Cytokine polarized, alternatively activated bone marrow neutrophils drive axon regeneration

Abstract

The adult central nervous system (CNS) possesses a limited capacity for self-repair. Severed CNS axons typically fail to regrow. There is an unmet need for treatments designed to enhance neuronal viability, facilitate axon regeneration and ultimately restore lost neurological functions to individuals affected by traumatic CNS injury, multiple sclerosis, stroke and other neurological disorders. Here we demonstrate that both mouse and human bone marrow neutrophils, when polarized with a combination of recombinant interleukin-4 (IL-4) and granulocyte colony-stimulating factor (G-CSF), upregulate alternative activation markers and produce an array of growth factors, thereby gaining the capacity to promote neurite outgrowth. Moreover, adoptive transfer of IL-4/G-CSF-polarized bone marrow neutrophils into experimental models of CNS injury triggered substantial axon regeneration within the optic nerve and spinal cord. These findings have far-reaching implications for the future development of autologous myeloid cell-based therapies that may bring us closer to effective solutions for reversing CNS damage.

Extended Data Figure 1. Kinetics of gene expression in BMN during polarization. Polarized and unpolarized Ly6G+ BMN express canonical markers of granulopoiesis

a, RNA was extracted from Ly6G⁺ BM cells at serial time points during culture in the presence or absence of polarizing factors. Levels of transcripts encoding IL-4Ra (IL4r), arginase-1 (Arg1), mannose receptor (Mrc1), and F4/80 (Adgre1), were measured by RT-qPCR and normalized to Actb. Data are shown as fold increase over mean normalized levels in unpolarized BMN. Data are presented as mean values ± SEM. b, Normalized levels of myeloperoxidase (Mpo) and neutrophil elastase (Ela) mRNA, extracted from Ly6G⁺ BM cells, following a 24 hour polarization under the indicated conditions (n=5). Transcript levels were also measured in mature peripheral blood neutrophils, isolated from naïve mice, for an additional control. (n=5). Data are presented as mean values ± SEM. c, Polarized and unpolarized BMN were analyzed by flow cytometry. Histograms showing MFI of MPO (left) and Elastase (right), gating on CD11b⁺Ly6G⁺ cells (n=5 mice per group). a, Data shown from one experiment, representative of two independent experiments. b, c, Data are shown from one experiment, representative of four independent experiments. d, Flow cytometric gating strategy for assessing the purity of MACS purified Ly6G⁺ BM cells

Extended Data Figure 2. GFAP expression in retina from mice subjected to ONC injury and i.o. injection of unpolarized or polarized BMN versus PBS.

Retina were harvested on day 4 following ONC injury and i.o. injection of either IL-4/G-CSF polarized BMN, unpolarized BMN, or PBS. Each panel shows a representative retinal cross-section obtained from an individual mouse, stained with antibodies against GFAP (green). Scale bar, 100 μm. Data shown from one experiment, representative of two independent experiments

Extended Data Figure 3. Representative western blot of HBEGF protein in BMN lysates

Representative western blot, analyzing lysates from unpolarized BMN (left) and IL-4/G-CSF polarized BMN (right), displaying total protein (stain-free gel) and HBEGF band on a PVDF membrane. Data shown from one experiment, representative of two independent experiments

Extended Data Figure 4. IL-4/G-CSF polarized Ly6G+ BMN accumulate in the sciatic nerve at 4 hours, and the spinal cord at 24 hours, following SCI and i.n. BMN injection

a, Representative section of a sciatic nerve harvested 4 hours after SCI injury and i.n. injection of IL-4/G-CSF polarized BMN that were derived from tdTomato (tdT)-reporter mice. Donor BMN are identified as tdT⁺ (red) and Ly6G⁺ (white) (scale bar, 50 μm). b, Representative flow cytometric pseudocolored dot plot of mononuclear cells isolated from sciatic nerves 4 hours after SCI injury and i.n. injection of either IL-4/G-CSF polarized tdT⁺ BMN (left) or PBS (right), gating on CD11b⁺Ly6G⁺ cells. c, Representative flow cytometric pseudocolored dot plot of spinal cord mononuclear cells isolated 24 hours following SCI injury and i.n. injection of either IL-4/G-CSF polarized tdT⁺ BMN (left) or PBS (right), gating on CD11b⁺Ly6G⁺ cells. a-c, Data shown from one experiment, representative of two independent experiments.

