As discussed before in this blog mammalian central nervous system fails to regenerate because of a combination of extrinsic and intrinsic inhibitory factors. So far, and despite different approaches based on attempts to neutralize those inhibitory cues, recreate permissive environments or transplant stem cells, there is no promising therapy for spinal cord injury patients yet. Therefore it is important to understand how other vertebrates do to regenerate their central nervous system. Among these, zebrafish are able to regenerate their spinal cord after injury. A recent study from the laboratory of Melitta Schachner reports on the role of syntenin-a in promoting spinal cord regeneration in zebrafish (http://www.ncbi.nlm.nih.gov/pubmed/23607754).
Syntenins are scaffolding proteins identified as syndecan-binding proteins and that contain multiple PDZ domains to bind the cytoplasmic domain of a variety of transmembrane proteins. Syntenins have been involved in cytoskeleton signalling, protein trafficking, cell adhesion, migration, activation of transcription factors and cytoskeletal reorganization in some cancer lines. Based on microarray data syntenin-a is one of the genes upregulated by 11 days post-injury. qPCR analyses confirm this upregulation of syntenin-a at 6 and 11 days after SCI (spinal cord injury), that correspond to the chroninc response phase of the spinal cord to the injury. However, no upregulation is detected during the early acute phase response (4-12h after SCI). Whole-mount in situ hybridizations validate this upregulation as there is a significant increase of cells expressing syntenin-a adjacent to the lesion site 6 days after SCI. Both neurons and glial cells upregulate the expression of syntenin-a. Next and in order to determine whether this upregulation of syntenin-a is required for locomotor function recovery after SCI the authors knocked-down this gene with morpholinos. After silencing syntenin-a both the duration of the movement and the total distance moved was significantly reduced compared to controls at 6 weeks post-SCI, suggesting an important role of syntenin-a in locomotor recovery. By performing retrograde and anterograde labelling the authors showed how there is significant decrease in the number of neurons, regrown axons and newly formed synapses at the lesion site, after knocking down syntenin-a.
After SCI in zebrafish there is an early phase of glial cell accumulation and migration to the injury site. Later, there is a second phase characterized by the formation of a glial bridge that promotes regeneration by facilitating the growth of the axons over the injury site. Syntenin-a is upregulated after 6 days post-SCI that correlates with the formation of the glial bridge supporting the idea that syntenin-a might have a role in the formation of this glial bridge and/or changes in glial morphology during this phase. In addition, as syntenin-a also plays a role in cell migration during zebrafish development it is also possible that this factor could function during glial cell migration during spinal cord regeneration.
Given the wide variety of partners and interactions in which syntenin-a may be involved it will be important to determine which specific processes mediate this beneficial effect of this protein during spinal cord regeneration.
In some previous posts I have commented on the pivotal role that epigenetic regulation may have on regeneration. It is known that in mammalian embryonic stem cells chromatin regulation is fundamental to control self-renewal, pluripotency and differentiation. Freshwater planarians are truly champions of regeneration and, remarkably, they carry out it through a population of adult pluripotent stem cells, the neoblasts. The neoblasts are the only proliferative cells in these animals and upon amputation they give rise to all the cell types needed to restore the missing parts. An important question is how similar are neoblasts and mammalian stem cells at the level of genetic programs and epigenetic regulators. Previous works have, in fact, highlighted a significant conservation between some of these factors as well as their functional relevance during planarian regeneration (http://www.ncbi.nlm.nih.gov/pubmed/23235145 , http://www.ncbi.nlm.nih.gov/pubmed/22543868).
Now, a paper from the laboratory of Qing Jing further expands our knowledge on the role of chromatin regulators during planarian regeneration (http://www.ncbi.nlm.nih.gov/pubmed/23629965). From the planarian draft genome they first identified about 210 genes that contained motifs common to conserved chromatin regulators. Upon RNAi silencing 12 of them gave a regeneration phenotype meaning that regeneration was, in fact, severely impaired. These genes belonged to at least six chromatin complexes: CAF1, BAF, NuRD, FACT, Cdk-activating kinase and MCm2-7 complex. All of them were enriched in the neoblast population and their silencing decreased the expression of typical neoblast markers such as smedwi-1.
