Control of Directed Cell Migration after Tubular Cell Injury by Nucleotide Signaling

Abstract

Acute kidney injury (AKI) is a common complication of severe human diseases, resulting in increased morbidity and mortality as well as unfavorable long-term outcomes. Although the mammalian kidney is endowed with an amazing capacity to recover from AKI, little progress has been made in recent decades to facilitate recovery from AKI. To elucidate the early repair mechanisms after AKI, we employed the zebrafish pronephros injury model. Since damaged cells release large amounts of ATP and ATP-degradation products to signal apoptosis or necrosis to neighboring cells, we examined how depletion of purinergic and adenosine receptors impacts the directed cell migration that ensues immediately after a laser-induced tubular injury. We found that depletion of the zebrafish adenosine receptors adora1a, adora1b, adora2aa, and adora2ab significantly affected the repair process. Similar results were obtained after depletion of the purinergic p2ry2 receptor, which is highly expressed during zebrafish pronephros development. Released ATP is finally metabolized to inosine by adenosine deaminase. Depletion of zebrafish adenosine deaminases ada and ada2b interfered with the repair process; furthermore, combinations of ada and ada2b, or ada2a and ada2b displayed synergistic effects at low concentrations, supporting the involvement of inosine signaling in the repair process after a tubular injury. Our findings suggest that nucleotide-dependent signaling controls immediate migratory responses after tubular injury.

1. Introduction

Acute kidney injury (AKI) is a common complication of severe human disease, associated with detrimental short- and long-term consequences [1]. Despite improved supportive care, the mortality of AKI has remained largely unchanged in recent decades. The mammalian kidney encompasses an extensive capacity to recover from AKI, but attempts to facilitate the repair processes have been generally unsuccessful. Mouse models of AKI suggest that surviving renal cells, undergoing de-differentiation followed by proliferation, are mainly responsible for replacing damaged cells. Labeling studies suggest that the repair process is mainly accomplished by resident cells [2,3,4,5]. However, how surviving renal cells sense surrounding tissue damage is largely unknown since it is difficult to assess immediate repair responses in vivo due to technical limitations. In contrast, the embryonal kidney of developing zebrafish, which are transparent during the first days post fertilization, can be monitored by high-resolution time-lapse imaging in combination with fluorescent microscopy to identify early repair programs ensuing after AKI. In two-day-old zebrafish embryos, a laser-induced injury is repaired by a directed migratory program that instructs tubular epithelial cells on the anterior side of the injury to reverse their migratory pattern until a connection with cells on the posterior side of the injury is made and tubular patency has been re-established [6,7].

Release of ATP and ATP metabolites represents a key signal emitted by damaged cells [8]. ATP can trigger chemotaxis through activation of purinergic receptors [9]; however, ATP also serves as a source for adenosine production through consecutive phosphohydrolysis by the ecto-nucleotidases CD39 and CD73 [10].

The purinergic receptors consist of two subfamilies: purinergic P1 and P2 receptors. The purinergic P1 receptors encompass the four G protein-coupled adenosine receptors A1, A2A, A2B and A3. While A1 and A3 couple to heteromeric proteins of the Gα_i/o family and inhibit cAMP, A2A and A2B stimulate cAMP levels through Gα_s [11]. The low-affinity adenosine receptor A2B appears to be only activated in stressful conditions, when adenosine levels increase to micromolar concentrations [12]. While A1 improves renal function after ischemia–reperfusion (IR) injury, A2A receptor activation protects organs after IR injury by inhibiting inflammatory responses [11]. In contrast to the other three adenosine receptors, stimulation of A3 appears to worsen IR injury [10]. The purinergic P2 receptors encompass the ligand-gated ionotropic cation channels P2X and the G protein-coupled metabotropic P2Y family, comprising eight receptors. P2Y receptors are expressed in all segments of the nephron [13]. The P2Y receptors can be divided in two subgroups based on their coupling to Gα_q/11 (P2Y_1,2,4,6,11), activating the PLC β/IP₃ pathway, or to Gα_i/o (P2Y_12,13,14), inhibiting adenylyl cyclase to lower cAMP [14].

