Genetic deficiency of the miR-15a/16-1 cluster promotes long-term recovery of sensorimotor and cognitive functions in mice after TBI. The overall experimental design of this study is shown in Supplemental Figure 1 (supplemental material available online with this article; https://doi.org/10.1172/jci.insight.178650DS1). To explore the effect of miR-15a/16-1 genetic deletion on neurobehavioral function after TBI, miR-15a/16-1–KO and WT mice were subjected to a controlled cortical impact (CCI) (4.0 mm diameter; centered 0.5 mm anterior and 2.0 mm lateral to Bregma) or sham operation; then, rotarod, adhesive tape removal, and foot-fault tests were performed to evaluate the long-term sensorimotor function before and 3, 5, 7, 14, 21, and 28 days after the CCI operation. Compared with their corresponding sham controls, miR-15a/16-1–KO and WT mice exhibited severe sensorimotor deficits in the rotarod test (Figure 1A), adhesive tape–removal test (Figure 1, B and C), and foot-fault test (Figure 1, D and E) 3–28 days after TBI. In contrast, genetic deletion of the miR-15a/16-1 cluster significantly increased the staying time on the rod in the rotarod test (Figure 1A), decreased the time to touch or remove the tape in the adhesive tape–removal test (Figure 1, B and C), and reduced the foot-fault rate of the forepaw/hindpaw in the foot-fault test (Figure 1, D and E) 3–28 days after TBI when compared with WT mice, suggesting that miR-15a/16-1 genetic deficiency significantly improved posttrauma recovery of sensorimotor function. Of note, no significant differences were detected in sensorimotor function between miR-15a/16-1–KO and WT mice under sham conditions, as shown by similar performances in the rotarod test (Figure 1A), adhesive tape–removal test (Figure 1, B and C), and foot-fault test (Figure 1, D and E).

Genetic deletion of the miR-15a/16-1 cluster partially preserves sensorimotor and cognitive functions in mice after TBI. Experimental TBI was induced in miR-15a/16-1–KO and WT mice by unilateral controlled cortical impact, followed by a 30 days survival period. Long-term sensorimotor function was evaluated in TBI mice and sham controls at the indicated time points (–1, 3, 5, 7, 14, 21, and 28 days after operation). (A) Time to fall in the rotarod test. (B and C) The time to touch and time to remove the tape in the adhesive tape–removal test. (D and E) Representative forepaw/hindpaw foot-fault images and the forepaw/hindpaw foot-fault rate in the foot-fault test (red circle: foot fault). Long-term cognitive function was examined in TBI mice and sham controls by the Morris water maze (MWM) test (22–27 days) and the passive avoidance test (29–30 days). (F) Representative swimming track plots. (G) Latency to find platform in the learning phase. (H) Time spent in the target quadrant in the memory phase. (I) Average swimming speed in MWM test. (J) Graphic schedule of the passive avoidance test. (K) Latency to enter the dark box. Data are presented as mean ± SD, n = 10–12/group. Statistical analyses were performed by 1-way/2-way ANOVA and Tukey’s post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001 versus TBI + WT group.

To investigate whether genetic deletion of the miR-15a/16-1 cluster alters cognitive function after TBI, we performed the Morris water maze (MWM) test and the passive avoidance test in miR-15a/16-1–KO and WT mice after TBI. In the MWM test, TBI mice (miR-15a/16-1 KO and WT) exhibited longer time to find the platform (Figure 1, F and G; learning) and a shorter time spent in the target quadrant (Figure 1, F and H; memory) than their corresponding sham controls, indicating declined posttrauma learning and memory capabilities. In comparison with WT controls, miR-15a/16-1–KO mice were able to spend less time to find the platform (Figure 1, F and G) and spend more time in the target quadrant (Figure 1, F and H) after TBI, suggesting significantly improved learning and memory abilities. Interestingly, there was no significant difference in swimming speed among all experimental groups (Figure 1I), indicating that posttrauma learning and memory outcomes in the MWM test were not affected by swimming speed. The passive avoidance test was performed as shown in Figure 1J. Sham-operated mice seldom crossed over the door to the dark environment after aversive footshock stimuli, whereas the TBI mice in both miR-15a/16-1–KO and WT groups presented a shorter waiting time before entering the dark environment versus their corresponding sham controls, indicating impaired memory ability after brain trauma (Figure 1K). Compared with WT controls, miR-15a/16-1–KO mice had a longer waiting time in a bright environment after TBI, indicating better memory ability (Figure 1K). No significant difference was detected in cognitive function between miR-15a/16-1–KO and WT mice under sham conditions, as evidenced by similar performance patterns in the MWM test (Figure 1, F–I) and passive avoidance test (Figure 1, J and K). Taken together, these results suggest that genetic deletion of the miR-15a/16-1 cluster significantly promotes the recovery of sensorimotor and cognitive functions in mice after CCI-induced brain trauma.

