Transient inflammatory response mediated by interleukin-1β is required for proper regeneration in zebrafish fin fold

1Progress in regenerative medicine depends on unraveling the mechanisms underlying tissue and organ regeneration. The zebrafish is a powerful model species for investigation of regeneration mechanisms because it exhibits high regenerative capacity. Zebrafish can regenerate complex structures such as fins, heart, brain, retina, and other tissues. The caudal fin of zebrafish, in particular, has been used as a model for analyzing epimorphic regeneration, a type of regeneration in the urodele limb and fish fin. During epimorphic regeneration, two characteristic tissues, the wound epidermis and the blastema, are induced in response to tissue amputation, and their coordinated actions regulate cell proliferation and morphogenesis and thus lead to tissue regeneration. Studies conducted using the caudal fin regeneration model have identified numerous genes and molecular signaling pathways that are critical for fin regeneration.

2In addition to the classical regeneration model developed using the adult caudal fin, a zebrafish larval fin fold model has been developed, allowing identification of additional genes and signaling pathways required for regeneration. Specifically, the fin fold regeneration model has facilitated genetic analysis of the regeneration mechanism by exploiting a number of mutant zebrafish resources, because several zebrafish lethal mutants survive beyond 7 days post fertilization (dpf)—an adequate period for assaying tissue regeneration—because of nutrient supply from the yolk.

3Previously, we reported that zebrafish mutants, cloche (clo) and tal1/scl, showed a unique regenerative defect: these mutants could not regenerate their fin fold because of apoptosis of the regenerative cells. Our analyses of the clo mutant revealed that a diffusible survival factor from myeloid-lineage cells is required to prevent apoptosis of the primed regenerative cells. That study suggested that the regenerative cells are sensitive to apoptosis during tissue regeneration, but the mechanism of this sensitization and the identity of the survival factor remain unknown.

4Myeloid cells are widely recognized to play a role in protection against pathogens and microorganisms after tissue injury. Neutrophils are first recruited to sites of inflammation, and these cells encounter pathogens and phagocytose and kill microorganisms by producing reactive oxygen species and/or antibacterial proteins. Subsequently, macrophages infiltrate the inflammation site, produce pro- or anti-inflammatory cytokines, remove tissue debris, secrete growth factors, and support tissue restoration.

5Several recent studies have used the zebrafish model and addressed the roles of myeloid cells in response to tissue injury: Li and coworkers showed that knockdown of macrophage differentiation delayed fin fold regeneration and resulted in formation of an abnormal fin fold featuring large vacuoles. Furthermore, Petrie and coworkers showed that genetic ablation of macrophages in adult zebrafish impaired fin outgrowth and often induced an abnormal fin. These studies suggest a crucial role for macrophages in tissue regeneration; however, the mechanism by which myeloid cells affect regeneration remains to be elucidated.

6Here, we show that a pro-inflammatory cytokine gene, interleukin 1 beta (il1b), is aberrantly upregulated in the clo mutant, and that the resulting excessive and prolonged inflammation causes apoptosis of the regenerative cells. Moreover, although myeloid cells are widely regarded as the main cells that secrete pro-inflammatory cytokines such as Il1b and thereby promote inflammation, we demonstrate that epidermal cells surrounding the amputated tissue are the source of Il1b in both wild-type (WT) and clo larvae. Notably, il1b expression in the regenerating epidermis is usually quenched by the action of macrophages, and a lack of this anti-inflammatory effect of the macrophages results in apoptosis. Lastly, we show that Il1b signaling plays an indispensable role in normal tissue regeneration by activating expression of regeneration-induced genes. Thus, our study highlights the function of Il1b and the inflammatory response in tissue regeneration. A proper level of il1b expression and the inflammatory response, which are tuned by macrophages, are necessary for progression of tissue regeneration.

