TRPV4-Mediated Detection of Hyposmotic Stress by Skin Keratinocytes Activates Developmental Immunity

Successful development of the vertebrate immune system has been associated with the appearance of hematopoietic stem cells (HSCs), expansion of immunocompetent cells, and establishment of lymphoid organs (1). Myeloid cells of the innate immune system, namely macrophages and neutrophils, are formed early in development, whereas lymphopoiesis has been suggested to initiate much later, following emergence of definitive HSCs (2, 3). Myeloid cells respond to a set of exogenous microbes and endogenous danger-associated molecular patterns (DAMP), called microbial-associated molecular pattern (MAMP) and DAMP, respectively (4, 5). Such molecules are recognized by diverse innate receptors, leading to the production of powerful messenger molecules, such as cytokines and eicosanoids (6, 7). However, how vertebrates manage to mount a protective innate immune response before the definitive immune cellular components are fully developed remains an ongoing challenge (8–11). Until birth (or hatching), all animals are sterile; hence, little or no self-defense is needed. Given the limited exposure to Ags before specimens are released to their final living environment, some immune elements could be dispensable. Soon after birth, >50% of all human newborns develop a prominent but transitory rash caused by reaction of neonatal skin to commensal flora. Emerging evidence indicates that such bacterial skin interaction may activate release of IL-6 by the host, which is likely to contribute to amelioration of the rash, exemplifying how important an appropriate innate immune response against a particular stimulus is on the maturation of the neonatal immune system (12). In this regard, we and others have also reported the need of commensal microbes at birth/hatching to mediate a transient immune response and a delayed immune compensatory adaptation in part mediated by epigenetic changes (6, 13). This early colonization by commensal microbes seems to have a crucial impact for the development of the immune system and the health status during adulthood (14–16). We show in this work that newly hatched germ-free (GF) zebrafish embryos sense the hyposmolarity of the aquatic environment to rapidly mount a protective innate immune response. Pharmacological and genetic inhibition experiments show that skin keratinocytes through a transient potential receptor vanilloid 4 (TRPV4)/Ca²⁺/TGF-β–activated kinase 1 (TAK1)/NF-κB signaling pathway were responsible for both sensing the hyposmolarity of the aquatic environment and mediating immune effector mechanisms. Taken together, these results reveal osmotic surveillance as a crucial mechanism in the regulation of innate immunity in the absence of tissue damage and full barrier integrity and point out to the embryo skin as the first immune organ with full capacities to mount an innate immune/inflammatory response even in the absence of functionally mature immune cells.

Adult wild type obtained from the Zebrafish International Resource Centre, myd88^{hu3568/hu3568} (17), tg(krt18:RFP)^zf9 (18), and tg(NF-κB:eGFP)^sh235 (19) zebrafish embryos were maintained according to the Guidelines of the European Union Council (Directive 2010/63/EU) and the Spanish RD 53/2013. Experiments and procedures were performed as approved by the Bioethical Committees of the University of Murcia (approval number 537/2011). In each experiment, immediately after mating, embryos were derived GF, as described (6), or conventionally raised (CONR) following standard procedures (https://zfin.org/zf_info/zfbook/zfbk.html).

To elucidate the impact of osmotic shock on the embryonic inflammatory process at the time of losing their protective chorions, dechorionation was performed over 40 embryos of 24 h postfertilization (hpf) (Prim-5 stage) or 48 hpf (Long-pec stage) with two replicates each treatment. Embryos were aseptically selected under a stereomicroscope (Leica) and transferred by means of sterile plastic Pasteur pipettes to six-well plates systematically filled with sterile egg water (EW) solution (15 mOsm/l) prepared with sea salts, as described in the Zebrafish Book (https://zfin.org/zf_info/zfbook/zfbk.html) (20). Isosmotic and hyperosmotic saline solutions (280 and 400 mOsm/l), respectively, were prepared using increased salt concentrations of the basal EW formula. The osmolarity was determined with a vapor pressure osmometer (Vapro, Wescor). All working solutions were adjusted to a pH 7.2.