Genetic deletion of the miR-15a/16-1 cluster reduces brain tissue loss and neuronal death in mice after TBI. Next, we examined the volume of brain tissue loss in miR-15a/16-1–KO and WT mice 30 days after TBI by using MAP2 IF staining (Figure 2A). As demonstrated in Figure 2, A–C, genetic deletion of the miR-15a/16-1 cluster in mice protects against trauma-induced brain damage, showing smaller volume or cross-sectional areas of brain tissue loss (slice 4, 6) than WT controls 30 days after TBI. Neuronal death in cortical and hippocampal regions was further examined by Cresyl violet (CV) histological staining and NeuN (a neuronal nuclear marker) immunostaining in miR-15a/16-1–KO and WT mice 30 days after TBI. CV/NeuN staining images were taken from the cerebral cortex (CTX), hippocampal CA1 (CA1), and CA3 regions as shown in Figure 2D. Compared with sham controls, fewer CV-stained (Figure 2, E–G) and NeuN+ (Figure 2, I–K) neurons were detected in different layers of the perilesional CTX and CA1 areas of TBI mice. However, genetic deletion of the miR-15a/16-1 cluster in mice dramatically reduced posttrauma neuronal death, showing more CV-stained (Figure 2, E–G) and NeuN+ (Figure 2, I–K) neurons in the perilesional CTX and CA1 regions when compared with WT controls. Of note, there are no significant differences in neuronal density in the perilesional CTX, CA1, and CA3 regions in miR-15a/16-1–KO and WT mice under sham conditions. Also, no obvious neuronal loss in the CA3 region was detected in the experimental groups (Figure 2, H and L). Moreover, we conducted Pearson correlation analyses and showed that neuronal densities in the perilesional CTX and CA1 regions, but not in the CA3 region, were positively correlated with sensorimotor and cognitive functions in all experimental mice (Figure 2M and Supplemental Tables 2 and 3). These results suggest that the neuronal numbers in the perilesional CTX and CA1 regions were highly associated with posttrauma sensorimotor and cognitive dysfunctions and improvements. A higher neuronal density in the perilesional CTX and CA1 regions of miR-15a/16-1–KO mice may contribute to better sensorimotor and cognitive recoveries after TBI compared with WT controls.

Genetic deficiency of the miR-15a/16–1 cluster reduces brain tissue loss and neuronal death in mice after TBI. MAP2 immunostaining was conducted in miR-15a/16-1–KO and WT mice 30 days after TBI to measure brain tissue loss. (A) Representative images of MAP2 immunostaining. Scale bar: 0.5 cm. (B and C) Quantitative analysis of volume or cross-sectional areas in brain tissue loss. Neuronal loss was examined by CV histological staining and NeuN immunofluorescence staining 30 days after TBI. (D) Coordinates and perilesional brain regions for CV and NeuN staining. (E) Representative images of CV staining in the pericontusional CTX, CA1, and CA3 regions. Scale bars: 100 μm. (F–H) Quantitative analysis of CV-stained cells in the pericontusional CTX, CA1, and CA3 regions. (I) Representative images of NeuN immunostaining in the pericontusional CTX, CA1, and CA3 regions. (J–L) Quantitative analysis of NeuN-immunopositive cells in the pericontusional CTX, CA1, and CA3 regions. Data are presented as mean ± SD, n = 6/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001 versus TBI + WT group. (M) Correlation analysis between sensorimotor or cognitive outcomes and CV-stained/NeuN+ neurons in the pericontusional CTX, CA1, and CA3 regions (n = 6/group, Pearson correlation analysis).

Genetic deletion of the miR-15a/16-1 cluster alleviates long-term white matter damage in mice after TBI. White matter lesions are considered a critical reason for cognitive impairment after TBI (4). Therefore, we examined white matter integrity by using Luxol fast blue (LFB) histological staining and myelin basic protein (MBP, myelin marker)/neurofilament heavy polypeptide (clone SMI32, injured axon marker) double-IF staining in the perilesional CTX, external capsule (EC), and striatum (STR) 30 days after TBI. As shown in Figure 3, the relative OD value of LFB (Figure 3, A–E) and the fluorescence intensity of MBP (Figure 3, F–I, and Supplemental Figure 2A) were significantly decreased in the perilesional CTX, EC, and STR 30 days after TBI in both miR-15a/16-1–KO and WT mice, indicating a significant loss of myelin in posttrauma brains compared with their corresponding controls. Additionally, the fluorescence intensity of SMI32 (Figure 3, J–L, and Supplemental Figure 2A) and the SMI32/MBP ratio (Supplemental Figure 2, B–D) were significantly increased in the perilesional CTX, EC, and STR 30 days after TBI, suggesting increased axonal damage in the posttrauma brains. Compared with WT controls, miR-15a/16-1–KO mice exhibited an increased OD value of LFB (Figure 3, A–E), increased MBP fluorescence intensity (Figure 3, F–I, and Supplemental Figure 2A), and decreased SMI32 fluorescence intensity and SMI32/MBP ratio (Figure 3, J–L, and Supplemental Figure 2, A–D) in the pericontusional CTX, EC, and STR, indicating that genetic deletion of the miR-15a/16-1 cluster effectively reduced white matter damage by alleviating the demyelination process and axonal damage in posttrauma brains. There were no significant differences in MBP, SMI32 fluorescence intensity, or SMI32/MBP ratio in the pericontusional CTX, EC, or STR in miR-15a/16-1–KO or WT mice under sham conditions.

