Interleukin‐1β influences functional regeneration following nerve injury in mice through nuclear factor‐κB signaling pathway

Summary

This study focuses on investigating the role of interleukin‐1β (IL‐1β) in functional regeneration following nerve injury in mice. A microarray‐based mRNA profiling study was used to analyze the expression level of IL‐1β in peripheral nerve regeneration. Quantitative real‐time polymerase chain reaction and Western blot were applied to assess the IL‐1β expressions of C57BL/6J‐crush and C57BL/6J‐crush+IL‐1β mice at different post‐injury time‐points after the standard sciatic nerve crush injury. The outcomes of nerve regeneration were evaluated by behavioral tests. IL‐1β was found to be up‐regulated in peripheral nerve regeneration and significantly raised on the 3rd day and returned to normal levels on the 14th day after nerve injury. Compared with C57BL/6J‐crush+IL‐1β mice, the nerve regeneration of C57BL/6J‐crush mice was worse after nerve crush injury. IL‐1β increased mechanical sensitivity and stimulated amplitude. IL‐1β could benefit the recovery of sciatic nerve crush injury by facilitating nerve regeneration.

Abbreviations

GEO

Gene Expression Omnibus

GO

Gene Ontology

IL‐1β

interleukin‐1β

KO

knockout

NF‐κB

nuclear factor κB

PCA

principal component analysis

RT‐qPCR

real‐time quantitative PCR

SD

standard deviation

SFI

sciatic functional index

Introduction

The nervous system is complex but of great importance to the human body as it controls behaviors. Therefore, any damage from physical trauma or disease‐related degeneration may easily cause neural injury and lead to serious dysfunction in sensation, movement, and even worse, thinking and living.1 Neural repair and regeneration are essential for treatment after neural injury.

Many factors and pathways have been reported to have connection with neural regeneration, including immune cells, the inflammasome, and particular cytokines. Some experiments have shown that immune cells can affect synapses and damage neuronal survival,2 and inflammasomes like AIM2 could mediate the death of nerve cells.3 Tumor necrosis factor‐α (TNF‐α) is thought to have protective effects on neurons,4 and interleukin‐1β (IL‐1β) is a pro‐inflammatory cytokine that strongly affects immunity or other physiological activities. It has been proved that IL‐1β can cause painful symptoms and induce cell cycle arrest and apoptosis in neural precursor cells.4, 5, 6 Temporin et al. reported that IL‐1β could promote regeneration of sensory nerves after sciatic nerve injury,5 and they later showed that it was through the p38 mitogen‐activated protein kniase (MAPK) pathway that IL‐1β promoted neurite outgrowth.6 Nadeau et al.7 discovered the co‐function of IL‐1β and TNF‐α in injury recovery but the underlying mechanisms remained unknown. Wang et al.8 noted that the multipotent neural precursor cells were influenced by the stress‐activated kinase/c‐jun N‐terminal kinases (SAPK/JNK) pathway activated by IL‐1β, proposing an involvement of IL‐1β in neural regeneration.

The nuclear factor κB (NF‐κB) and related NF‐κB signaling pathway regulate abundant number of genes by influencing their transcription. In early studies, it had already been noted that NF‐κB played multiple roles in the central nervous system,9 varying from inflammatory control to cell death inducer, and more importantly, it was also a vital part of neural recovery, showing its dual effect on cells. Similar to IL‐1β, NF‐κB shared many of the same functions in neural injuries. For example, the research by Armstrong et al.10 showed its ability to enhance neurite outgrowth in Schwann cells. They found that after NF‐κB inhibition, the neurites of Schwann cells were not only fewer in number but also shorter compared with the control group, indicating that NF‐κB strongly promoted the growth of neurites. Moreover, Zhang et al.11 discovered that the NF‐κB signaling pathway was able to induce the early differentiation of neural stem cells, which was of great importance for repairing and regenerating nervous tissues. However, it was not only the NF‐κB signaling pathway that was implicated in the neural regeneration process. Temporin et al. mentioned the p38 MAPK pathway instead of the NF‐κB signaling pathway as the mediator of IL‐1β‐induced neurite extension. Both the PI3K/Akt pathway and the Toll‐like receptor signaling pathway were also thought to be associated with neural repair and regeneration,12, 13 making it a complicated network.

In this study, bio‐informatics analysis not only showed NF‐κB to be the most neural‐repair‐related pathway but also revealed IL‐1β as a differentially expressed protein in peripheral nerve injured mice, and its expression was dynamic during neural regeneration. Hence, we investigated the role of IL‐1β in the process of neural functional regeneration.