Due to degradation as well as rapid uptake by equilibrative nucleoside transporters, extracellular adenosine has a short half-life. Human tissues contain two ADA isoenzymes, ADA1 and ADA2, that convert adenosine to inosine. ADA is present in virtually all tissues; and in addition to its activity in the cytoplasm, it can associate with CD26 or adenosine receptors to function as an ecto-enzyme [15,16]. Adenosine deaminase (ADA) deficiency causes severe combined immunodeficiency (SCID), characterized by life-threatening infections from bacteria, viruses, and fungi [17].

We hypothesized that damaged tubular epithelial cells release ATP and ATP metabolites to signal damage to neighboring cells. We tested this hypothesis by systematically eliminating components of the ATP/adenosine signaling cascade by gene knockdown. We intentionally used translation (TBM)- and splice-blocking (SBM) morpholino oligonucleotides (MO) to screen putative pathway involved in damage recognition and control of directional cell migration for two reasons: first, each potential pathway is encoded by multiple family members, requiring several simultaneous gene deletions to determine the contribution of a single pathway; second, we intended to circumvent the capacity of zebrafish to compensate germline gene mutations [18,19]. Our approach revealed that overlapping components of this signaling cascade are involved in the migratory response triggered by a laser-induced tubular damage. Furthermore, we found that manipulation of purinergic signaling can potentially enhance migration-based pronephros repair.

2. Results

2.1. Adenosine Receptors Support the Repair of Laser-Induced Zebrafish Pronephros Injuries

Speculating that ATP released from injured tubular cells is rapidly metabolized to adenosine by the ecto-nucleotidases CD39 and CD73, we analyzed the involvement of the zebrafish adenosine receptors in the repair process after a laser-induced tubular injury. We found increased adora1a and adora1b expression in pronephros cells adjacent to the injury suggesting a rapid upregulation of these receptors in response to injury (Figure 1A). Splice- and translation-blocking MOs to deplete these two receptors significantly delayed the repair process in 2-day-old zebrafish embryos, suggesting that both adenosine receptors are required for a normal repair response (Figure 1B). We next examined the contribution of zebrafish adora2aa and adora2ab to the repair process after laser-induced pronephros injury. Both adenosine receptors were upregulated after injury (Figure 1C). Depletion of either adora2aa or adora2ab by SBMs delayed the repair; the differences were statistically significant for adora2aa, but not for adora2ab (Figure 1D); similar non-significant results were obtained for adora2b (Figure S1A). While both A2A and A2B receptors stimulate adenylyl cyclase (AC) to generate the second messenger cAMP, adenosine A1 receptors couple to G_i/o and inhibit AC (Figure S1B). Despite these opposing effects on AC activity, both receptor types are required for the migration-based repair process.