miR-15a/16–1 genetic deletion partially preserves long-term white matter integrity in mice after TBI. Luxol fast blue (LFB) histological staining and MBP (green)/SMI32 (red) double-immunofluorescence staining were used to evaluate white matter integrity in miR-15a/16-1–KO and WT mice 30 days after TBI. (A) Representative images of LFB staining in the pericontusional CTX, EC, and STR regions. Scale bars: 100 μm. (B) Coordinates and brain regions for LFB, MBP/SMI32, and Caspr/Nav1.6 staining. (C–E) Quantitative analysis of relative OD values from LFB staining in the pericontusional CTX, EC, and STR regions. (F) Representative images of MBP/SMI32 immunostaining in the pericontusional CTX, EC, and STR regions. Scale bar: 50 μm. (G–I) Quantitative analysis of MBP fluorescence intensity in the pericontusional CTX, EC, and STR regions. (J–L) Quantitative analysis of SMI32 fluorescence intensity in the pericontusional CTX, EC, and STR regions. (M) The node of Ranvier (NOR) was examined by Caspr (red)/Nav1.6 (green) double-immunofluorescence staining. Representative images of Caspr/Nav1.6 immunostaining in the pericontusional EC area. Scale bars: 50 μm (top), 5 μm (bottom). (N) A representative image showing the composition and normal structure of the NOR. (O–P) Quantitative analysis of the number of NOR (O), paranode length (P), and the length of the paranode gap (Q). Data are presented as mean ± SD, n = 6/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001 versus TBI + WT group. (R) Correlation analysis of sensorimotor or cognitive outcome and white matter integrity (n = 6/group, Pearson correlation analysis).

The normal structure of the node of Ranvier (NOR) is essential to maintain signaling conduction along myelinated axons (19). Therefore, contactin-associated protein (Caspr, a paranodal cell adhesion protein)/Nav1.6 (a nodal sodium channel protein) dual immunostaining was performed to examine the structure of the NOR (Figure 3N) in the perilesional EC 30 days after TBI. As shown in Figure 3, M–Q, the number of NORs (Figure 3, M and O) and paranode length (Figure 3P) were decreased, and the length of the paranode gap was increased in posttrauma brains compared with sham controls. Of importance, genetic deletion of the miR-15a/16-1 cluster in mice caused a significantly higher amount of NORs and longer paranode length in the perilesional EC region after TBI than WT controls, implying improved function of the NOR (Figure 3, O and P). However, genetic deletion of the miR-15a/16-1 cluster failed to alter the paranode gap in posttrauma brains (Figure 3Q).

As demonstrated in Figure 3R and Supplemental Tables 4 and 5, further correlation analyses showed a positive/negative correlation between the structural parameters of white matter (LFB, MBP, SMI32, SMI32/MBP ratio, and Caspr/Nav1.6) and neurobehavioral outcomes (sensorimotor and cognitive functions) in all experimental mice; this indicated that posttrauma white matter structural injury or loss in mouse brains was highly associated with sensorimotor and cognitive impairments. Improved white matter integrity may causatively contribute to better neurobehavioral function in miR-15a/16-1–KO mice after TBI compared with WT controls.

Genetic deletion of the miR-15a/16-1 cluster decreases activation of astrocytes and microglia cells and decreases infiltration of blood-borne immune cells in mice after TBI. Since astrocytic activation is a symbolic event in posttrauma brains (2), we examined astrocytic activation in white/gray matter 3 days after TBI by using glial fibrillary acidic protein (GFAP) IF staining (Figure 4, A–C). We did not observe any significant differences in astrocytic activation in miR-15a/16-1–KO or WT mice under sham conditions. Compared with sham controls of 2 mouse strains, CCI induced considerable reactive astrocytes in the perilesional CTX, EC, and STR regions in miR-15a/16-1–KO and WT mice (Figure 4, A–F). However, significantly reduced GFAP+ astrocytes were detected in these brain regions of miR-15a/16-1–KO mice 3 days after TBI in comparison with WT controls, indicating that genetic deletion of the miR-15a/16-1 cluster dramatically reduces reactive astrocytes in posttrauma brains. Additionally, we examined the effect of miR-15a/16-1 on the long-term astrocytic activation after TBI. The results show that genetic deletion of miR-15a/16-1 reduced GFAP+ cells in the perilesional CTX, EC, and STR regions when compared with WT controls 30 days after TBI (Supplemental Figure 3, A–D).