7To understand the molecular pathway leading to apoptosis of regenerative cells in the clo mutant, we performed a transcriptome analysis of amputated larval tail tissues of WT and clo mutant zebrafish at approximately 6 hr post amputation (hpa), just before the stage at which apoptosis is first detectable in the clo mutant. Compared with WT larvae, the clo larvae showed marked upregulation of regeneration-induced genes such as junba, junbb, matrix metallopeptidase 9 (mmp9), fibronectin 1b (fn1b), and fgf20a (Figure 1A and Supplementary file 1). In addition to these genes, expression of a group of genes including il1b and tumor necrosis factor-b (tnfb), which are involved in the inflammatory response, was also upregulated in the clo mutant during fin fold regeneration. Excluding enzymes and uncharacterized proteins, il1b was the most upregulated gene. il1b expression in the clo mutant was also confirmed using reverse transcription polymerase chain reaction (RT-PCR) analysis (Figure 1B). Notably, prostaglandin2a (ptgs2a), a downstream gene in Il1b signaling, was also upregulated in the clo mutant (Figure 1A), which clearly indicated that Il1b signaling was activated in the clo mutant.

8Next, we performed in situ hybridization (ISH) analysis of il1b to reveal its spatiotemporal expression during regeneration. At 3 hpa, strong il1b expression was observed in both WT and clo mutant larvae (Figure 1C and D). In WT larvae, il1b was expressed in a group of scattered cells that appeared to be myeloid cells (Figure 1C; arrowheads) and also expressed at the distal edge of the fin fold. After 3 hpa, il1b expression was rapidly diminished in WT larvae, and most larvae displayed very weak or no il1b expression at 6 hpa. In contrast to the expression in WT, il1b expression in the clo mutant was maintained at 6 hpa and was still detectable at 12 hpa, indicating a prolonged inflammatory reaction in the clo mutant.

9Bacterial infection is recognized to trigger an inflammatory response that accompanies the induction of il1b expression. To investigate whether the same mechanism underlies the induction of il1b expression by fin fold amputation and bacterial infection, we injected zebrafish with Salmonella typhosa lipopolysaccharide (LPS), a reagent whose injection mimics bacterial infection. WT larvae injected with LPS displayed systemic induction of il1b in cells that appeared to be myeloid cells (Figure 1E). By contrast, the clo mutant larvae injected with LPS showed few il1b-positive cells, which indicated that tissue injury and bacterial infection induce il1b expression through distinct mechanisms.

10The absence of myeloid cells in the clo mutant suggests that the il1b is expressed in non-myeloid cells. Accordingly, examination of tissue cryosections of the clo fin fold after il1b ISH analysis suggested that il1b was expressed in epidermal cells (Figure 2A).

11To further investigate the identity of the il1b-expressing cells, we generated a transgenic (Tg) line, Tg(il1b:egfp), by introducing the egfp-nitroreductase fusion gene into the il1b locus of a bacterial artificial chromosome (BAC) clone, CH211-147H23. The larvae of Tg(il1b:egfp) showed faint EGFP expression in the entire epidermis and weak EGFP expression in the pectoral fins and at the end of the notochord (Figure 2—figure supplement 1A; arrowheads). After fin fold amputation, egfp expression was induced in a similar pattern to endogenous il1b expression (Figure 2B; Figure 2—figure supplement 1B), although EGFP fluorescence was detected at later stages of regeneration than il1b mRNA, probably as a result of the relatively longer half-life of the protein EGFP. LPS injection also induced EGFP fluorescence in myeloid cells (Figure 2—figure supplement 1C), which indicated that the il1b:egfp transgene recapitulated endogenous il1b expression in myeloid cells and also in the injured fin fold.

12Next, using the Tg(il1b:egfp) reporter line, we confirmed that the EGFP-positive cells in the transgenic line mostly colocalized with the epithelial marker E-cadherin (Figure 2C), in both WT and the clo mutant; this finding indicated that il1b expression was principally induced in epithelial cells in response to tissue amputation. In addition to detecting il1b expression in the injured tissue, we detected il1b expression in WT larvae in migrating cells that are likely to be the myeloid cells (Video 1).

13Excessive inflammation has been suggested to directly or indirectly induce apoptosis. Therefore, because elevated expression of the pro-inflammatory cytokine Il1b leads to chronic inflammation, we suspected that excessive il1b expression in the clo mutant could be a cause of the apoptosis of regenerative cells.

14First, we used the Tg(il1b:egfp) reporter line to examine the relationship between il1b-expressing cells and apoptotic cells. Although many of the il1b-expressing cells did not overlap with the apoptotic cells that were mainly distributed in the mesenchyme, the cells lay in close apposition (Figure 3A).

