^−/− mice at 10 dpl, normalized to vehicle-treated monocyte transplanted control. ^* P = 0.0131, one sample t test against value of 1. n = 4–6 mice/group. (J) Mean percentage of Olig2⁺ cells which are CC1⁺ per field ± S.E.M. in C59- or vehicle-treated monocyte transplanted Ccr2^−/− mice at 10 dpl. ^*P = 0.4217; unpaired 2-tailed Student t test. n = 3–5 mice/group. (K) Representative immunofluorescent labeling of oligodendrocyte lineage cells (Olig2+, green) which are mature (CC1+, magenta) in Vehicle-treated and C59-treated monocyte transplanted mice at 10 dpl, counterstained with Hoechst. Scale bar, 20 μm. (L) Graphical abstract of results: Classical monocytes support oligodendrocyte differentiation and myelin protein expression, yet impede myelin production during remyelnation in a Wnt-dependent manner. Csf1r-iCre; Porc^fl/fl were 8–14 weeks old. Ccr2^−/− in transplant experiments were 8–16 weeks old. Source data may be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003073.g004

Second, we aimed to specifically target monocytic Wnt activity. We asked whether we could influence the efficiency of remyelination by intravenously supplementing classical monocytes – with and without ex vivo Wnt inhibition – to lesioned Ccr2^−/− mice, in which endogenous circulating monocytes are significantly reduced (Fig 4F). To track injected monocytes, we isolated cells from the bone marrow of adult mice in which myeloid cells express Green Fluorescent Protein (GFP) (Csf1r-eGFP), via magnetic-activated cell sorting (S3F Fig). To inhibit the Wnt pathway, we treated isolated cells ex vivo with a small molecule inhibitor of Porcupine, C59, prior to extensive washing and tail vein injection of 500,000 cells in 100 µL of PBS. GFP+ cells were observed in lesions of mice injected with either vehicle-treated or C59-treated monocytes, with no significant difference in densities between groups (Figs 4G and S3G). GFP+ cells were IBA1 negative, indicating maintenance of monocyte identity without macrophage differentiation (S3H Fig), consistent with our findings in wild type and Ccr2^RFP/+ mice (S1H and S1I Fig). C59 impaired Wnt signaling as demonstrated by reduced Axin2 staining (32 ± 3% Axin2+ GFP+ /total GFP+ cells versus 60 ± 3% in vehicle control, P = 0.0023; Fig 4G). Injection of C59-treated monocytes into Ccr2^−/− mice significantly increased remyelination at 10 dpl compared to vehicle-treated monocyte injection, evidenced by increased MAG signal (Fig 4H and 4I). However, there was no significant difference in oligodendrocyte differentiation between treatment groups (Fig 4J and 4K). Overall, these findings reveal that monocyte-derived Wnt decreases the efficiency of remyelination, independently of impacting oligodendrocyte differentiation.