Methods

Bioinformatics analysis

Affymetrix microarray platform GPL21035 and microarray data GSE74087 used for validation were obtained from Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/). This data set included 12 tissue specimens, including four groups randomly depending on the time‐point of 0, 3, 7, and 14 days after peripheral nerve injury. R language and KEGG database were used for analyzing the differentially expressed genes and dysregulated pathways. The threshold used to screen up‐regulated and down‐regulated mRNA in analysis of variance (anova) between four groups was P < 0·05. The thresholds used in two different group comparisons were log₂ (FC) > 1 and log₂ (FC) < −1 (P < 0·05), respectively.

Gene ontology classification analysis

David database (https://david-d.ncifcrf.gov/) was used to explore more biological knowledge, especially the biologically rich information of biased genes. We screened out 181 biased genes in 3 days post‐crush samples compared with 0 days post‐crush samples with |log₂foldchange| > 4·5 and adjust P < 0·001 and 194 biased genes in 14 days post‐crush samples compared with 3 days post‐crush samples with |log₂foldchange| > 3 and adjust P < 0·05. We submitted these two gene lists respectively and analyzed both of them under Mus musculus background for functional annotation and gene ontology (GO) classification biological process (BP), cellular composition (CC), molecular function (MF). Graphical displays based on analysis results were implemented by the goplot package in R (http://wencke.github.io/). Fifteen GO terms with lowest adjust P were selected, five terms each for MF, CC and BP results. goplot integrated and visualized the information from the 15 GO terms for plotting high‐quality charts.

Animals

Interleukin‐1β knockout female mice, on a C57BL/6J background (obtained by Prof. Fabio Martinon, Epalinges, Switzerland), were compared with wild‐type littermates. Body weight, fertility and viability were similar among different genotypes. Female mice aged 8–12 weeks and weighing 20–30 g were used. The housing and surgical conditions strictly followed national and institutional guidelines. The facilities for housing, surgery, analgesia and assessment were specific pathogen‐free and accredited by Association for Assessment and Accreditation of Laboratory Animal Care. Animal protocols were approved by the Shanghai Jiao Tong University Affiliated Sixth People's Hospital.

Animal surgery

By injecting ketamine (25 mg/kg) and xylazine (50 mg/kg), the mice were anesthetized before surgery (day 0). During the surgery, a consistent pressure was applied to the right sciatic nerve at the mid‐thigh level using No. 5 Jeweler forceps for 30 seconds. The injured site was marked by a 10‐0 Ethilon suture (Micro suture Ethicon, Somerville, NJ) which passed through the epineurium, but did not constrict. For the sham‐operated mice, the nerve crush surgery was not conducted. but the corresponding sites of sham‐operated mice were marked by an epineurial suture. After the surgery, the mice were woken and kept in the postoperative care room separately for recovery.

Functional assessment of sciatic functional index

At 0, 3, 7, and 14 days after surgery, walking track analysis was performed based on Benitez et al.14 The lengths of the third toe to its heel (PL), the first to the fifth toe (TS), and the second toe to the fourth toe (IT) were measured on the experimental side (E) and the contralateral normal side (N) in each rat. The Sciatic Function Index (SFI) in each animal was calculated using the following formula: SFI = −38·3 × (EPL–NPL)/NPL + 109·5 × (ETS–NTS)/NTS + 13·3 × (EIT–NIT)/NIT–8·8.

Behavioral tests

Behavioral tests were conducted on mice before nerve crush or IL‐1β injection to provide a baseline. After nerve crush surgery and IL‐1β injection, the behavioral tests were conducted again at the selected times. The behavioral tests were designed as blind tests to the people who conducted them.

Mechanical sensitivity

The tactile sensitivity response was used and measured as the direct pressure stimulus required to elicit foot withdrawal in non‐restrained conditions. Mice subjected to surgery were tested in the morning (08.30–11.30 h) whereas those subjected to intrathecal injection were tested during the day (08.30–17.00 h). Forty minutes before the test, the mice were first habituated to the testing apparatus. Then, they were stimulated by being pressed with a 2 von Frey filament (0·02 g; Stoelting, Wood Dale, IL) to the point of 30° bending against the plantar surface of the ipsilateral hind paw. This stimulation was for the mechanical allodynia, which increased their response to a non‐noxious stimulus. Each filament was used for three sets of test, with 10 stimulations in each set. The interval of each set of stimulations was approximately 10 min so that the possible sensitization was decreased. The numbers of paw withdrawals out of 30 stimuli were recorded as allodynia. The mechanical sensitivity of each mouse was calculated as the 50% threshold force needed for paw withdrawal.