Most of the paper is focussed on the characterization of a homologue of the HP1 (heterochromatin protein 1) family. This is a conserved family whose function in stem cells is not well characterized yet. Planarians possess two HP1 genes, but only one of them, HP1-1 seems required for neoblast function. When ectopically expressed in NIH3T3 cells HP1-1 is localized in the nucleus. Different experiments show that HP1-1 is expressed in neoblasts that also co-express smedwi-1. Upon its silencing a normal blastema seems to be initially formed at 1 day of regeneration but after that it does not grow and in fact it regresses, the animals curl and finally die in the typical pattern associated to neoblast depletion. The function of HP1-1 is equally required in intact non-regenerating planarians to sustain homeostatic cell turnover. Through BrdU labelling and double stainings with early and late neoblast progeny markers the authors conclude that HP1-1 is important for neoblast self-renewal and its silencing leads to a failed neoblast proliferative response that results in their premature differentiation.
In order to better characterize the mechanism through which HP1-1 functions the authors followed two strategies: i) they did microarray analyses to identify genes up- and down-regulated after HP1-1 silencing, and ii) they functionally characterized candidate genes that can bind HP1-1 and whose silencing gives similar defective regeneration phenotypes. By doing it, they found two genes of the FACT complex, SSRP1 and Spt16 that co-localize with HP1-1 in neoblasts. In fact, SSRP1 coimmunoprecipitates with HP1-1. As SSRP1 seems to be important for transcription, the authors suggest that HP1-1 and SSRP1 may cooperate to activate gene transcription during regeneration. Finally the authors performed additional microarray analyses to characterize genes missregulated after silencing SSRP1. When comparing the profiles obtained after HP1-1 and SSRP1 RNAi they identified 85 genes shared in both lists that are decreased upon their silencing. One of them is Mcm5 and chromatin immunoprecipitation (ChIP) suggest that HP1-1 protein binds the promoter region of Mcm5. Overall these results allow the authors to suggest that HP1-1 and SSRP1 activate Mcm5 during transcription and that is necessary to support proliferation and self-renewal of planarian stem cells.
These results will be important to understand how adult pluripotent stem cells are maintained and how respond to amputation. This work highlights also how planarian stem cells are regulated by conserved chromatin factors found also in mammalian ES cells and how studying their function in these flatworms can provide novel information on their function on stem cell biology.
In the field of regeneration two important and recurrent questions await for a satisfactory and “definitive” answer: 1) are there regeneration-specific genes that could be missed in species with poor regenerative capabilities? and, 2) up to what extend regeneration is a mere recapitulation of embryonic development? Most probably there will not be a common and unique answer for these questions that can be applied to all those classical models of regeneration.
In a recent paper from the laboratory of Elly Tanaka the authors have followed a high-throughput comparative transcriptomic approach to try to answer those questions in the axolotl limb regeneration paradigm (http://www.ncbi.nlm.nih.gov/pubmed/23658691). In this work the authors have carried out a comparative analysis of the transcriptional profiles of regenerating limbs, healing severe wounds and developing limb buds at different time points. Remarkably, the time course analysed has been impressively extensive with samples taken at 0, 3, 6, 9, 12, 24, 36, 54, 72, 120, 168, 288 and 528 hours after wounding or amputation.
Whereas most animals trigger a proper wound healing response after injury, only in few cases this wound healing is proceeded by the activation of a successful regeneration program. Therefore, by comparing the gene expression response to severe non-regenerative wounding (without amputation) to that triggered by amputation the authors sought first to identify common and distinct genes activated in those two contexts. One first remarkable result from these analyses is that they identified a molecular tripartite program that parallels the three phases of limb regeneration previously described based on morphological and cellular observations. Those phases comprise, basically: wound healing, blastema formation and growth of the new limb (with the establishment of a limb development program). This last phase resembles very much the normal growth of the limb bud during embryonic development. At the molecular level, the authors identified three distinct phases: 1) an initial phase up to 12 hours in which the genetic program activated after amputation resembles very much the activated program triggered by non-regenerative wounding; 2) from 24 up to 72 hours the amputation gene profile starts to diverge from the wound samples and, therefore, 24 hours represent the time point at which the regeneration-specific gene program is first discernible from the wound-healing program. And finally, 3) the amputated samples from 120 to 528 hours cluster more closely with the developing limb bud than to their corresponding wound samples. Thus, the 120-528 hours limb blastema would establish a limb development-like program.
