V-ATPase proton pumping activity is required for adult zebrafish appendage regeneration

Cell proliferation during blastema formation is controlled by several signalling pathways, including Fgf, Wnt/β-catenin and Retinoic acid (RA). We assessed the link between V-ATPase and these signalling pathways by comparing the expression of mkp3, wnt10a and aldh1a2 during proximal and distal regeneration in control fins (AB WT fish) and in atp6v1e1b knocked down fins (vivo-MO delivered 2 hpa). In the controls, mkp3 and aldh1a2 expression domain was wider at proximal stumps, at both 24 and 48 hpa (Fig. 5 compare F–G; J–K; N–O; R–S). This pattern was somewhat similar to what was observed for the V-ATPase. On the contrary, in V-ATPase knocked down fins, aldh1a2 and mkp3 were absent at 24 hpa at either amputation planes, demonstrating that V-ATPase affects the expression of both genes (Fig. 5H–I; P–Q). At 48 hpa, these transcripts' expression was re-established, likely as a consequence of the vivo-MO transient effect (Fig. 5L–M; T–U). Expression of wnt10a remained unchanged in control and V-ATPase knocked down fins for all time points, in proximal and distal stumps alike (File S6). These results show that V-ATPase is required for the normal expression of at least aldh1a2 and mkp3, establishing a molecular link between V-ATPase and two important signalling pathways that control cell proliferation, Fgf and RA.

Proper tissue innervation is another important factor for normal fin regeneration [39]. As such, it was decided to test if any relation to the H+ pump existed. For that, atp6v1e1b knocked down fins (vivo-MO delivered 2 hpa) were immunostained for the axonal marker acetylated α-tubulin. 24 hours after the proximal amputation of non-treated controls, nerve axons extended mainly along intra-ray regions, forming bundles parallel to the proximal-distal axis of the fin. Actually, these bundles almost reached the amputation plane, after an initial retraction from the stump that typically occurs upon amputation (Fig. 6A). At 48 hpa, the control stumps were innervated by axons sproutings (Fig. 6B). After V-ATPase knockdown, fins were less innervated at 24 hpa and axons remained distant from the amputation plane (Fig. 6C). Later, at 48 hpa, axons were beginning to invade the regenerating tissue in proximal stumps (Fig. 6D). Similar results were found in distal regenerates (File S7). These results show that V-ATPase is required for normal innervation of the regenerating fin.

Our data showed that H+ efflux is specifically established during appendage regeneration in an adult vertebrate. The efflux was first detected several hours upon amputation, and it was present for at least 5 days after wound closure. Previous reports in vertebrates including humans have showed that regeneration is accompanied by an endogenous electric current (EC) for several days after wound healing [13] [14] [40], and this EC seems to be controlled by the activity of specific ion transporters [14] [17]. Since the wound is already closed when an efflux is detected, we propose that H+ efflux contributes to the regeneration-specific endogenous EC, probably together with sodium flux, which is a crucial component of regeneration-specific EC, and other ions such as potassium and chloride [13] [14].

The molecular source of regeneration-specific ion fluxes has only recently begun to be unveiled: the H+ pump V-ATPase and the voltage-gated sodium channel NaV1.2 are essential for the regeneration of Xenopus larvae tail [18] [22], and the H+,K+-ATPase is required for planarian head regeneration [19]. Nevertheless, evidence in other models for conservation of these mechanisms, especially in adult vertebrate models, is still lacking. In this work, we have showed for the first time that the V-ATPase is upregulated in the regenerating tissue of the adult zebrafish caudal fin. We took advantage of the fact that different manipulations (eg. injection of drugs or morpholinos, amputation planes) can be held on separate parts of one single fin to provide an extra level of control to our experiments that could not be examined in other models such as the Xenopus larval tail [18]. We demonstrate that the onset and magnitude of V-ATPase expression varied with the amputation plane along the PD axis and H+ efflux followed a similar pattern. Besides, inhibition of this H+ pump after either proximal or distal amputation decreased H+ efflux. Thus, V-ATPase contributes to the regeneration-specific H+ efflux in the adult zebrafish caudal fin.

In the atp6v1e1b knocked down fins, the H+ efflux decreased at 24 hpa but increased from 48 hpa onwards compared to the control. This phenotype reversal probably reflects the decreasing inhibitory effect of the MO with time [41]. Two interesting facts were that the efflux increase at 48 and 72 hpa closely resembled the control efflux level at 24 and 48 hpa respectively, and for each time point, efflux intensity was always smaller in distal stumps than in proximal ones. Taken together, our data shows that during regeneration V-ATPase activity and the consequent H+ efflux are tightly controlled at the genetic level in order to reach a specific intensity that depends on the amputation position. These observations suggest that the V-ATPase has an active role in regeneration and is involved in the formation of position-dependent cues.

