Temporal gene expression changes after acute and delayed ventral root avulsion-reimplantation

Abstract

Background:

In a model of injured spinal motor neurons where the avulsed spinal nerve is surgically reimplanted, useful regrowth of the injured nerve follows, both in animal experiments and clinical cases. This has led to surgical reimplantation strategies with subsequent partial functional motoric recovery. Still, the ideal time point for successful regeneration after reimplantation and the specific genetic profile of this time point is not known.

Objective:

To explore the temporal gene expression of the whole genome in the ventral spinal cord after reimplantation at different time points after avulsion.

Methods:

Totally 18 adult rats were subjected to avulsion of the left L5 root only (N = 3), avulsion followed by acute spinal reimplantation (N = 3), avulsion followed by 24 h (N = 3) or 48 h (N = 3) delayed reimplantation. Animals were allowed to survive 24 h after their respective surgery whereafter the ventral quadrant of the spinal cord at the operated side was harvested, processed for and analysed with Affymetrix Rat Gene ST 1.0 array followed by statistical analysis of gene expression patterns

Results:

Specific gene expression patterns were found at different time points after avulsion and reimplantation. Over all, early reimplantation seemed to diminish inflammatory response and support gene regulation related to neuronal activity compared to avulsion only or delayed reimplantation. In addition did gene activity after avulsion-reimplantation correspond to regeneration-associated genes typical for regeneration in the peripheral nervous system.

Conclusions:

Our study reveal that genetic profiling after this kind of injury is possible, that specific and distinct expression patterns can be found with early reimplantation being favourable over late and that regenerative activity in this kind of injury bears hallmark typical for peripheral nerve regeneration. These findings can be useful in elucidating specific genetic expression typical for successful nerve regeneration, hopefully not only in this specific model but in the nervous system in general.

Keywords

Ventral root avulsion ventral root reimplantation spinal cord nerve regeneration gene expression microarray

1 Introduction

The lack of successful nerve regeneration in the injured central nervous system (CNS) is a major challenge in neuroscience. Though, already in the dawn of modern neuroscience did Ramon y Cajal demonstrate that certain injured CNS neurons have the capacity to extend new sprouts (Ramón y Cajal, 1928), findings that has been reinvestigated in modern time with findings of peripheral nerves as a potential supporter of regeneration if injured CNS neurons (Aguayo, Vidal-Sanz, Villegas-Perez, & Bray, 1987; Benfey & Aguayo, 1982; Bray, Vidal-Sanz, & Aguayo, 1987; Cheng, Cao, & Olson, 1996; David & Aguayo, 1981; Richardson, McGuinness, & Aguayo, 1980). Other strategies to overcome the lack of regenerative capacity of injured neurons in the CNS has been to transplant nerve growth supporting cells such as olfactory ensheathing glia (Li, Field, & Raisman, 1997; Ramon-Cueto, Cordero, Santos-Benito, & Avila, 2000), to inhibit components of the CNS scar of potential importance for unsuccessful regrowth of nerves such as extracellular matrix molecules (Bradbury et al., 2002) and components of myelin degradation (Chen et al., 2000; Schnell & Schwab, 1990), or to inhibit the establishment of cells that could have nerve growth inhibitory functions in the CNS scar (Dias et al., 2018).

On spinal cord level special interest has been focused on the capacity of injured motorneuron axons to regrow to ventral roots, a finding already described by Cajal (Ramón y Cajal, 1928) and later confirmed (Carlstedt, Cullheim, Risling, & Ulfhake, 1989; Linda, Cullheim, & Risling, 1992; Risling, Cullheim, & Hildebrand, 1983). Specific prerequisites such as a more permissive scar characterized by expression of both neurotrophin (Frisen et al., 1998; Frisen et al., 1992) and angiogenic (Skold et al., 2000) receptors but also nerve growth supportive extracellular matrix molecules (Risling, Fried, Linda, Carlstedt, & Cullheim, 1993) has been described (see (Cullheim et al., 2002) for review). These basic scientific findings on the capacity of regrowing motoneuron axons is the foundation behind the clinical attempts to replant avulsed spinal roots that has shown successful nerve regrowth with gain of lost function (Carlstedt, Grane, Hallin, & Noren, 1995; Carlstedt, Misra, Papadaki, McRobbie, & Anand, 2012).

So far the vast majority of the basic scientific findings behind the biological response on ventral root avulsion and reimplantation has been performed acutely i.e. the avulsion of the root is followed by acute reimplantation (Cullheim, Carlstedt, & Risling, 1999; Gu et al., 2004). Naturally, in the clinical setting that can not, due to practical circumstances, be the case and normally root avulsion injuries are followed by reimplantation at earliest up to a week after injury though clinical experience now indicate that reimplantation should not be delayed and that early reimplantation is favourable over late reimplantation (Carlstedt, 2009; Carlstedt & Havton, 2012).

This means that there is a lack of basic biological knowledge on what characterizes the response in the spinal cord subjected to root avulsion and reimplantation at different time points after root avulsion and even more so what characterizes the optimal time point for reimplantation and regeneration. This study aims at bridging this gap in knowledge and broaden or understanding on what differs in biological response after root avulsion and reimplantation at different time points. To do this we perform not only acute avulsion and reimplantation but also avulsion followed by reimplantation delayed 24 and 48 hours. The collected ventral quarter of the spinal cord at the injured level, containing both ventral horn/motoneurons and the site of the reimplantation are analysed with microarray technique regarding genome wide gene regulation. We can show that different time points of reimplantation are characterized by distinctly different responses in gene activity and find indications of an early reimplantation being favourable over a late regarding regenerative activity.