A second interesting observation is that after amputation and non-regenerative wounding the behaviour of G1/S-phase genes and G2/M genes follow a similar dynamics in both context. Thus, the expression of G1/S genes rises to an initial peak at 72 hours whereas G2/M-associated genes peak at 120 hours. As this response is similar after amputation and wounding it seems reasonable to consider that this early proliferative response is associated to tissue injury. However, the samples coming from amputation display a second phase of expression of genes associated to proliferation that correlates to the phase in which the blastema is being formed. Thus, whereas after wounding there is just an initial proliferative response a second one appears to be specific for regeneration. As the authors point out these two waves of proliferation, one associated to injury and the second associated to regeneration and blastema formation resembles very much to what it has been previously described in freshwater planarians (http://www.ncbi.nlm.nih.gov/pubmed/20599901).
Next, the authors identified a set of 194 “regeneration-specific” genes, many of which fall into the GO categories, at the level of cellular function, of: 1) cellular stress, 2) chromatin associated factors, 3) epithelial function and differentiation, and 4) limb development module. Whole-mount in situ hybridizations validate the expression of many of those genes in the wound epidermis and the mesenchyme of the blastema. Finally, they wanted to identify those genes that would be present in the regenerative blastema but not in the limb bud. The reason for that is that although the growth of the blastema (once it is formed) into a new limb resembles very much the development of a limb, the initial phases of the blastema formation are quite different from the differentiation of the embryonic limb bud. Through comparative analyses they identified 20 genes that are expressed at low levels in the limb bud and are upregulated specifically after amputation. Many of those genes fall into the GO category of epidermal development and differentiation, which may highlight important functional differences between the epidermis of the blastema and the limb bud.
In summary the authors have identified here a set of regeneration-specific genes. Further analyses will determine the function of those genes during the phase of blastema formation in limb regeneration and may provide novel insights into this amazing process, not only in amphibians but also in other regeneration models.
Upon injury inflammation is one of the first responses that are triggered. Modulating the inflammatory response may have important consequences to enhance the regenerative capabilities in different systems. Thus, too little or too much inflammation can inhibit or alter processes such us wound healing or cell proliferation that are required for proper repair and regeneration. As previously mentioned in this blog, the failure of mammalian CNS to regenerate depends in great measure on the inhibitory action of myelin and the glial scar. This is also true for the axons of the retinal ganglion cells (RGC), that after optic nerve injury normally die by apoptosis. Previous works have shown that the JAK/STAT3 and PI3K/Akt signalling pathways can enhance the regenerative response of RGC axons. Similar enhancement can be seen after inflammation stimulation, that does not only promote axonal regrowth but also plays a neuroprotective function as it increases the survival rate of RGCs upon injury.
In a recent paper from the laboratory of Dietmar Fischer the authors show how interleukin-6 is an important mediator of RGCs regeneration upon inflammatory stimulation (http://www.nature.com/cddis/journal/v4/n4/abs/cddis2013126a.html) in rats. IL-6 is a cytokine that is lowly expressed in the CNS but strongly upregulated after ischemia, trauma or axotomy in the peripheral nervous system. Here, the authors first show that IL-6 is upregulated in the retina after optic nerve crush or inflammatory stimulation. In retinal cell cultures IL-6 increases neurite growth not only in permissive substrates but also in inhibitory substrates as CNS myelin extracts. Moreover, the presence of a IL-6 antibody in the medium blocks this IL-6-stimulated neurite outgrowth. Also, IL-6 exerts a neuroprotective role on the cultured RGCs, although this effect is lower compared to those exerted by other neuroprotective and axon-promoting factors such as CNTF (astrocyte-derived ciliary neurotrophic factor). Consistent with this function of IL-6, the authors show how a IL-6 receptor is expressed in retinal cells as well as in Müller cells. When cells are cultured with an antibody against IL-6 receptor, the growth-promoting effect of IL-6 is significantly reduced. Also, when this receptor is activated by another cytokine it induces neurite growth at a similar level as IL-6 does. These results indicate that the stimulation of the IL-6 receptor is sufficient to promote neurite growth.