V-ATPase is ubiquitously expressed in eukaryotic cells and is not only required for several housekeeping functions that depend on pH or membrane potential. Additionally, it is strongly upregulated in particular cell types where it is necessary for specific processes such as bone formation/resorption and renal acidification [42]. In the amputated adult zebrafish caudal fin, V-ATPase was upregulated in the regenerating tissue during blastema formation and early regenerative outgrowth, and its inhibitions impaired regeneration. Therefore, V-ATPase seems to be required for the regenerative process in some role other than housekeeping functions.

In the Xenopus tadpole tail, the V-ATPase-mediated H+ extrusion is necessary for cell proliferation in the blastema-like regeneration bud [18]. Here, we show that V-ATPase inhibition decreases regeneration by decreasing proliferation in the mature blastema (48 hpa), but not during early blastema formation (24 hpa). Interestingly, H+ efflux, mediated by ion transporters such as sodium/proton exchanger 1 (NHE1) and V-ATPase, is a major contributor for cell migration by promoting extracellular matrix degradation [43] [44]. As the blastema arises from intra-ray cells that proliferate and migrate distally [3] [4], V-ATPase could be required for the migration of blastema-forming cells from the intra-ray mesenchyme into the regenerating region.

An alternative view is that cells migrate to the blastema but become unable to proliferate in the absence of V-ATPase. In fact, V-ATPase has been associated to proliferation in other regeneration models [18] [23]. Fgf signalling is required for cell proliferation in the blastema but not in cell populations below the amputation plane [38], and the same influence on cell proliferation was found when V-ATPase activity was inhibited. Besides, V-ATPase knockdown inhibited the expression of mkp3, a target of Fgf signalling, demonstrating that this H+ pump interferes with this signalling pathway. V-ATPase was also required for the expression of aldh1a2, a key molecule for RA synthesis. It is known that both RA and V-ATPase mediate cell survival by upregulating bcl2 expression [45] [46], so it would be interesting to investigate if the two molecules act as part of a single, common pathway. Importantly, it was recently demonstrated that RA and Fgf signalling, together with Wnt/β-catenin, regulate each other in a positive reciprocal manner and modulate the overall rate of regeneration [47]. Considering this, V-ATPase seems to be another component of this interconnected molecular net by affecting RA and Fgf signalling in a direct or indirect manner. Additionally, an effect of V-ATPase on other signalling pathways, including Wnt/β-catenin, cannot be excluded. Although wnt10a was not affect by the pump's inhibition, it has been showed in other systems that V-ATPase is required for Wnt/β-catenin signalling downstream of Wnt ligand [24].

Inhibition of V-ATPase decreased the amount of axons in the fin and inhibited axonal growth into the regenerating tissue. It is possible that this hindered innervation was a direct result of the decrease of V-ATPase mediated H+ efflux since it is known that endogenous electric currents and associated electric fields can control the amount of nerve sprouting and the direction of axonal growth into the regenerating region in other models [47]. Besides, in Xenopus froglet limb regeneration, nerve supply is not necessary for blastema formation but is required for blastema outgrowth, by controlling cell proliferation, cell survival and expression of Fgf signalling genes in the blastema [39]. Our results showed a similar effect of V-ATPase on blastema cells. Taken all, V-ATPase effect on blastema outgrowth may be mediated by nerve supply, though independent routes for V-ATPase and nerve supply effects cannot be excluded.

Although larval fin fold of atp6v1e1b hi577atg-/- zebrafish regenerated normally in the absence of V-ATPase activity, our results clearly show that V-ATPase is required for regeneration in adult fish appendages. In fact, despite the reported conservation of molecular events during regeneration of adult fins and larval fin fold [5] [28], recent in vivo cell tracing experiments have showed that the fin fold blastema does not have a specific function for proliferation, and in that way it is not a classical blastema as observed in the adult system [48]. Thus, regeneration of the adult and larval caudal appendages in zebrafish are distinct processes. On the other hand, some remarkable parallels have been observed between regeneration of the adult zebrafish caudal fin and Xenopus tadpole limb and tail buds [49], including conserved molecular pathways and dependence on blastema-restricted cell proliferation. Particularly, our work and others [18] demonstrate that, in both models, V-ATPase is required for adequate nerve supply and cell proliferation in the blastema. Taken all, we propose that the V-ATPase has a conserved role in regeneration events that depend on a blastema with specific proliferative function.