Thermal sensitivity

The thermal hyperalgesia of the mice was determined by Hargreaves thermal hyperalgesia test and tail‐flick test after the mechanical allodynia. Before the Hargreaves thermal hyperalgesia test, the mice were accustomed to the testing chambers with the PAW Thermal Stimulator System 106 (UCSD, San Diego, CA) for 40–60 min. In the test, the paw withdrawal of mice in response to radiant heat generated by the thermal system was recorded as the thermal nociceptive latency. After the Hargreaves thermal hyperalgesia test, the mice were then subjected to the tail‐flick test after 30‐min rest. Two‐thirds of their tails were immersed in hot water (49°) to measure the thermal nociception, which was defined as the latency of tail withdrawal from the hot water. For each mouse, we recorded three sets of paw withdrawal and tail‐flick latency data to calculate the average. In general, the Hargreaves test was used to determine the thermal sensitivity of the hind paw area of unrestrained animals, which were affected by the injury to the spinal nerve, whereas the tail‐flick test was used to determine the general thermal sensitivity of the animal.

Intrathecal administration of IL‐1β

The mice was anesthetized using isoflurane (induced by 4% isoflurane and maintained by 2% isoflurane) in 100% O₂. The intrathecal injection was performed with a Hamilton syringe (10 μl) attached to a 30G needle, which was inserted into the distal sciatic nerve under the epineurium. The injected solution was either (i) 500 ng of soluble IL‐1β in the form of recombinant mouse IL1β protein (ab219437; Abcam, Cambridge, MA) or (ii) 500 ng of bovine serum albumin control (Sigma‐Aldrich, St. Louis, MO). They were prepared in sterile phosphate‐buffered saline and the injection volume was 5 μl. The experimenter who conducted the injection was blinded to the type of injected solutions. Mice were allowed to recover from anesthesia (mice were usually woken within 5 min post‐injection) and then subjected to a von Frey filament test for mechanical sensitivity post‐injection.

Western blot analysis

After 0, 3, 7, and 14 days of the nerve crush injury, the distal nerve specimens of C57BL/6 mice were harvested and digested with T‐PER™ reagent (Pierce, Waltham, MA) with phosphatase and protease inhibitors to extract tissue proteins. Then, 30 μg of proteins was resolved on 10% sodium dodecyl sulfate–polyacrylamide gels. After being transferred to polyvinylidene difluoride membranes, they were blocked with 5% skim milk (supplemented with Tween‐20 and phosphate‐buffered saline) and then incubated with primary antibodies, that is, rabbit anti‐IL‐1β (ab9722; Abcam), rabbit anti‐GAPDH (ab181602; Abcam), anti‐TAK1 antibody (EPR5984) (ab109526; Abcam), anti‐NF‐κB p65 antibody (ab16502; Abcam), anti‐TNF‐α antibody (ab6671; Abcam) at 4° overnight. Further incubation was performed with goat anti‐rabbit IgG H&L (horseradish peroxidase) (ab205718; Abcam) secondary antibodies for 60 min. Finally, the blots were developed with ECL™ Western Blotting Systems (Amersham Pharmacia Biotech, Little Chalfont, UK). The FluorChem 8900 imaging system and the alphaeasefc software (Alpha Innotech Corp., Santa Clara, CA) were applied to detect and quantify the protein bands, respectively.

Real‐time quantitative polymerase chain reaction

Tissue perfusion was performed to remove the blood in the sciatic nerve. Distal segments of crushed sciatic nerves (1 cm) were isolated and harvested from the mice of each group 0, 3, 7, and 14 days post‐surgery. The total RNA of tissue samples was isolated by the RNeasy Mini kit (Qiagen, Duesseldorf, Germany) and reverse transcriptase by the High‐Capacity cDNA Reverse Transcription (ABI 4368814; Applied Biosystems, Foster City, CA). The Applied Biosystems^® 7500 Real‐Time PCR System (Applied Biosystems) was used to perform real‐time quantitative polymerase chain reaction (RT‐qPCR) with the primers designed by Qiagen (Tables 1 and 2). The amplification was carried out by the Power SYBR Green PCR Master Mix (ABI 4367659; Applied Biosystems). The cycling steps were as follows: reverse transcription (30 min, 50°), denaturation (2 min, 95°), and 40 amplification cycles (15 seconds at 95° and 1 min at 60°). The mRNA levels were calculated by 2^−ΔΔCt with GAPDH as internal control.

