Genome reprogramming during zebrafish retina regeneration
In previous posts in this blog I have discussed how zebrafish are able to regenerate their retinas from the dedifferentiation of quiescent supportive Müller glia (MG) that re-enter the cell cycle and give rise to a cycling population of multipotent progenitors (MGPC) that differentiate into the different retina cell types. Previous studies have suggested that dynamic changes in the DNA methylation landscape can have a function in the transition from MG to MGPC. How much similar is this cellular reprogramming that occurs during regeneration to the reprogramming required to transform somatic cells into induced pluripotent stem cells (iPS) is an interesting question for the field of regenerative medicine. During iPS generation there is an increased DNA demethylation of the promoter regions of pluripotency genes that correlates with an increase in their expression. Similarly, the expression of pluripotency genes as well as other regeneration-associated genes increases during the transition from MG to MGPC.
In a recent paper from the laboratory of Daniel Goldman the authors wonder up to what extend changes in the DNA methylation landscape are important for the transition from MG to MGPC (http://www.ncbi.nlm.nih.gov/pubmed/24248357). First, they checked how the expression of several regulators of DNA methylation changed in MGPCs (from injured retinas) compared to MG (from uninjured retinas). They found an increased expression of genes associated with both DNA demethylation and methylation, suggesting that the regulation of DNA methylation may be important for MGPC formation. Next, they showed how the induction of DNA demethylation perturbed the migration and differentiation of MGPC-derived progeny. They took then a genomic approach to compare the DNA methylation landscapes between MG from uninjured animals and MGPCs from injured retinas 4 dpi (between 2 and 7 days post-injury there is an asymmetric division and proliferative amplification of MGPCs). They compared the methylation levels of 611,434 individual cytosines within the CpG context and found 9,554 differentially methylated bases (DMBs) that represented an overall difference of 1,54%. Of those changes, 54% corresponded to increased methylation events and 46% to decreased ones. Most DMBs were localized in intergenic and intronic regions with few of them localized in promoter regions.
From 2 to 4 dpi the DNA methylation landscape shifted from one that was predominantly driven by demethylation to one with increasing levels of methylation, which could correlate to the MGPCs getting ready to enter into differentiation. As expected, a decrease methylation in the promoter regions correlated with an increased gene expression. However, and surprisingly, no DMBs were found in the promoters of pluripotency and retina regeneration-associated genes. Even more remarkably was the fact that when the authors checked the levels of methylation of the promoters of those genes in quiescent MG they found that they were also hypomethylated. So, these data made the authors hypothesize that pluripotency and regeneration-associated genes might be poised for activation in quiescent MG, implying that MG would require only limited reprogramming and would be more stem-like than thought before. Because the regenerative capacity of mouse MG is very limited the authors checked the methylation state of mouse MG. Remarkably, they found that the promoters of pluripotency and regeneration-associated genes showed low levels of methylation, as observed in zebrafish.
In summary, this study shows how the global DNA methylation landscape changes during the transition from MG to MGPCs pointing out an important regulatory function during this reprogramming. However, pluripotency and other genes required for regeneration appeared to be hypomethylated already in MG suggesting that they could be already preprogrammed for a regenerative response. These genes could be then regulated by other events such as histone modifications or transcription factors, indicating that changes in the DNA methylation status would be required but it would not be sufficient for MG reprogramming. The fact that mouse MG share this hypomethylated landscape in pluripotency genes opens the door to search for strategies to facilitate their reprogramming into progenitor cells that could enhance the poor regenerative abilities of the mammalian retina.
Expression of Hox genes during polychaete regeneration
Hox genes code for transcription factors that play an essential role in the regionalization of the anterior-posterior axis in animals. Thus, they function by providing specific identities to the different segments of the body plan, which translates in the differentiation of the correct structures in each segment. Well-known and astonishing examples of homeotic transformations include those that transform Drosophila antennas into legs or flies with two pairs of wings. Hox genes have been highly conserved through evolution and, for example, Hox genes from flies can replace the function of their vertebrate orthologs.