In vivo focal CNS white matter demyelination

Experiments were performed according to ARRIVE2 guidelines under a project license issued to VEM (PP8451532) under the Animals (Scientific Procedures) Act 1986 and reviewed then approved by the Animal Welfare Ethical Review Body of the University of Edinburgh and the United Kingdom Home Office, or were approved by the Animal Care Committee at St. Michael’s Hospital (protocol numbers 230, 231, 232) in accord with the Canadian Council on Animal Care (CCAC). Mice were provided food and water ad libitum and exposure to 12-h light/dark cycles. Demyelinated lesions were induced in the corpus callosum of adult male (8−17 weeks) C57BL/6J mice (Jackson Laboratory 000664), Ccr2^−/− mice (B6.129S4-Ccr2tm1lfc/J, Jackson Laboratory 004999), Csf1r-iCre;Porc^fl/fl and Porc^fl/fl mice [41] (Jackson Laboratory 021024), and Ccr2^RFP/+ mice (B6.129[Cg]-Ccr2tm2.1lfc/J, Jackson Laboratory 017586), by stereotaxic injection of 2 μL of 1% lysolecithin (LPC; vol/vol, Sigma-Aldrich L4129) using a Hamilton syringe at coordinates from Bregma: −1.2 mm caudal, −0.5 mm lateral, −1.4 mm ventral, as previously described [4,14,18,23]. Mice were sacrificed at various days post-lesion (dpl). To confirm our gating strategy to distinguish between microglia and monocytes, LPS was injected stereotaxically into the striatum (coordinates from Bregma: 0.0 mm caudal, −2.0 mm lateral, 2.5 mm ventral), and mice were sacrificed at 3 days post-injection. Mice used for FACS or flow cytometry were intracardially perfused with ice-cold 1X phosphate-buffered saline (PBS; Sigma-Aldrich, D8537). Mice used for immunofluorescence were intracardially perfused with PBS, then 4% paraformaldehyde (PFA, wt/vol; Sigma-Aldrich, P6148 or BioShop Canada, PAR070); brains were post-fixed overnight and cryoprotected in sucrose (BioShop Canada, SUC700) before OCT embedding (VWR, 361603E or Fisher, 23-730-571) and storage at −20 °C.

Immunofluorescent staining of mouse tissue

Frozen coronal brain sections (10 μm) were air-dried for 15 min and blocked for 1 hour room temperature with PBS containing 5% horse serum (Gibco, 2,605,008) and 0.3% Triton-X 100 (Fisher Scientific, 10254583). For Axin2, Olig2, and PU.1 staining, heat-activated antigen retrieval was carried out by microwaving sections on a medium setting in 10 mM citrate buffer (pH 6.0; Vector laboratories, H-3300) and heating for 30 min at 60 °C in an oven. Sections were washed in 1X TBS (Fisher Scientific, BP24711) with 0.001% Triton-X-100 (Fisher Scientific, 10254583) prior to incubation with primary antibodies. For myelin and RFP stains, sections were incubated with ice-cold methanol at −20 °C for 10 min and washed in PBS prior to primary antibody incubation. Primary antibodies were applied overnight at 4 °C in a humid chamber and included rat-anti-MBP (Millipore, AB5,864, 1:250), rabbit anti-Axin2 (1:50; Abcam, ab18582 or 1:200; Abcam, ab32197), mouse anti-MAG (1:100; Sigma-Aldrich, MAB1567), rabbit anti-RFP (1:100; Abcam, ab62341 or ab185921), rabbit anti-Olig2 (1:100; Sigma-Aldrich, AB9610), goat anti-IBA1 (1:200; Abcam, ab5076), rat-anti-PU.1 (1:100; Biolegend, 681202), and mouse anti-APC (CC1; 1:500, Abcam, ab16794). For Axin2, Olig2, and RFP, primary antibodies were amplified using tyramides. Briefly, after washing in 1XTBS with 0.001% Triton-X to remove the primary antibody, sections were incubated with the species-specific HRP-conjugated secondary antibody (Vector Laboratories, anti-rabbit VECTMP740115 or anti-mouse VECTMP740215) before being washed again in 1XTBS with 0.001% Triton-X and incubated for 10 min in a humid chamber with either Opal 520 (1:100; Akoya Biosciences, FP1487001KT), Opal 570 (1:100; Akoya Biosciences, FP1488001KT), or Opal 650 (1:100; Akoya Biosciences, FP1496001KT). After a final washing step, the sections were co-stained with additional primary antibodies listed above, after which sections were incubated with Alexa fluor-conjugated secondary antibodies (1:500; Life Technologies) for 2 h at 20–25 °C in a humid chamber. Sections were then counterstained with Hoechst (Sigma-Aldrich, B2261 or Thermo Fisher Scientific, H1339), washed in PBS, and cover slipped using Fluoromount-G (Invitrogen, 00-4958-02). Tissue was imaged using an Olympus 3i Spinning Disk confocal microscope, a Leica SP8 confocal microscope, a Zeiss LSM 900, or a Zeiss wide field microscope.