2 Material and methods

2.1 Ventral root avulsion – acute and delayed reimplantation

All animal experiments were approved by the Stockholm animal welfare ethics committee (N22/11). 18 adult Sprague-Dawley rats were anesthetized by Isoflurane inhalation and the lumbosacral spinal cord was exposed. The left L5 ventral root was identified and avulsed by gentle traction of the root. In 3 of the animals the root was not replanted (avulsion only) and in 3 of the animals the root was directly replanted into the lateral funiculus at the level of avulsion. In 6 animals the avulsed root was marked with suture and left whereafter the wounds were closed. 3 of these animals were reoperated after 24 hours, the avulsed root identified and replanted into the lateral funiculus of the spinal cord. 3 of the animals were reoperated in the same manner after 48 hours and the avulsed root replanted into the lateral funiculus. All animals were left to survive for 24 h after their respective surgery whereafter they were euthanized with 0.5 ml pentobarbital (40 mg/ml) and the inferior vena cava was cut open. The lumbosacral spinal cord was rapidly dissected out, meninges and rootlets removed. Thereafter the spinal cord segment L5 was immersed in RNAlater® (Qiagen, Crawley, West Sussex, UK). Lumbar spinal cord specimens from 6 animals not undergoing surgery were collected and treated accordingly and used as controls. For microarray analysis, samples comprising the left ventral quadrant (lesion side) were used. Data from animals in the control group, the avulsed group and the group undergoing acute reimplantation has partly been described in a previous publication (Risling et al., 2011).

2.2 Gene array analysis

RNA samples were analysed at the Karolinska Institutet core facility for Bioinformatics and Expression Analysis (BEA), where target preparation and hybridization to the microarray were completed. RNA was labeled with biotin to produce the final target according to Affymetrix standard procedures. Labeled cRNA was then hybridized to the Affymetrix Rat Gene ST 1.0 array. Each of the 27,342 genes are represented on the array by approximately 26 probes spread across the full length of the gene. After probing and scanning, the quality of the images was checked. All arrays passed the Affymetrix quality control check. The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (Edgar, Domrachev, & Lash, 2002) and are accessible as GEO Series accession number GSExxx (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSExxx). The.CEL files stem from different batches but used the same affymetrix protocol (RaGene-1_0-st-v1). The Xh and 0h data has in part been presented in an earlier publication (Risling et al., 2011). The 24 h and 48 h data was acquired during a later time. The current analysis was based on a combined gene-level normalization and signal summarization of the different time points, and thus provides unique temporal insights not obtained in the previous publication, justifying a new publication. Normalization, data quality control and signal summarization by RMA of the.CEL files was performed by Affymetrix Expression Console software v.1.4.1 and AltAnalyze v.2.1.0 (Emig et al., 2010). Coherent sample expression distribution was attained (Figs. 2 and 3). Gene level differential expression analysis, volcano plots and chromosome distribution analysis was done by Affymetrix Transcriptome Analysis Console v.3.1.0.5. Principal component analysis was done by AltAnalyze. Lineage correlations was done by lineage profiler in AltAnalyze, in order to identify the most likely cell types and tissues represented in the RNA samples. The samples were compared to a curated lineage identifier database provided by AltAnalyze and resulted in Z scores calculated from the distribution of Pearson correlation coefficients, specifically for each sample and was visualized as a hierarchically clustered heat map. Correlation based Z scores are visualized as positive associations (red) and negative associations (blue). Differentially expressed genes were matched and visualized for biological relevance in WikiPathways (Kelder et al., 2012) (Kutmon et al., 2016). A heatmap for the most important wiki pathways was created by AltAnalyze. Correlation based Z scores are visualized as positive associations (red) and negative associations (blue). A similar comparison was done in Affymetrix Transcriptome Analysis Console. The pathway “Spinal cord injury” (WP2433) was highly enriched and used for temporal comparison of the general injury patterns. The injury studied in this work is situated at the border between the CNS and PNS and sometimes also referred to as a longitudinal spinal cord injury (Carlstedt, 2010) making a comparison with post traumatic reactions in spinal cord injuries reasonable. Genes with significant differential regulation, resulting from Affymetrix Transcriptome Analysis Console (fold change 1.4 and p < 0.05) were analysed in DAVID bioinformatics database (Huang, Sherman, & Lempicki, 2009). Functional annotation clusterings with enrichment scores >1.3 (corresponding to p < 0.05) were compared for the groups. Comparisons were made between each time point and controls. Regulated genes (up and down) were analysed for common transcription factors by AltAnalyze based on GO-Elite enrichment transcription factor analysis. The association with regenerative transcription control programs for CNS and PNS was tested by Gene Set Enrichment Analysis (GSEA) (Subramanian et al., 2005). In GSEA groups of genes that share common biological function or regulation are analysed. Genes are ranked based on the correlation between their expression and the class distinction (comparison between two groups). GSEA determines whether the genes in a gene set are randomly distributed throughout the genome or primarily found at the top or the bottom of the genome. The primary statistic, the enrichment score (ES) is calculated and reflects the degree to which a gene set is overrepresented at the extremes of the entire ranked list, corresponding to a weighted Kolmogorov-Smirnov statistic. The ES is then normalized (NES) for each gene set to account for the size of the set, allowing for comparison between gene sets. Gene sets were collected from a recent publication describing regenerative transcriptional programs observed after PNS but not CNS injury, named Magenta, Pink, Purple, Darkred and Greenyellow (Chandran et al., 2016). The temporal changes of NES for each gene set were compared.