Chorions were removed with sharply pointed forceps (type Dumont No. 5) by piercing the chorion gently and—after internal pressure adjustment—enlarging the cleft carefully by means of the forceps. Lifting the chorion and turning it upside down allows the embryo to drop out of its chorions, as previously reported (21).

Morpholino antisense oligonucleotides targeting the zebrafish TRPV4 (5′-GTTACAAAGAAAAAGAGTCCAGAAC-3′) (22) or a standard (STD) control (5′-CCTCTTACCTCAGTTACAATTTATA-3′) were obtained from Gene Tools (Philomath, OR). Using a microinjector (Narishige IM-300), zebrafish embryos at the one-cell stage were microinjected into the yolk sac with STD or TRPV4 morpholinos at 8 pg/egg maximum, or 400 pg/egg antisense mRNA or encoding human wild-type TRPV4 (full length isoform A, accession number NM_021625) or its pore-dead mutant M680D (23). In some experiments, embryos were also microinjected with 6.5 pg/egg phenol-extracted genomic DNA from the bacterium Vibrio anguillarum, to activate TLR9 signaling (24).

For chemical treatments, zebrafish embryos were set up in six-well plates (Corning) and bath immersed at 28.5°C for 4 h in hyposmotic or hyperosmotic EW containing 0.5 μM TAK1 kinase inhibitor, 5Z-7-oxozeaenol (5Z), 100 μM NF-κB activator inhibitor (NAI), 1 μM 4α-phorbol 12,13-didecanoate (4α-PDD), 10 μg/ml HC-067047, 20 μM N-(p-amylcinnamoyl) anthranilic acid (ACA), 25–100 nM U73122, 10 μM 17-octadecynoic acid (17-ODYA), 500 nM 5,6-epoxyeicosatrienoic acids (EET), 30 μM BAPTA-AM (AM), and 30 and 100 μM BAPTA free acid (Free-Acid) purchased from Sigma-Aldrich, Calbiochem, Santa Cruz Biotechnology, Tocris Bioscience, or Enzo Life Sciences. After the bath treatment, all embryos were collected in 1.5 ml Eppendorf tubes containing 500 μl TRIzol reagent (Invitrogen) and immediately frozen at −80°C for further analysis.

Tail fin amputation was performed in wild-type larvae, as previously described (25).

Total RNA was isolated using TRIzol reagent, according to the manufacturer’s specifications, and was treated with DNase I (Invitrogen). cDNA was synthesized with the Superscript III First-Strand Synthesis system (Invitrogen), and real-time PCR was performed on a ABI PRISM 7500 cycler with SYBR Green PCR core reagents (Applied Biosystems) and primer pairs specific for cDNA of defensin β–like-1 (defbl1), il1b, cxcl8a, egfp, il10, lysozyme (lyz), prostaglandin-endoperoxide synthase 1 (ptgs1), ptgs2a, ptgs2b, tnfa, and ribosomal protein S11 (rps11) transcripts. The primer sets were as follows: defbl1, 5′-CAGGACTGCCATCATCTGAA-3′ (forward) and 5′-CTCCTTGTCTGCAAACACCA-3′ (reverse); il1b, 5′-GGCTGTGTGTTTGGGAATCT-3′ (forward) and 5′-TGATAAACCAACCGGGACA-3′ (reverse); cxcl8a, 5′-GTCGCTGCATTGAAACAGAA-3′ (forward) and 5′-CTTAACCCATGGAGCAGAGG-3′ (reverse); egfp, 5′-CACATGAAGCAGCACGACTT-3′ (forward) and 5′- AGTTCACCTTGATGCCGTTC-3′ (reverse); il10, 5′-ATTTGTGGAGGGCTTTCCTT-3′ (forward) and 5′-AGAGCTGTTGGCAGAATGGT-3′ (reverse); lyz, 5′-TGGCAGTGGTGTTTTTGTGT-3′ (forward) and 5′-TCAAATCCATCAAGCCCTTC-3′ (reverse); ptgs1, 5′-TTTTGCTGCTGAGTGTGTCC-3′ (forward) and 5′-CGAACACAGATCCCTTGGTT-3′ (reverse); ptgs2a, 5′-TGGATCTTTCCTGGTGAAGG-3′ (forward) and 5′-GAAGCTCAGGGGTAGTGCAG-3′ (reverse); ptgs2b, 5′-CCCTCATGCCTGATGATTTT-3′ (forward) and 5′-CCACCCTTAACACTGCTGGT-3′ (reverse); tnfa, 5′-GCGCTTTTCTGAATCCTACG-3′ (forward) and 5′-TGCCCAGTCTGTCTCCTTCT-3′ (reverse); and rps11, 5′-ACAGAAATGCCCCTTCACTG-3′ (forward) and 5′-GCCTCTTCTCAAAACGGTTG-3′ (reverse). Reactions were run in triplicate, and samples were normalized to rps11 content in each sample using the Pfaffl method (26). In all cases, each PCR was repeated at least twice.