Tissue processing and electron microscopy

Mice were intracardially perfused with 4% PFA (wt/vol) and 2% glutaraldehyde (vol/vol, Sigma-Aldrich G5882) in 0.1 M phosphate buffer. Tissue was then post-fixed overnight in 1% glutaraldehyde in 0.1 M phosphate buffer at 4 °C. About, 1 mm tissue sections were processed into araldite resin blocks. Semi-thin (1 μm) sections were cut using a microtome and stained using a solution of 1% toluidine blue and 2% sodium borate prior to imaging on a Zeiss Axio microscope using brightfield to identify areas containing lesioned corpus callosum. Ultra-thin sections (60 nm) were subsequently cut of the lesioned corpus callosum, and stained with uranyl acetate and lead citrate before imaging on a JEOL Transmission Electron microscope. Axon size and g-ratios were measured using ImageJ (FIJI) software with a minimum of 100 axons measured per animal, and number of myelinated axons was counted using ImageJ (FIJI) software with a minimum of 10 images per animal analyzed.

Myeloid cell detection and isolation from focal CNS remyelinating lesions

Monocytes and microglia were isolated from focal remyelinating lesions of corpus callosum as previously described [18,61]. Briefly, following perfusion with ice-cold PBS, brains were extracted and stored on ice in Hanks’ Balanced Salt Solution (HBSS) without Ca²⁺ and Mg²⁺ (Gibco, 14,175,095) with 10% FBS (Gibco, 10,500,064). Lesioned corpus callosum was dissected out from a 2 mm coronal section then homogenized manually, with a minimum of 40 passes using a 2 mL Dounce (DWK Life Sciences, 10,530,762). Following filtration (75 μm filter; Pierce, 87,791) cells were spun at 600 g for 5 min with brake. Cells were resuspended in 100% fetal bovine serum (FBS; Gibco, 10,500,064) and 33% isotonic Percoll (1:10; SLS, 17,089,102 or Cytiva, 17-0891-01). 1 mL of 10% FBS was carefully overlaid to create a concentration gradient, and the cells were spun for 15 min at 800 g at 4 °C without brake. The cell pellet was washed in FACS buffer (1% FBS, 1 µm EDTA in PBS) and spun for 10 min at 600 g at 4 °C, then resuspended in FACS buffer. Cells isolated from lesions were added to a 96-well plate and centrifuged at 400 g for 3 min. Cells were incubated in anti-mouse CD16/32 Fc-block (1:200; BD, 553142 or 1:400; BioLegend, 101319) and viability dye (1:1,000; eBioscience, 65086613) diluted in FACS buffer and incubated on ice for 20 min. Fluorescently-conjugated antibodies were added directly to the samples without washing and incubated on ice for 30 min. Antibodies included anti-mouse CD11b-PeCy7 (1:100; eBioscience, 25-0112-81 or 1:1,000, BioLegend, 101216, Clone M1/70), anti-mouse CD45-BV605 (1:300; Biolegend, 103139, Clone 30-F11), Ly6G-PerCP Cy5 (1:200, BioLegend, 127615, Clone 1A8) or Ly6G-PE/Dazzle 594 (1:600, BioLegend, 127648, Clone 1A8), Ly6C-FITC (1:600; BD, 553104, Clone AL-21), CD3-APC (1:200, BioLegend, 100235, Clone 17A2 or A700; 1:1,000, BioLegend, 100216, Clone 17A2) and F4/80-APC (1:600; BioLegend, 123116, Clone BM8). Cells were centrifuged at 400 g for 3 min resuspended in FACS buffer and passed through a 30 μm filter (Sysmex, 04-004-2326) prior to sorting (BD FACSAria Fusion) or analysis by flow cytometry (BD LSR Fortessa or CytoFLEX-LX). Cells were sorted into 1.5 mL Eppendorf tubes coated with 100% FBS then kept on ice. FlowJo was used for post-acquisition data analysis. To facilitate visualization of monocyte numbers in lesions, 3 mice were pooled for Ccr2^RFP/+ lesion analysis, and 2 mice were pooled repeated 3 times for a total of 6 mice/group for wild-type sham PBS-injected, wild-type LPC-injected, and Ccr2^−/− lesion analyses.

