Zebrafish (Danio rerio) were reared and kept in our facilities according to standard procedures (20). The fish lines used in this work were the csf1ra^j4blue (panther) mutant (21); Tg(mpeg1:Dendra2) (22) and TgBAC(mpx:GFP)ⁱ¹¹⁴ (23). Homozygous csf1ra^j4blue adults are viable and were identified and sorted from their phenotypically wild type siblings by their pigmentation, because of the lack of xantophores (21), and they were incrossed to obtain homozygous mutant larvae. All procedures complied with the “Guidelines for the Use of Fishes in Research Use” of the American Fisheries Society (Guidelines for the use of fishes in research. American Fisheries Society, Bethesda, Maryland. www.fisheries.org) and were approved by the Animal Ethics Committee of the University of Chile (document approval: 2015-04-20).

At 3 days post fertilization or dpf (72 h post fertilization or hpf), transgenic Tg(mpeg1:Dendra2) larvae were mounted in 0.8% low melting point agarose solution with 0.01% MS-222 (Sigma, St. Louis, MO, USA). Time lapse imaging of a portion of the tail, considered as the section of the larvae posterior to the anus, were performed using an Olympus IX81 epifluorescence microscope (Olympus, Tokyo, Japan) with a 10x zoom, every 2 min for a total of 3 h. The average speed for the imaged macrophages was calculated using the Manual Tracking plugin in the ImageJ software (NIH, Bethesda, ML, USA).

For tail fin amputation, 3 dpf larvae were anesthetized with MS-222 and amputated with a sterile scalpel. The transection was performed by using the posterior section of the ventral pigmentation gap in the tail fin as a reference, and immediately after amputation larvae were rinsed and incubated in E3 medium at 28°C. Images of recruited neutrophils and macrophages to the damage site (up to ~150 μm from the amputation site) were captured using an Olympus MVX10 stereomicroscope or an Olympus IX81 epifluorescence microscope, and analyzed using ImageJ software. The quantification of peripheral and CHT macrophages in non-amputated larvae was performed according to their location in the tail (Figure 1A). For the normalization of recruited macrophages, the number of recruited macrophages was divided by the total number of macrophages located in the tail, i.e., the sum of peripheral macrophages, CHT macrophages, and recruited macrophages.

At 3 dpf and before tail fin amputation, groups of macrophages in fish harboring the photoconvertible reporter Tg(mpeg1:Dendra2) localized at a distance between 400 and 600 μm from the transection line were photoconverted from green to red using a Zeiss Axiovert 200 M fluorescence microscope (Carl Zeiss, Jena, Germany) equipped with a Mercury lamp. The photoconversion was performed exposing a portion of the tail to a 385 nm laser for 40 s, using a 40x zoom. After photoconversion, individuals were amputated and immediately mounted in 0.8% low melting point agarose solution with 0.01% MS-222. Time lapse of the whole larval tail was performed in the Olympus IX81 microscope equipped with a 4x zoom, starting from 30 min after amputation and imaging every 5 min for a period of 24 h. Data analyses, including the average speed of photoconverted macrophages, the calculation of its initial distance and the time of their arrival to the damage site was performed using ImageJ.

For partial macrophage depletion, a dilution of clodronate liposomes (henceforth referred to as lipo-clodronate) (24, 25) was injected in the circulation valley of 54 hpf zebrafish larvae, thus allowing the spread of lipo-clodronate in the whole larva through the bloodstream. Control larvae were injected with a same dilution of PBS-loaded liposomes (lipo-PBS). Eighteen hours post injection (72 hpf or 3 dpf), larvae were sorted and amputated.

At 3 and 5 days post amputation or dpa (6 and 8 days post fertilization, respectively), larvae were mounted in 1% low melting point agarose, and regenerating tail fins were imaged in bright field using the Olympus MVX10 microscope, with a 5x zoom. Tail fin area was measured from the anterior section of the ventral tail fin gap to the end of the regenerating fin, as previously described (26), using ImageJ software. Tail fin areas were calculated and expressed in square millimeters (mm²).