3 Results

Volcano plotting showed relatively small changes at avulsion only, downregulation of genes at acute reimplantation, and general upregulation at 24 h and 48 h (Fig. 1). Genes were spread across all chromosomes, and showed no chromosomal clustering (Fig. 2). The sample expression distribution after normalization and transformation passes the threshold for minimal variance (Fig. 2). Principal component analysis shows separate distributions for each group with a clear separation between avulsion only (Xh) and acute reimplantation (0 h) but with some overlap in the 24 h and 48 h groups (Fig. 3).

Fig. 1

Overview of the experimental model. 3 rats were subjected to avulsion only, 3 animals to avulsion followed by direct reimplantation of the avulsed root, 3 animals to avulsion followed by a 24 h delay of the reimplantation and 3 animals to avulsion followed by a 48 h delay of the reimplantation. To the left is shown volcano plots for the respective group showing relatively small changes in animals subjected to avulsion only, a downregulation of gene regulation in animals subjected to acute reimplantation, and a general upregulation at 24 h and 48 h. X-axis: fold change, y-axis: p-value. Colored dots mark genes above the threshold for further analysis: p < .05, FC > 1.4. Genes are spread across all chromosomes. Green: downregulation, red: upregulation.

Fig. 2

Whole transcriptome overview. The sample expression distribution after normalization and transformation passes the threshold for minimal variance. Principal component analysis shows separate distributions for each group with an overlap in the 24 h and 48 h groups. Chromosome distribution analysis shows equal distribution of differentially expressed genes across the chromosomes. Acute reimplantation caused mostly down-regulation of genes (green color), while at 24 h and 48 h mostly up-regulation of genes (red color).

Fig. 3

Time dependent inflammatory activation and fibrosis in the spinal cord after VRA. Lineage correlations of protein coding genes confirming high correlation in all groups for CNS (fetal spinal cord, oligodendrocyte progenitor, visual cortex, brain, neurons, neural progenitor). At 24 h and 48 h, high lineage correlations for inflammation (microglia) and fibrosis (osteoblasts, cardiac fibroblasts, fibroblasts, embryonic fibroblasts) are found. Note the high lineage correlation for neurotransmission (dopaminergic progenitor) at 24 h and 48 h, not found in control, Xh and 0 h. Note the higher correlation for inflammation and fibrosis in acute reimplantation (0 h) compared to no implantation after avulsion (Xh).

Lineage correlations of protein coding genes confirmed high correlation in all groups for CNS (fetal spinal cord, oligodendrocyte progenitor, visual cortex, brain, neurons, and neural progenitor). At 24 h and 48 h, high lineage correlations for inflammation (microglia) and fibrosis (osteoblasts, cardiac fibroblasts, fibroblasts, embryonic fibroblasts) were seen. A high lineage correlation for neurotransmission (dopaminergic progenitor) at 24 h and 48 h, was not found in control, avulsion (Xh) and acute reimplantation (0 h). A higher correlation for inflammation (microglia) and fibrosis (fibroblasts) was found after acute reimplantation (0 h) compared to no implantation after avulsion only (Xh) (Fig. 4).

Fig. 4

The inflammatory activation after VRA and reimplantation is mediated by interleukine- and lymphocyte genes. Pathway analysis of most enriched WikiPathways. Reimplantation at 0 h does not cause changes in major inflammatory pathways IL-3, IL-4, IL-5, IL-6, T-cell receptor signaling and B-cell receptor signaling compared to avulsion only (Xh). At 24 h and 48 h, these inflammatory pathways are equally activated along with signaling pathways (cardiovascular, Alpha6-Beta4 integrin, Kit receptor, G13, G-protein).

Pathway analysis of the most enriched WikiPathways showed that acute reimplantation at 0 h (0 h vs Control) did not cause changes in major inflammatory pathways IL-3, IL-4, IL-5, IL-6, T-cell receptor signaling and B-cell receptor signaling compared to avulsion only (Xh vs Control). At 24 h and 48 h (24 h vs Control and 48 h vs Control), these inflammatory pathways were equally activated along with signaling pathways (cardiovascular, Alpha6-Beta4 integrin, Kit receptor, G13, G-protein) (Fig. 4).

Temporal comparison of the highly enriched WikiPathway “Spinal cord injury” (WP2433) showed a time dependent inflammatory upregulation at 24 h and 48 h. Acute reimplantation (0 h) compared to avulsion only (Xh) caused upregulation of genes related to leucocyte adhesion (selp), and monocyte chemoattractants (Ccr2, Ccl2), possibly related to a higher influx of inflammatory cells into the area (Fig. 5a-d). Red boxes indicate upregulation/activity and green boxes downregulation/non activity. Genes related to neuronal injury response and axonal regeneration (Zfp36, Fkbp1a, Erg1) were upregulated in acute reimplantation compared to avulsion only and with a trend to further upregulation at 24 h and 48 h. At 24 h and 48 h cell cycle re-entry genes were also strongly upregulated in reimplantation compared to avulsion only.

Fig. 5A-D

Acute reimplantation leads to adhesion- and chemoattractant gene activation. Temporal comparison of highly enriched WikiPathway “Spinal cord injury” (WP2433) showing a time dependent inflammatory upregulation at 24 h and 48 h. Acute reimplantation compared to avulsion only caused upregulation of genes related to leucocyte adhesion (selp), and monocyte chemoattractants (Ccr2, Ccl2), involved in influx of inflammatory cells and also were genes related to activation of microglia and macrophages strongly upregualted in especially 24 h and 48 h compared to acute reimplantation but especially to avulsion only. Also were genes related to neuronal injury response and axonal regeneration (Zfp36, Fkbp1a, Erg1) upregulated in acute reimplantation compared to avulsion only and further upregulated at 24 h and 48 h. Cell cycle re-entry genes were also strongly upregulated in both acute reimplantation, 24 h and 48 h compared to avulsion only.