miR-15a/16–1 genetic deletion alleviates TBI-induced glial activation and infiltration of peripheral immune cells in mouse brains. Experimental TBI was induced in miR-15a/16-1–KO and WT mice, and then astrocytic activation and microglial polarization were examined in brain sections at 3 days after surgery. (A–C) Representative images of GFAP (green, an astrocyte marker)/DAPI (blue) immunofluorescence staining in the perilesional CTX, EC, and STR regions. (D–F) Quantitative analysis of GFAP+ cells in the perilesional CTX, EC, and STR areas. (G and H) Representative images of Iba-1 (green)/CD16/32 (red) and Iba-1 (green)/CD206 (red) double-immunofluorescence staining in the perilesional brain regions. (I and J) Quantitative analysis of Iba-1+CD16/32+ and Iba-1+CD206+ cells. The infiltration of peripheral neutrophils and macrophages was examined in perilesional brain regions. (K and L) Representative images of NeuN (green)/Ly-6B (red) and F4/80 (green)/DAPI (blue) immunostaining in the perilesional brain regions. (M and N) Quantitative analysis of Ly-6B+ neutrophils and F4/80+ macrophages in perilesional brain regions. Data are presented as mean ± SD, n = 6/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s post hoc test. ***P < 0.001 versus TBI + WT group. Scale bars: 100 μm (top), 20 μm (bottom).

Next, we examined the effect of miR-15a/16-1 genetic deletion on microglial/macrophage polarization by coimmunostaining Iba-1 with either CD16/32 (a proinflammatory microglial/macrophage marker) or CD206 (an inflammatory-resolving microglial/macrophage marker) in mouse brain sections at 3 days after TBI (Figure 4, G and H). Increased numbers of Iba-1+CD16/32+ and Iba-1+CD206+ microglia/macrophages were observed in the perilesional brain regions of miR-15a/16-1–KO and WT mice 3 days after TBI compared with their corresponding sham controls (Figure 4, H–J). Intriguingly, miR-15a/16-1–KO mice exhibited significantly decreased proinflammatory microglia/macrophages and increased inflammatory-resolving microglia/macrophages 3 days after TBI in comparison with WT controls (Figure 4, H–J). These data suggest that miR-15a/16-1 genetic deletion in mice promotes cerebral microglia/macrophage polarizing from a proinflammatory phenotype to an inflammatory-resolving phenotype in acute stage of brain trauma. The long-term effects of miR-15a/16-1 on microglia/macrophage polarization were also examined. We showed that genetic deletion of miR-15a/16-1 reduced the Iba-1+CD16/32+ cells in perilesional brain regions when compared with WT controls (Supplemental Figure 3, E and G). However, no significant Iba-1+CD206+ cells were detected between miR-15a/16-1–KO mice and WT controls either before or 30 days after TBI (Supplemental Figure 3, F and H).

Blood-brain barrier disruption and resultant infiltration of peripheral immune cells into the parenchyma occur in the acute phase of TBI and exacerbate neuroinflammatory reactions in posttrauma brains (20). To examine whether genetic deletion of the miR-15a/16-1 cluster affects peripheral infiltration of neutrophils and macrophages to posttrauma brains, we conducted immunostaining of NeuN/Ly-6B (a neutrophil marker) and F4/80 (a marker relatively specific to monocyte-derived macrophages) in mouse brain sections at 3 days after TBI. As shown in Figure 4, K–N, CCI induced abundant Ly-6B+ neutrophils and F4/80+ macrophages in the pericontusional brain regions of miR-15a/16-1–KO and WT mice in comparison with their corresponding sham controls. In contrast, genetic deletion of the miR-15a/16-1 cluster significantly reduced infiltration of peripheral neutrophils and macrophages after TBI compared with WT controls (Figure 4, K–N). Taken together, these results indicate that genetic deletion of the miR-15a/16-1 cluster in mice alleviates the inflammatory cell infiltrate in posttrauma brains.