RNA extraction and ultra-low input RNA sequencing

The average cell yield was 5,777 microglia and 38 monocytes per mouse. We sequenced an average of 4 ng of RNA for microglia and 1.6 ng of RNA from monocytes, using ultra-low input RNA sequencing which is optimized for samples with as little as 0.5 ng; this was sufficient for our purposes as we detected high expression of myeloid cell-associated genes in both microglia and monocyte populations and differential expression of genes known to be enriched in microglia versus monocytes. Isolated cells were centrifuged at 800 g for 5 min, resuspended in 350 μL RLT Plus buffer with β-mercaptoethanol (QIAGEN, 1053393), and centrifuged at 8,600 g for 2 min using QIAshredder tubes (QIAGEN, 79656). RNA was extracted using the AllPrep DNA/RNA/miRNA kit (QIAGEN, 80224) as per the manufacturer’s instructions. Quantity and quality of RNA were determined using the Bioanalyzer 2100 (Agilent) and RNA 6000 Pico kit (Agilent, 5067−1513) as per the manufacturer’s instructions. cDNA production and library preparation were performed by BGI (Hong Kong). Briefly, double-stranded cDNA was prepared according to Ovation RNA-Seq System V2 (Tecan/NuGen, 7102−32), as per the manufacturer’s instructions. The amplified cDNA was combined with End Repair Mix (QIAGEN, Y9140-LC-L) and incubated at 20 °C for 30 min. End-repaired DNA was purified with AxyPrep Mag polymerase chain reaction (PCR) clean-up kit (Axygen, MAG-PCR-CL-50), then combined with A-Tailing Mix (Enzymatic) at 37 °C for 30 min. Adaptors (Invitrogen) were ligated to the Adenylate 3′ ends DNA, incubated at 16 °C for 16 h, and purified with AxyPrep Mag PCR clean-up kit. The adapter-ligated DNA fragments were selected based on the insert size and enriched by PCR amplification; the PCR products (final library) were purified with AxyPrep Mag PCR clean-up kit. The final library was quantitated in two ways: The average molecule length was determined using the Bioanalyzer 2100 using the DNA 1000 kit (Agilent, 5067−1504), and the library was quantified by real-time quantitative PCR (qPCR) (TaqMan Probe). The qualified libraries were amplified using cBot (Illumina) to generate the cluster on the Flow Cell (HiSeq 4000, Illumina). An average of 40M clean 100 paired-end reads were obtained per sample. Data was processed to remove adaptors, and low-quality reads were removed from raw reads.

Bioinformatics

Raw data analysis was conducted by Fios Genomics (Edinburgh, UK). Briefly, data were pre-processed and aligned to the mouse genome (GRCm38) using STAR aligner, and number of mapped read pairs per gene was quantified based on GENCODE vM12 annotation. Quality control of samples was conducted using the array Quality Metrics package in Bioconductor. Data were normalized using trimmed mean of M values (TMM) and transformed using variance modeling at the observational level (VOOM) to Log2-counts per million (Log2- CPM) or Fragments per kilobase of transcript per million mapped reads (FPKM). Linear modeling was used for group comparisons using Limma and incorporating precision weights from VOOM to identify DEGs (P < 0.01). Functional analysis was conducted using GO term analysis, Venny2.1, and Ingenuity Pathway Analysis (IPA; QIAGEN).