Genes that had a fold change of >1.4 and p < 0.05 in the initial comparison between groups in the Affymetrix Transcriptome Analysis Console software, were further analysed in the Database for Annotation, Visualization and Integrated Discovery (DAVID). In DAVID, the annotation clustering stringency was set to the highest level, and resulted in annotation clusters, meaning that the genes were connected to a possible biological function. The “Enrichment Score” for further listing in Fig. 6A-B was set to 1.3, which corresponds to p < 0.05. Colored clusters mean they were found at both time points and could be compared (blue arrows). Clusters highlighted with red borders were connected to neurotransmitter activity. Clusters highlighted in blue borders were connected to inflammation.

Fig. 6a-b

Neurotransmission-related gene expression is time and reimplantation dependent. Functional Annotation Clustering by the Database for Annotation, Visualization and Integrated Discovery (DAVID). Genes passing threshold p < 0.05 and fold change >1.4. Rattus norwegicus background. Annotation clustering stringency highest. Enrichment Score cut-off 1.3, corresponding to p < 0.05. Colored clusters mean they are found at both time points and may be compared (blue arrows). Clusters highlighted with red borders are connected to neurotransmitter activity. Clusters highlighted in blue borders are connected to inflammation. In 7a can be seen that acute reimplantation tends to induce genes related to neurotransmission (left) compared to avulsion only (right) were genes related to inflammation are activated. At reimplantation at 24 h (7b left) there is still a tendency to more pronounced activation of genes related to neurotransmission, a situation changed at reimplantation at 4 h (7b right) when genes mainly related to inflammation and antigene presentation were activated.

Neurotransmission-related gene expression was dependent on time and reimplantation, and neurotransmission genes were highly enriched at 0 h and 24 h, but not at 48 h and in avulsion only. Neurotransmission-related genes were absent if no reimplantation occurred, and absent at delayed reimplantation at 48 h. Thus, reimplantation-induced neurotransmission activity not seen after avulsion only and this activity was still seen at 24 h but not after 48 h since neurotransmission genes were highly enriched at 0 h and 24 h, but not at 48 h and in Xh (avulsion only).

Immune response related clusters were found at Xh and increasingly at 24 h and 48 h but not at 0 h, acute reimplantation. Apoptosis related clusters were increasing at 24 h and 48 h, but not found at Xh and 0 h. Immune response related clusers were found at Xh and increasingly at 24 h and 48 h but not at 0 h. Apoptosis related clusters were increasing at 24 h and 48 h, but not found at Xh and 0 h (Fig. 6A-B). These results were combined for comparison purposes (Fig. 7).

Fig. 7

Neurotransmission-related genes absent if no reimplantation occurred, and at prolonged reimplantation at 48 h, suggesting time window for reimplantation. DAVID annotation clusters temporal comparison for selected enrichment groups. Enrichment groups of special interest from Fig. 5 has been selected and summarized. Neurotransmission genes are highly enriched at 0 h and 24 h, but not at 48 h and in avulsion only. Immune response related clusers are found at Xh and increasingly at 24 h and 48 h but not at 0 h, acute reimplantation. Apoptosis related clusters are increasing at 24 h and 48 h, but not found at Xh and 0 h.

A temporal comparison of main transcription factor targets of regulated genes showed that injury-related gene regulation was related to few specific transcription factors. Acute reimplantation (0 h) compared to no reimplantation (Xh) caused an elevated response caused by NFkB, HCC-2GM, SP1, miR373 and an array of “immune response” transcription factors. The main regulation at 24 h was mediated by Foxp3, NFkB, miR34, miR155 and miR124. At 48 h additional transcription factors HCC-G2M and myogenin was added.

The previous findings by Chandran et al of modules of regeneration-associated genes (RAGs) (Abe & Cavalli, 2008)) typical for nerve regeneration in the peripheral nervous system (Chandran et al., 2016) were used for comparison of our microarray findings. A Gene Set Enrichment Analysis based on these gene modules sets (Magenta, Pink, Purple, Darkred, Greenyellow (Chandran et al., 2016)) showed that neuronal regeneration in the spinal cord following nerve root implantation bore genotypic hallmarks of peripheral nerve injury regeneration. Gene sets magenta, pink and purple showed a temporal variability and had statistically significant, concordant differences between injury and controls. No regenerative activation was observed at avulsion only (Xh), where all gene sets showed negative normalized enrichment scores (Fig. 8).

Fig. 8

Neuronal regeneration in the spinal cord following nerve root implantation bears the genotypic hallmarks of peripheral nerve injury regeneration. For comparison the five modules of genes identified by Chandran et al (Chandran et al., 2016) that relates to regeneration after injury in the peripheral nervous system were tested against our material with Gene Set Enrichment Analysis (GSEA). This produced enrichment plots (first half of Fig. 8) and especially Gene sets magenta, pink and purple showed a temporal variability and had statistically significant, concordant differences between injury and controls. In the lower half of the figure is seen a NES, Normalized Enrichment Score (y-axis) and time (x-axis) that summarizes the findings shown in the upper half of Fig. 8. No regenerative activation was observed at avulsion only (Xh), where all gene sets showed negative normalized enrichment scores whereas especially Magenta and Pink had a high score at acute and delayed reimplantation (24 h and 48 h) and Purple at delayed reimplantation (24 h and 48 h).