Intranasal delivery of the miR-15a/16-1 antagomir reduces TBI-induced sensorimotor and cognitive decline in mice. The deficiency of miR15a/16-1 does not completely prevent most of the endpoints assessed as compared with uninjured miR-15a/16-1–KO shams. However, based on the above-described findings on genetic deletion of miR-15a/16-1–mediated intermediate brain protection in mice after TBI, we further explored the potential therapeutic role of pharmacological inhibition of cerebral miR-15a/16-1 in TBI mice by intranasal administration of the miR-15a/16-1 antagomir. As shown in Supplemental Figure 4, we examined the nose-to-brain delivery efficiency of the miR-15a/16-1 antagomir. A total of 1 nmol miR-15a/16-1 antagomir was intranasally delivered to C57BL/6J mice 2 hours after TBI. The relative expression levels of miR-15a and miR-16-1 were measured in mouse brains before TBI and 1, 2, 3, 5, and 7 days after TBI. Compared with nontreated mice, we found that a single 1 nmol dose of the miR-15a/16-1 antagomir was able to significantly reduce cerebral expression of miR-15a and miR-16-1 up to 7 days after administration under sham conditions (Supplemental Figure 4, A–C). Under TBI conditions, the expression of miR-15a in mouse brains was also significantly decreased up to 7 days after intranasal delivery (Supplemental Figure 4D), whereas cerebral expression of miR-16-1 was significantly reduced at 1–3 days and was modestly inhibited at 5–7 days after intranasal delivery (Supplemental Figure 4E). Afterward, C57BL/6J mice were intranasally delivered with 1 nmol of miR-15a/16-1 antagomir or its control antagomir at 2 hours, 5 days, 10 days, 15 days, 20 days, and 25 days after TBI to evaluate the therapeutic effect of the miR-15a/16-1 antagomir on TBI (Supplemental Figure 5).

A panel of sensorimotor neurobehavioral tests (rotarod test, adhesive tape–removal test, and foot-fault test) were conducted in miR-15a/16-1 antagomir–treated and control antagomir–treated mice before TBI and 3, 5, 7, 14, 21, and 28 days after TBI. Both control- or antagomir-treated mice exhibited remarkable sensorimotor deficits 3–28 days after TBI in comparison with their corresponding sham controls (Figure 5, A–E). However, miR-15a/16-1 antagomir–treated mice had a prolonged staying time to fall in the rotarod test (Figure 5A), reduced time to touch (Figure 5B) or remove the tape (Figure 5C) in the adhesive tape–removal test, and declined forepaw/hindpaw foot-fault rates in the foot-fault test (Figure 5, D and E) when compared with control-treated mice, indicating that nose-to-brain delivery of the miR-15a/16-1 antagomir can effectively rescue TBI-caused sensorimotor deficits in mice. We also examined long-term cognitive function in the miR-15a/16-1 antagomir–treated and its control-treated mice by using the MWM test and passive avoidance test. In the MWM test (Figure 5, F–I), control antagomir–treated mice had declined spatial learning and memory functions after TBI when compared with sham controls. However, mice treated with the miR-15a/16-1 antagomir showed significantly improved spatial learning and memory functions in comparison with the control antagomir–treated mice, as evidenced by showing a shorter time to find the platform in the learning phase and a longer time to explore in the target quadrant in the memory phase. There were no significant differences in swimming speed among all experimental groups (Figure 5I). In addition, miR-15a/16-1 antagomir–treated mice also showed an increased latency to the dark box after TBI compared with control antagomir–treated mice in the passive avoidance test (Figure 5J). Taken together, these results suggest that intranasal delivery of the miR-15a/16-1 antagomir promotes recovery of sensorimotor and cognitive functions in mice after TBI. Interestingly, under sham conditions, both miR-15a/16-1 antagomir–treated and control antagomir–treated mice exhibited similar performances in all neurobehavioral tests.

Long-term intranasal delivery of the miR-15a/16–1 antagomir promotes neurobehavioral recovery and reduces brain tissue loss and neuronal death in mice after TBI. C57BL/6J mice were subjected to experimental TBI and intranasal treatment of the miR-15a/16-1 antagomir or control antagomir at 2 hours, 5 days, 10 days, 15 days, 20 days, and 25 days after CCI surgery. Sensorimotor function was examined in experimental mice at the indicated time points (–1, 3, 5, 7, 14, 21, and 28 days after surgery). (A) Time to fall in the rotarod test. (B and C) The time to touch and time to remove the tape in the adhesive tape–removal test. (D and E) The forepaw foot-fault rate and hindpaw foot-fault rate in the foot-fault test. Cognitive function was evaluated by the MWM test (22–27 days) and the passive avoidance test (29–30 days). (F) Representative swimming track plots. (G) Latency to find platform in the learning phase. (H) Time spent in the target quadrant in the memory phase. (I) Average swimming speed in the MWM test. (J) Latency to enter the dark box. Data are presented as mean ± SD, n = 10/group. Statistical analyses were performed by 1-way/2-way ANOVA and Tukey’s post hoc test. MAP2 and NeuN immunostaining were performed to examine brain tissue/neuronal loss. (K and L) Representative images of MAP2 immunostaining and quantitative analysis of brain tissue loss volume (n = 10/group, unpaired t test). (M) Representative images of NeuN immunostaining in the pericontusional CTX, CA1, and CA3 regions. Scale bar: 100 μm. (N–P) Quantitative analysis of NeuN+ cells in the pericontusional CTX, CA1, and CA3 regions (n = 6/group, 1-way ANOVA & Tukey’s test). *P < 0.05 and ***P < 0.001 versus TBI + control antagomir group. (Q) Correlation analysis between sensorimotor or cognitive outcome and CV-stained or NeuN+ neurons in the pericontusional CTX, CA1, and CA3 regions (n = 6/group, Pearson correlation analysis).