15Next, to demonstrate the role of Il1b in the apoptosis occurring in the clo mutant, we knocked down il1b expression using an antisense morpholino oligonucleotide (MO). The il1b MO1 targeted to the splice sites effectively inhibited il1b mRNA splicing (Figure 3B). Notably, il1b knockdown in the clo mutant resulted in a reduction in apoptosis of regenerative cells, whereas in the clo mutant injected with the standard control MO (std MO), no reduction in apoptosis was detected (Figure 3C and D). A similar anti-apoptosis effect of il1b knockdown was also obtained using another MO against il1b (Figure 3—figure supplement 1).

16To examine whether excessive inflammation is a cause of the apoptosis in the clo mutant, we used dexamethasone (Dex), a synthetic glucocorticoid that has been reported to suppress inflammation and il1b expression. Our results showed that Dex treatment abolished il1b expression in the clo mutant starting from an early stage of regeneration (Figure 3E), which indicated that the treatment effectively suppressed inflammation and il1b expression. In the Dex-treated clo larvae, the regenerative cells were rescued from apoptosis, much as in the il1b-knockdown larvae (Figure 3F and G). Collectively, these data support the notion that excessive Il1b signaling and inflammation in response to tissue injury represent a cause of apoptosis in the clo mutant.

17We have previously shown that the presence of myeloid cells is essential for survival of regenerative cells. Myeloid cells such as neutrophils and macrophages have been shown to transiently migrate to the injured fin fold. Actually, we observed that the myeloid cell accumulation in the injured fin fold reached a maximum between 3 and 6 hpa (Figure 4A and B). The timing coincides with the stage at which the initial il1b expression immediately after wounding was downregulated (Figure 1C and D). This suggests that myeloid cell migration could play a role in attenuating il1b expression, although we cannot conclude that myeloid cell migration is necessary for il1b downregulation.

18Myeloid cells comprise several different cell types, including neutrophils and macrophages, but the cell type required for the survival of regenerative cells remains to be determined. Here, we used csf3r, irf8, and spi1b (pu.1) MOs, which inhibited the differentiation of neutrophils, macrophages, and both types of cells, respectively, and examined whether the depletion of neutrophils or macrophages induces apoptosis of regenerative cells as in the clo mutant. The efficacy of csf3r knockdown was confirmed by Sudan Black staining of neutrophils and ISH analysis using the mpx probe. In both the csf3r morphants and the spi1b morphants, fewer neutrophils were detected than in the control morphants (Figure 4—figure supplement 1A,B,D). In the irf8 morphants, the expression of apoeb, a marker for microglia—the resident macrophages in the central nervous system—was drastically decreased (Figure 4—figure supplement 1C). The decrease of macrophages was also evident in the wounded fin fold as revealed by lcp1 ISH analysis (Figure 4—figure supplement 1E). None of these morphants showed any morphological abnormalities. These data indicated that the irf8 MO effectively inhibited macrophage differentiation.

19To examine apoptosis, we performed TUNEL staining in the aforementioned morphants after fin fold amputation, and the results showed that irf8 morphants exhibited aberrant apoptosis of the regenerative cells, as did the spi1b morphants (Figure 4C). Importantly, similar numbers of TUNEL-positive cells were observed in the irf8 and the spi1b morphants (Figure 4D). In sharp contrast to the irf8 morphants, the csf3r morphants showed no increase in the apoptosis of regenerative cells. These results suggest that macrophages are responsible for supporting the survival of regenerative cells.

20We next investigated the mechanism by which macrophages support the survival of regenerative cells. First, we examined il1b expression at 6 hpa in the spi1b, csf3r, and irf8 morphants. Whereas normal il1b expression was detected in the csf3r morphants, as in WT larvae, il1b expression remained elevated in the spi1b and irf8 morphants, as in the clo mutant (Figure 5A and B). This result suggests that macrophages attenuate il1b expression and thereby prevent apoptosis of regenerative cells in WT larvae. Furthermore, the apoptosis induced in the spi1b or irf8 morphants was rescued following il1b knockdown or Dex treatment (Figure 5C–E), which suggests that elevated il1b expression in the spi1b and irf8 morphants is responsible for the observed apoptosis.