Real-time quantitative polymerase chain reaction (RT-qPCR) on mouse cells

Real-time (RT) qPCR was run using BioRad Custom PrimePCR plates, as per the manufacturer’s instructions. Briefly, cDNA was synthesized using 5x iScript Advanced Reaction Mix and iScript Advanced Reverse Transcriptase (BioRad, 1,708,890) at 46 °C for 20 min then 95 °C for 1 min. cDNA samples were mixed with 2x iTaq Universal SYBR-green Supermix (BioRad, 1725120) and iScript Reverse Transcriptase (BioRad, 1708840). The RT-qPCR was performed at 50 °C for 10 min, 95 °C for 1 min, 95 °C for 15 sec (40 cycles), and 60 °C for 60 sec (40 cycles). Data was analyzed using CFX manager and is represented as 2^−∆Ct.

Bone-marrow-derived monocyte isolation and transplantation

Monocytes were isolated from the bone marrow of 8–12 week-old MacGreen (Csf1r-eGFP) male mice using magnetic cell sorting (MACS) according to the manufacturer’s instructions (Miltenyi Biotec), and cells were kept on ice or at 4 °C throughout. Briefly, bone marrow cells were collected from the femur and tibia bones of MacGreen mice by flushing the shaft with a syringe and needle containing buffer solution (0.5% bovine serum albumin; Miltenyi Biotec, 130-091-376, and 2 mM EDTA; Thermo Fisher, AM9260G, in PBS). Cells were blocked and stained using a monocyte biotin-antibody cocktail (Miltenyi Biotec) for 5 min at 4 °C. Cells were subsequently washed with buffer solution and centrifuged for 10 min at 300 g. The cell pellet was resuspended in buffer containing anti-biotin microbeads (Miltenyi Biotec, 130-091-256) and incubated for 10 min at 4 °C before proceeding to magnetic separation. Separation was carried out by applying the cell suspension to an LS column (Miltenyi Biotec, 130-042-401) attached to a MidiMACS magnetic separator (Miltenyi Biotec, 130-042-302) and collection of the flow-through containing the enriched monocyte fraction. Monocytes were then treated with 10 nM C59 (Tocris, 5,148) or DMSO vehicle, diluted in PBS for 25 min at 4 °C. Cells were centrifuged for 10 min at 300 g and resuspended in PBS prior to injection into the tail vein of Ccr2^−/− mice. Each mouse received 1 injection containing 500,000 cells in 100 μL of PBS. Identity of isolated cells following MACS was confirmed using flow cytometry. Cells were incubated in anti-mouse CD16/32 Fc-block (Clone 93, 1:200; BD, 55,314) diluted in FACS buffer, and incubated on ice for 10 min. Fluorescently-conjugated antibodies were added directly to the samples and incubated on ice for 30 min. Antibodies included anti-rat CD11b-BV650 (1:100; BioLegend, 101239), anti-mouse CD115-BV605 (1:100; BioLegend, 135517), rat-anti Ly6G-BV510 (1:200; Biolegend, 127633), anti-rat Ly6C-PB (1:100; Biolegend, 128014), and rat-anti CD45-PE/Dazzle 594 (1:100; Biolegend, 103139). Cells were centrifuged at 400 g for 5 min and resuspended in FACS buffer then passed through a 30 μm filter prior to sorting (BD FACS Aria Fusion).

Data-mining published MS monocyte transcriptomes

Datasets were obtained from a published study of single-nuclei RNA sequencing of MS chronic active (i.e., “mixed”) lesions [16]. In their study, the authors subclustered immune cells (CSF1R+, C3+, CD74+, RUNX1+) and lymphocytes (PTPRC+, CD2+, SKAP1+), then further subclustered these cells. Two clusters were annotated as monocytes based on upregulation of CD83 and lower expression (negative Z scores) of microglia-enriched genes SALL1 and TMEM119 compared to all immune cells, including microglia. Peripheral immune cell (PBMC) signatures were further mapped onto a multimodal peripheral blood mononuclear cell reference dataset to confirm identities (https://satijalab.org/seurat/v4.0/reference_mapping.html); the majority of the monocytes expressed CD14 rather than CD16, therefore representing classical monocytes. We mined the transcriptomes of both clusters for genes that are known to be regulated by Wnt or regulate Wnt signaling.

