Neural regeneration is successfully achieved in vertebrate models such as zebrafish and some amphibians. Among non-chordate Deuterostomes, echinoderms can efficiently regenerate their nervous system. A recent paper from the laboratory of José E. García-Arrarás (http://www.ncbi.nlm.nih.gov/pubmed/23597108) describes the important role of radial glial cells in this process. Studying how similarly zebrafish and amphibians regenerate their nervous system compared to the non-chordate echinoderms may provide insights into: i) up to what extend the cellular and molecular mechanisms and events during neural regeneration have been evolutionary conserved among Deuterostomes, and ii) trying to improve the poor regenerative abilities of the mammalian CNS.
Echinoderm and chordate radial glia are similar morphologically and immunocytochemically. In adult echinoderms, radial glial cells retain their proliferative capabilities suggesting a role in neurogenesis. The authors here use sea cucumbers as a model. These animals possess 5 radial nerve cords (RNC) that are joined at the oral side of the body. Each RNC is formed by two bands, the ectoneural and the hyponeural neuroepithelia separated by a thin layer of connective tissue. These neuroepithelia are formed by radial glial cells and interspersed neurons. In their paradigm the authors cut one of the RNC at the mid-body level and analyze how both truncated ends regenerate and get re-connected.
Thus, RNC regeneration is divided into 4 stages: 1) early post-injury phases (days 1-2), 2) late-post injury phase (days 6-8), 3) growth phase (days 8-12) and 4) late regenerate (+21 days). During the early phase the radial glial cells close to the wound start dedifferentiating. These cells lose their basal processes while their cell bodies maintain the epithelial organization in the apical region. During the late phase, the coelomic epithelium migrates over the injury site sealing the wound. Also, at this stage, there is an expansion of the region of the neuroepithelium where the radial glial cells dedifferentiate. Those dedifferentiated radial glial cells do not only show changes in morphology but also show a much-reduced expression of a typical material that is recognized by antibodies against Reissner’s substance (RS), which is typically produced by glial cells also in vertebrates. Next, during the growth phase, the two truncated ends of the RNC at each side of the wound start growing towards each other. The ectoneural and hyponeural bands of the RNC form separate tubular rudiments that grow parallel to each other. Finally, in the late regenerate phase the continuity of the RNC is restored. New radial glial cells re-differentiate adopting their typical morphology and producing again RS-like material. Also, this new neural cord gets populated with neuronal cell bodies and processes.
Then, the authors combined BrdU pulse-chase experiments with labelling with radial glia specific markers. In adult non-regenerating CNS of sea cucumbers a basal level of BrdU incorporation had been previously reported. Now, after injury, no significant increase in BrdU incorporation takes place during the early phase. However, at the late post-injury phase there is a significant increase mainly in the ectoneural part of the RNC. During the growth phase this increase in BrdU-positive cells peaks in both ectoneural and hyponeural cords. From this stage the levels start decreasing until reaching again typical basal levels in the late regenerate phase. Remarkably, most of the BrdU positive cells (90-100%) are positive also for a radial glia-specific marker, indicating the dominant role of radial glia in cell division during RNC regeneration. It is also worthy to mention that whereas in the intact CNS the BrdU-positive cells are randomly scattered during neural regeneration most of the BrdU-positive cells localize in the dedifferentiated glia at the tip of the regenerate.
Finally, the authors show that at the very late regenerate phase (day 51 post-injury) 45% of the BrdU-positive cells are not labelled with the specific radial glia-specific marker and that some of these cells express now different neuronal markers. This suggests that a percentage of the progeny of radial glia differentiate into neurons.
In summary, this paper reports how radial glia is involved not only in mediating the bridging of the wound gap and axonal outgrowth but also acts as a source of new neurons during RNC regeneration in sea cucumber. This leading role for the radial glia resembles very much the role of the chordate radial glia in those regeneration-competent species, which opens the possibility that these conserved mechanisms for neural regeneration are also present in non-regenerating vertebrates but need to be somehow awaken.
In previous posts I have discussed the positive role that ROS species have during tail regeneration in Xenopus as well as the important role played by apoptosis to trigger head regeneration in Hydra. Now a recent paper from the laboratory of Sophie Vriz links both processes during adult caudal final regeneration in zebrafish (http://www.ncbi.nlm.nih.gov/pubmed/23803955).
First, the authors show how whereas wounding induced a transitory ROS production with a peak 30 minutes post-lesion that disappeared after 2 hours, fin amputation induced a much prolonged production of ROS peaking at about 12-16 hours post-amputation (hpa). ROS production was not detected after 24 hours of regeneration. Then, and to try to see the functional relevance of this regeneration-specific extended ROS production the authors used different inhibitors of ROS. Thus, inhibiting NOX (NADPH oxidase) activity from 12 to 24 hpa seemed sufficient to reduce the size of the regenerated fin at 72 hours. This reduced regeneration was accompanied by a decrease in the expression of klf4 (a progenitor cell fate marker) and dio3 (involved in progenitor cell proliferation). The expression of another progenitor cell fate marker as it is myc was not affected.