All steps were performed at room temperature, unless otherwise indicated. Tg(krt18:RFP) larvae were fixed in 4% paraformaldehyde, dehydrated in methanol/PBS, and stored in 100% methanol at −20°C. For staining, larvae were rehydrated in methanol/PBS and 0.1% Tween 20 (PBT), washed in PBT, incubated with 150 mM Tris-HCl (pH 9), followed by heating at 70°C for 15 min (27). After heating, larvae were washed in PBT and dH₂O. Subsequently, to enhance tissue permeabilization, larvae were incubated with acetone and washed in dH₂O and in PBT (5 min each). Samples were blocked in PBT + 1% DMSO supplemented with 5% FBS and 2 mg/ml BSA for 2 h. Then, embryos were incubated overnight at 4°C with primary Abs diluted (1:200) in blocking buffer, washed in PBT + 1% DMSO (15 min each), and blocked again for 2 h. Secondary Ab staining was made for 2 h at 1:500 dilution in blocking buffer; embryos were then washed in PBT and stored in Vectashield (Vector Laboratories) until image acquisition. Primary Abs were used at 1/200: rabbit anti-zfTRPV4 (OST00072G; GeneTex) at 1:200; rabbit anti-human p63 (SC8343; Santa Cruz Biotechnology); mouse anti-RFP (MA5-15257; Thermo Scientific); and mouse monospecific anti–zfIL-1β (epitope AEGSAPHLVLKE) (Abmart). Alexa Fluor 594 (A11032) (H+L) and Alexa Fluor 488 (A11008) goat anti-mouse or anti-rabbit IgG (H+L) (Life Technologies) were used as secondary Abs at 1:500. Confocal immunofluorescence images were acquired with a confocal microscope (LEICA TCS-SP2; Leica) using a NA 0.70/20× dry objective. Z-series were acquired using a 210- to 300-μm pinhole. Two-dimensional and three-dimensional maximum intensity projections and animation videos were made using ImageJ (http://rsb.info.nih.gov/ij/).

Salmonella enterica serovar Typhimurium (ST) was used for all inoculations according to the following recipe. ST was grown overnight at 37°C in Luria–Bertani (LB) broth at 250–300 rpm (28). The next morning, inocula was diluted 1/5 in LB broth, with 0.3 M NaCl, and incubated at 37°C until 1.5 OD at 600 nm was achieved. At 24 and 48 hpf, half of the embryos in each experimental batch were dechorionated and subjected to hyposmotic shock in 15 mOsm/l EW for 4 h. Thereafter, the remaining eggs not subjected to an extended osmotic shock were dechorionated and, together with previously dechorionated embryos, were anesthetized in EW containing 0.16 mg/ml tricaine. Each embryo was inoculated with 10 bacteria into the yolk sac following a microinjection procedure. Larvae were allowed to recover in EW at 28.5°C and monitored for clinical signs of disease or mortality over a 5-d period. The numbers of bacteria injected per larvae were frequently checked by determining the number of CFU/injection volume in LB agar plates.