Extended Data Figure 5. IL-4/G-CSF polarized CD34+ bone marrow cells, derived from individual donors, reproducibly induce neurite outgrowth of human cortical neurons.

a, Gating strategy for flow cytometric analysis and sorting of IL-4/G-CSF polarized CD34⁺ human BM cells. b, Mean neurite length of the longest neurite grown by primary cortical neurons following 24 hour culture with polarized or unpolarized CD34⁺ BM cells. Each plot represents an experiment performed using BM cells derived from a unique BM donor. c, Mean neurite length of the longest neurite grown by primary cortical neurons following 24 hour culture with CD34⁺ or CD33⁺CD15⁺ cells FACS sorted from IL-4/G-CSF polarized BM cells. Each plot represents an experiment performed using BM cells derived from a unique BM. b, c, Data are presented as mean values ± SEM. Each symbol represents a single neuron with the n value listed below each treatment condition. Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test.

Figure 1.. IL-4/G-CSF polarized murine bone marrow neutrophils (BMN) exhibit an immature, alternatively activated phenotype.

a-d, Ly6G⁺ neutrophils were isolated from naïve C57BL/6 bone marrow (BM) cells and cultured with or without polarizing factors for 24 hours. a, RNA was extracted from the cultured BMN (n=5/ group), or from peripheral blood neutrophils (n=4) directly ex vivo. Levels of transcripts encoding IL-4 receptor α chain (Il4r), arginase-1 (Arg1), mannose receptor (Mrc1), and F4/80 (Adgre1), were measured by RT-qPCR and normalized to β-Actin (Actb). Data are shown as fold increase over mean normalized levels in unpolarized BMN. Each symbol represents data from a single mouse. Data are shown from one experiment representative of four independent experiments. b, d, Ly6G⁺ BM cells, cultured under the indicated conditions (n=5/ group), and Ly6G⁺ peripheral blood neutrophils (n=4), were analyzed by flow cytometry. b, Cell surface expression of indicated molecules on Ly6G⁺ cells. Upper panels, mean fluorescence intensity (MFI) of cell surface IL4Rα and F4/80; % of CD101⁺ cells. Each symbol represents data from a single mouse. Lower panels, representative histograms. Data shown are representative of five independent experiments. c, Cytospins of Ly6G⁺ BMN, following culture under the indicated conditions, stained with Wright-Giemsa solution. Scale bar,10 μm. Data shown are representative of three independent experiments. d, Representative sample and overlay t-SNE plots generated by combining flow cytometric data from analysis of polarized and unpolarized BMN, as well as naïve peripheral blood neutrophils. Data shown from one experiment, representative of four independent experiments. a, b, Data presented as mean values ± SEM. Statistical significance determined using one-way ANOVA followed by Dunnett’s post hoc test.

Figure 2.. IL-4/G-CSF polarized BMN enhance RGC survival and axon regrowth.

a-c, Ly6G⁺ BMN, that had been polarized under the indicated conditions, were co-cultured with primary RGCs for 24 hours. RGCs and BMN were either plated in direct contact with one another (a, b), or separated across a transwell (c). Some RGCs were cultured with recombinant ciliary neurotrophic factor (CNTF) as a positive control, or in media alone as a negative control. a, Representative images of RGCs from each group (scale bar = 20 μm). b,c Length of the longest neurite grown per neuron. Each dot represents a single neuron. b, n=3 mice per condition; RGC counted per group: n=274 (media), 426 (CNTF), 830 (unpolarized), 586 (IL-4), 696 (G-CSF),1211 (IL-4+G-CSF). c, n=3 mice per condition; RGC counted per group: n= 491 (media), 397 (CNTF), 502 (unpolarized), 482 (IL-4), 496 (G-CSF), 446 (IL-4+G-CSF). b, c, Data presented as mean values ± SEM. Data shown are from one experiment, representative of three independent experiments. d-g, Mice were injected i.o. with IL-4/G-CSF BMN, unpolarized BMN, or PBS on days 0 and 3 post-ONC. CTB tracer was administered i.o. on day 12, and mice were euthanized 2 days later. d, Representative images of optic nerves from each group. Scale bar, 200 μm. e, Density of CTB+ regenerating axons in longitudinal optic nerve sections, counted at serial distances from the crush site. n=19 (PBS), 20 (unpolarized BMN), and 18 (IL-4+G-CSF BMN). f, Area under the curve (AUC) from the data shown in (e). Each dot represents an individual nerve from (e). e, f, Data are presented as mean values ± SEM. Data are representative of five independent experiments. g, Frequency of viable Brn3a⁺ RGCs per mm² in retinal whole mounts. n=8 mice (PBS), 9 (unpolarized BMN), and 10 (IL-4+G-CSF BMN). Data presented as mean values ± SEM. Data are from one experiment, representative of two independent experiments. e, Statistical significance was determined using the unpaired two-sided t-test with Welch correction. *P, compared with PBS; ^#P, compared with unpolarized BMN. b, c, f, g, Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test.

