CCR5 Is a Therapeutic Target for Recovery after Stroke and Traumatic Brain Injury

Abstract

We tested a newly described molecular memory system, CCR5 signaling, for its role in recovery after stroke and traumatic brain injury (TBI). CCR5 is uniquely expressed in cortical neurons after stroke. Post-stroke neuronal knockdown of CCR5 in pre-motor cortex leads to early recovery of motor control. Recovery is associated with preservation of dendritic spines, new patterns of cortical projections to contralateral pre-motor cortex, and upregulation of CREB and DLK signaling. Administration of a clinically utilized FDA-approved CCR5 antagonist, devised for HIV treatment, produces similar effects on motor recovery post stroke and cognitive decline post TBI. Finally, in a large clinical cohort of stroke patients, carriers for a naturally occurring loss-of-function mutation in CCR5 (CCR5-Δ32) exhibited greater recovery of neurological impairments and cognitive function. In summary, CCR5 is a translational target for neural repair in stroke and TBI and the first reported gene associated with enhanced recovery in human stroke.

Keywords: MOCA; NIHSS; astrocyte; axon; axonal sprouting; dendritic spine; microglia; motor; premotor.

Conflict of interest statement

DECLARATION OF INTERESTS

The authors declare no competing interests.

Copyright © 2019 Elsevier Inc. All rights reserved.

Figures

Figure 1.

CCR5 Is Differentially Expressed in Neurons after a Cortical Stroke Data represent detection of CCR5 transcripts in sections of mouse cortex through FISH (A–E) and quantitative gene expression following FACS isolation (F and G). (A) Representative images from healthy mouse cortex show absence of co-localization of CCR5 transcripts (red, column 2) with TUBB3+ve (yellow) neurons (column 3). Insets represent digitally magnified field of view from the image. Inset in column 2 shows DAPI+ve (blue) nuclei and negligible CCR5 (red). Schematic on left represents location of field of view. Scale bar, 50 μm. (B) Microglia (CX3CR1+ve, yellow) co-localize with CCR5 (red) as seen in column 3 in healthy cortex. Scale bar, 50 μm. (C) CCR5 expression co-localizes with TUBB3+veneurons and non-neuronal DAPI+ve nuclei at 12 h after stroke. Scale bar, 100 μm. (D) Column 1 represents a projection image (3D visualization, Imaris) of neurons at higher magnification (TUBB3+ve, yellow; DAPI+ve, blue) that express CCR5 (red) at 12 h after stroke. Columns 2 and 3 represent images from column 1 that were processed with spot detection feature from Imaris (Bitplane software) as a visual aid for co-localization of TUBB3, CCR5 (column 2) and DAPI (column 3). Scale bar, 10 μm. (E) CCR5 expression at 7 days after stroke; additional images are in Figure S1. CCR5 is expressed in neurons at lower abundance in TUBB3+ neurons compared to 12 h and non-neuronal DAPI+ve cells. Columns 2 and 3 were processed similar to (D). Scale bar, 20 μm. (F) Representative dot plots from FACS show gating strategy used to isolate allophycocyanin (APC)-expressing NCAM+ve neurons isolated from peri-infarct cortex. Similar gating strategy was used for isolation of CD11b+ve cells. (G) Data on CCR5 gene expression quantified from FACS-isolated cells from naive and 12 h, 7 days,14 days, and 28 days post stroke. Gene expression data are from four different observations for NCAM pre-stroke, NCAM 12 h, CD11b pre-stroke, CD11b 12 h, and three different observations from all other groups. Data are mean ± SEM.

Figure 2.

CCR5 kd Induces Early and Sustained Motor Recovery after Stroke and Improves Cognitive Decline after TBI (A) Images represent expression of shCCR5 AAV in cortex at 9 days post injection in naive and post-stroke tissue. shCCR5 AAV targets neurons. Viral GFPexpression (green) co-localizes with NeuN (red) but does not co-localize with an astrocytic marker, GFAP (red) or microglial marker, IBA-1 (red). Asterisk denotes stroke site. Scale bar, 100 μm. (B) Schematic of experimental timeline. shCCR5AAV injected anterior to the prospective stroke site in pre-motor cortex at 3 days prior to stroke. This enables delivery of viral knockdown 3–5 days after stroke as quantified in (C). Motor performances were assessed at baseline and 1–9 weeks after stroke. (C) Quantification of CCR5 transcript expression in GFP+ve neurons FACS isolated from animals with shCCR5 AAV or control AAV at 4 days after stroke. Data show downregulation of CCR5 expression in neurons transduced with shCCR5AAV; p = 0.04. Data are mean ± SEM; n = 3. (D and E) Animals with neuronal knockdown of CCR5 (shCCR5AAV) show improved motor performances in gridwalk (D) and cylinder (E) tests at early and late time points after stroke. Data are mean ± SEM; n = 11 for stroke + control AAV and n = 10 per group for all other conditions. (F and G) Pharmacological inhibition of CCR5 with maraviroc produces motor recovery after stroke in gridwalking (F) and use of forelimb (G); n = 10 animals per group. Additional data are in Figure S2A. (H and I) Maraviroc produces sustained motor recovery even after treatment cessation. n = 10 animals per group for gridwalk (H); n = 9 stroke + maraviroc and n = 8 stroke + vehicle for cylinder test (I). (J and K) Administration of maraviroc 1 month after stroke produces motor recovery in chronic stroke as assessed in the grid walk test (J). No statistical differences observed in the cylinder test (K); n = 9 animals per group. (L and M) shCCR5AAV or pharmacological inhibition with maraviroc improves cognitive function after TBI (closed head injury [CHI]). Animals with neuronal CCR5 kd in hippocampal CA1 and CA3 (L) or treatment with maraviroc (M) show improved performances in the Barnes maze test. n = 6 for maraviroc; n = 8 for all other groups. (N and O) Treatment with shCCR5AAV (N) or maraviroc (O) leads to improved performances in the novel object-recognition task after CHI when compared to CHI + control AAV. Patterned bars indicate baseline performance prior to CHI. n = 6 for maraviroc; n = 8 for all other groups. Additional data presented in Figures S2B and S2C. For (F)–(N), data are mean ± SEM.