For calcium imaging experiments, HeLa cells were transfected with human wild-type or M680D mutant TRPV4, as described previously (29). Isosmotic bath solutions contained 140 mM NaCl, 2.5 mM KCl, 1.2 mM CaCl₂, 0.5 mM MgCl₂, 5 mM glucose, and 10 mM HEPES (pH 7.3). The 30% hyposmotic solution (220 mOsm/l) was obtained by using 100 mM NaCl instead of 140 mM NaCl. Cytosolic Ca²⁺ signals, relative to the ratio (340/380) measured prior to cell stimulation, were obtained from cells loaded with 4.5 μM fura-2 AM (Invitrogen), as previously described (29).

For statistical analyses, p values were calculated by paired Student t tests, multiple comparisons with ANOVA and post hoc Tukey test, or log rank test to calculate the differences in survival, unless stated otherwise in figure legends. A p value <0.05 was considered significant.

We previously found that microbiota-derived MAMP acting through TLRs and epigenetic modifications of the promoters of genes encoding proinflammatory mediators regulate developmental immunity in zebrafish embryos (6). In this work, we studied whether osmotic stress is also able to regulate developmental immunity in zebrafish. Forty-eight hpf zebrafish embryos raised in GF conditions, to prevent TLR activation, showed increased transcript levels of il1b at 4 h posthatching in EW, which has an osmolarity of 15 mOsm/l (Fig. 1A). This behavior suggested that, in addition to the impact of commensal microbes observed in CONR embryos (Fig. 1A), the osmotic insult occurring after hatching also contributes to the activation of innate immunity. To extend this preliminary observation, we investigated whether the induction of il1b in newly hatched embryos was dependent either on the external medium osmolarity or the developmental stage. Dechorionated GF embryos at Prim-5 (24 hpf) and Long-pec (48 hpf) developmental stages were left in hypo and hyper (∼400 mOsm/l) osmotic media (Fig. 1B, 1C). Notably, the previously observed induction of il1b triggered by a hyposmotic shock was not found in newly hatched embryos maintained under hyperosmotic conditions. Surprisingly, independently of the profound differences between embryos at these two distant developmental stages, in which macrophages are hardly present at Prim-5 (30) and neutrophils are completely absent (31), the same il1b expression profile was found under hyper- and hyposmolarity conditions (Fig. 1C). Therefore, we reasoned that hyposmolarity-induced il1b expression might be mediated through a developmental stage-independent mechanism in zebrafish embryos, with capacity to immediately trigger innate immunity after hatching. To test this idea, we exposed dechorionated Prim-5 embryos to a gradient of osmolarities (from 15 to 465 mOsm/l) (Fig. 1D). il1b mRNA levels showed a biphasic response, increasing until reaching a peak at 100 mOsm/l and then decreasing with higher osmolarities to levels similar to nondechorionated embryos. Embryos that remained inside their protective chorions did not show any change (Fig. 1D). Curiously, however, no changes were observed in the mRNA levels of genes encoding antimicrobial effectors (Supplemental Fig. 1).

To explore whether the observed activation of il1b expression would be mediated by classical TLR signaling, we assessed the effects of hyposmotic shock on CONR animals harboring a mutation on the myeloid differentiation primary response gene 88 (myd88) (17), which is known to be required by all TLRs but TLR3 in mammals (32). We observed that expression of il1b in dechorionated myd88^−/− or their wild-type sibling embryos was similarly affected after hatching in hyposmotic medium (Fig. 1E). Furthermore, dechorionated myd88^−/− embryos showed lower il1b transcript levels than myd88^+/+, suggesting that commensal recognition by Myd88-dependent receptors is abrogated in the mutant embryos (Fig. 1E). Therefore, these observations clearly pointed out that the induction of il1b in GF embryos after hatching is not mediated by pattern–recognition receptors, at least those from the TLR family engaged with the adaptor molecule Myd88. To extend these studies further, we explored the Long-pec embryo (i.e., 48 hpf) to identify which cells might be responsible for the production of IL-1β. Thus, we performed a double immunostaining against P63, a basal keratinocyte marker, and Il1b in whole GF and CONR embryos with or without their protective chorions. The results unequivocally revealed that keratinocytes were the cells expressing Il1b at this developmental stage, and, although GF embryos with their protective chorions had a low expression of Il1b in keratinocytes, it was significantly more pronounced in dechorionated embryos subjected to hyposmotic shock (Fig. 1F, 1G, Supplemental Fig. 2, Supplemental Videos 1, 2). Thus, hyposmolarity triggers an inflammatory response in the skin of newly hatched zebrafish embryos. These experiments conclusively show that cells expressing IL-1β in developing embryos are keratinocytes, although we could not rule out that primitive myeloid cells present at these early developmental stages may contribute to the production of this important proinflammatory cytokine.