4 Discussion

In this study we can demonstrate specific gene expression patterns at different time points after avulsion and reimplantation of ventral spinal roots. After early reimplantation the inflammatory response was lower than after avulsion only or after delayed reimplantation. Furthermore did an early reimplantation result in a higher activity in genes related to neuronal activity compared to avulsion only or delayed reimplantation. We can also show that the gene activity after avulsion-reimplantation correspond to regeneration-associated genes typical for regeneration in the peripheral nervous system.

Successful regeneration of injured axons in the vertebrate CNS is an exception to the rule, the examples as such few and the medical challenge connected to this fact an important focus in the neurosciences (Carlstedt et al., 1995; Cheng et al., 1996; Curcio & Bradke, 2018; David & Aguayo, 1981; Ramón y Cajal, 1928; Richardson et al., 1980).

The avulsion-reimplantation model in the rodent has led to the development of a clinical method applied especially after brachial plexus injuries with reimplantation of avulsed ventral roots (Carlstedt et al., 1995; Carlstedt et al., 2012). Naturally, such a surgical intervention is complicated to perform acutely due to practical circumstances, though most of the basic science behind the clinical method is based on acute avulsion/reimplantation studies (Carlstedt et al., 1989; Linda et al., 1992; Risling et al., 1983). Therefore, we found it of interest to study the genetic regulation in the spinal cord if a delayed reimplantation, more similar to the clinical situation, was undertaken. To the best of our knowledge this is the first time that a genome wide screening has been undertaken in such a model, even though others have studied delayed reimplantation of avulsed spinal nerve roots. Gu and colleagues showed in rats that if the reimplantation of avulsed cervical roots was delayed 2 weeks there was still a survival of over 50% of the motoneurons at the lesion side at 20 weeks after injury, that axon did regrow and that forelimb function to some extent recovered but no specific analysis of genetic regulation was performed (Gu et al., 2005). Wu and colleagues reported in similar studies that over 80% of the motoneurons survived even at 3 weeks delayed reimplantation (Wu et al., 2004). These findings has to be put in the context of earlier findings that avulsion of spinal roots in adult rats lead to as much as 45% motoneuron death after 7 days and over 80% motoneuron death after 14 days (Koliatsos, Price, Pardo, & Price, 1994). The distance from the nerve injury to the spinal cord surface can be one important issue that explains this difference since there is big difference in neuron survival if the nerve is truly separated from the spinal cord at the spinal cord surface or more distant. For example did Martin et al show that large spinal motoneurons where reduced by 30% at 21 days if the injury was more distal (sciatic nerve avulsion (Martin, Kaiser, & Price, 1999). Yet others has demonstrated, in injured hypoglossal nerve that is it not as much the proximo-distal placement of the injury as the proximity of the nerve stumps that is essential for motoneuron survival (Tornqvist & Aldskogius, 1994). In our model of nerve root avulsion both aspects of described importance for increased motoneuron death after injury are added since the avulsion injury is proximal (at the actual spinal cord surface) and the distal nerve stump is not in contact with the proximal part. On the other hand is our model of reimplantation an actual reconnection of the avulsed root to its more proximal part in the spinal cord and thus do we in this study compare two extremes and their gene expression.

One major finding in this work is the shift in gene activity from immune response to neurotransmission, when comparing avulsion only to acute reimplantation, and thereafter a gradual shift back to activity in immune response genes combined with a lesser activity in neurotransmission at later reimplantation time points (Fig. 7). It is known that avulsion injury to the spinal cord ventral roots and the resulting degeneration of motoneurons is closely related to inflammatory cell infiltration and that after 2–3 weeks post-operatively approximately 50–75% of the axotomized cells disappear. (Olsson, Piehl, Swanberg, & Lidman, 2005). In our previous work on gene regulation after acute reimplantation we could show activity in genes related to neurite growth and development after reimplantation compared to avulsion only (Risling et al., 2011), findings that are confirmed in the present study and expanded to the findings of persistent activity in genes related to neurotransmission, adhesion and chemoattractant related genes at 24 h delayed reimplantation but not at 48 h (Figs. 5 and 6).

It is well known that avulsion of ventral roots lead to inflammation. How this relates to nerve regeneration is not fully understood. Both effects beneficial for nerve regeneration (David, Bouchard, Tsatas, & Giftochristos, 1990; Lu & Richardson, 1991) as well as detrimental for nerve survival and regeneration has been discussed (Giulian, Corpuz, Chapman, Mansouri, & Robertson, 1993). The immune response after nerve injury is complex system with immune molecules, such as MHC I, involved in the synaptic stripping after peripheral nerve injury (Oliveira et al., 2004) were synaptic input is detached from motoneurons of the spinal cord, probably as a mechanism involved the neuron focusing on survival rather than signaling (for review see (Spejo & Oliveira, 2015)). Our findings could indicate that such mechanisms are being involved after avulsion only, where processes such as antigen presentation via MHC, lymphocyte and mononuclear proliferation are activated while acute reimplantation is not followed by such activity but rather with processes involved in synaptic vesicle production and transmission (Fig. 6A). The same type of genes (neurotransmitter activity, glutamate receptor activity) are still activated if reimplantation is delayed 24 h but not if delayed 48 h. If this regulation, induced by reimplantation up to 24 h after avulsion, is also followed by a re-establishment of stripped synapses is not known but an interesting hypothesis worth investigating further.

To further elucidate the potential role of our findings a temporal comparison with the WikiPathway “Spinal cord injury” (WP2433) (shown in Fig. 5a-d) was performed. This comparison revealed that genes related nerve regeneration and to neuronal injury response (Zfp36, Fkbp1a, Erg1) were upregulated in acute reimplantation compared to avulsion only and further upregulated at 24 h and 48 h but also that genes related to inflammatory cell adhesion and chemoattraction were strongly upregulated in 24 h and 48 h delayed reimplantation compared with acute reimplantation.