21To further demonstrate that excessive il1b expression induces apoptosis, we generated and used the construct pTol2(hsp70l:mCherry-T2a-il1b), which we hereafter refer to as pTol2(hsp70l:il1b). The precise cleavage site of the zebrafish Il1b pro-peptide has not yet been identified, but based on comparing the zebrafish and human sequences, we estimated that the human IL-1β cleavage site (D116) corresponds to the threonine residue at the 124th amino acid (a.a.) position in zebrafish Il1b (Figure 6—figure supplement 1A). Furthermore, to obtain constitutive secretion of this Il1b protein lacking a canonical signal-peptide sequence, the Il1b sequence after a.a. 125 was fused with the human IL-1β receptor antagonist signal sequence and placed under the control of the promoter of heat shock protein 70l (Figure 6A). In the obtained Tg(hsp70l:il1b) line, mCherry fluorescence and il1b expression were observed after heat shock treatment (Figure 6B; Figure 6—figure supplement 1B), but not after stress caused by fin fold amputation (Figure 6—figure supplement 1C).

22When TUNEL staining was performed after heat shock induction at 2 days post amputation (dpa), several TUNEL-positive cells were detected in the injured fin fold (Figure 6C and D), which indicated that the primed regenerative cells were susceptible to the excess Il1b signal. Furthermore, systemic and aberrant apoptosis was also observed when il1b was overexpressed in WT embryo from 12 to 24 hr post fertilization (hpf), at which stage myeloid cells have not fully differentiated (Figure 6—figure supplement 1D). This indicated that the excessive Il1b signaling potentially induces apoptosis not only in fin fold but also in other parts of the body. In addition to inducing the aberrant apoptosis, il1b overexpression suppressed the extension of the fin fold (Figure 6E and F).

23Our data thus far suggest that excessive Il1b-mediated inflammation negatively affects the survival of regenerative cells and tissue regeneration in the clo mutant. However, inflammation has also been suggested to be essential for complete tissue regeneration. To examine the role of tissue inflammation and Il1b during normal fin fold regeneration, we used Dex and il1b MO1 and tested whether the inflammation mediated by Il1b affects normal fin fold regeneration.

24To assess the role of inflammation in tissue regeneration, we administered Dex to WT larvae and assessed fin fold regeneration. The Dex-treated larvae displayed apparent retardation of regeneration at 3 dpa (Figure 7A and B) and a statistically significant reduction in cell proliferation in the distal fin fold region (Figure 7C and D). Notably, the expression of regeneration-induced genes such as junba and fgf20a was downregulated in the Dex-treated larvae (Figure 7E). These data suggest that inflammation is required for activation of regeneration-induced gene expression which is a prerequisite for normal regeneration.

25Next, we used il1b knockdown in WT zebrafish to demonstrate the role of Il1b signaling in normal regeneration: il1b knockdown induced a similar phenotype to Dex treatment (Figure 7F–I), although the phenotype was slightly milder than that of the Dex-treated larvae, probably because of lower penetrance of the MO knockdown. Importantly, similar to the Dex-treated larvae, the il1b morphants showed attenuated expression of junba and fgf20a (Figure 7J). Collectively, these results suggest that the inflammatory reaction mediated by Il1b plays a necessary role in normal fin fold regeneration.

26Lastly, we tested whether il1b overexpression induces a marked increase in expression of the regeneration-induced genes junba, junbb, and fn1b in uninjured tissue (Figure 7K). As expected, il1b overexpression induced the ectopic expression of fn1b, junba, and junbb, both in the uncut fin fold (Figure 7L) and in several other tissues such as the pectoral fins. Importantly, the apoptosis induced by il1b overexpression was only detectable at 2 days after multiple heat shocks (Figure 6C and D), possibly because of the anti-inflammatory function of macrophages in WT. But, induction of junba, junbb, and fn1b was observed at 12 hr after two heat shocks, a stage at which the apoptosis induced by il1b overexpression has not taken place (Figure 7—figure supplement 1), indicating that expression of the regeneration-induced genes was not caused by an indirect regenerative response to cell death. These data indicate that Il1b signaling is required tostimulate expression of regeneration-induced genes and regulate the initiation of fin fold regeneration.