2.2. Adenosine Degradation Represents an Essential Component of the Repair Process

Repair of a laser-induced pronephros wound critically depends on overruling the posterior-to-anterior collective cell migration that is characteristic for the developing pronephros of zebrafish embryos [20]. While cells on the posterior side of the gap increase their track speed, cells on the anterior side of the pronephros injury reverse the migratory direction, and continue to migrate in an anterior-to-posterior direction until the gap is closed and the patency of the pronephros is re-established [6,7]. Extracellular adenosine, generated from released ATP, is degraded to inosine by adenosine deaminase (ADA), affecting local adenosine concentrations [16]. Since ADA might support the formation of adenosine gradients leading to differential activation of adenosine receptors, we examined the role of ADA isoenzymes during the zebrafish pronephros repair process. Increased expression after laser-mediated injury was observed for all four zebrafish isoenzymes, ada, ada2a, ada2b, and adal (Figure 2). While depletion of zebrafish ada (SBM, 0.3 mM, and TBM 0.1 mM) and ada2b (TBM, 0.3 mM) significantly delayed the repair process (Figure 2A,C and Figure S2), depletion of zebrafish ada2a (TBM, 0.4 mM) and adal (TBM, 0.1 mM) had only a marginal effect (Figure 2B,D). Since all four isoenzymes were expressed after injury, it is likely that the depletion of one isoenzyme can be compensated by the upregulation of another family member. Combining ada and ada2b MOs at concentrations (0.1 mM) that did not affect the repair process in comparison to control embryos (ctrl MO, 0.2 mM), resulted in a significant repair delay (Figure 3A). Similar results were obtained for combining ada2a (0.15 mM) and ada2b MO (0.1 mM) (Figure 3B). While ada2a alone at MO concentrations of 0.4 mM had no effect, it significantly delayed the repair process in combination with low concentrations of ada2b. These results suggest that the zebrafish adenosine deaminases exert overlapping functions after pronephros injury. Furthermore, control of local adenosine concentrations by ada family members appears to support the repair process. To assess whether disruption of adenosine degradation affects the migratory response, we analyzed track speed and cell displacement of pronephros cells involved in the repair process by high-resolution video microscopy. In control embryos (ctrl MO, 0.3 mM), the laser-induced gap was almost completely repaired within 4 h, while the gap persisted in ada2b-exposed embryos (ada2b, TBM 0.3 mM) (Figure S3A). Analyses of track speed and cell displacement revealed a significant reduction in directed cell migration both in the proximal as well as in the distal segments of the pronephros (Figure S3B). Thus, adenosine signaling through adenosine receptors and degradation of adenosine by adenosine deaminases appear to be essential components of the migratory repair response in zebrafish embryos.

2.3. The ATP-Sensing Purinergic P2RY2 Receptor Is Required for a Normal Migratory Response after Pronephros Injury

Single-cell RNA sequencing revealed expression of zebrafish p2ry2 along the entire pronephros [21], while members of the p2x and other members of the p2y gene family were expressed at lower levels (Figure 4A). Three different genes encode for zebrafish p2ry2 (p2ry2.1, p2ry2.2, and p2ry2.3). All three p2ry2 genes were upregulated in response to injury (Figure 4B–D). The combined knockdown, using three different TBMs, resulted in a significant repair delay (Figure 4E). Similar, albeit statistically not significant results were obtained with CRISPR/Cas9-mediated gene targeting, using guide RNAs against all three p2ry2 genes (Figure 4F and Figure S4). Time-lapse imaging revealed that a significantly reduced track speed of the neighboring tubular epithelial cells involved in the repair process was the likely cause for the defective repair process (Figure 4G,H). Notably, renal cells of embryos micro-injected with the control MO (ctrl MO) reassumed cellular contacts within 6 h after laser ablation, while the laser-induced gap between the anterior and posterior end of the injured pronephros remained clearly detectable (Figure S5, Movies S1 and S2).

2.4. The Adenosine Pathway Is Required for the Pronephros Repair Process

To highlight the importance of the adenosine pathway for the pronephros repair process, we targeted the entire pathway by combining low MO concentrations of ada2b SBM (0.1 mM), adora1b TBM (0.05 mM), p2ry2.1 (0.05 mM), p2ry2.2 (0.05 mM) and p2ry2.3 (0.05 mM). The combined knockdown of components of the entire pathway had a synergistic effect and strongly suppressed the pronephros repair process (Figure 5A). Thus, the activation of the adenosine pathway is required for the normal migratory response after pronephros injury.