Our study does not include any addition of external growth supporting factors but do solely study the natural response in avulsion, acute and delayed reimplantation. The root avulsed and replanted can be considered a peripheral nerve and as such has growth supporting capacity in it self (Abe & Cavalli, 2008; Benfey & Aguayo, 1982) which is confirmed by studies that did not show any additional growth support when adding peripheral nerve grafts compared to reimplant the avulsed root it self (Su et al., 2013). In an attempt to refer our findings to typical genetic regulation in regeneration after peripheral nerve injury previous findings by Chandran et al of modules of regeneration-associated genes (RAGs) in peripheral nerve regeneration were used as comparison (Chandran et al., 2016). Chandran et al identified five modules strongly associated with regeneration as follows: two modules (Magenta [394 genes] and Pink [74 genes]) whose genes are upregulated, and three (Purple [194 genes], Darkred [52 genes], and Greenyellow [53 genes]) whose genes are downregulated after nerve injury (Fig. 8), all of which were conserved in a third, independent peripheral nerve injury dataset (Costigan et al., 2002). We could show high activity in three of these moduls in our models after reimplantation (Magenta, Pink and Purple) but not in avulsion only (Fig. 8, first part). To further annotate module function at a general level, we applied gene ontology (GO) enrichment analyses, which showed enrichment (Benjamini-corrected p values <0.05) for several GO categories in the upregulated RAG co-expression modules (Magenta and Pink) that are functionally associated with neuronal regeneration. Significant clusters included regulation of transcription, neuron differentiation, inflammation, stimulus related, signaling related, and cell proliferation/growth/migration. One of the intrinsic molecular mechanisms contributing to the regenerative process is the retrograde transport of injury signals to the cell body of the neuron, leading to expression of regeneration-associated genes (Abe & Cavalli, 2008). For example, injured PNS axons activate RAGs such as Atf3, Jun, Hsp27, Sprr1a, Gap43, and genes involved in the JAK-STAT3 pathway, whereas injury to CNS axons does not result in the activation of these RAGs (Afshari, Kappagantula, & Fawcett, 2009).

The tissue material studied in this work consists of the affected ventral quadrant of the spinal cord. This means that all cells in this region are included in our microarray analysis. This approach has both strengths and limitations. One strength is that this way of analysing the material includes all reactions and interactions between cells in the area after the different kinds of injury and timepoints and thus investigate the ventral horn as a whole system. Despite this whole tissue analysis approach, where activity in one special cell sort could be expected to be obscured by activity in a more abundant cell sort, we can clearly see distinct regulation patterns at the different time points after injury. Coupled with the whole genome analysis it gives us a possibility to tell both on tissue system level (ventral spinal cord) and gene system level (whole genome) what trends in gene regulation that are in action at different time points after avulsion/reimplantation. On the other hand can we not from this whole tissue analysis tell exactly which cell or group of cells that are involved in the reactions we find. A natural next step in analysing gene expression in this kind of injury would be to utilize techniques for single cell analysis (Nichterwitz, Benitez, Hoogstraaten, Deng, & Hedlund, 2018)

In summary we have in study shown that it is possible with genome wide expression analysis to show differences in expression of gene activity at different time points after avulsion and reimplantation of ventral spinal roots. Furthermore do we show that the gene activity shifts during this time period – from a non regenerative, inflammatory dominated gene expression at avulsion only to a regeneration and neurotransmission dominated expression at early time points of reimplantation and back to a more non regenerative, inflammatory dominated expression at later reimplantation time points. This is in line with previous results which show that early reimplantation improves functional outcome after ventral roo avulsion. In addition to this we show that the regenerative pattern of this centrally located nerve cells do hold typical genetic hallmarks for regeneration of the peripheral nervous system. We conclude that the method used in this work can be of value when exploring factors that can be of potential benefit in supporting regeneration of injured nervous cells.

References

Abe,

, & Cavalli,

(2008). Nerve injury signaling. Current Opinion in Neurobiology, 276–283.

Afshari,

F.T.

, Kappagantula,

, & Fawcett,

J.W.

(2009). Extrinsic and intrinsic factors controlling axonal regeneration after spinal cord injury. Expert Reviews in Molecular Medicine, 11, e37.

Aguayo,

A.J.

, Vidal-Sanz,

, Villegas-Perez,

M.P.

, & Bray,

G.M.

(1987). Growth and connectivity of axotomized retinal neurons in adult rats with optic nerves substituted by PNS grafts linking the eye and the midbrain. Annals of the New York Acadmy of Sciences, 495, 1-9.

Benfey,

, & Aguayo,

A.J.

(1982). Extensive elongation of axons from rat brain into peripheral nerve grafts. Nature, 296(5853), 150–152.

Bradbury,

E.J.

, Moon,

L.D.

, Popat,

R.J.

, King,

V.R.

, Bennett,

G.S.

, Patel,

P.N.

, et al. (2002). Chondroitinase ABC promotes functional recovery after spinal cord injury. Nature, 416(6881), 636–640.

Bray,

G.M.

, Vidal-Sanz,

, & Aguayo,

A.J.

(1987). Regeneration of axons from the central nervous system of adult rats. Progress in Brain Research, 71, 373–379.

Carlstedt,

(2009). Nerve root replantation. Neurosurg Clinics of North America, 20(1), 39–50, vi.

Carlstedt,

(2010). Perspectives on the treatment of the longitudinal spinal cord injury. Frontier in Neurology, 1, 11.