One important property of the blastema is positional memory, which instructs both the amount and the rate of regeneration so that the missing structures are replaced in the correct 3D pattern and the process is completed simultaneously, regardless the level of amputation along the PD axis [38] [50]. Position-dependent regeneration rate is regulated by the level of expression of several molecules during regeneration [50]. In zebrafish, those include Fgf signalling and msxb, which have enhanced proximal expression compared to distally amputated fins [37] [38]. RA is also a major instructor of positional information in several vertebrates, but its role in zebrafish appendage regeneration has proven difficult to assess [51]. We showed that aldh1a2 has stronger proximal expression compared to distal stumps, adding new evidence that agree with a role for RA in positional memory in zebrafish. Moreover, our results showed that V-ATPase and H+ efflux follow a position-dependent pattern with increased proximal expression, while other regeneration markers, such as wnt10a, maintain a similar expression regardless the amputation level. Besides, V-ATPase knockdown decreased proliferation in the blastema and inhibited aldh1a2 and mkp3 expression. Altogether, these data agree with a role for the V-ATPase in position-dependent regeneration rate, by affecting blastema proliferation through the modulation of at least two essential signalling pathways, Fgf and RA.

To further investigate the role of V-ATPase in position-dependent regeneration rate, we knocked down atp6v1e1b at different time points after proximal and distal amputation. The most dramatic regeneration reduction was obtained when the gene knockdown approximated the different onset of H+ efflux at proximal and distal positions, around 3 and 12 hpa, respectively. In addition, the decrease in the regenerate area was more pronounced proximally, demonstrating that regions of higher regeneration rate have stronger dependence on V-ATPase activity. This suggests that H+ efflux triggers some important mechanism that needs to be activated earlier in the highly proliferating proximal wounds.

Overall, our results suggest that V-ATPase H+ pumping activity is part of the signals that translate position into adequate cell behaviour, including position-dependent proliferation rate. In this way, V-ATPase has a role in positional memory system. We propose that some unknown position-instructor signal upstream of V-ATPase sets the level of expression for this H+ pump according to the level of amputation along the PD axis. Then, localized V-ATPase H+ pumping activity in the blastema generates pH and/or voltage domains within the regenerating tissue, that will act, directly or indirectly (for example, via inhibition of nerve supply), as positive regulators of RA and Fgf signalling, ultimately affecting cell proliferation in the blastema. As V-ATPase gene expression is activated earlier and more strongly in proximal stumps than in distal ones, pH/voltage domains would be more intense and maintained for a greater period of time after proximal amputation. Consequently, V-ATPase activity would exert a stronger influence on gene expression in proximal stumps than in distal ones, including Retinoic Acid and Fgf signalling, ultimately setting a higher regeneration rate in proximal amputated fins.

In 2007, a study from Adams et al. [18] demonstrated that the V-ATPase is necessary for Xenopus tadpole tail regeneration. In the present study, our findings further support that the V-ATPase is important to adult vertebrate regeneration. Particularly, our results indicate that the V-ATPase H+ pumping activity orchestrates with major molecular signalling pathways to control position-dependent regeneration rate. Understanding of this and other ion-driven mechanisms underlying adult regeneration may open way for new therapeutic strategies, both in regenerative and developmental medicine and in cancer therapy.

The authors thank the Zebrafish International Resource Center for the atp6v1e1hi577aTg (AB) zebrafish line, and Fish Facility technicians Maysa Franco and Liliana Carvalho for support with animal care. We also thank Alan Shipley (Applicable Electronics LLC, USA) for SIET technical support and Moisés Mallo for reagents and scientific discussion.

JJM was supported by a fellowship (SFRH/BD/45131/2008) from Fundação para a Ciência e Tecnologia (FCT), Portugal. ACC was supported by a postdocoral fellowship (SFRH/BPD/29957/2006) from Fundação para a Ciência e Tecnologia (FCT), Portugal. This work was supported by the following grants: european FP6-Cells into Organs, FCT (Grant PTDC/BIA-BCM/100867/2008 and POCTI-ISFL-4-664) and Ministerio de Ciencia e Innovación, Spain (Subprograma Ramon y Cajal, reference RYC-2008-02753 and grant BFU2009-12279). JRL is supported by the Subprogram Ramón y Cajal from Ministerio de Ciencia e Innovación, Spain (RYC-2008-02753). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.