2.5. Activation of Purinergic Signaling Promotes Pronephros Repair

Diquafosol is a P2RY2 receptor agonist that has been approved for treating dry eyes disease in several countries [22]. Since P2RY2 receptor signaling appears to ameliorate ischemia–reperfusion injuries [23,24], we tested whether Diquafosol can accelerate the repair process. Diquafosol tetrasodium, dissolved in water, was non-toxic (Figure 5B). To test the effect of Diquafosol on the repair process, it was added at 0.5 and 1.0 mg/mL (1.14 mM) three hours before the injury, and continued until the end of the observation period. Diquafosol at 0.5 mg/mL only slightly facilitated the repair process, while Diquafosol at 1.0 mg/mL exerted a significant effect (Figure 5C, Movies S3 and S4). Prior depletion of the p2ry2 (TBM, 0.6 mM) prevented the effect of Diquafosol (Figure 5D). Thus, manipulation of the P2ry2 signaling pathway can be utilized to accelerate pronephros repair processes.

3. Discussion

Acute kidney injury (AKI), often caused by a combination of ischemia and toxic injuries, remains an enormous medical challenge and socio-economic burden [25,26,27,28,29]. Despite progress in understanding the underlying pathophysiology, AKI continues to cause high morbidity and mortality, accounting for about 1.7 million deaths per year [30]. Although the kidney possesses an amazing capacity to recover from severe AKI, repair is often incomplete, resulting in detrimental long-term complications [1]. To improve immediate and long-term prospects, tremendous efforts have been made to characterize the mechanisms that control renal regeneration. Aiming to identify cells responsible for repairing the damaged kidney, cell labeling and genetic cell fate-tracing experiments have revealed that resident renal epithelial cells re-enter the cell cycle to replace damaged cells, involving both canonical mitosis and endoreplication [2,3,31,32,33]. Two-photon in vivo microscopy has provided additional insight into the inflammatory response ensuing after AKI [34]. However, it has not yet been possible to track individual tubular cells after kidney injury to characterize the immediate adaptive responses after cell damage and tissue necrosis.

In the embryonal zebrafish kidney, injuries are repaired by directed cell migration. While cells on the posterior side of the injury increase the speed of the ongoing posterior-to-anterior collective cell migration, cells on the anterior side reverse their direction, and only resume the posterior-to-anterior collective cell migration after the injury-induced gap has been bridged and the patency of the tubular lumen has been re-established. However, it is currently unknown how tubular epithelial cells sense damage and how the migratory response is coordinated. Single-cell RNA sequencing and microarray analysis revealed upregulation of P1 and P2 purinergic family members in the pronephros of two-day-old zebrafish embryos [7,21], suggesting that ATP and ATP metabolites are involved in the signaling events in response to injury.

We found that depletion of the high-affinity adenosine receptors adora1a, adora1b, and adora2aa significantly delayed the repair process. A1 and A2A/2B signal through different G proteins with opposing effects on adenylyl cyclase and cAMP production in mammalian cells; however, heteromeric interaction between adenosine receptors has been demonstrated at the structural level, resulting in important control of neurotransmitter release [35], suggesting that “biased agonism” occurs in zebrafish embryonal kidneys in response to injury, a concept to explain complex G protein-coupled receptor signaling [36]. Depletion of low-affinity zebrafish adora2b did not affect the repair process, suggesting that the adenosine signaling events likely occur within a nanomolar concentrations.

Since adenosine is rapidly removed by equilibrative nucleoside transporters or degraded to inosine by extracellular adenosine deaminases (ADA), we next depleted ADA family members. Depletion of zebrafish ada and ada2b impaired the repair after injury suggesting that a precise control of adenosine levels or formation of adenosine gradients is mandatory for a normal repair response. Since the combination of low concentrations of ada/ada2b and ada2a/ada2b MOs augmented the repair defect, adenosine deaminase family members seem to act in partially complementary and/or redundant pathways.