Figure 3.. IL-4/G-CSF polarized neutrophils exhibit a transcriptome indicative of alternative activation and a pro-regenerative phenotype.

Purified Ly6G⁺ BMN, harvested 24 hours after culture with or without the polarizing factors indicated, were analyzed by bulk RNA sequencing. a, Principal component analysis (PCA) plot. Each symbol represents data collected from one biological replicate. (n=3 mice/ group). b, c Heat maps depicting scaled expression of selected genes associated with classical activation (b, left panel), alternative activation (b, right panel), or IL-4 signaling (c, left panel), or genes encoding growth factors (c, right panel). d, e, Volcano plots illustrating the differences in gene expression between unpolarized BMN and IL-4/G-CSF polarized BMN (d), and between G-CSF polarized BMN and IL-4/G-CSF polarized BMN (e). d, e, Volcano plots were generated using a Negative Binomial GLM with Wald test.

Figure 4.. IL-4/G-CSF polarized BMN enhance axon regeneration, in part, via the production of IGF-1 and HBEGF.

a, b Ly6G⁺ BMN were cultured with IL-4+G-CSF or without cytokines (Unp). a, IGF-1 levels in 24 hour conditioned media were measured via ELISA (n=7). b, HBEGF levels were measured in cell lysates via western blot (n=8). Band intensity was normalized to total protein. Data are shown as fold increase over unpolarized cells. a, b, Each dot represents one biological replicate. Pooled from 2 independent experiments. c, d, RGCs were co-cultured with IL-4/G-CSF polarized BMN, with/without anti-IGF-1 and PD153, for 24 hours. c, Representative images of RGCs. Scale bar, 20 μm. d, Length of the longest neurite grown per neuron. Each dot represents a single RGC. RGCs were cultured with CNTF as a positive control, or in media alone as a negative control. (n=3 mice per condition, RGC counted: n=140 (media), 151 (CNTF), 188 (unpolarized), 121 (IL-4+G-CSF/ No Ab), 142 (IL-4+G-CSF/ αIGF1+PD153), 135 (IL-4+G-CSF/Isotype+Vehicle)). Data presented as mean values ± SEM. Statistical significance determined by one-way ANOVA followed by Dunnett’s post hoc test. Data are representative of two independent experiments. e-h, On days 0 and 3 post-ONC, mice were injected i.o. with PBS (n=8), or with IL-4/G-CSF polarized BMN plus either anti-IGF-1/PD153 (n=15), or control antibody/vehicle (n=16). CTB tracer was injected on day 12 and optic nerves and retina harvested 2 days later. e, Representative images of CTB⁺ optic nerves. Scale bar, 200 μm. f, Density of CTB+ axons at serial distances from the crush site. Pooled from two independent experiments. g, Area under the curve from data in (f). h, Frequency of Brn3a+ RGCs per mm² in retinal wholemounts from mice injected i.o. with PBS (n=7), IL-4/G-CSF polarized BMN and control antibody/vehicle (n=9), or IL-4/G-CSF polarized BMN and αIGF1/PD153 (n=10). Data shown from one experiment, representative of two independent experiments. g, h, Each dot represents an individual nerve or retina. a, b, f-h, Data presented as mean values ± SEM. a, b, f, Statistical significance determined by unpaired two-tailed t-test with Welch correction. g, h, Statistical significance determined by one-way ANOVA followed by Dunnett’s post hoc test. f, *P compared with PBS; ^#P compared with αIGF-1+PD153.

Figure 5.. IL-4/G-CSF polarized neutrophils drive the regeneration of spinal cord axons.