27It has been suggested that myeloid cells play crucial roles in the inflammatory responses of injured tissues. The pro-inflammatory cytokines (such as Il1b) produced by myeloid cells, including macrophages, induce early responses against infection or injury. Thus, myeloid cells are considered to trigger inflammation by providing pro-inflammatory cytokines. Here, we demonstrated that the pro-inflammatory cytokine Il1b is provided by epidermal cells in response to tissue injury. We showed that apoptosis occurred if the Il1b action at the injury site was prolonged, and that macrophages were responsible for attenuation of il1b expression and resolution of inflammation. Furthermore, our data suggest that normal il1b expression and inflammatory response are necessary in tissue regeneration to activate expression of regeneration-induced genes. Thus, our study has revealed that Il1b signaling and tissue inflammation act as a double-edged sword: they are required for regeneration, but in excess, they impair tissue regeneration (Figure 8).

28Myeloid cells are considered to be the principal producers of Il1b, although certain studies have reported il1b expression in melanoma cells and human keratinocytes. In this study, we showed that il1b is expressed by epidermal cells around the injury site. Notably, LPS injection induced il1b expression in myeloid cells, but not in the injured epidermal cells, suggesting that il1b induction by tissue injury occurs through a mechanism distinct from the mechanism underlying induction in myeloid cells following bacterial infection. Although it remains unknown how tissue injury stimulates il1b expression, a subject for future study, a substance released from disrupted cells or a mechanical signal could be responsible for inducing il1b expression.

29We also showed that in the clo mutant apoptosis is caused by prolonged il1b expression and inflammatory response. The precise mechanism by which Il1b induces apoptosis is unclear, but Tnf signaling could be a downstream mediator for triggering apoptosis. Tnfα signaling is known to stimulate apoptosis through activation of Caspase 8. Tissue inflammation has been suggested to induce tnfα expression, and zebrafish tnfb, one of the zebrafish homologs of tnfα, was upregulated in the clo mutant (Figure 1A); therefore, Tnfα signaling potentially serves as a direct mediator for inducing apoptosis in the clo mutant. Alternatively, the clo apoptosis might occur because of ER stress signaling: Il1b administration into primary rat β-cells and MIN6 cells was reported to increase ER stress through intracellular release of Ca2+ and to induce apoptosis in a JNK-dependent manner. Thus, excessive ER stress evoked by Il1b around the injury site could induce apoptosis in the clo mutant.

30In this study, we further demonstrated that macrophages play a crucial role in fin fold regeneration by attenuating il1b expression and supporting survival of regenerative cells. Recent studies have shown that knockdown of macrophage differentiation delayed fin fold regeneration and resulted in an abnormal fin fold featuring large vacuoles, and genetic ablation of macrophages in adult zebrafish was demonstrated to affect fin ray patterning and fin growth. These studies suggest that macrophages play a critical role in tissue regeneration, but the precise function of the cells remains unknown. Our study has, for the first time, elucidated a role of macrophages in attenuating il1b expression and inflammation during tissue regeneration. In accordance with our observation, macrophage-depleted mice and salamanders were found to display elevated il1b expression, which suggests that the macrophage function of attenuating il1b transcription is conserved in other vertebrate species.

31Our observation of macrophage accumulation in the injured fin fold suggests that the macrophage migration could play a role in attenuating il1b expression and quenching inflammation. However, we cannot conclude that the macrophage migration is necessary for attenuating the il1b expression, because our preceding study showed that an activity that rescues the clo apoptosis exists in the WT body extract prior to tissue amputation. We previously suggested that the effect of myeloid cells is mediated by a factor that is diffusible and heat-stable, but its identity remains unknown. The factor could be an anti-inflammatory protein/molecule such as IL-10, Tgf-β, or a lipid mediator. It is speculated that signals such as nFkB and/or STAT3, which are activated by tissue injury, are involved in inducing il1b expression. The anti-inflammatory molecules bind to cell surface receptors to activate intracellular pathways and antagonize the nFkB and/or STAT3 signals to attenuate il1b transcription.

32Most importantly, our study suggests that inflammation mediated by Il1b is also necessary for normal fin regeneration. Previously, we showed that genes such as fn1b, junba, and junbb were induced in response to tissue injury, but the upstream mechanism that regulates the expression of these genes remains unknown. Here, we have demonstrated that Il1b signaling is necessary and sufficient for triggering expression of regeneration-induced genes. Although the primary factors involved in tissue regeneration, tentatively named alarmins, remain to be identified, Il1b could represent one of the key upstream mediators for initiating expression of regeneration-induced genes and advancing tissue regeneration.