Figure 3.

CCR5 kd Produces Motor Recovery through CREB and DLK Signaling (A) Schematic on workflow for identification of downstream signaling targets. (B–G) Protein levels for potential downstream targets quantified from FACS-isolated neurons at 7 days post stroke. Neurons transduced with shCCR5AAV show a significant increase in CREB (p = 0.05), pCREB (B; p = 0.013), and DLK (C; p = 0.009) compared to neurons with control AAV after stroke. Protein levels for Erk-p42 p44 (D), GAP43 (E; p = 0.08), JNK1, JNK2 (F), and p38MAPK (G; p = 0.09) did not statistically differ between treatment and control groups; n = 4 stroke + shCCR5AAV and n = 5 stroke + control AAV. Data are mean ± SEM. (H) Representative images for pCREB immunostain in post-stroke tissue from animals with shCCR5AAV (top panel) or control AAV (bottom panel). Imaging parameters for pCREB photmicrographs were kept constant between groups. Neurons (NeuN, red) with shCCR5 AAV (green) show higher expression of pCREB (grayscale) compared to control AAV, compatible with protein expression data in (B). Scale bar, 10 μm. (I and J) Knockdown of DLK abrogates motor recovery induced through CCR5 kd. Data on motor performances assessed with gridwalk (I) and cylinder (J) tests. n = 8 control AAV + stroke, and n = 10 for all other groups. Data are mean ± SEM.

Figure 4.

CCR5 kd Reduces Dendritic Spine Loss and Increases Survival Fraction of Dendritic Spines in Pre-motor Cortex following Stroke (A) Representative images of dendritic spine dynamics after stroke, or stroke + treatment. Stroke causes spine loss (red arrows). Some spines re-emerge at 12 days after stroke (green arrows). Few new spines are added (yellow arrows). (B and C) Quantification of spine changes at 4 days (B) and 12 days (C) following stroke. n = 5 stroke alone, n = 4 stroke + shCCR5AAV, and n = 3 stroke + control AAV. Data are mean ± SEM.

Figure 5.. CCR5 kd Induces Axonal Sprouting in Contralateral Cortex

Maps represent Cartesian coordinates of BDA-labeled axons. Axes represent approximate stereotactic coordinates lateral and anterior to bregma. BDA injection placed in pre-motor cortex 10 weeks post stroke. BDA+ve projections from injection site mapped in ipsilateral (A and B) and contralateral (C and D) cortices. Blue represents projections from control condition, red from treated conditions, and royal blue from overlap between conditions. Polar plots on the right of each map aid in visualizing spatial differences between conditions plotted. n = 5 animals per group; Hoteling’s T2 test. Additional data presented in Figures S3 and S4. (A and B) BDA-labeled projections mapped in ipsilateral cortex. A large proportion of axons overlaps between stroke + shCCR5 AAV (red) and stroke + control AAV (sky blue) (A). Treatment from either condition did not produce unique projections; p = 0.062. Statistical non-significance confirmed with additional testing that showed that 95% prediction ellipses under bivariate normal distribution have 93.5% overlap, suggesting spatial distribution of axons did not differ in a statistically meaningful way. Treatment with maraviroc (B, red) resulted in a band of unique projections that occupied peri-infarct cortex compared with stroke + vehicle (sky blue; p = 0.001). (C and D) Cortical axons projecting from ipsilateral pre-motor cortex mapped in contralateral cortex. (E) Overview of (A) and (C) in mouse brain showing extent of axonal projections in contralateral hemisphere as a result of CCR5 kd.

Figure 6.

CCR5 kd Reduces Astrocyte Reactivity and Dampens Post-stroke Macrophage Recruitment (A) Images represent differential immunoreactivity for astrocytic marker GFAP and macrophage or microglial marker IBA-1 in different treatment conditions at 7 days after stroke. Astrocytic reactivity is notably reduced in animals with maraviroc or shCCR5 AAV treatment. Dotted line denotes infarct border; asterisk denotes infarct. Scale bar, 250 μm. (B) Filled contour plots show pixel intensity heatmap for quantification of changes in spatial reactivity across infarct to peri-infarct zones. Axes represent distance from midline in the medial-lateral (x axis) and dorsal-ventral (y axis) positions. See also data for overall pixel intensity in Figure S6. n = 5 stroke + shCCR5 AAV, stroke + maraviroc and n = 4 stroke alone, stroke + control AAV. (C–E) CCR5 kd dampens macrophage recruitment. Representative dot plots from FACS analysis show events gated for Ly6G (neutrophils; upper-right quadrant), Ly6Clow (macrophages; lower-right quadrant), and Ly6Chigh (reactive monocytes) from animals treated with stroke + control AAV or stroke + CCR5 AAV at 7 days post stroke quantified in (D) and stroke + maraviroc or stroke alone at 1 month post stroke quantified in (E); n = 4 stroke + control AAV and n = 5 for all other groups. Data are mean ± SEM.