ATP can directly bind to P2YR2 family members. Our single-cell RNA sequencing results suggested that zebrafish p2ry2 is expressed in the pronephros. In situ hybridization after injury revealed an upregulation of all three p2ry2 variants (p2ry2.1, p2ry2.2 and p2ry2.3). The combined knockdown by either MOs or CRISPR reduced the tracking speeds of pronephros cells involved in the repair response, resulting in repair delays, while stimulating P2ry2 resulted in accelerated repair. The ligand-gated ionotropic cation channel P2X4 and P2X7 appear to exacerbate ischemic AKI by triggering pro-inflammatory cytokine production and NLRP3 inflammasome activation [37,38]. Similarly, G protein-coupled metabotropic P2Y14, expressed in intercalated cells of the nephron, generate chemoattractant cytokines in response to uridine glucose, resulting in renal inflammation [39]. In contrast, P2RY2 receptor ameliorates renal fibrosis after subtotal nephrectomy [40], alleviates cerebral ischemia–reperfusion injury [24], and confers the beneficial effects of uridine-5′-triphosphate after myocardial infarction [41]. Our results support a potentially unique involvement of P2RY2 in recovery from ischemia.

4. Materials and Methods

4.1. Zebrafish Lines Maintenance and Treatment

All animal work has been conducted according to the relevant national guidelines (Regierungspräsidium, Freiburg, Germany). Zebrafish lines were maintained as previously described [42]. All studies were performed in the Tg(wt1b:GFP); Tg(cdh17:GFP) transgenic line [7]. The p2ry2 agonist Diquafosol (MedChemExpress, Monmouth Junction, USA) was diluted in Danieau’s buffer to the described final concentrations. Zebrafish larvae were added to the Diquafosol solution 3 h before laser ablation and were kept in Diquafosol for 24 h after ablation.

4.2. Data Analysis and Visualization

The RNAseq data are publicly available and the analysis has been previously described [21]. The heat map was generated in Excel. The repair timecourse graphs and the statistical analysis were performed in GraphPad Prism. The utilized statistical tests are mentioned with the respective experiment.

4.3. Laser Ablation, Image Acquisition and Migration Quantification

Laser-induced ablations were performed as previously described [7]. Briefly, 2-day-old zebrafish larvae were embedded in 1% low melting agarose in glass-bottom dishes. Cell ablations were performed with a 2-photon laser (Chameleon) attached to an LSM 880 Observer confocal microscope (Carl Zeiss, Jena, Germany). In total, 80 µm of the pronephros was ablated. For WISH, larvae were fixed in methanol 2 h post ablation. Confocal images were recorded with a C-Apochromat 40×/1.2 objective (Carl Zeiss, Jena, Germany). Time-lapse video microscopy was carried out at the LSM 880 microscope. Z-stacks of the injury site were recorded every 10 min. The 3D reconstruction, track speed and cell displacement were calculated in Imaris (Bitplane, Zürich, Switzerland). The tubular repair was monitored with a Leica MZ16 epifluorescent stereo microscope (Leica, Solms, Germany).

4.4. Whole-Mount In Situ Hybridization (WISH)

Whole-mount in situ hybridization (WISH) was performed as previously described [21]. RNA probes against p2ry2.1, p2ry2.2, p2ry2.3, ada, ada2a, ada2b, adal, adora1a, adora1b, adora2b, adora2aa and adora2ab were generated from a cDNA library from 1–2-day-old embryos. For that, gene-specific sequences were amplified by PCR and cloned into a pCRII-Topo vector (Invitrogen, Carlsbad, USA). DIG-labelled anti-sense RNAs were transcribed from linearized vectors using T3 or T7 RNA polymerases (Roche, Mannheim, Germany).