Amputated larvae were incubated in 5 mM of 5-Bromo-2′-deoxyuridine (BrdU, EMD Chemicals, San Diego, CA, USA) diluted in E3 medium with 1% DMSO, from 6 to 24 h post amputation (hpa). After incubation, larvae were fixed in 4% paraformaldehyde (PFA) and then stored at −20°C in methanol. For BrdU detection, larvae were rinsed several times with PBS-Triton X-100 0.3% (PBS-Tx), then permeabilized with proteinase K 20 μg/mL for 30 min at room temperature, refixed with PFA 4% per 20 min at room temperature and washed with PBS-Tx. To increase permeability, larvae were incubated in cold acetone for 7 min at −20°C, then rinsed with PBS-Tx. For BrdU epitope exposure, samples were incubated in HCl 2N for 20 min at room temperature and then rinsed several times with PBS-Tx. Larvae were incubated in blocking solution (BSA 10 mg/ml, goat serum 2%, DMSO 1%, Triton X-100 0.1% in PBS) for at least 45 min and room temperature, and then incubated with 1:500 of mouse monoclonal anti-Bromodeoxyuridine Bu20a (M0744, Dako, Glostrup, Denmark) overnight at 4°C. After primary antibody incubation and washes with PBS-Tx, samples were incubated with 1:2,000 of Alexa Fluor 488 goat anti-mouse IgG (A11029, Invitrogen, Eugene, OR, USA) for 2 h at room temperature in darkness, and rinsed several times with PBS-Tx. Stained larvae were imaged in a Zeiss LSM 510 Meta confocal Microscope (Carl Zeiss), and BrdU dots were quantified using ImageJ software. As the ventral pigmentation gap area contains a group of proliferating cells, we restricted the quantification of BrdU+ dots to the region immediately adjacent to the damaged site.

For cell death quantification, TUNEL analysis was performed in 24 hpa larvae previously fixed in 4% PFA and dehydrated with methanol, using the ApopTag red in situ apoptosis detection kit (Merck Millipore, Temecula, CA, USA). Stained larvae were imaged using a confocal microscope and analyzed using ImageJ.

For ROS detection in the damage site, 6 hpa larvae were rinsed several times with Calcium-free HBSS medium (Gibco, Carlsbad, CA, USA) and then incubated in 50 μM of the ROS sensor 2′7′-dichlorohydrofluorescein diacetate (H₂DCF-DA, Sigma) for 15 min at 28°C in darkness. After incubation, larvae were rinsed in E3 medium and mounted in 0.8% low melting point agarose with MS-222. Images were acquired using the Olympus MVX10 stereomicroscope, with a 6.3x zoom. Fluorescence quantifications were obtained using ImageJ, and the relative fluorescence in the damage site was calculated as the ratio between the fluorescence in the damage site by the fluorescence in the tail (~500 μm far from the damage site) of the same individual.

After 6 and 24 hpa, panther and wild-type larvae were euthanized by MS-222 overdose, larval tails were collected, and RNA was isolated using the RNAqueous-Micro kit (Ambion, Lithuania). cDNA from isolated RNA samples were generated using ImProm-II Reverse Transcription system (Promega, Madison, WI, USA), and qPCR was performed in a Stratagene Mx3000P system using Brilliant II SYBR Green as fluorescent detector (both from Agilent Technologies, Cedar Creek, TX, USA). Primers used were: il1b Fwd 5′-TGGACTTCGCAGCACAAAATG-3′ and Rev 5′-CGTTCACTTCACGCTCTTGGATG-3′ (27); tnfa Fwd 5′-CAAGGCTGCCATCCATTTAACAGG-3′ and Rev 5′-TCAGTTCAGACGTGCAGCTGAT-3′; tgfb1a Fwd: 5′- CAACCGCTGGCTCTCATTTGA-3′ and Rev 5′- ACAGTCGCAGTATAACCTCAGCT-3′ (19); il10 Fwd 5′- ACAGTCCCTATGGATGTCACGTCA-3′ and Rev 5′- GCATTTCACCATATCCCGCTTGAG-3′; ef1a Fwd 5′-AGAAGGAAGCCGCTGAGATG-3′ and Rev 5′-TGTCCAGGGGCATCAATAAT-3′ (27). For analysis, target gene expression was calculated using the 2^−ΔCt method (28), using ef1a as a reference housekeeping gene.