As oxidative stress has been related to MAP kinase activation and apoptosis, the authors wanted to search whether ROS was also involved in those processes during caudal fin regeneration. First, they observed that the normal activation of the JNK pathway at 6 hpa was reduced about at 50% after NOX inhibition, suggesting that ROS has a role in the early activation of the JNK pathway. Second, the inhibition of ROS production reduced also the number of apoptotic cells during the first 18 hpa. Then, the authors checked in more detail the cell death and apoptotic responses after amputation. Remarkably they found a bimodal response with two peaks of TUNEL positive cells (also active caspase-3 positive cells) at about 5 hpa and 15-18 hpa mainly concentrated in the stump epithelium and very few cells in the mesenchyme. In contrast, after wounding, and in agreement with the only early peak observed for ROS production, cell death increased and peaked around 6 hpa without a second peak later. To further characterize the function of apoptosis on regeneration the authors inhibited this process during fin regeneration. Remarkably, blocking the 2nd wave of apoptosis (from 12 to 72 hpa) was sufficient to impair regeneration. In fact, inhibiting the first apoptotic wave (from 4 to 10 hpa) neither blocked regeneration nor affected the 2nd apoptotic wave. These results suggest that the 2nd apoptotic wave was the one required for regeneration. However, these effects on regeneration affect only the size of the regenerate because those small blastemas appeared to be normally patterned.
Next, the authors looked how proliferation was affected after JNK and apoptosis inhibition. In control animals, at 24 hpa the mitotic cells were mainly localized in the inter-ray epidermis of the stump. The inhibition of ROS or apoptosis reduced the number of mitotic cells by 50%. In contrast to what it has been described in other models JNK pathway did not seem to be involved in the induction of apoptosis during zebrafish caudal fin regeneration. So it seems that apoptosis and JNK work in parallel downstream of ROS production to induce cell proliferation during regeneration.
Finally the authors checked whether the inhibition of apoptosis or JNK pathway impairs cellular reprogramming or signalling molecules required for blastema growth. Inhibiting the second wave of apoptosis reduced the expression of klf4 and dio3, as NOX inhibition does. However, JNK inhibition did not seem to affect these markers. At the level of signalling pathways, apoptosis inhibition reduced the expression of fgf20, sdf1 and enhanced the expression of igf2b and wnt10a. On the other hand, inhibiting the JNK pathway reduced the expression of sdf1, wnt5b and igf2b. In summary, the authors found out that that the production of ROS during regeneration is important to induce apoptosis and JNK pathway that would work in parallel to promote epidermal cell proliferation and blastema formation during zebrafish caudal fin regeneration.
Different mechanisms are used to specify new segments during, for example, somitogenesis in vertebrates and Drosophila body segmentation. Other segmented animals as some annelids keep adding segments as they grow. In a recent paper the laboratory of Shigeo Hayashi describes how new segments are added during the regeneration of the tail in the polychaete Perinereis nuntia (http://www.ncbi.nlm.nih.gov/pubmed/23608458). In many models the new posterior segments are added from a segment-addition zone.
In this study the authors work on the species P. nuntia that, in laboratory conditions, adds a new posterior segment every 4 days. This new segment is added between the last posterior segment and the terminal end, called pygidium. During tail regeneration, a new pygidium is formed and then new segments are added initially at a rate of 1 segment per day. First, the authors show how during normal growth of segments it appears a single row of cells with high PCNA (proliferating nuclear antigen) expression at the border between the last segment and the pygidium. Because PCNA expression is high in late G1-S phase these observations suggest that this single row of cells are in synchronous G1-S phase and has been named by the authors as “zone of cell-cycle synchronization” (ZCS). In addition, and in contrast to the quite random orientation of the mitoses in more anterior regions of the segment, the cells in the ZCS divide along the AP axis with the division axis almost perpendicular to the segmental furrow. The authors also show how changes in PCNA staining are related to changes in chromatin structure based on labelling of the histone modifications H3K9me2 and H3K9, K14ac. Blastema cells are characterized by high H3K9me2 and proliferation rate, whereas when cells differentiate exit the cell cycle and display high level of acetylated histone H3.
Next, the authors follow the regeneration of the new segments by analysing the expression of wingless (wg) and hedgehog (hh). These two genes are expressed in rows of cells at the anterior and posterior side of the segmental furrow, respectively. These cells also overlap with the ZCS of cells with high PCNA expression. Remarkably, the expression of tcf, a nuclear effector of Wg signalling is expressed in a segmental pattern peaking at the hh stripe and decreases in a posterior gradient, suggesting an important role of Wg signalling during segmentation. In fact, treatment with LiCl, which mimics overexpression of Wg signalling produces various regeneration defects. Thus, the row of cells with high PCNA expression at the posterior border of the newly formed segment is disorganized and the animals show a decrease in segment number and segments wider than in controls.
From these results the author suggest a model in which the last row of the most posterior segment serves as the source of Wg that coordinates cell-cycle entry and recruitment of anterior pygidium cells that will initiate the formation of a new segment in the ZCS. Thus, and in contrast to other models, it appears that in P. nuntia segmentation is initiated by cell cycle synchronization and row-by-row addition of cells to the posterior end of the new segment.