To fully understand the function of osmotic shock on inducing inflammation, quantitative RT-PCR analysis of a selected array of key immune-related genes was performed with mRNA from whole Prim-5 (24 hpf) embryos dechorionated or not, and incubated in hyposmotic medium for 4 h (Fig. 2A–H). As expected, many genes whose expression was strongly increased were those involved in a proinflammatory response, tnfa being one of the most highly induced genes (Fig. 2D). Strikingly, transcript levels of gene encoding anti-inflammatory Il10 (Fig. 2F) and the antimicrobial effectors Defbl1 and Lyz were not different from embryos with their protective chorions (Fig. 2G, 2H). Conversely, the mRNA levels of Ptgs2a, Ptgs2b, Tnfa, and Cxcl8a were significantly induced in Prim-5 embryos by a short hyposmotic shock (Fig. 2B–E).

To determine whether the hyposmotic shock has a role in the protection of newly hatched animals against bacterial infection, Prim-5 that had been dechorionated immediately prior to challenge or 4 h earlier were microinjected in their yolk sacs with 10 CFU ST or PBS as control. ST-infected and immediately dechorionated embryos were more susceptible than their sibling embryos dechorionated 4 h in advance (Fig. 2I). Collectively, these data suggested a critical role for the innate immune response triggered by hyposmotic shock after hatching by ultimately increasing embryo protection against infection.

To explore the molecular mechanisms by which hyposmotic shock modulates immunity in newly hatched embryos, a pharmacological inhibition approach was used. We assessed in Prim-5 embryos, dechorionated or not, the effect of 5Z, a specific inhibitor of the key transcription factor TAK1, which is known to be indispensable for the signaling by a wide set of membrane-bound cell receptors, including TLRs. The 5Z inhibitor (4-h incubation) on dechorionated Prim-5 embryos exposed to hyposmotic medium caused a strong reduction of IL-1β transcript levels (Fig. 3A). As expected, TAK1 inhibition with 5Z resulted in lower il1b mRNA levels in CONR MyD88 mutant embryos and their wild-type siblings as well (Fig. 3B). We also found that the impact of osmotic insult in the transcript levels of a wide set of immune-related genes was also inhibited by pharmacological inhibition of TAK1 (Fig. 3C–G). However, tnfa expression seemed to be regulated by an alternative signaling pathway (Fig. 3E). Notably, similar results were obtained for all genes analyzed with the NAI (Fig. 3C–G). Moreover, the specificity of the inhibitors was further confirmed through decreased GFP transcript levels on NF-κB:GFP reporter animals (Fig. 3D). Collectively, these results demonstrated that osmotic insult triggers il1b expression through the TAK1/NF-κB signaling axis in skin keratinocytes and strongly suggest a osmo/mechanoreceptor, rather than a pattern–recognition receptor, which acts upstream of the TAK1 complex by sensing the osmotic insult.