Table 1.

Pathway expression (days 0 and 3)

Gene	Day 0			Day 3
	1	2	3	1	2	3
Cxcl2	5·134298	5·839012	5·149091	10·89226	9·74502	12·58793
Ccl4	6·874272	7·024928	6·892002	11·25103	11·04391	11·64885
Tnfsf14	5·36769	5·540975	5·677262	9·492923	9·22511	9·275947
Tnf	6·034796	6·597973	6·468856	9·651752	9·581875	10·35747
Il1b	4·979124	5·306719	4·879144	8·595093	7·650094	9·10135
Bcl2a1c	2·800707	3·167893	2·54516	6·433361	6·677304	5·099856
Bcl2a1a	9·636209	9·784777	9·665188	12·84726	13·00328	12·62412
Bcl2a1d	10·1271	10·23648	10·10738	13·28743	13·44154	13·08737
Card11	2·233452	2·079174	2·089147	5·048716	5·830245	4·643776
Gadd45b	10·25238	10·19241	10·50927	13·26365	13·12494	13·43197
Pidd1	5·564383	5·559582	5·375891	7·614161	7·70655	7·973311
Syk	9·774614	9·792648	9·928868	12·03361	12·24072	11·66015
Btk	5·035231	5·132156	6·033528	7·546746	7·739262	7·178598
Blnk	10·05096	10·20921	10·67543	12·20863	12·52732	11·95832
Cd40	7·527667	7·753478	7·577158	9·709321	9·887747	8·922036
Ptgs2	7·10643	7·380504	6·299036	8·591159	8·827305	8·577044
Lyn	6·328576	6·651527	6·643281	8·319334	8·430282	7·960258
Plcg2	11·02271	10·94715	11·02774	12·70707	12·72997	12·38605
Ltb	6·931468	7·009225	6·99311	8·375578	8·952195	8·325581
Tnfsf11	1·767861	1·722215	1·941551	3·136357	3·214639	3·755228
Lck	5·685245	5·661881	5·742423	7·070881	7·657104	6·99578
Prkcb	8·959749	9·226438	9·434144	10·80223	10·87275	10·27634
Cd14	11·81686	11·86428	11·94824	13·39377	13·14985	13·14529
Tnfrsf11a	8·44293	8·687594	8·802733	10·06444	10·11274	9·746023
Ticam2	4·322858	4·101683	4·520437	5·713049	5·732997	5·184902
Tlr4	7·077783	7·280162	7·684455	8·455549	8·605726	8·566688
Bcl2	6·742627	7·108817	7·068429	8·186391	7·796517	8·337928
Lat	7·956663	8·133876	8·218188	9·065353	9·443107	9·155
Tradd	11·41899	11·47056	11·52607	12·56027	12·50743	12·63184
Il1r1	10·76802	10·69712	10·67846	11·95484	11·77678	11·54019
Traf1	8·390327	8·214742	8·555956	7·302961	7·533649	7·23255
Card10	12·40544	12·41198	12·52367	11·15928	11·33003	10·95238
Cxcl12	13·52992	13·40028	13·25269	12·29643	12·03642	11·87155
Lbp	11·65036	11·54983	11·47157	10·3194	10·39067	8·386344
Prkcq	11·2049	11·26783	11·10364	8·915763	8·831329	9·055496
Ccl21a	14·72317	15·66212	15·27306	11·58151	12·95066	10·29879

Table 2.

Pathway expression (days 3 and 14)

Gene	Day 3			Day 14
	1	2	3	1	2	3
Cxcl2	10·89866	9·742982	12·59512	6·557565	7·239256	7·720634
Ptgs2	8·602508	8·836594	8·58808	5·879209	5·555844	5·464685
Gadd45b	13·27184	13·13403	13·44215	10·41918	10·4322	10·37902
Il1b	11·04193	9·904755	11·55568	7·660795	7·877797	8·519448
Cd14	13·40375	13·1583	13·15386	11·61989	11·86109	11·66754
Bcl2 l1	14·18724	14·26581	14·2746	12·96643	13·02177	13·06578
Btk	7·588307	7·771983	7·228397	6·493622	6·317517	6·493622
Ube2i	8·167091	8·038935	7·549398	6·902843	6·705478	6·868959
Tnfrsf11a	10·0641	10·11311	9·743717	8·792222	8·935765	8·997525
Ikbkg	12·05114	11·94331	12·05258	10·91917	10·96052	10·9766
Cd40	9·616263	9·833286	8·748529	8·466914	8·344889	8·26084
Csnk2a2	4·727717	3·888681	4·582063	5·437501	5·002944	5·863331
Card11	5·10551	5·879209	4·691102	6·548549	6·28516	6·100251
Prkcq	8·922851	8·840042	9·060908	10·07375	9·882413	10·13586
Tnfrsf13c	3·940237	4·045245	3·208267	5·388088	5·916598	5·974206