a-c, DRG neurons were co-cultured with Ly6G⁺ BMN for 24 hours, either in direct contact (a, b), or separated across a transwell (c). Other DRG neurons were cultured with NGF as a positive control, or in media alone as a negative control. a, Representative images. Scale bar, 100 μm. b, c Length of the longest neurite grown by each neuron. Each dot represents one neuron. Data presented as mean values ± SEM. b, n=4 mice per condition; DRG counted: media (n=472), NGF (n=675) unpolarized (n=658), IL-4 (n=360), G-CSF (n=435), IL-4+G-CSF (n=631). Data shown from one experiment, representative of three independent experiments. c, n=3 mice per condition; DRG counted: media (n=708), NGF (n=942), unpolarized (n=1227), IL-4 (n=1125), G-CSF (n=1244), IL-4+G-CSF (n=1236). Data shown from one experiment, representative of two independent experiments. d-g, Spinal cord (SC) axon regeneration in mice following dorsal column transection and different treatments. In one group, mice were injected in the left sciatic nerve with unpolarized BMN, and in the right sciatic nerve with IL-4/G-CSF polarized BMN (n=10 mice). For an additional negative control, other mice were injected in both sciatic nerves with PBS (n=8). For a positive control, mice were subjected to bilateral conditioning injury (CI) 5 days prior to SC injury (n=7). All mice were injected with Alexa Fluor 680-conjugated or Texas red-conjugated dextran in the left and right sciatic nerve, respectively, 8 weeks following injury and 10 days prior to SC harvest. f, Distance between the lesion center and rostral tip of the longest regenerating axon. The vertical line in each box plot represents the median, the box indicates the interquartile range, and the whiskers indicate the minimum and maximum values. g, Total volume of tracer positive axon voxels measured within 800 μm of each side of the lesion (n=8). Each symbol represents one individual mouse. Data presented as mean values ± SEM. f, g Data shown are from one experiment, representative of two independent experiments. b,c, f, Statistical significance determined by one-way ANOVA followed by Dunnett’s post hoc test. g, Statistical significance determined via two-tailed unpaired t-test with Welch correction.

Figure 6.. IL-4/G-CSF polarized CD34+ BM cells display a transcriptome indicative of alternatively activated, immature neutrophils.

CD34⁺ human BM cells were cultured in media alone, or in the presence of recombinant human IL-4 and/or G-CSF, for 48 hours. a-d, RNA was extracted from cells in each group for bulk RNA sequencing. a, PCA plot. Each symbol represents 1 biological replicate, comprised of cells derived from a single unique donor (n=3). b, Heat maps depicting scaled expression of selected genes associated with neuroprotection/axonal regeneration (left), alternative activation (middle), and granulopoiesis (right). c, d, Volcano plots illustrating differences in gene expression comparing unpolarized and IL-4/G-CSF polarized human BM cells (c), or G-CSF polarized and IL-4/G-CSF polarized human BM cells (d). Volcano plots were generated using a Negative Binomial GLM with Wald test.

Figure 7.. IL-4/G-CSF polarized CD34+ BM cells contain a pro-regenerative neutrophil subset.

a, b Human cortical neurons were co-cultured for 24 hours with polarized or unpolarized human CD34⁺ BM cells. BM cells derived from unique donors were cultured independently of one another. Additional neurons were cultured with recombinant NGF as a positive control, or in media alone as a negative control. a, Representative images of cortical neurons following culture with BM cells from each group. Scale bar, 20 μm. b, Length of the longest neurite grown by each cortical neuron in each experimental group. Each dot represents a single neuron. (n= 6 unique donors per set of conditions; neurons counted: n=397 (media), 314 (NGF), 1028 (unpolarized), 789 (IL-4), 909 (G-CSF), 835 (IL-4+G-CSF). c, d, IL-4/G-CSF polarized or unpolarized CD34⁺ human BM cells were analyzed by flow cytometry. c, Representative pseudo-color density plots showing expression of CD34 and CD33, gated on viable CD45⁺CD3⁻CD19⁻ cells (left and middle panels). Histogram showing CD15 expression, gated on CD34⁺ cells or CD34⁻CD33⁺ cells, among the IL-4/G-CSF polarized human BM cells (right panel). d, Percentage of CD45⁺CD19⁻CD3⁻ cells that are CD33⁺ in each group (left panel). CD15 MFI, gating on CD33⁺ cells (right). (n=3 unique donors). Data shown are from one experiment, representative of two independent experiments. e, Human cortical neurons were co-cultured for 24 hours with either CD34⁺ or CD34⁻CD33⁺ cells, which were FACS sorted from IL-4/G-CSF polarized human BM cells. Additional neurons were cultured with recombinant NGF as a positive control, or in media alone as a negative control. Data are shown as the length of the longest neurite grown by the human cortical neurons. Each dot represents a single neuron. (n=2 unique donors; neurons counted: n=256 (media), 271 (NGF), 467 (CD34+), 598 (CD34−CD33+). Data are from one experiment, representative of two independent experiments. b, d, e, Data are presented as mean values ± SEM. Statistical significance was determined by one-way ANOVA followed by Dunnett’s post hoc test.