Transient receptor potential cation channels participate in the transduction of both physical and chemical stimuli. Among them, TRPV4 is widely expressed, including the skin, and participates in the transduction of a variety of stimuli, such as warm temperatures, hyposmolarity, and endogenous lipids (33). Immunostaining for Trpv4 and RFP of krt18:RFP transgenic embryos showed that skin keratinocytes expressed Trpv4 at high levels (Fig. 4A, Supplemental Video 3). To determine whether keratinocyte Trpv4 sensed the hyposmotic shock, embryos were dechorionated in hyperosmotic medium to inhibit osmotic shock (Fig. 4B). Under this condition, no statistically significant increased levels of il1b mRNA were found, but further incubation with the TRPV4 agonist 4α-PDD (34) resulted in marked increased levels of il1b mRNA (Fig. 4B). Following a similar strategy, dechorionated embryos were incubated for a short period in hyposmotic medium, but supplemented with the highly specific TRPV4 antagonist HC067047 (35), which was able to impair the hyposmolarity-induced il1b expression (Fig. 4C). To confirm the chemical inhibition, a previously validated TRPV4 splice-blocking morpholino (22) was used. This morpholino targets the end of the exon 8, 4 bp upstream the intron 8 (ENSDART00000138186), resulting in an in-frame deletion of the whole seventh-coding exon (93 bp) and therefore the loss of most of the second transmembrane domain (22). In addition, we found that this morpholino promoted the degradation of mature trpv4 mRNA (Supplemental Fig. 3A). No developmental abnormalities were observed at the doses of Trpv4 morpholino tested (data not shown). Dechorionated Trpv4-deficient embryos incubated in hyposmotic solutions showed a strong inhibition of il1b levels, confirming the involvement of Trpv4 on sensing hyposmolarity in keratinocytes (Fig. 4D). Moreover, we quantified the expression of other genes encoding key mediators of inflammation, namely, Ptgs2a and Ptgs2b, and found a strong inhibition as well (Supplemental Fig. 3B, 3C). Notably, hyposmolarity had no comparable effect on the transcript levels of antimicrobial effectors in Trpv4-deficient animals; even the addition of a strong stimulus, represented in this study by bacterial DNA, did not produce any change (Fig. 4E, Supplemental Fig. 3D), which suggests that the expression of antimicrobial effectors was independent of both hyposmolarity/Trpv4 and pathogen-associated molecular patterns/TLR signaling. Curiously, Trpv4 deficiency resulted in lower il1b mRNA levels in Prim-5 embryos either in the presence or absence of bacterial DNA (Fig. 4F), suggesting a crosstalk between Trpv4 and TLR signaling. Importantly, Trpv4 deficiency fully abrogated the resistance to ST infection of embryos exposed to hyposmotic medium (Figs. 2I, 4G). These results were consistent with our previous findings and give support to the fact that developmental immunity is triggered after hatching by hyposmolarity through a TRPV4/TAK1/NF-κB signaling pathway operating in skin keratinocytes.

Lipid signaling through activated phospholipases has been suggested as a critical regulator of inflammation. In addition, osmotic (36) and mechanical (37) sensitivity of TRPV4 depend on phospholipase A2 (PLA2) activation and the subsequent production of the arachidonic acid metabolites, EETs, by cytochrome P450 (CYP) epoxygenase. Reports also exist claiming EET-independent TRPV4 activation by membrane stretch in excised patches from oocytes (38). To analyze whether PLA2 and CYP epoxygenase were required to trigger immunity in newly hatched embryos, their enzymatic activities were pharmacologically inhibited with ACA and 17-ODYA, respectively. Both inhibitors failed to significantly inhibit hyposmolarity-induced il1b expression (Fig. 5A). To rule out that PLA2 was immediately translocated upon TRPV4 activation, Prim-5 embryos were dechorionated in hyperosmotic solution to inhibit the osmotic shock, and after 2 h they were switched to hyposmotic medium with or without the PLA2 inhibitor (Fig. 5B). Consistent with our previous results, PLA2 inhibition failed to affect hyposmolarity-induced il1b expression. Although we observed no participation of the PLA2/CYP pathway in hyposmolarity-activated, TRPV4-mediated increased IL-1β mRNA, direct application of 5,6-EET triggered a marked increased in IL-1β mRNA that was blocked with the TRPV4 inhibitor HC067047 (Fig. 5C), suggesting that zebrafish TRPV4 channel can also be activated by EET.

As it is known that, under conditions of low PLA2 activation, hyposmolarity also uses extracellular ATP-mediated activation of phospholipase C (PLC)-inositol trisphosphate signaling to support TRPV4 gating (39), we next inhibited PLC using U73122, but, unfortunately, PLC inhibition drastically induced il1b transcript levels (Supplemental Fig. 4).