Statistical analysis

All quantitative values were presented as the mean ± standard deviation (SD) of at least three repeated individual experiments. Unpaired Student's t test was used to achieve differences between groups, and the differences among the groups of samples were accomplished by one‐way anova. All the statistical analyses were made using graphpad prism v6.0 (GraphPad Software, Inc., San Diego, California). A value of P < 0·05 was considered statistically significant.

Results

Differential genes were found by microarray analysis

Through principal component analysis, we implemented visual clustering analysis on samples for these four groups respectively to check the data quality for each group (Fig. 1a). anova results visualized by heat map of 12 samples from four groups demonstrated that there was a significant difference between 0 days post‐crush compared with 3 days post‐crush samples, and the expression in the 14 days post‐crush group was similar to the 0 days post‐crush group. This led us to further detect the significantly biased genes in these three groups (Fig. 1b). The top 20 genes (10 up‐regulated and 10 down‐regulated) of the most significant P signal were displayed by heat map for 0 days post‐crush compared with 3 days post‐crush samples as well as 3 days post‐crush compared with 14 days post‐crush samples, respectively (Fig. 1c,d). In order to identify differentially regulated levels of enriched pathway, dysregulated enriched pathways ranked by their normalized enrichment score (NES) value based on gene set enrichment analysis (GSEA) analysis results were plotted for 0 days post‐crush samples compared with 3 days post‐crush samples. DNA Replication, NF‐κB signaling pathway and Cell Cycle were found to be among the seven most up‐regulated pathways at the 3rd day, but were down‐regulated at day 14 (Fig. 1e,f).

Figure 1.

Microarray analysis of samples at 0, 3, 7, and 14 days post‐crush. (a) Principal component analysis (PCA) map of four groups. (b) Heat map of dysregulated genes in 12 samples from four groups after anova analysis. (c) Heat map of top 20 genes (10 up‐regulated and 10 down‐regulated) with most significant P signal in 3 days group compared with 0 days group. (d) Heat map of top 20 genes (10 up‐regulated and 10 down‐regulated) with most significant P signal in 14 days group compared with 3 days group. (e) Ranking plot of top seven dysregulated pathways by NES value based on enrichment results in 3 days group compared with 0 days group. (f) Ranking plot of top seven dysregulated pathways by their NES value based on enrichment results in 14 days group compared with 3 days group.

The function of the gene set is obtained by GO terms enrichment analysis

The Z‐score bar plot demonstrated that cell cycle, an increasing term in biological processes, had the lowest P value in 3 days post‐crush samples compared with 0 days post‐crush samples. For the cellular component function, the extracellular region, an increasing term, displayed the most significant P signal. For the molecular function, hormone activity displayed the most significant P signal and it was indicated as a decreasing term by Z‐score (Fig. 2a). The selected 15 GO term descriptions and their corresponding GO term ID in 3 days post‐crush compared with 0 days post‐crush samples are listed (Fig. 2b). Similarly, gobubble also illustrated the relationship between enriched GO terms and the Z‐score level. We noticed that the size of the bubble dot for the cell cycle term in BP was largest, which indicated that this term contained the largest amount of significant biased genes. The extracellular region contained the largest amount of significant biased genes and simultaneously demonstrated the lowest P value in CC (Fig. 2c). Moreover, general gene expression distribution in each term and the related Z‐score value were represented by gocircle. Extracellular region, an increasing term, had the largest number of up‐regulated genes compared with other terms (Fig. 2d). A Z‐score bar plot displayed that neutrophil chemotaxis, with the lowest P value, was a decreasing term in biological process, in 14 days post‐crush samples compared with 3 days post‐crush samples. For the cellular component function, the extracellular region displayed the most significant P signal and was a decreasing term, which was contrary to the result in 3 days post‐crush versus 0 days post‐crush samples. In the 14 days post‐crush samples compared with 3 days post‐crush samples, for the molecular function, CXCR chemokine receptor binding displayed the most significant P signal and it was indicated as a decreasing term by Z‐score (Fig. 2e). The 15 selected GO term descriptions and their corresponding GO term ID are listed in Fig. 2(f). We found that the size of bubble dot for the calcium ion binding term in MF was largest, which indicated that this term contained the largest amount of dysregulated genes. Extracellular region contained the largest amount of significant biased genes and simultaneously demonstrated the lowest adjust P value in CC (Fig. 2G). Calcium ion binding, an increasing term in MF, had the largest number of up‐regulated genes compared with other terms (Fig. 2h).