Intranasal delivery of miR-15a/16-1 antagomir reduces brain tissue loss and neuronal death in mice after TBI. We also examined brain tissue loss and neuronal death in miR-15a/16-1 antagomir or control antagomir–treated mice 30 days after TBI. MAP2 immunostaining (Figure 5K) showed that miR-15a/16-1 antagomir–treated mice exhibited smaller volume (Figure 5L) and cross-sectional areas (Supplemental Figure 6A; slice 1, 6) of brain tissue loss 30 days after TBI in comparison with control antagomir–treated mice, indicating that intranasal administration of the miR-15a/16-1 antagomir effectively reduced brain tissue loss in mice after TBI.

Neuronal density was further examined by CV staining (Supplemental Figure 6C) and NeuN immunostaining (Figure 5M) in miR-15a/16-1 antagomir and control antagomir–treated mice 30 days after TBI. Two groups of mice exhibited reduced CV-stained (Supplemental Figure 6, D and E) and NeuN+ (Figure 5, N and O) neurons in the perilesional CTX and CA1 compared with their corresponding sham controls. In contrast, miR-15a/16-1 antagomir treatment increased CV-stained and NeuN+ neurons in the pericontusional CTX and CA1 (Figure 5, N and O, and Supplemental Figure 6, D and E) and reduced neuronal death in mouse brains after TBI in comparison with control antagomir–treated mice. Mice with treatment of the miR-15a/16-1 antagomir showed no obvious changes in neuronal density of the CA3 region after TBI (Figure 5P and Supplemental Figure 6F). Both miR-15a/16-1 antagomir– and control antagomir–treated mice showed no changes in the neuronal density of CTX, CA1, and CA3 regions under sham conditions.

As demonstrated in Figure 5Q and Supplemental Tables 6 and 7, correlation analyses further show that neuronal densities in the pericontusional CTX and CA1 regions, but not in the CA3, were positively correlated with sensorimotor and cognitive functions, suggesting that neuronal numbers in the perilesional CTX and CA1 regions were highly associated with posttrauma sensorimotor and cognitive dysfunctions and improvements. Intranasal delivery of the miR-15a/16-1 antagomir reduced neuronal death in the pericontusional CTX and CA1 that may contribute to better sensorimotor and cognitive recovery after TBI.

Intranasal delivery of the miR-15a/16-1 antagomir alleviates long-term white matter injury in mice after TBI. Moreover, we examined white matter integrity in miR-15a/16-1 antagomir–treated mice 30 days after TBI. LFB histological staining and MBP/SMI32 dual immunostaining were conducted to examine white matter integrity in the pericontusional CTX, EC, and STR regions (Figure 6B). After TBI, both control- and antagomir-treated mice exhibited severe myelin loss when compared with their corresponding sham controls, showing a reduced LFB OD value and MBP fluorescence intensity in the pericontusional CTX, EC, and STR regions (Figure 6, A and C–I). However, intranasal delivery of the miR-15a/16-1 antagomir significantly reduced myelin loss after TBI in comparison with control antagomir–treated mice, showing both an enhanced LFB OD value and MBP fluorescence intensity in these brain regions (Figure 6, A and C–I). Additionally, control antagomir–treated mice exhibited an increased fluorescence intensity of SMI32 (Figure 6, F–L, and Supplemental Figure 7A) and the SMI32/MBP ratio (Supplemental Figure 7, B–D) in pericontusional CTX, EC, and STR 30 days after TBI, suggesting increased axonal damage in the posttrauma brains. In contrast, miR-15a/16-1 antagomir–treated mice showed decreased SMI32 fluorescence intensity and an SMI32/MBP ratio (Figure 6, J–L, and Supplemental Figure 7, A–D) in the pericontusional CTX, EC, and STR regions, suggesting that treatment with the miR-15a/16-1 antagomir alleviates axonal damage in mouse brains after TBI.