Our next effort was to determine the role of intracellular Ca²⁺ in the hyposmolarity-triggered immunity in newly hatched embryos. Results revealed no significant difference in IL-1β transcript levels between embryos hatched in the physiologic zebrafish EW and Milli-Q water (6 mOsm/l) (Fig. 5D). Using two potent Ca²⁺ chelators, the membrane-permeant BAPTA-AM and the nonpermeant BAPTA free acid, we confirmed that intracellular, but not extracellular, Ca²⁺ was involved in the rapid increase of il1b mRNA levels mediated by hyposmolarity (Fig. 5E). To rule out the possibility that extracellular Ca²⁺ was not completely chelated, BAPTA, free acid concentration was increased to 100 μM in zebrafish EW, but il1b transcript levels still increased (Fig. 5F), further suggesting that extracellular Ca²⁺ was not required. To further confirm the involvement of TRPV4 in the hyposmolarity-induced IL-1β response, we used the pore-dead human TRPV4-M680D mutant (23). We demonstrated that TRPV4-M680D behaved as dominant negative by partially inhibiting Ca²⁺ influx mediated by wild-type human TRPV4 in HeLa cells incubated in a hyposmotic solution (Fig. 5G). Consistent with the dominant-negative effect of TRPV4-M680D in human cells, we observed that overexpression of the human TRPV4-M680D mRNA inhibited the hyposmolarity-induced il1b transcript levels in Prim-5 embryos, whereas it did not affect the transcript levels of the lyz gene (Fig. 5H). In addition, expression of human wild-type TRPV4 was able to restore hyposmolarity-induced il1b expression in newly hatched Trpv4-deficient embryos and augmented the response of control embryos (Fig. 5I). Together, these results confirmed the specificity of Trpv4 morpholino and the full activity of human TRPV4 in zebrafish embryos. Besides, these data indicate that hyposmolarity activates immunity in newly hatched embryos through a TRPV4/Ca²⁺/TAK1/NF-κB signaling pathway in keratinocytes independently of PLA2/CYP epoxygenase-mediated production of EET (Fig. 6).

Extensive progress has been made in understanding the molecular pathways orchestrating inflammation and immunity by different myeloid cell lineages. In contrast, how developing organisms, which have not fully achieved immune system maturation, mount immune reactions and respond against extrinsic insults is poorly understood. In this work, we have demonstrated a previously unrecognized function for skin keratinocytes as osmotic sensors with a critical role on activating innate immunity and providing protection to bacterial infection in newly hatched zebrafish embryos. Zebrafish has emerged as a good model to study human diseases and immunological mechanisms that are hardly accessible in more classical vertebrate models (6, 40–44). In our experimental setting, immediately after hatching, keratinocytes recognize a new hyposmotic environment, a mechanism in which the TRPV4 channel plays an important role. Upon a short activation, genes encoding key proinflammatory mediators were induced through a canonical TAK1/NF-κB signaling cascade, which leads to increased resistance against the pathogenic bacterium ST in a MyD88-independent manner (Fig. 6). Surprisingly, noncanonical arachidonate metabolites derived from PLA2 and CYP epoxigenase were not involved in the osmotic sensitization of zebrafish keratinocyte via the TRPV4 channel. In contrast, PLA2 and noncanonical arachidonate metabolites, via a 5-oxo-ETE receptor, have been shown to mediate hyposmolarity-induced leukocyte chemotaxis and rapid leukocyte recruitment to sites of tissue damage in zebrafish (45). These differences might be explained by the absence of tissue damage and, therefore, full barrier integrity in newly hatched embryos compared with a local wounded tissue where DAMPs are rapidly released, such as ATP that activates purinergic receptors and signals through alternate pathways (46, 47). Unfortunately, we failed to determine the contribution of PLC downstream TRPV4 activation, because pharmacological inhibition of PLC drastically induced IL-1β transcript levels, probably reflecting the essential role of PLC-δ1 in skin stem cell lineage commitment (48), which results in the induction of inflammation in PLC-δ1–deficient mice (49).