33Given that macrophages are required for resolving inflammation, an impairment or reduction of this function will lead to the diseases that accompany chronic inflammation, such as rheumatoid arthritis, inflammatory bowel disease, and auto-inflammatory diseases. Therefore, elucidation of the regulatory mechanism of the Il1b signal and the action of anti-inflammatory substances released from macrophages will enhance our understanding of the pathogenic processes of chronic-inflammation diseases and lead to development of suitable therapeutics.

34Zebrafish of a WT line, which has been maintained in our facility for >10 years through inbreeding, were housed in a recirculating system in a 14 hr day/10 hr night cycle at 28.5°C. When necessary, larvae were incubated in egg water (0.06% artificial marine salt, 0.0002% methylene blue) containing 0.003% phenylthiourea to prevent pigment formation. The zebrafish mutant strain clom39 (RRID:ZFIN_ZDB-ALT-980203-381), the fgf20a enhancer-trap line HGn21A, the BAC Tg(il1b:egfp) line, and Tg(hsp70l:il1b) were used. The clom39 mutant was genotyped as previously described. Zebrafish were subject to heat-shock induction at 38°C for 1 hr in a small water bath and then cooled to 28.5°C over time. Fin fold fin amputation was performed as previously described. Briefly, zebrafish larvae at 2 dpf were anesthetized with 0.04% 3-amino benzoic acid ethylester (tricaine) in egg water and their fin fold was amputated using a surgical razor blade. For reproducible quantification, the fin fold was carefully amputated at sites just posterior to the notochords. For quantification of fin fold regeneration in Figure 6F, Figure 7B and G, the length from the notochord end to the posterior tip of fin fold was measured.

35Larval posterior tissues (~1 mm from the distal end, approximately the tissue posterior to the yolk extension) from uncut WT, amputated WT, and amputated clo mutant, which were obtained from more than five independent batches of incrosses of clo heterozygotes, were collected on dry ice and stored at −80°C. Total RNAs from approximately 500 posterior tissue samples were extracted using TRIzol (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. RNA sequencing analysis was performed by Dragon Genomics (Kyoto, Japan). Briefly, purified and fragmented PolyA+ RNAs were prepared from 3 μg each of the total RNAs using the TruSeq RNA sample preparation kit (Illumina, San Diego, CA), and the respective cDNAs were synthesized according to the manufacturer’s instructions. The cDNAs fused with an adaptor on both ends were PCR-amplified (15 cycles) and used for sequencing analysis (Illumina HiSeq2000). The obtained sequence reads (60927373 reads for WT uncut, 59284653 reads for amputated WT at 6 hpa, 58507894 reads for amputated clo mutant at 6 hpa) were aligned to the zebrafish reference sequence using Bowtie software (version 0.12.7; RRID:SCR_005476). The normalized expression level of the respective genes, RPKM (reads per kilobase of exon model per million mapped reads), was calculated using ERANGE software (version 3.2; RRID:SCR_005240).

36MOs (Gene Tools, Philomath, OR) were dissolved in Danieau solution. Fertilized zebrafish eggs were dechorionated by incubating them with 2% pronase (Roche) and then microinjected at the 1–2 cell stages. The following MOs were used in this study:

37irf8 MO: 5ʹ-AATGTTTCGCTTACTTTGAAAATGG-3ʹ

38csf3r MO1: 5ʹ-ATTCAAGCACATACTCAC-TTCCATT-3ʹ

39csf3r MO2: 5ʹ- GAACTGGCGGATCTGTAAAGACAAA −3ʹ

40spi1b MO: 5ʹ-GATATACTGATACTCCATTGGTGGT-3ʹ

41il1b MO1: 5ʹ-CCCACAAACTGCA-AAATATCAGCTT-3ʹ

42il1b MO2: 5ʹ- AAACGTAAAATAACTCACCATTGCA −3ʹ

43std MO: 5ʹ-CCTCTTACCTCAGTTACAATTTATA-3ʹ

44The csf3r MO2 (0.25 mM) and the rest of MOs (1 mM) were injected at approximately 0.5 nl per egg. Under these conditions, no apparent side effect was observed. Injected embryos were incubated in 0.3× Niu-Twitty solution.