Student's t-tests were performed when comparing two conditions, whereas two-way ANOVA with a Tukey post-test were used to analyze grouped data. Both statistical tests were performed using GraphPad Prism 6.0 software (GraphPad, San Diego, CA, USA).

Supplemental materials include: videos of macrophage recruitment in photoconverted reporter larvae (Supplemental Video 1); analyses of neutrophils in panther (Supplemental Figure S1) and after the injection of a 1:50 dilution of clodronate liposomes (Supplemental Figure S2); quantifications of macrophage number and tail fin regeneration following a 1:10 injection of clodronate liposomes (Supplemental Figure S3).

At 72 hpf, zebrafish macrophages can be classified according to their location as CHT-resident or peripheral tissue resident (Figure 1A). We performed time lapse imaging in the tail of 72 hpf Tg(mpeg1:Dendra2) zebrafish individuals to compare the spontaneous movement kinetics of CHT-resident and peripheral tissue-resident macrophages in the intact animal. We found that the average speed of peripheral tissue-resident macrophages is higher compared to CHT-resident ones (averages ± standard deviations (SDs) of 155.6 ± 25.3 and 114.2 ± 23.2 μm/h, respectively; Figure 1B), suggesting that both populations may be primed to react in a different fashion when an inflammatory response is triggered. Thus, we sought to analyze whether these populations are recruited in differential fashion to the site of a tissue injury (tail fin amputation). In order to track peripheral and CHT macrophages during damage response, we used the Tg(mpeg1:Dendra2) reporter and photoconverted groups of macrophages in either region before damage, allowing us to track them and to analyze their recruitment, and behavior during the inflammatory response triggered by tail fin amputation. We imaged the tail of photoconverted/damaged larvae every 5 min for 24 h, starting at 30 min post amputation (Supplemental Video 1; Figure 1C). In line with the findings under steady state conditions, we observed that the average speed of photoconverted peripheral tissue-resident macrophages is higher compared to photoconverted CHT-resident macrophages (186.8 ± 43.2 vs. 106.3 ± 14.0 μm/h; Figure 1D). Next, we analyzed the migration of peripheral and CHT macrophages to the damage site, analyzing the time of arrival at the damage site for both populations. We found that periphery-derived macrophages are recruited earlier than CHT-derived macrophages (9.7 ± 4.7 and 14.5 ± 3.6 h, respectively; p < 0.01). Moreover, when we calculated the speed of travel of macrophages (the ratio between the initial distance of the cells to the site of damage and their time of arrival) we found that peripheral tissue-resident macrophages are recruited faster compared to CHT-resident macrophages (73.5 ± 40.0 vs. 41.3 ± 17.1 μm/h, respectively; Figure 1E). These results indicate that peripheral tissue-resident macrophages and CHT-resident macrophages have different motility in homeostasis and show differences in migration behavior when an inflammatory response is triggered.

It has been previously described that missense mutations in the colony stimulating factor 1 receptor gene negatively affect the number of macrophages, most importantly, the peripheral tissue-resident population (29, 30). We used the previously described csf1ra^j4blue mutant allele (21), whose homozygotes (referred to from here on as panther) exhibit a significant reduction in the pool of tissue-resident macrophages during larval stages (16, 31). We combined the csf1ra^j4blue carriers with the macrophage reporter line Tg(mpeg1:Dendra2), to obtain panther mutants with Dendra2-labeled macrophages. In the tail of panther individuals, we observed a ~40% overall reduction of macrophages compared to wild type animals (Figures 2A,B). A more detailed analysis revealed that panther larvae had a reduction in the CHT-resident macrophage population of ~20% compared to wild type, whereas the pool of peripheral tissue-resident macrophages is reduced by ~60% in panther compared to wild type fish (Figure 2C). We did not observe differences in the number of neutrophils when we analyzed panther mutants in the TgBAC(mpx:GFP)ⁱ¹¹⁴ transgenic reporter background (Supplemental Figure S1A). These findings are in line with previous reports (16, 32) and indicate that, although the overall macrophage population is reduced, the pool of peripheral tissue-resident macrophages is more strongly affected in panther than the CHT-resident population.