The function of Hox genes has been (and still is) largely studied during the embryonic development of a large number of invertebrate and vertebrate species. During regeneration, specially in those cases in which large structures or body regions need to be re-grown, it is easy to imagine that Hox genes should have an important function in the re-establishment of the positional identity within the regenerate. However, and in contrast to the observations that conserved signalling pathways such us the BMP, Wnt/beta-catenin, Hedgehog and FGFs play important functions during regeneration, many less functional studies have been reported on the role of Hox genes during regeneration. Thus, whereas several families of genes containing an homeobox domain (pbx, prox, pitx, six, Rx, msx,…) have been functionally characterized during regeneration in planarians, tunicates or amphibians, most of the studies on Hox genes and regeneration focus on their changes in expression during this process. Some of those studies were important for example to show how Hoxc10L is not expressed during axolotl forelimb development but it is upregulated during regeneration, pointing to a “regeneration-specific” expression of this gene (http://www.ncbi.nlm.nih.gov/pubmed/11150241). Also, few years ago it was reported that Hoxc13 orthologs are important for zebrafish tail regeneration where are required for blastema cells proliferation and growth (http://www.ncbi.nlm.nih.gov/pubmed/17437127).
Now, a recent paper from the laboratory of Elena Nokikova and Milana Kulakova reports on the expression of 10 Hox genes during regeneration in the polychaete Alitta (Nereis) virens (http://www.ncbi.nlm.nih.gov/pubmed/23638687). During postlarval development the Hox genes are mainly expressed in a posterior-to-anterior gradient mode. These annelids can regenerate their posterior body end. After amputation the terminal pygidial structures and prepygidial growth zone (GZ) form first and then the new segments appear sequentially. The authors divided the Hox genes into four groups depending on their changes in expression during regeneration. Three genes (early genes: Nvi-Lox5, Nvi-Lox2 and Nvi-Post2) were rapidly upregulated in the nervous system near to the amputation plane by 4 hpa (hours post amputation). Then, Nvi-Hox5 and Nvi-Hox7 (middle genes) expression patterns in the nervous system were reorganized by 10 and 18 hpa, respectively, before active proliferation in the GZ started. Next, two Hox genes, Nvi-Hox2 and Nvi-Hox3 (middle genes), that are not expressed in a graded manner but specifically found in the GZ were upregulated de novo by 10 hpa. Nvi-Hox2 first appeared in two bilateral domains at the amputation plane. By 2 dpa (days post amputation) this gene was expressed in the mesoderm and ectoderm of the area between the forming pygidium and the last body segment. By 7 dpa it was detected in the mesoderm of the GZ and the mesoderm and ectoderm of the newly formed segments. Nivi-Hox3 was also first seen by 10 hpa and as regeneration proceeds got restricted to the ectoderm of the GZ. Finally, there are three genes (late genes: Nvi-Hox1, Nvi-Hox4 and Nvi-Lox4) whose expression patterns changed at late stages of regeneration once proliferation and organogenesis were under their way.
Based on all these expression patterns and dynamics the authors divided the regeneration process in two phases: during the first 48 hpa the expression patterns were reorganized inside the new body boundaries. The second phase (that overlaps with the first one) started around 24 hpa when the blastema was formed. Most of the Hox genes were highly expressed within the blastema and the rudiment of the terminal structures that were evident by 3 dpa.
Because during postlarval development Hox genes are expressed in a anterior-to-posterior gradient in segments that are morphologically similar, the authors suggest that Hox genes at this stage provide the positional information needed to determine the position of body parts. Thus, after amputation the expression of most of the Hox genes of the early and middle groups is reorganized so the last body segments adjacent to the amputation plane acquire the Hox pattern typical of the posterior body end. Remarkably the blastema seems to be formed after the Hox genes have been reorganized to the new body proportions.
Future functional experiments should help to determine the exact role of the Hox genes during not only the re-patterning of the body during polychaete regeneration but also in other models such as planarians or amphibians.