45Total RNAs were extracted from larval tail tissues (200 for each) posterior to the yolk extension using TRIzol and further purified using the RNeasy kit (Qiagen, Venlo, Netherlands). Among the purified total RNAs, aliquots that corresponded to 90 larval tails were used for cDNA synthesis. The cDNAs were synthesized using the Thermoscript RT-PCR kit (Thermo Fisher Scientific) with random hexamers as the primer. The synthesized cDNAs were diluted to 0.1 μg/μl and stored at −30°C. PCR was performed according to a standard procedure using Paq5000 DNA polymerase (Agilent Technologies, Santa Clara, CA), and the products were analyzed using 2% agarose gels. The following primers were used for PCR:

46il1b Fw: 5ʹ- GCAGAGGAACTTAACCAGCT −3ʹ

47il1b Rv: 5ʹ- TGCCGGTCTCCTTCCTGA −3ʹ

48actb1 Fw: 5ʹ- ATGGATGAGGAAATCGCTGCCCTGGTCGTTGACAA −3ʹ

49actb1 Rv: 5ʹ- AGAGAGAGCACAGCCTGGATGGCCACATACATGGC −3ʹ

50Whole-mount ISH analysis was performed as described previously, except that 5% polyvinyl alcohol was included in the buffer during the color reaction. After detection of the ISH signal, the samples were fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for color preservation, equilibrated with 80% glycerol, and then mounted on glass slides and photographed. The antisense junba, junbb, fn1b, apoeb, mpx, and lcp1 (l-plastin) probes used here were described previously. The egfp probe was synthesized from the construct pCS2-egfp, which contained the full-length egfp coding sequence. The il1b probe was directly synthesized from the il1b PCR product using the T7 promoter (underlined in the primer sequence).

51il1b probe Fw: 5ʹ-ATGGCATGCGGGCAATATGA-3ʹ

52il1b probe Rv: 5ʹ-TAATACGACTCACTATAGGGCTAGATGCGCACTTTATCCT-3ʹ

53For preparing tissue sections, samples were incubated in 20% sucrose in PBS for overnight at 4°C, embedded in Tissue-Tek compound (Miles) and 16 μm sections were obtained using a cryostat.

54S. typhosa LPS (Sigma–Aldrich, St. Louis, MO) was dissolved in water at 10 mg/ml and stored in aliquots at −30°C. A solution containing LPS (5 mg/ml) and phenol red (0.5%) was injected at 0.5 nl per larva into the pericardial cavity of anesthetized larvae at 2 dpf.

55The construct for generating Tg(hsp70l:mCherry-2a-il1b) was prepared by replacing the creERt2 sequence of pT2(hsp70l:mCherry-2a-creERt2) with the il1b cassette. The il1b cassette was prepared by PCR-amplifying the il1b sequence a.a. 125–273. The signal sequence of the human IL-1β receptor antagonist and the HA tag were fused at the N-terminus and C-terminus, respectively (underlined in the primer sequences).

56Fw il1b primer: 5ʹ-GCTAGCATGGAAATCTGCAGAGGCCTCCGCAGTCACCTAATCACTCTCCTCCTCTTCCTGTTCCATTCAGAGACGATCTGCAAAAACGTCTTGCAATGCACGATTTGCG-3ʹ

57Rv il1b primer: 5ʹ-AGATCTCTAAGCGTAGTCTGGGACGTCGTATGGGTAGATGCGCACTTTATCCTGCAGCTCGAAG-3ʹ

58The construct was injected into fertilized eggs. To identify germ-line transmission, the F0 fish were crossed with each other or with WT, and the embryos produced were screened for mCherry fluorescence after heat shock.

59To generate Tg(il1b:egfp), the iTol2 cassette was first introduced into the BAC clone CH211-147H23 using BAC recombineering. Next, the egfp-nitroreductase cassette was introduced into the site of the initiation codon of il1b. Lastly, a mixture of purified BAC DNA (125 ng/μl) and the transposase mRNA (25 ng/μl) was injected into 1-cell-stage zebrafish embryos (1 nl/embryo). The Tg was screened for EGFP expression in the lens, which is driven by the crystalline alpha promoter, and further tested for EGFP induction in response to fin fold amputation.