We analyzed the recruitment of macrophages in panther mutants after tail fin amputation (Figure 2D) and we found a reduction in the number of macrophages recruited to the damage site from 6 to 24 hpa (Figure 2E). As the decreased recruitment of macrophages could be attributed to the overall reduction of macrophages in panther mutants, we divided the number of recruited macrophages of each larva by the total number of macrophages in the tail of the same larva, to obtain a normalized number of recruited macrophages. After this correction, we still found a reduced recruitment of macrophages in panther mutants at 6 and 12 hpa, when compared to wild type (Figure 2F). In contrast to macrophages, neutrophil recruitment to the damaged site was indistinguishable between panther and wild type individuals (Supplemental Figure S1B). These results support previous findings and further demonstrate that peripheral tissue-resident macrophages are recruited earlier compared to CHT-resident macrophages after tail fin amputation in zebrafish larvae.

As we found impaired recruitment of macrophages in panther mutants, we wanted to determine if this observation could be replicated by reducing the global macrophage pool in zebrafish larvae or if the effect was specific to panther mutants. To this end, we used different concentrations of clodronate liposomes to reduce the total macrophage number in larvae to a level comparable with those in panther mutants. This method has been used to decrease macrophage number in zebrafish larvae in a local or a systemic fashion (11, 18, 32, 33). We injected lipo-clodronate at 54 hpf in the circulation valley, allowing its distribution throughout the whole larva. At 72 hpf, 18 h after injection of the liposomes, we quantified the number of macrophages in the tail of lipo-clodronate larvae and lipo-PBS controls. We observed that a global reduction of ~40% of macrophages in the tail is achieved when injecting ~5 nl of 1:50 dilution of lipo-clodronate compared to an equal dilution of lipo-PBS (Figures 3A,B). Importantly, when we analyzed the number of CHT-resident and peripheral tissue-resident macrophages, we found a similar reduction in both populations of ~40% (Figure 3C), indicating that the administration of 1:50 lipo-clodronate in circulation affects all macrophages, irrespective of their location in the body. We confirmed that lipo-clodronate injection did not affect neutrophil numbers in zebrafish larvae (Supplemental Figure S2A), as previously reported (18). Next, we analyzed macrophage recruitment after tail fin amputation in 1:50 lipo-clodronate injected larvae at 72 hpf (18 hpi) and we observed a decrease in the number of macrophages recruited at the injured site at all-time points from 6 to 48 hpa (Figures 3D,E). However, these differences went not significant when we normalized all numbers by the number of total macrophages in the tail at their respective timepoints (Figure 3F). As in panther larvae, we did not detect differences in neutrophil recruitment after Lipo-clodronate treatment (Supplemental Figure S2B). Thus, our results indicate that the kinetics of macrophage recruitment is not affected when both peripheral and CHT-derived macrophages are depleted homogeneously. In contrast, macrophage recruitment to a wound is impaired by specific depletion of peripheral macrophages, revealing a specific role for these cells in the inflammatory response.