Carlstedt,

, Cullheim,

, Risling,

, & Ulfhake,

(1989). Nerve fibre regeneration across the PNS-CNS interface at the root-spinal cord junction. Brain Research Bulletin, 22(1), 93–102.

10.

Carlstedt,

, Grane,

, Hallin,

R.G.

, & Noren,

(1995). Return of function after spinal cord implantation of avulsed spinal nerve roots. The Lancet, 346(8986), 1323–1325.

11.

Carlstedt,

, & Havton,

(2012). The longitudinal spinal cord injury: Lessons from intraspinal plexus, cauda equina and medullary conus lesions. Handbook of Clinical Neurology, 109, 337–354.

12.

Carlstedt,

, Misra,

V.P.

, Papadaki,

, McRobbie,

, & Anand,

(2012). Return of spinal reflex after spinal cord surgery for brachial plexus avulsion injury. Journal of Neurosurg, 116(2), 414–417.

13.

Chandran,

, Coppola,

, Nawabi,

, Omura,

, Versano,

, Huebner,

E.A.

, et al. (2016). A systems-level analysis of the peripheral nerve intrinsic axonal growth program. Neuron, 89(5), 956–970.

14.

Chen,

M.S.

, Huber,

A.B.

, van der Haar,

M.E.

, Frank,

, Schnell,

, Spillmann,

A.A.

, et al. (2000). Nogo-A is a myelin-associated neurite outgrowth inhibitor and an antigen for monoclonal antibody IN-1. Nature, 403(6768), 434–439.

15.

Cheng,

, Cao,

, & Olson,

(1996). Spinal cord repair in adult paraplegic rats: Partial restoration of hind limb function. Science, 273(5274), 510–513.

16.

Costigan,

, Befort,

, Karchewski,

, Griffin,

R.S.

, D’Urso,

, Allchorne,

, et al. (2002). Replicate high-density rat genome oligonucleotide microarrays reveal hundreds of regulated genes in the dorsal root ganglion after peripheral nerve injury. BMC Neuroscience, 3, 16.

17.

Cullheim,

, Carlstedt,

, & Risling,

(1999). Axon regeneration of spinal motoneurons following a lesion at the cord-ventral root interface. Spinal Cord, 37(12), 811–819.

18.

Cullheim,

, Wallquist,

, Hammarberg,

, Linda,

, Piehl,

, Carlstedt,

, et al. (2002). Properties of motoneurons underlying their regenerative capacity after axon lesions in the ventral funiculus or at the surface of the spinal cord. Brain Research Reviews, 40(1-3), 309–316.

19.

Curcio,

, & Bradke,

(2018). Axon Regeneration in the Central Nervous System: Facing the Challenges from the Inside. Annual Review of Cell and Developmental Biology.

20.

David,

, & Aguayo,

A.J.

(1981). Axonal elongation into peripheral nervous system “bridges” after central nervous system injury in adult rats. Science, 214(4523), 931–933.

21.

David,

, Bouchard,

, Tsatas,

, & Giftochristos,

(1990). Macrophages can modify the nonpermissive nature of the adult mammalian central nervous system. Neuron, 5(4), 463–469.

22.

Dias,

D.O.

, Kim,

, Holl,

, Werne Solnestam,

, Lundeberg,

, Carlen,

, et al. (2018). Reducing pericyte-Derived scarring promotes recovery after spinal cord injury. Cell, 173(1), 153–165 e122.

23.

Edgar,

, Domrachev,

, & Lash,

A.E.

(2002). Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Research, 30(1), 207–210.

24.

Emig,

, Salomonis,

, Baumbach,

, Lengauer,

, Conklin,

B. R.

, & Albrecht,

(2010). AltAnalyze and DomainGraph: Analyzing and visualizing exon expression data. Nucleic Acids Research, 38(Web Server issue), W755–W762

25.

Frisen,

, Risling,

, Korhonen,

, Zirrgiebel,

, Johansson,

C.B.

, Cullheim,

, et al. (1998). Nerve growth factor induces process formation in meningeal cells: Implications for scar formation in the injured CNS. Journal of Neurosci, 18(15), 5714–5722.

26.

Frisen,

, Verge,

V.M.

, Cullheim,

, Persson,

, Fried,

, Middlemas,

D.S.

, et al. (1992). Increased levels of trkB mRNA and trkB protein-like immunoreactivity in the injured rat and cat spinal cord. Proceedings of the National Academy of Sciences of the United States of America, 89(23), 11282–11286.

27.

Giulian,

, Corpuz,

, Chapman,

, Mansouri,

, & Robertson,

(1993). Reactive mononuclear phagocytes release neurotoxins after ischemic and traumatic injury to the central nervous system. Journal of Neuroscience Research, 36(6), 681–693.

28.

Gu,

H.Y.

, Chai,

, Zhang,

J.Y.

, Yao,

Z.B.

, Zhou,

L.H.

, Wong,

W.M.

, et al. (2005). Survival, regeneration and functional recovery of motoneurons after delayed reimplantation of avulsed spinal root in adult rat. Experimental Neurology, 192(1), 89–99.

29.

Gu,

H.Y.

, Chai,

, Zhang,

J.Y.

, Yao,

Z.B.

, Zhou,

L.H.

, Wong,

W.M.

, et al. (2004). Survival, regeneration and functional recovery of motoneurons in adult rats by reimplantation of ventral root following spinal root avulsion. European Journal of Neuroscience, 19(8), 2123–2131.

30.

Huang

da W.

, Sherman,

B.T.

, & Lempicki,

R.A.

(2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols, 4(1), 44–57.

31.

Kelder,

, van Iersel,

M.P.

, Hanspers,

, Kutmon,

, Conklin,

B.R.