Long-term intranasal delivery of the miR-15a/16–1 antagomir partially preserves white matter integrity in mice after TBI. C57BL/6J mice were subjected to experimental TBI and intranasally treated with the miR-15a/16-1 antagomir or control antagomir at 2 hours, 5 days, 10 days, 15 days, 20 days, and 25 days after CCI surgery. White matter integrity was examined by LFB staining and MBP/SMI32 immunostaining 30 days after surgery. (A) Representative images of LFB staining in the perilesional CTX, EC, and STR regions. Scale bar: 100 μm. (B) Coordinates and brain regions for LFB, MBP/SMI32, and Caspr/Nav1.6 staining. (C–E) Quantitative analysis of relative OD values from LFB staining in the pericontusional CTX, EC, and STR regions. (F) Representative images of MBP/SMI32 immunostaining in the pericontusional CTX, EC, and STR regions. Scale bar: 50 μm. (G–I) Quantitative analysis of MBP fluorescence intensity in the pericontusional CTX, EC, and STR regions. (J–L) Quantitative analysis of SMI32 fluorescence intensity in the pericontusional CTX, EC, and STR regions. (M) The node of Ranvier (NOR) was examined by Caspr (red)/Nav1.6 (green) double-immunofluorescence staining. Representative images of Caspr/Nav1.6 immunostaining in the pericontusional EC area. Scale bars: 50 μm (top), 5 μm (bottom). (N–P) Quantitative analysis of the number of NOR (N), paranode length (O), and the length of the paranode gap (P). Data are presented as mean ± SD, n = 6/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001 versus TBI + control antagomir group. (Q) Correlation analysis of sensorimotor or cognitive outcome and white matter integrity (n = 6/group, Pearson correlation analysis).

We further examined the structure of the NOR by coimmunostaining Caspr with Nav1.6 in miR-15a/16-1 antagomir–treated mice 30 days after TBI. Our results document that the number of NORs (Figure 6, M and N) and paranode length (Figure 6O) in the pericontusional EC region decreased in both control- and antagomir-treated TBI brains compared with their corresponding sham controls. However, miR-15a/16-1 antagomir–treated mice showed a significantly higher number of NORs and longer paranode length after TBI than control antagomir–treated mice, implying improved function for the NOR (Figure 3, N and O) in posttrauma brains after miR-15a/16-1 inhibition. No significant differences in the length of the paranode gap were found among all experimental groups (Figure 6P).

As demonstrated in Figure 6Q and Supplemental Tables 8 and 9, further correlation analyses showed a positive/negative correlation between the structural parameters of white matter (LFB, MBP, SMI32, SMI32/MBP ratio, and Caspr/Nav1.6) and neurobehavioral outcomes (sensorimotor and cognitive functions) in all experimental mice. This indicates that greater white matter integrity in miR-15a/16-1 antagomir–treated mice may causatively contribute to the improved neurobehavioral function after TBI.

Intranasal delivery of the miR-15a/16-1 antagomir reduces the activation of astrocytes and microglia cells and reduces infiltration of blood-borne immune cells in mice after TBI. We further tested whether the miR-15a/16-1 antagomir affects activation of astrocytes in mouse brains after TBI. As demonstrated in Figure 7, A–C, both control- and antagomir-treated mice showed a higher percentage of GFAP+ reactive astrocytes in the perilesional CTX, EC, and STR regions 3 days after TBI compared with their corresponding sham controls. However, treatment with the miR-15a/16-1 antagomir significantly reduced the percentage of reactive astrocytes in the perilesional CTX, EC, and STR regions following TBI (Figure 7, D–F).

miR-15a/16–1 antagomir alleviates TBI-induced glial activation and infiltration of peripheral immune cells in mouse brains. C57BL/6J mice were subjected to experimental TBI and intranasally treated with the miR-15a/16-1 antagomir or control antagomir at 2 hours after surgery. Astrocytic activation and microglial polarization were examined in brain sections at 3 days after CCI surgery. (A–C) Representative images of GFAP (green, an astrocyte marker)/DAPI (blue) immunofluorescence staining in the perilesional CTX, EC, and STR regions. (D–F) Quantitative analysis of GFAP+ cells in the perilesional CTX, EC, and STR areas. (G and H) Representative images of Iba-1 (green)/CD16/32 (red) and Iba-1 (green)/CD206 (red) double-immunofluorescence staining in the perilesional brain regions. (I and J) Quantitative analysis of Iba-1+CD16/32+ and Iba-1+CD206+ cells. (K and L) The infiltration of peripheral neutrophils and macrophages was examined in perilesional brain regions. Representative images of NeuN (green)/Ly-6B (red) and F4/80 (green)/DAPI (blue) immunostaining in perilesional brain regions. (M and N) Quantitative analysis of Ly-6B+ neutrophils and F4/80+ macrophages in perilesional brain regions. Data are presented as mean ± SD, n = 6/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s post hoc test. ***P < 0.001 versus TBI + control antagomir group. Scale bars: 100 μm (top), 20 μm (bottom).

Next, we examined the effect of the miR-15a/16-1 antagomir on microglia/macrophage polarization by coimmunostaining Iba-1 with either CD16/32 or CD206 in mouse brain sections at 3 days after TBI (Figure 7, G and H). Both control- and antagomir-treated mice showed increased numbers of Iba-1+CD16/32+ and Iba-1+CD206+ microglia/macrophages in the perilesional brain regions 3 days after TBI compared with sham controls of 2 treatment groups (Figure 7, I and J). However, miR-15a/16-1 antagomir–treated mice exhibited significantly decreased proinflammatory microglia/macrophages and increased inflammatory-resolving microglia/macrophages (Figure 7, I and J) after TBI. These data indicate that the miR-15a/16-1 antagomir promotes posttrauma cerebral microglia/macrophage polarization from a proinflammatory phenotype to an inflammatory-resolving phenotype.