60Zebrafish larvae were fixed with 4% PFA in PBS for 2 hr at room temperature (RT) or overnight at 4°C, washed thrice with PBS containing 0.1% Triton X-100 (PBTx), dehydrated with methanol, and stored at −30°C. Samples were rehydrated with PBTx and then incubated (overnight, 4°C) with anti-E-cadherin (1:1000; BD Biosciences, Franklin Lakes, NJ; RRID:AB_397580) and anti-EGFP (1:1000; Nacalai Tesque, Kyoto, Japan; RRID:AB_10013361) antibodies in blocking buffer (5% serum and 0.2% bovine serum albumin in PBTx). After extensive washing with PBTx at RT, the samples were incubated (overnight, 4°C) with anti-mouse Alexa Fluor 568 and anti-rat Alexa Fluor 488 antibodies (both 1:1000; Invitrogen). After washing with PBTx, the tail regions were isolated and mounted in 80% glycerol containing 2.5% 1,4-diazabicyclo[2.2.2]octane (Nacalai Tesque) as an anti-fading reagent. Fluorescence images were acquired using a confocal microscope (FV-1000, Olympus, Tokyo, Japan).

61Zebrafish larvae of the BAC Tg(il1b:egfp) line at 1 hpa were placed on a 2% agarose-gel stage in egg water and embedded in 0.7% low-melting agarose. Time-lapse images were acquired once every 15 min using the confocal microscope equipped with a 20× water-immersion objective lens; the acquired images were z-stacks containing 20 optical slices.

62Larvae were fixed with 4% PFA for 2 hr at RT or overnight at 4°C, dehydrated with methanol, and stored at −30°C. Apoptosis was examined using an in situ apoptosis detection kit (Roche, Basel, Switzerland). Briefly, samples were rehydrated with PBTx, treated with 10 μg/ml Proteinase K in PBTx (5 min, RT), washed with PBTx, and refixed with 4% PFA in PBS (20 min). The samples were further incubated (15 min, on ice) in a freshly prepared 0.1% sodium citrate buffer containing 0.1% Triton X-100, washed with PBTx, and reacted with the TUNEL reaction mixture at 37°C for 1.5 hr. The reaction was terminated by washing with PBTx. The samples were mounted in 80% glycerol and fluorescence images were acquired using confocal microscopy. The TUNEL-positive cells were quantified by counting the number of cells in the area posterior to the notochords.

63Dex (D-2915; Sigma) was dissolved in dimethyl sulfoxide (DMSO) at 100 mM and stored at −30°C. The Dex solution was diluted to 100 μM with egg water and administered to zebrafish larvae at least 1 hr before fin fold amputation.

64Larvae were fixed with 4% PFA for 1 hr at RT, rinsed with PBS, and incubated in 0.03% Sudan black B (Sigma). After extensive washing with 70% ethanol and rehydration with PBS containing 0.1% Tween 20 (PBT), samples were mounted in 80% glycerol.

65Proliferating cells were labeled with 5-bromo-2-deoxyuridine (BrdU, 5 mM) during 0–12 hpa. The labeled larvae were fixed with 4% PFA for 2 hr at RT or overnight at 4°C, dehydrated with methanol, and stored at −30°C. Immunochemical detection was performed as described and BrdU-positive cells were quantified from the acquired confocal images. Similar to the quantification of TUNEL-positive cells, the BrdU-positive cells were quantified by counting the number of cells in the area posterior to the notochords.

66Most assessments of phenotypes and expression patterns were replicated in at least two independent experiments with comparable results. Larvae were collected from independent crosses, and experimental processing (injection, heat shock, and/or staining) was carried out on independent occasions. Exceptions to this include data presented in Figure 1E, Figure 7L, Figure 2—figure supplement 1, and Figure 6—figure supplement 1B–D. In each of these cases, multiple larvae were processed, and the obtained phenotypes were the same in all or most cases. The n is reported within the respective figures.

67Data are presented as means ± SEM. Statistical analyses were performed using Microsoft Excel 2013. For normally distributed data, differences were analyzed using Student’s t tests; p<0.05 was considered to be statistically significant.