Figure 2.

Gene Ontology (GO) terms enrichment analysis results. (a) Z‐score colored barplot of GO‐BP, GO‐CC, GO‐MF in 3 days post‐crush compared with 0 days post‐crush samples. (b) Table of GO term description and their corresponding GO term ID in 3 days group compared with 0 days group. (c) GOBubble of enriched GO terms and the P signal level. The size of bubble dot denoted the amount of genes in a certain term. (d) GOCircle of general gene expression distribution in each term and the related Z‐score value. The outer ring denoted the subset displaying expression of dysregulated genes for each term. The height of colored bars in the inner ring represented the level of fold change value. Extracellular region contained the largest number of up‐regulated genes compared with other terms. (e) Z‐score colored barplot of GO‐BP, GO‐CC, and GO‐MF in 14 days post‐crush samples compared with 3 days post‐crush samples. (f) The table of selected GO term description and their corresponding GO term ID in14 days group compared with 3 days group. (g) GOBubble of enriched GO terms with Z‐score level. The dot size indicated the term containing the amount of dysregulated genes. (h) GOCircle of general gene expression distribution in each term and the related Z‐score value.

NF‐κB signaling pathway was activated in nerve regeneration after nerve injury

Based on microarray and bioinformatics analysis, the differentially expressed genes and dysregulated pathways were identified. ridgeplot indicated that NF_Kappa_B signaling pathway was activated 3 days post‐crush compared with 0 days post‐crush, which was in accordance with the results by dot plot (Fig. 3a,b); however, it was identified as a suppressed pathway 14 days post‐crush compared with 3 days post‐crush, which indicated that the NF_Kappa_B signal pathway was activated in nerve regeneration process after nerve injury and was returning to normal levels at 14 days (Fig. 3c,d). The results of the enrichment map analysis were presented in the form of an enrichment network to demonstrate cross‐referencing genes contained in multiple pathways. The dark colored node of the NF_Kappa_B signaling pathway indicated its high enrichment significance. The connections representing overlapping parts indicated that the NF_Kappa_B signaling pathway contained cross‐referencing genes with four other pathways (Fig. 4a). gseaplot further confirmed that the majority of biased genes contained in the NF_Kappa_B signaling pathway were up‐regulated 3 days post‐crush compared with 0 days post‐crush, whereas most biased genes contained in NF_Kappa_B signaling pathway were down‐regulated at 14 days (Fig. 4b,d).

Figure 3.

NF_Kappa_B signaling pathway was activated in nerve regeneration after nerve injury. (a) Dotplot of enriched pathways in 3 days post‐crush samples compared with 0 days post‐crush samples. The color intensity denoted the enrichment level of pathways. Gene ratio was counted by the proportion of biased genes in the whole gene set. The size of dot was directly proportional to the amount of gene contained in the corresponding pathway. NF_Kappa_B signaling pathway was activated. (b) Ridgeplot of enriched pathways 3 days post‐crush samples compared with 0 days post‐crush samples. Color intensity of the peaks indicated the enrichment significance for each pathway. NF_Kappa_B signaling pathway was also indicated as activated. (c) Dotplot of enriched pathways in 14 days post‐crush compared with 3 days post‐crush samples. NF_Kappa_B signaling pathway was suppressed. (d) Ridgeplot of enriched pathways 14 days post‐crush compared with 3 days post‐crush samples. NF_Kappa_B signaling pathway was also indicated as suppressed.

Figure 4.

Interleukin‐1β (IL‐1β) was up‐regulated in nerve regeneration. (a) Enrichment network of multiple pathways. The connections represented cross‐referencing genes contained in pathways. The dark colored node indicated high enrichment significance of a certain pathway. The NF_Kappa_B signaling pathway contained cross‐referencing genes with other four pathways. (b) Gseaplot of NF_Kappa_B signaling pathway in the 3 days group. The majority of biased genes contained in it were up‐regulated. (c) Heat map of dysregulated mRNAs contained in NF_Kappa_B signaling pathway in 3 days group. IL‐1β was up‐regulated. (d) Gseaplot of NF_Kappa_B signaling pathway in 14 days group. The majority of biased genes contained in it were down‐regulated. (e) Heat map of dysregulated mRNAs contained in NF_Kappa_B signaling pathway in 14 days group. IL‐1β was down‐regulated.