Previous studies have highlighted the importance of early recruited macrophages during zebrafish tail fin regeneration in both adult and larval stages (10, 11). We sought to determine if the impaired macrophage recruitment observed after tail fin amputation in panther mutants affects regeneration. We evaluated regeneration after amputation by measuring the tail fin area of regenerating fins in wild type and mutant 3 and 5 dpa fish. We found significantly impaired regeneration in panther mutants at both timepoints, when compared to wild type fish (Figures 4A,B). Again, as the number of macrophages is differentially affected in panther individuals depending on their place of residence in homeostasis (CHT vs. periphery), we wanted to determine if the regeneration phenotype in panther larvae is a consequence of the specific reduction in the peripheral tissue-resident macrophages or if it is due to the overall reduction of macrophages. We thus carried out tail fin amputation and regeneration analysis in 1:50 lipo-clodronate injected embryos and compared them to Lipo-PBS controls. As was the case for macrophage recruitment, we found no differences in regeneration efficiency between the lipo-clodronate and lipo-PBS conditions at 3 and 5 dpa (Figures 4C,D). Since panther mutants exhibit a strong reduction in the pool of peripheral macrophages (~60%; Figure 2C), we attempted to replicate this degree of reduction by injecting embryos with an increased concentration of lipo-clodronate. This effect was achieved by using a 1:10 dilution of lipo-clodronate; in injected fish, we observed a reduction in the peripheral macrophage pool similar to that of panther mutants (Supplemental Figures S3A–C). Under these conditions, we now observed impaired tail fin regeneration at 3 dpa in 1:10 lipo-clodronate injected fish compared to controls (Supplemental Figures S3D,E). Altogether, our results show that specifically decreasing the peripheral macrophage pool impairs tail fin regeneration after amputation in zebrafish larvae.

To further characterize the impaired tail fin regeneration observed in panther mutants, we analyzed both cell death and cell proliferation after injury in these animals compared to wild type fish. We performed TUNEL staining at 24 hpa and we found a significant increase in the number of dead cells in panther mutants (Figures 5A,B). Furthermore, cell proliferation, measured by BrdU incorporation in the tail from 6 to 24 hpa, was reduced in panther mutants (Figures 5C,D). Thus, the impaired regeneration displayed by panther mutants can be explained by elevated cell death events and loss of proliferative ability.

Next, we sought to understand how recruited peripheral tissue-resident macrophages contribute to tissue regeneration. One of the initial responses after tissue damage is the production of ROS, which promote leukocyte recruitment (34), and potentiate the inflammatory response (35). To analyze ROS accumulation in the damage site, we used the sensor H₂DCF-DA, which is catabolized to the fluorescent compound 2′7-dichlorofluorescein (DCF) through intracellular esterase activity and oxidation (36). We found that DCF accumulation in the damage site is higher in panther compared to wild type larvae at 6 hpa (Figures 6A,B), indicating that ROS production after damage is increased in panther animals.

New evidence indicates that pro-inflammatory cytokines play a major role during zebrafish tissue regeneration. Hasegawa et al. demonstrated that recruited macrophages downregulate il1b expression in the tissue during inflammation after amputation of the tail fin fold, thus controlling inflammation, decreasing cell death and promoting its regeneration (37). In a similar experimental setting, Nguyen-Chi et al. observed that the Tnfa/Tnfr1 signaling promotes macrophage recruitment and stromal proliferation, and that this factor is required for the regeneration of the tail fin fold (11). We analyzed the expression of both markers, il1b and tnfa mRNA, in the tail region using quantitative RT-PCR in panther and wild type animals. We found that il1b mRNA levels, but not tnfa mRNA, are increased in panther mutants at 6 hpa (Figure 6C). However, this difference in il1b expression became indistinguishable between panther and wild type at 24 hpa (Figure 6C). In mammals, macrophages suppress pro-inflammatory signaling through IL-10 and TGFβ (8, 9); hence, we sought to analyze the expression of these anti-inflammatory cytokines in our experimental setting. We did not find differences in il10 expression after damage, but we observed an impaired tgfb1a expression in panther larvae at 24 hpa, when compared to wild type (Figure 6C). Our results show that panther mutants, which display decreased numbers of peripheral macrophages, exhibit an increased pro-inflammatory environment, and an impaired anti-inflammatory response after tail fin amputation, which correlates with a reduced proliferation and an increased cell death in the injured site (Figure 5), as well as impaired tail fin regeneration (Figures 4A,B). Therefore, we propose that peripheral macrophages contribute to tail fin regeneration through the negative regulation of the inflammatory environment in the damage site.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.