Homeostatic signalling interferes with fin regeneration in male zebrafish
There are several cases in which regenerative capabilities vary depending on the regenerative stage of the organisms. Some examples of this loss of regeneration include that of limbs in post-metamorphic amphibians and the heart in one week-old mice. Now a recent paper from the laboratory of Kenneth Poss describes how the pectoral fins in zebrafish show a sexually dimorphic response to amputation and links that to the maintenance of male-specific structures and the regulation of the Wnt signalling pathway (http://www.ncbi.nlm.nih.gov/pubmed/24135229).
In a previous paper the same laboratory had found that pectoral fin regeneration is often impaired in males whereas it proceeds normally in females (http://www.ncbi.nlm.nih.gov/pubmed/22079110), being such impairment partially rescued by overactivating the Wnt/b-catenin pathway. Regeneration failure correlates with sexual maturity indicating an age- and sex-dependent loss of regenerative abilities for the male zebrafish pectoral fins. In this new paper the authors characterize in more detail the molecular basis of such sexual dimorphism and focus on the study of Dkk1, an inhibitor of the Wnt/b-catenin signalling pathway, which is well-known for being required for a successful regeneration in many contexts and species, including zebrafish fins. dkk1 was found to be expressed at higher levels in male pectoral fins that in females. By using a transgenic line the authors found that dkk1 displayed a sexual dimorphic expression being detected in what they called epidermal tubercles (ETs) in the anteromedial rays of the pectoral fins of males. Because similar structures had been described in other fishes and associated to mating and spawning, the authors first addressed the role of ETs in spawning. What they found was that zebrafish males used the pectoral fin to grasp the female abdomen to stimulate egg laying. In fact, males with amputated pectoral fins appeared mostly incapable of stimulating an efficient spawning.
Next, they analysed the role of androgens in the development of ETs. Remarkably, an androgenic factor was able to induce the development of ETs in the pectoral fins of females; conversely, the application of an androgen inhibitor in males decreased the number and definition of ETs. These changes were accompanied with an increase or decrease of dkk1 expression, respectively. In order to characterize the role of the Wnt/b-catenin pathway in the formation of ETs they crossed a transgenic line bearing a Wnt signalling reporter with the line with labelled dkk1-expressing cells. Following the maturation of ETs they found 3 distinct types of ETs: immature structures where Wnt signalling is active and dkk1 is not expressed, intermediate cells expressing dkk1 and active Wnt signalling and mature ETs expressing dkk1 and no active Wnt signalling. Because the balance of activated/inhibited Wnt signalling is important for the induction and/or patterning of epidermal appendages the authors further analysed the role of Wnt/b-catenin on ETs formation. Experiments inhibiting or overactivating Wnt signalling suggested that that this pathway has a positive role in the formation of new ETs. In fact, ETs go through a continuous cell turnover from a basal layer cell population(s) between adjacent ET units, being this renewal dependent on local dkk1 regulation and Wnt activation.
During fin regeneration, at 2 days post amputation (dpa) Wnt signalling is activated in early blastema cells. As regeneration proceeds Wnt pathway keeps active in the proliferative distal blastema, whereas dkk1 is expressed in more proximal differentiating cells. This dynamics was observed in different fins in both females and males, including female pectoral fins and posterior regions of male pectoral fins. Defects in male pectoral fin regeneration mostly appeared when amputation was performed across a region containing ETs. Further analyses on male pectoral fin regeneration revealed that the wound was capped by epidermal cells expressing high levels of dkk1 with little proliferative activity underneath. This was in contrast to females in which dkk1 was not expressed in the blastema cells at this stage (2 dpa). Also, whereas in females axin2, a known target of active Wnt signalling was detected in blastema cells, no expression was found in males clearly indicating a failure in Wnt activation in those regenerating males. As regeneration proceeded males appeared to achieve different degrees of recovery. After several weeks of regeneration males were then classified in three groups depending on whether the defects in their regenerated pectoral fins were mild or severe. Remarkably, males with severe defects were incapable of stimulating spawning during mating.