IL‐1β was up‐regulated in nerve regeneration

We screened out the up‐regulated and down‐regulated mRNAs of NF_Kappa_B signaling pathway at 0, 3, and 14 days after surgery. Thirty mRNAs were found up‐regulated at 3 days, and 11 mRNAs were found down‐regulated at 14 days (Fig. 4c,e). Eight mRNAs were found to be up‐regulated at 3 days after surgery and restored at 14 days after surgery, including IL‐1β and, simultaneously, it was at the upstream location in NF_Kappa_B signaling pathway (Fig. 5a,b). Sciatic nerve tissues were harvested at 0, 3, 7, and 14 days after surgery. The qPCR and Western blot results showed that the IL‐1β was significantly raised 3 days after surgery and returned to normal levels at 14 days after surgery (Fig. 6a,b), which was in accordance with the results of bioinformatics analysis. Those results implied the vital role that IL‐1β played in nerve regeneration.

Figure 5.

Interleukin‐1β (IL‐1β) was up‐regulated in nerve regeneration. (A) Venn diagram of three data sets. A total of eight mRNAs were found to be the intersection of up‐regulation of mRNAs contained in NF_KAPPA_B signaling pathway in the 3 days group and down‐regulation of mRNAs contained in NF_KAPPA_B signaling pathway in the 14 days group, including IL‐1β. (b) IL‐1β was at the upstream location in NF_KAPPA_B signaling pathway.

Figure 6.

Interleukin‐1β (IL‐1β) highly expressed in nerve regeneration. (a) Expression of IL‐1β at different post‐injury time‐points (0, 3, 7, 14 days) in C57BL/6J mice (n = 5, **P < 0·01). (b) Protein expression of IL‐1β at different post‐injury time‐points (0, 3, 7, 14 days) in C57BL/6J mice (n = 5, **P < 0·01).

IL‐1β increased mechanical hypersensitivity and promoted nerve regeneration

The mechanical allodynia of C57BL/6J‐crush mice was attenuated compared with that of C57BL/6J‐crush+IL‐1β mice 3–14 days after crush surgery (Fig. 7a,b). Similarly, the thermal hypersensitivity reflected by both Hargreaves and tail‐flick tests was also decreased in C57BL/6J‐crush mice compared with the C57BL/6J‐crush+IL‐1β mice (Fig. 7c,d). Before surgery, SFI values in all groups were close to zero. After the nerve crush, the mean SFI decreased to −100 due to the complete loss of sciatic nerve function in all animals. With the recovery of the sciatic nerve after surgery, the SFI value gradually increased. However, the nerve recovery capacity of C57BL/6J‐crush mice was significantly lower than that in C57BL/6J‐crush+IL‐1β mice (Fig. 7e).

Figure 7.

Interleukin‐1β (IL‐1β) contributes to nerve recovery. (a) Mechanical allodynia with von Frey filaments (0·02 g). (b) The mechanical sensitivity of each mouse was calculated as the 50% threshold force needed for paw withdrawal. (c) Thermal hyperalgesia using the Hargreaves test. (d) Thermal hyperalgesia using the tail‐flick test. (e) Diagrammatic representation of effects on the sciatic nerve function index (SFI). Significant differences between the ‘C57BL/6J‐crush’ and ‘C57BL/6J‐crush+IL‐1β’ groups at each time point are indicated by asterisks

IL‐1β actives NF‐κB signaling pathway in neurally injured mice

To determine the expression of genes in the NF‐κB signaling pathway during nerve injury, we examined the expression of TAK1, P65 and TNF‐α in the C57BL/6J‐crush group by Western blot assay (Fig. 8a). The results showed that the expression of these genes increased first and then decreased. To further confirm the effect of IL‐1β on NF‐κB, we also examined the expression of TAK1, P65, and TNF‐α in the C57BL/6J‐crush+IL‐1β group (Fig. 8b), and the results were consistent with previous findings. However, compared with the C57BL/6J‐crush group, the expression of the corresponding protein in the C57BL/6J‐crush+IL‐1β group increased more significantly, which also indicated that IL‐1β can indeed affect the NF‐κB signaling pathway.