The effects of IL-6 on RGCs regeneration is mediated through the activation of the JAK/STAT3 and PI3K/Akt pathways. Thus, IL-6 promotes the phosphorilation of STAT3 and, on the other side, drug-mediated inhibition of the of these two signalling pathways blocks the regeneration response induced by the treatment with IL-6. Finally, the authors also check the effects of IL-6 in vivo. To do that they injected IL-6 protein after optic nerve crush and saw the activation of the JAK/STAT3 pathway as well as the upregulation of some genes associated to regeneration such as Sprr1a, galanin and gap-43. Retinal cell cultures from rats that went through optic nerve crush and posterior IL-6 injections show a significant increase in neurite outgrowth, suggesting that IL-6 transforms those RGCs into a regenerative state.
In summary, IL-6 is an important element for optic nerve regeneration mediated by inflammatory stimulation. Further studies will be necessary to see up to what extend this system can become a therapeutical target to promote optic nerve regeneration in mammals.
In some previous posts I have commented on the role that matrix metalloproteinases (MMPs) play during regeneration. Among others, one important aspect in which MMPs play an important function is in allowing a type of wound healing that reduces the presence of a fibrotic scar that may block the regenerative response. In fact, a scarless healing is very important for creating this regeneration-permissive environment. Although adult mammals do not really regenerate well, at the fetal stage this is quite different (at least for certain tissues and structures). Thus, fetal wounds in the skin heal without any fibrotic scar and, consequently, a normal skin architecture and function can be regenerated. However, the adult skin when wounded elicits the formation of a fibrotic scar that blocks regeneration. Among the factors that appear to control the transition from a scarless healing to a fibrotic scar response we may find the development of the immune system, the ratio between different isoforms of TGF-b expressed, or the presence of PDGF (platelet-derived growth factor).
MMPs are another important factor in the differences between how mammalian fetal and adult skin respond to wounding. A recent paper from the laboratory of Yong Li shows how the regeneration of soft tissues can be enhanced by MMP1 during digit regeneration in mice (http://www.ncbi.nlm.nih.gov/pubmed/23527099). During wound healing, MMPs are required to promote ECM (extracellular matrix) degradation that is necessary also to favour the migration of cells into the injury site. In this paper, the authors analyse the effects of applying exogenous MMP1 into adult mice digits amputated through the middle phalanx bones. In terms of the final size of the digits regenerated in control and MMP1-treated amputated digits the authors did not find any significant difference. However, whereas the wounds were almost completely closed by day 10 after MMP1 treatment, control wounds needed more time to heal. This indicates that MMP1 has a positive impact on the healing of soft tissues in those amputated digits. Next, the authors show how this faster healing correlates with an improved revascularization after MMP1 treatment. In addition, and as innervation is important in many regenerating contexts, such as limb regeneration, the authors checked here the amount of NCAM present in the peripheral nerve fibers and neuromuscular junctions. They also found that NCAM increases in those MMP1-treated amputated digits.
As said above, one of the factors that block regeneration in may cases is the excess of fibrotic connective tissue deposited in the ECM space of wounded tissues. It is known that MMP1 digest, for example, collagens types I and III that contribute to the fibrotic scar. This collagenase activity of MMP1 is repressed by TGF-b1 and, in fact, TGF-b1 is more abundant in the adult skin than in the fetal one. Also, knocking down TGF-b1 reduces scar formation. Here, the authors show how MMP1 treatment reduces significantly the deposition of collagen in the ECM, and therefore the fibrotic scar. Finally, the authors use Sca-1 (Stem cell antigen-1) that labels hematopoietic and other types of progenitor cells and found an enrichment in the number of positive cells in the MMP1-treated amputated digits.
In summary, the authors show, by using different markers, that MMP1 promoted a faster wound closure, the regeneration of soft tissues and a decrease in scar tissue formation. However, no beneficial action of MMP1 is observed for the bones or the final size of the regenerated digit. Future work should help to characterize better the targets of MMP1 and try to find factors to enhance the regeneration of the skeletal tissues.