4.5. CRISPR/Cas9 Gene Targeting

Small guiding RNAs targeting p2ry2.1, p2ry2.2 and p2ry2.3 were designed using the online tool ChopChop (http://chopchop.cbu.uib.no, accessed on 21 January 2021). CRISPR/Cas9 gene targeting was carried out as previously described [7]. Briefly, eight sgRNAs targeting p2ry2.1, p2ry2.2 and p2ry2.3 were injected simultaneously in 1-cell-stage zebrafish embryos together with TrueCut Cas9 protein (Thermo Fisher Scientific, Vilnius, Lithuania). The efficiency of each sgRNA was evaluated by PCR followed by Sanger sequencing (not shown). Injected embryos were kept at 28 °C in Danieau’s solution and subjected to laser-induced injury at 48 hpf. The following sgRNAs were used in this study:

p2ry2.1_prm_gRNA1 5′-AGAGAGACTCGAATAACATG AGG-3′;

p2ry2.1_prm_gRNA2 5′-GTGTAGGTATGCAAACCGAG AGG-3′;

p2ry2.1_exon2_gRNA1 5′-ACGAGATGAAACGAGCACGA CGG-3′;

p2ry2.1_exon2_gRNA3 5′-CAGGGGTCGAGTGATCTTAT AGG-3′;

p2ry2.2_exon_gRNA1 5′-GAATCGGGGCCTGTAAGATG AGG-3′;

p2ry2.2_exon_gRNA3 5′-GTTGATCCTCCGCCAACTCG AGG-3′;

p2ry2.3_exon_gRNA1 5′-GAAACTAACCAGAGGTCGTG TGG-3′;

p2ry2.3_exon_gRNA2 5′-TTGCTTGATCCAGGTAGCGG AGG-3′;

4.6. Morpholino Oligonucleotides (MOs) Gene Targeting

MOs were designed and obtained from Gene Tools, Philomath, USA. The following MOs were used in this study:

p2ry2.1-TBM 5′- TCGTGATCCAGATATAGATACTTTC-3′ (this study);

p2ry2.2-TBM 5′- GTTGTTAAATGCTGCCATCCTGATG-3′ (this study);

p2ry2.3-TBM 5′-TCATTCATTTTCCTTCACTTAGTCT-3′ (this study);

ada-TBM 5′-TCCATTCATTTCAGCCATTGTGTTG-3′ (this study);

ada-SBM 5′-AACAACAGGACACCACTAACCTTAG-3′ (this study);

ada2a-TBM 5′-TGCATGTCTGTAAGGTAATTCAACC-3′ (this study);

ada2b-TBM 5′-GCTTATGCTACTCATTGCTCCCAGC-3′ [43];

adal-TBM 5′-AAAGAGATCCGCTTCGGTGTCCATC-3′ [44];

adora1a-SBM 5′-ATTAAAATCTTATTACCTCATTGGT-3′ (this study);

adora1b-TBM 5′-GAGAGATCCTCGGGCATTCTTGCAC-3′ [45];

adora2b-TBM 5′-CAATGGCGATGTAGAGCGAATCCAT-3′ [46];

adora2aa-SBM 5′-AGAAACACCCTTCACTCACCTAAGC-3′ [47];

adora2ab-TBM 5′-GTGCTATCAACCAGTGTGAAAGGAT-3′ [48];

p53-MO 5′-GCGCCATTGCTTTGCAAGAATTG-3 [49];

Standard control (ctrl) MO 5′-CCTCTTACCTCAGTTACAATTTATA-3’.

All MOs were co-injected with p53-MO to reduce unspecific effects [49]. A volume of 4 nl of MO diluted in 100 mM KCl, 0.1% phenol red and 10 mM HEPES (pH 7.5) was injected in zebrafish embryos at the 1-cell stage. Injected embryos were kept at 28 °C in Danieau’s solution.

5. Conclusions

To obtain insight into the mechanism(s) that orchestrate the immediate events following a pronephros injury in zebrafish embryos, we studied the effects of ATP and ATP metabolites based on their ability to overrule the posterior-to-anterior collective cell migration that characterizes embryonal pronephros development until a laser-induced tissue gap is repaired (Figure 6). Depleting purinergic receptors delayed the repair process, revealing important roles for members of the adenosine receptor family, the ATP-degradation pathway, and for the P2 purinergic receptor P2ry2. Agonistic stimulation of the P2ry2 receptor accelerated the repair process, which might be applicable to ameliorate human acute kidney injury.