Figure 8.

Interleukin‐1β (IL‐1β) affects protein expression in nuclear factor‐κB (NF‐κB) signaling pathway. (a,b) The protein expression in the ‘C57BL/6J‐crush’ group was determined by Western blot *P < 0·05, vs. the 0 day group.

Discussion

In this study, we examined genes with differential expression after neural injury using bioinformatics analysis. IL‐1β was screened out as it was obviously up‐regulated. Multiple experiments were performed to confirm that IL‐1β could facilitate neural regeneration after injury.

First, bioinformatics as well as animal experiments showed an obvious increase of IL‐1β expression in the injured group in contrast to the control group. A similar phenomenon was observed by Lin et al.;15 that after nerve root injury, more cells were positive for IL‐1β. Interestingly, the expression level of IL‐1β was slightly reduced at day 14 after high expression in the first 3 days, exhibiting the dynamic process of IL‐1β expression in response to nerve injury. This variation was reported by several studies in different kinds of neural injuries. For instance, Shamash et al.16 investigated the cytokines in Wallerian degeneration and found IL‐1β had the highest expression 1 day after injury but then rapidly reduced.

Analogously, Ryoke et al.17 noticed that IL‐1β was remarkably higher in the sciatic nerve injured group than the control group and reached a maximum in the first day after injury. Similarly, a peak of IL‐1β expression in the sciatic nerve was reported by Bizette et al.18 to appear by the second day after injury, which was a little later than the result of Ryoko et al. Taken together, though the peak of IL‐1β expression varied in different cases, it was a common rule that IL‐1β would instantly increase in the first few days after injury and thereafter gradually decrease.

The NF‐κB signaling pathway was also put forward according to the bioinformatics analysis because it was enriched at the beginning of neural injury and decreased together with IL‐1β. In accordance of the KEGG database, IL‐1β could regulate the expression of several genes through the NF‐κB signaling pathway like TNF‐α and COX2, which were also considered to have connections with neural regeneration.19, 20 However, it had thought that IL‐1β contributed to neurite outgrowth through the p38 MAPK pathway and the janus kinase/signal transducer and activator of tran‐ions (JAK/STAT) pathway,6, 21 implicating that the regenerating process might be highly complicated. Our research mainly focused on the function of IL‐1β, and its underlying mechanism remains to be explored.

It was affirmed by animal experiments that IL‐1β could promote neural recovery both physically and functionally. Still, there were several limitations in our study. We proved the effectiveness of IL‐1β in the repair of nerve injury in C57BL/6J‐crush mice, but its therapeutic effect in normal mice without genetic defects remained unrevealed; the latter is what is more important in developing future treatment for patients with nerve damage. Horie et al.22 observed an augmented neuronal survival rate and neural regeneration in the IL‐1β‐treated group. In the study by Korompilias et al.,23 rats treated with rhIL‐1β showed faster recovery in motor function than those treated with saline solution during the first week after sciatic crush injury. This partially solved our problems, but it brings out another limitation. We did not further study the suitable dose of IL‐1β or the time of applying IL‐1β to achieve the best therapeutic benefits, the change of which may greatly vary the results. Interestingly, there had also been disputes that IL‐1β was pernicious for neural regeneration instead. Taking telmisartan for example, it was reported to have promotion on nerve outgrowth and functional recovery by inhibiting IL‐1β, which was contrary to our findings.24 SFI has proven to be one of the indicators for evaluating neurological recovery.25 This reminds us that inhibiting the expression of IL‐1β in further study is also needed to revalidate the effect of IL‐1β in nerve injury repair. There might be other regulation methods that telmisartan or other IL‐1β inhibitors are involved in, which may have led to the differences in our results, which required deeper investigation. Despite the many limitations in our research, the C57BL/6J‐crush mice showed significant decline in neural regeneration in comparison with the C57BL/6J‐crush+IL‐1β mice, suggesting that it was undeniable that the existence of IL‐1β truly helps better neural recovery.

In conclusion, our study revealed that IL‐1β was significantly beneficial for both physical and functional recovery in injured peripheral nerves, probably through the NF‐κB signaling pathway. This result extended our knowledge about the functions of IL‐1β as well as the whole process of neural regeneration and provided new possibilities for clinical treatments for nerve injuries.

Disclosures

The authors declare no conflict of interest.

Informed consent

Animal protocols were approved by the Shanghai Jiao Tong University Affiliated Sixth People's Hospital.