In addition, we also conducted immunostaining of NeuN/Ly-6B and F4/80 in mouse brain sections from miR-15a/16-1 antagomir–treated mice 3 days after TBI. As demonstrated in Figure 7, K–N, both control- and antagomir-treated TBI mice showed an abundance of Ly-6B+ neutrophils and F4/80+ macrophages in the perilesional brain regions in comparison with sham controls of 2 treatment groups. In contrast, treatment with the miR-15a/16-1 antagomir in mice significantly reduced the infiltration of peripheral neutrophils and macrophages after TBI (Figure 7, M and N). Taken together, these results indicate that intranasal delivery of the miR-15a/16-1 antagomir alleviates inflammation-mediating cells in posttrauma brains.

Genetic deletion and pharmacological inhibition of miR-15a/16-1 function reduce brain inflammatory mediators in mice after TBI. To further investigate the underlying mechanism of brain protective effects conferred by either genetic deletion or pharmacological inhibition of miR-15a/16-1 function after TBI, a panel of 40 inflammatory mediators was examined by using a commercial inflammatory antibody array kit (Supplemental Figure 8). Compared with sham controls, we found that 27 inflammatory mediators were significantly elevated in the perilesional brain regions 3 days after TBI (Figure 8, A–E). Genetic deletion of the miR-15a/16-1 cluster significantly reduced the protein levels of 22 of these cerebral inflammatory mediators in mice 3 days after TBI compared with WT controls (Figure 8, B–E). On the other hand, altered cerebral inflammatory mediators were also found in miR-15a/16-1 antagomir–treated mice 3 days after TBI. We uncovered 32 significantly elevated cerebral inflammatory mediators in mice 3 days after TBI compared with sham controls, and treatment with the miR-15a/16-1 antagomir dramatically reduced 28 of the inflammatory mediators (Figure 8, F–J). Among these cerebral inflammatory mediators, CD30L, granulocyte CSF (G-CSF), granulocyte/macrophage CSF (GM-CSF), IFN-γ, IL-1β, IL-2, IL-10, IL-12-p40/p70, IL-13, I-TAC, KC, Leptin, MIP-1α, MIP-1γ, Stromal cell–derived factor-1 (SDF-1), TNF-α, and soluble tumor necrosis factor receptor (sTNF RII) are suppressed by both genetic deletion and antagomir inhibition of miR-15a/16-1 function in mice after TBI.

Genetic deletion and pharmacological inhibition of miR-15a/16–1 function reduces the inflammatory burden in posttrauma brains. (A–E) miR-15a/16-1–KO and WT mice were subjected to experimental TBI or sham operation. A panel of 40 inflammatory mediators was examined in mouse brains at 3 days after CCI surgery. (A) A heatmap showing the mean expression levels of 40 inflammatory mediators in posttrauma brains from experimental groups. (B–E) Quantitative analysis of 40 brain inflammatory mediators in experimental groups. Data are presented as mean ± SD, n = 4/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s test or Kruskal-Wallis test & Dunn’s test. &,*P < 0.05, &&,**P < 0.01, and &&&,***P < 0.001, where & represents TBI + WT versus Sham + WT group and * represents TBI + KO versus TBI + WT group. (F–J)C57BL/6J mice were subjected to experimental TBI and intranasally treated with the miR-15a/16–1 antagomir or control antagomir at 2 hours after surgery. A panel of 40 inflammatory mediators was measured in mouse brains at 3 days after CCI surgery. (F) A heatmap showing the mean expression levels of 40 inflammatory mediators in posttrauma brains from experimental groups. (G–J) Quantitative analysis of 40 brain inflammatory mediators in experimental groups. Data are presented as mean ± SD, n = 4/group. Statistical analyses were performed by 1-way ANOVA and Tukey’s test or Kruskal-Wallis test and Dunn’s test. &,*P < 0.05, &&,**P < 0.01, and &&&,***P < 0.001 where & represents TBI + control antagomir versus Sham + control antagomir group and * represents TBI + miR-15a/16-1 antagomir versus TBI + control antagomir group.

Taken together, although TBI-induced injuries persist in both miR-15a/16-1–KO and antagomir-treated mice, the in vivo loss of miR-15a/16-1 function partially reduced the inflammatory burden in TBI brains via suppressing various cerebral inflammatory mediators. These downregulated cerebral inflammatory mediators might be the downstream targets of the miR-15a/16-1 cluster and may contribute to the improved histological and functional outcomes found in either miR-15a/16-1–KO mice or miR-15a/16-1 antagomir–treated mice after TBI.