, Evelo,

C.T.

, et al. (2012). WikiPathways: Building research communities on biological pathways. Nucleic Acids Research, 40(Database issue), D1301–1307.

32.

Koliatsos,

V.E.

, Price,

W.L.

, Pardo,

C.A.

, & Price,

D.L.

(1994). Ventral root avulsion:Anexperimental model of death of adult motor neurons. Journal of Comparative Neurology, 342(1), 35–44.

33.

Kutmon,

, Riutta,

, Nunes,

, Hanspers,

, Willighagen,

E.L.

, Bohler,

, et al. (2016). WikiPathways: Capturing the full diversity of pathway knowledge. Nucleic Acids Research, 44(D1), D488–494.

34.

Li,

, Field,

P.M.

, & Raisman,

(1997). Repair of adult rat corticospinal tract by transplants of olfactory ensheathing cells. Science, 277(5334), 2000–2002.

35.

Linda,

, Cullheim,

, & Risling,

(1992). A light and electron microscopic study of intracellularly HRP-labeled lumbar motoneurons after intramedullary axotomy in the adult cat. Journal of Comparative Neurology, 318(2), 188–208.

36.

Lu,

, & Richardson,

P.M.

(1991). Inflammation near the nerve cell body enhances axonal regeneration. Journal of Neuroscience, 11(4), 972–978.

37.

Martin,

L.J.

, Kaiser,

, & Price,

A.C.

(1999). Motor neuron degeneration after sciatic nerve avulsion in adult rat evolves with oxidative stress and is apoptosis. Journal of Neurobiology, 40(2), 185–201.

38.

Nichterwitz,

, Benitez,

J.A.

, Hoogstraaten,

, Deng,

, & Hedlund,

(2018). LCM-Seq: A method for spatial transcriptomic profiling using laser capture microdissection coupled with polyA-based RNA sequencing. Methods in Molecular Biology, 1649, 95–110.

39.

Oliveira,

A.L.

, Thams,

, Lidman,

, Piehl,

, Hokfelt,

, Karre,

, et al. (2004). A role forMHCclass I molecules in synaptic plasticity and regeneration of neurons after axotomy. Proceedings of the National Academy of Sciences of the United States of America, 101(51), 17843–17848.

40.

Olsson,

, Piehl,

, Swanberg,

, & Lidman,

(2005). Genetic dissection of neurodegeneration and CNS inflammation. Journal of the Neurological Sciences, 233(1-2), 99–108.

41.

Ramón y Cajal,

(1928). Degeneration and regeneration of the nervous system (RMMay, Trans. reprinted 1991 ed. Vol. No. 5). New York, Oxford: Oxford University Press.

42.

Ramon-Cueto,

, Cordero,

M.I.

, Santos-Benito,

F.F.

, & Avila,

(2000). Functional recovery of paraplegic rats and motor axon regeneration in their spinal cords by olfactory ensheathing glia. Neuron, 25(2), 425–435.

43.

Richardson,

P.M.

, McGuinness,

U.M.

, & Aguayo,

A.J.

(1980). Axons from CNS neurons regenerate into PNS grafts. Nature, 284(5753), 264–265.

44.

Risling,

, Cullheim,

, & Hildebrand,

(1983). Reinnervation of the ventral root L7 from ventral horn neurons following intramedullary axotomy in adult cats. Brain Research, 280(1), 15–23.

45.

Risling,

, Fried,

, Linda,

, Carlstedt,

, & Cullheim,

(1993). Regrowth of motor axons following spinal cord lesions: Distribution of laminin and collagen in the CNS scar tissue. Brain Research Bulletin, 30(3-4), 405–414.

46.

Risling,

, Ochsman,

, Carlstedt,

, Linda,

, Plantman,

, Rostami,

, et al. (2011). On acute gene expression changes after ventral root replantation. Frontiers in Neurology, 1, 159.

47.

Schnell,

, & Schwab,

M.E.

(1990). Axonal regeneration in the rat spinal cord produced by an antibody against myelin-associated neurite growth inhibitors. Nature, 343(6255), 269–272.

48.

Skold,

, Cullheim,

, Hammarberg,

, Piehl,

, Suneson,

, Lake,

, et al. (2000). Induction of VEGF and VEGF receptors in the spinal cord after mechanical spinal injury and prostaglandin administration. European Journal of Neuroscience, 12(10), 3675–3686.

49.

Spejo,

A.B.

, & Oliveira,

A.L.

(2015). Synaptic rearrangement following axonal injury: Old and new players. Neuropharmacology, 96(Pt A), 113–123.

50.

Su,

, Yuan,

, Qin,

, Yang,

, Wong,

W.M.

, So,

K.F.

, et al. (2013). Ventral root re-implantation is better than peripheral nerve transplantation for motoneuron survival and regeneration after spinal root avulsion injury. BMC Surgery, 13, 21.

51.

Subramanian,

, Tamayo,

, Mootha,

V.K.

, Mukherjee,

, Ebert,

B.L.

, Gillette,

M.A.

, et al. (2005). Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences of the United States of America, 102(43), 15545–15550.

52.

Tornqvist,

, & Aldskogius,

(1994). Motoneuron survival is not affected by the proximo-distal level of axotomy but by the possibility of regenerating axons to gain access to the distal nerve stump. J Neuroscience Research, 39(2), 159–165.

53.

Wu,

, Chai,

, Zhang,

, Gu,

, Xie,

, & Zhou,

(2004). Delayed implantation of a peripheral nerve graft reduces motoneuron survival but does not affect regeneration following spinal root avulsion in adult rats. Journal of Neurotrauma, 21(8), 1050–1058.