Although the TRPV4 is a plasma membrane Ca²⁺-permeable, nonselective cation channel (33), our pharmacological and genetic approaches point to the involvement of intracellular Ca²⁺ as the source of Ca²⁺ required for the activation of immunity in newly hatched embryos, which may be related to the reported link between TRPV4 activity and the machinery involved in intracellular Ca²⁺ release (39, 50). Regardless of the molecular mechanisms involved in the activation of TRPV4 and the source of Ca²⁺ release, aspects that deserve future studies, our results clearly demonstrate that a raise in intracellular Ca²⁺ is required downstream of TRPV4 to activate developmental immunity (Fig. 6). It is not unexpected that cytosolic Ca²⁺ would then activate TAK1, because we have recently demonstrated that transient receptor potential channel activation modulates intracellular Ca²⁺ during cell volume regulation (51), and this change in intracellular calcium appears to be crucial for the activation of TAK1 (52). In addition, it has also been shown that TAK1 might be activated by intracellular Ca²⁺ variations after membrane stretch (53). Therefore, our results are consistent with a TRPV4/Ca²⁺/TAK1/NF-κB axis in zebrafish skin keratinocytes that activates immunity in response to hyposmolarity.

An important observation of this study is that osmotic activation of TRPV4 after hatching regulates the expression of all proinflammatory genes analyzed, but not those encoding antimicrobial effectors or anti-inflammatory cytokines, as we recently observed for the activation of developmental immunity by MAMPs that signal through a TLR/NF-κB signaling pathway (6). Furthermore, IL-1β in this study was upregulated by a gradient of tonicities between iso to hypotonic, but not hypertonic (∼400 mOsm/l), pointing out the sensitivity of skin keratinocytes against a slight osmotic change. This exquisite sensitivity is compatible with the involvement of TRPV4 in mechano/osmotic responses (33). Accordingly, a number of disorders tightly linked to inflammatory responses, such as diabetes, obesity, and dry eye syndrome, to mention a few, are all associated to some extent with local and systemic changes in extracellular fluid osmolarity in the absence of apparent tissue damage (54, 55). Interestingly, we found a crosstalk between both signaling pathways that deserves further study, because it may have important consequences on chronic inflammatory or infectious diseases.

From 22 to 72 hpf, the temporal window over which zebrafish HSCs emerge from aortic endothelium, the only leukocytes present are primitive myeloid cells, namely macrophages and neutrophils (30, 31, 56). Neutrophils, and to a less extent macrophages, have been described as the main source of IL-1β and TNF-α (19, 43, 57), but at the Prim-5 stage macrophages are hardly detected and neutrophils are absent (30, 31, 56, 58). However, we demonstrate that if protective chorion is removed in advance, zebrafish has already developed yet various types of epidermal cells at the surface and inner layer (59) that cope with common stressors (60). Although our results do not completely rule out a role for primitive myeloid cells in the hyposmolarity-dependent induction of immunity after natural hatching, this is unlikely because TRPV4 and IL-1β were both mainly expressed by keratinocytes at both Prim-5 (24 posthatching) and Long-pec (48 hpf) developmental stages.

Overall, our work has uncovered a previously unappreciated role of skin in the activation of developmental immunity, a mechanism in which the TRPV4 channel plays a key role. These findings also reveal that drugs targeting TRPV4 channel or the associated signaling pathway are promising tools for reduction of skin inflammatory disorders, such as psoriasis and atopic dermatitis. Further studies should explore in detail how hyposmolarity regulates the maturation of the vertebrate immune system, and the crosstalk between skin and immune cells and between the TRPV4 and TLR signaling pathways during development.

We thank Inma Fuentes and Pedro J. Martínez for expert technical assistance, Dr. P. Pelegrín for helpful discussion, Prof. Li Zon for critical reading of the manuscript, and Drs. A. Meijer, H. J. Tsai, and S. A. Renshaw for providing the myd88 mutant, krt18:RFP, and NF-κB:eGFP lines, respectively.

The authors have no financial conflicts of interest.