Human monocyte isolation, RNA extraction, and qPCR

Participants with relapse-remitting or secondary progressive MS (using the 2017 revised McDonald criteria) were recruited from the MS clinics at NL Health (St John’s, NL, Canada) as part of the Health Research Innovation Team in Multiple Sclerosis study. Participants provided written consent with prior approval by the provincial Health Research Ethics Board. Age and sex-matched healthy control subjects were also recruited. Blood was drawn via venipuncture from four healthy controls, five patients with relapsing-remitting multiple sclerosis and six patients with secondary progressive MS; all participants were not on treatment, were female, and aged between 26 and 66 years old (S1 Table). Immediately after blood collection, PBMCs were isolated using SepMate PBMC isolation tubes (STEMCELL Technologies, 85460) and a Ficoll-Paque (Cytiva, 17-1440-03) density gradient. PBMCs were washed twice with PBS, counted, and cryopreserved in freezing media (20% DMSO, 10% DMEM, 70% FCS) and liquid nitrogen. CD14+ monocytes were positively selected to 95%–98% purity from thawed PBMCs using MACS anti-human CD14 microbeads, LS columns, and QuadroMACS separators (Miltenyi Biotec, 130-050-201, 130-042-401). The CD14+ cells were stored at −80 °C in 500 µL QIAzol Lysis Reagent (Qiagen, 79306). RNA was extracted using an RNeasy Micro kit (Qiagen, 74004). Briefly, 100 µL of chloroform (Fisher Scientific, AC423,555,000) was added to the sample, agitated for 2–3 min and centrifuged at 4 °C at 12,000 g for 15 min. The aqueous layer was removed and transferred to an RNase-free 1.5 mL centrifuge tube. The RNA was precipitated with 350 µL 70% ethanol and added to a spin column. The samples were centrifuged, washed, treated with DNase, and eluted with 14 µL RNase-free water as per the manufacturer’s instructions. A total of 200 ng RNA and 250 ng of random primers were used to synthesize complementary DNA (cDNA) using the M-MLV reverse transcriptase kit (Invitrogen, 28025013). The cDNA was stored at −80 °C. Gene-specific pre-amplification of the resulting cDNA was performed by mixing 5 μL of cDNA with custom primers (330,241, QIAGEN) before being placed in a real-time cycler (Biorad C1000 Thermal Cycler). Side reaction reducer was used to eliminate residual primers from the pre-amplification step before samples were stored at −20 °C. The PCR components mix for each sample was then generated using 102 μL of the amplified cDNA plus RT2 SYBR-green ROX qPCR Mastermix (330521 QIAGEN) before being loaded into a 384-well custom RT2 PCR array (330181, QIAGEN). Plates were run on QuantStudio 7 (Applied Biosystems) using SYBR as the method of detection. Following qPCR, genes were excluded from analysis if the difference in CT values between technical duplicates was greater than 1. Data is represented as fold difference in relative quantity over healthy control, calculated as 2^{ΔCt (target gene in control − MS)}/ Geometric mean of 2^{ΔCt (housekeeping gene in control − MS)}; the housekeeping genes selected were those stably expressed across groups (ACTB and HSP90AB1).

Statistics

Statistical analysis was performed using GraphPad Prism software. Gaussian distribution was analyzed using the Shapiro–Wilks test or Kolmogorov–Smirnov test. Data are presented as mean ± standard error of the mean (S.E.M.). Data were analyzed using unpaired Student t test, Wilcoxon paired t test, Mann–Whitney test, One-way analysis of variance (ANOVA) with Tukey’s post hoc test, or Two-way ANOVA with Sidak’s multiple comparisons test. Human monocyte qPCR data were analyzed using a Brown Forsyth Welch ANOVA with Dunnet’s T3 multiple comparisons post-test. Values of P ≤ 0.05 were considered statistically significant.