In summary, this paper describes the molecular basis of the sexual dimorphism shown by regenerating pectoral fins in female and male zebrafish. Thus, a fine balance between active and inactive Wnt signalling is pivotal for proper regeneration. In males, the high level of dkk1 required in ETs to maintain these basic structures for a successful mating appears to come at the expense of reducing their regenerative potential.
Neural progenitor cells in planarians
Planarians are really amazing creatures as they can regenerate a whole animal from a tiny piece of their bodies. Planarians can do that because they possess a population of adult pluripotent stem cells, called neoblasts. For many years one of the debates within the field has been how homogeneous or heterogeneous is this neoblast population. Thanks to a study from the laboratory of Peter Reddien we know now that at least a proportion of those neoblasts are real totipotent stem cells. But how do these neoblasts, the only proliferative cells in planarians, differentiate into the different cell-lineages? Are there cell type-specific progenitors in these animals? Recently, some studies from Peter Reddien clearly indicate that at least for some cell types, such as photoreceptors and the excretory system, specific progenitors exist.
One of most astonished abilities of planarians is that they can regenerate a complete central nervous system de novo from those undifferentiated neoblasts. For long time people have wondered whether neural progenitor cells exist in these animals and how they would behave during regeneration. Two very recent papers from the laboratories of Kerstin Bartscherer (http://www.ncbi.nlm.nih.gov/pubmed/24131630) and Bret Pearson (http://www.ncbi.nlm.nih.gov/pubmed/23903188) have shown that the transcription factors lhx175-1 and pitx were required for the regeneration of the serotonergic lineage. Importantly, these transcription factors were co-expressed in cells expressing Smedwi-1, a homologue of the PIWI proteins and a marker of planarian neoblasts, suggesting the existence of progenitor cells for the serotonergic lineage.
Now these findings have been further corroborated and expanded by the laboratory of Ricardo Zayas (http://www.ncbi.nlm.nih.gov/pubmed/24173799). In this study they carried out a genome-wide analysis of bHLH transcription factors in planarians. In other models bHLH factors play pivotal roles in neurogenesis, from fate commitment to cell migration. In planarians, the authors identified 44 genes predicted to code for a bHLH domain, of which 12 were expressed in the CNS and neoblasts. Because of their specific expression patterns they mainly focussed on three of them: coe (collier/olfactory-1/early B-cell factor), hesl-3 (hairy/enhancer of split) and sim (single-minded). By double labelling they first checked how these factors were co-expressed with different markers of specific neuronal populations: cholinergic, GABAergic, octopaminergic, dopaminergic and serotonergic neurons. Next, in intact non-regenerating animals and through elegant BrdU pulse-chase experiments, they detected proliferating cells expressing coe or sim close to the nervous system, which could be traced to the brain or ventral nerve cords. Over time, the number of such double-labelled cells increased, especially in the head region. Similarly, during anterior regeneration these populations of progenitor cells expressing coe or sim, seems to contribute to the regenerative blastema. Therefore, it seems that these transcription factors would be labelling progenitor cells for distinct neuronal populations.
In the second part of the paper the authors performed RNAi analyses of this set of bHLH transcription factors in order to determine their functions during regeneration. Remarkably, they did not find any strong phenotype after knocking-down proneural bHLH genes such as neuroD or acheate-scute. However, silencing of coe, hesl-3 and sim resulted in clear defects of the regenerating brain either in terms of its gross morphology and/or the number or localization of specific neuronal populations. On the other hand, in intact animals the silencing of hesl-3 and sim did not produced any detectable defect in the nervous system. This was different for coe, as after RNAi in intact animals these animals displayed an aberrant external morphology and they lost specific neuropeptidergic neurons (also lost in regenerating animals after coe RNAi).
In summary, these results suggest that coe, sim and hesl-3 may define progenitor cells committed to distinct neural fates and their function would be required for the differentiation of some neuronal cell types. As coe and sim are co-expressed with specific markers of different neuronal populations, their expression could be defining a set of multipotent progenitors.
regeneration in nature 