The Mineralocorticoid Receptor Plays a Crucial Role in Macrophage Development and Function

Abstract

Stress and the attendant rise in glucocorticoids (GCs) results in a potent suppression of the immune system. To date, the anti-inflammatory role of GCs, via activation of the glucocorticoid receptor, has been well-characterized. However, cortisol, the primary GC in both fish and humans, also signals through the high-affinity mineralocorticoid receptor (MR), of which the immunomodulatory role is poorly understood. Here, we tested the hypothesis that MR is a key modulator of leukocyte function during inflammation. Using transgenic MR knockout zebrafish with fluorescently labelled leukocytes, we show that a loss of MR results in a global reduction in macrophage number during key development stages. This reduction was associated with impaired macrophage proliferation and responsivity to developmental distribution signals, as well as increased susceptibility to cell death. Using a tail fin amputation in zebrafish larvae as a model for localized inflammation, we further showed that MR knockout larvae display a reduced ability to produce more macrophages under periods of inflammation (emergency myelopoiesis). Finally, we treated wild-type larvae with an MR antagonist (eplerenone) during definitive hematopoiesis, when the macrophages had differentiated normally throughout the larvae. This pharmacological blockade of MR reduced the migration of macrophages toward a wound, which was associated with reduced macrophage Ccr2 signalling. Eplerenone treatment also abolished the cortisol-induced inhibition of macrophage migration, suggesting a role for MR in cortisol-mediated anti-inflammatory action. Taken together, our work reveals that MR is a key modulator of the innate immune response to inflammation under both basal and stressed conditions.

The immune system can be characterized as a trait that provides high benefits but also incurs high costs; its activity is therefore tightly regulated by various endocrine factors. Indeed, glucocorticoids (GCs) are key drivers of the adaptative stress response and are well-characterized immune suppressants (1). The primary endogenous GC in both humans and fish is cortisol, and the principal role of this hormone is to redistribute energy post-stress, aiding in recovery and restoring homeostasis (2-4). During stress, cortisol production is upregulated through activation of the hypothalamus-pituitary-adrenal axis in mammals (3), analogous to the hypothalamus-pituitary-interrenal axis in teleosts (2). Cortisol exerts its effects on target tissues through activation of corticosteroid receptors, either the mineralocorticoid receptor (MR) or the glucocorticoid receptor (GR). Both of these receptors are ubiquitously expressed, ligand-activated transcription factors. The immune-suppressive effects of cortisol are thought to be primarily mediated by GR, and synthetic GR agonists have been used therapeutically for decades as anti-inflammatory drugs. However, little is known about the immune-modulatory role of MR, in particular, whether MR activation induces a different phenotype than GR activation and what the associated mechanisms of action may be (1, 5).

The molecular mechanisms underlying the anti-inflammatory effects of GC-GR signalling have been thoroughly investigated (1). GCs diffuse through the plasma membrane and bind to the cytosolic GR, which exists as a heterocomplex with chaperone proteins and immunophilins (1). On ligand binding, GR will translocate to the nucleus, where it will regulate the transcription rate of target genes through several mechanisms. It may interact directly with GR elements and promote gene transcription of anti-inflammatory genes (1, 6). Alternatively, it may restrict transcription by interacting with negative GR elements (7), or by interacting with other transcription factors, such as NF-kB or AP-1, thereby suppressing the transcription of pro-inflammatory genes (1). Indeed, the ability of GR to restrict NF-kB signalling via this transrepression mechanism is thought to be crucial for GR's anti-inflammatory action (1, 6). For example, monomeric GR, which can still transrepress NF-kB and AP-1 signalling, exhibits a potent anti-inflammatory effect, but it does not induce the negative metabolic side effects of chronic GC treatment (8).

A key difference between GR and MR is their respective affinities for cortisol. GR is a low-affinity receptor primarily activated under conditions of high cortisol, such as during stress or the diurnal peak. In contrast, MR has a 10-fold higher affinity to cortisol and is thought to be activated at basal levels of cortisol and is considered to be constitutively active (7). The molecular mechanisms of cortisol-MR action are thought to be highly similar to those underlying GR action, but they have not been studied as extensively, and it is largely unknown what the immunological profile of MR activation is. Although cortisol is the primary ligand for MR in fish, terrestrial vertebrates synthesize aldosterone, which is an MR ligand important for salt and water homeostasis. Given the high affinity of cortisol for MR, aldosterone-MR signalling is only possible via the action of the enzyme 11-β hydroxy-steroid-dehydrogenase 2 (11βHSD2), which is particularly active in the kidney and locally inactivates cortisol by converting it into cortisone (9). Importantly, monocytes and macrophages do not express 11βHSD2 (9), which suggests that cortisol-MR signalling predominates in these cells and may point to a conserved role for this receptor through evolution (10). Clinically, MR antagonists (MRAs) are used to treat hypertension, congestive heart failure, and other conditions associated with excess aldosterone activity, and recent work has described anti-inflammatory effects (5), making further characterization of MR modes of action in the immune system paramount.

In humans, it has been observed that MR is activated in cardiovascular disease (CVD) and that treatment of patients with an MRA reduces CVD morbidity and mortality (reviewed in (5, 11)). Although it might be initially assumed that this is due to the role of MR in the regulation of blood pressure and salt homeostasis, the effects of MRA on CVD appear to be independent of blood pressure levels (5), and it is thought that the anti-inflammatory action of MRAs may be playing a role, but little is known regarding the underlying molecular mechanisms. Furthermore, human data investigating the immunomodulatory effects of MR are scarce and restricted to in vitro studies (11), precluding any information regarding the role of MR at the systems level. Apart from its role in CVD, the inflammatory effects of MR have also been examined in mouse models of Duchenne muscular dystrophy, in which it was determined that a loss of MR resulted in a reduced number of infiltrating macrophages into the skeletal muscle (12). However, the mechanisms behind this pro-inflammatory MR effect have not been fully elucidated. Taken together, it can be hypothesized that MR is a key modulator of inflammation, in particular of macrophage function, promoting macrophage migration and a pro-inflammatory phenotype. This would contrast with the effects of GR activation, which results in potent immune suppression and an anti-inflammatory macrophage phenotype.

Overall, uncovering an immunological role of MR has been hampered by the lethality of a ubiquitous knockout of MR in mice from a loss of fluid and ion balance and eventual dehydration (13). Cell type-specific MR knockouts in mice have been applied to myeloid cells (MyMRKO), and macrophages lacking MR exhibited a transcriptional profile of alternative activation, pointing to a pro-inflammatory role of MR in these cells (10). The molecular targets of MR in driving this alternative macrophage phenotype are unknown, and reports in vitro are contradictory (14, 15). What has been consistently reported is that a loss of myeloid MR reduces the infiltration of macrophages under various inflammatory conditions (eg, in the muscle of mouse models of muscular dystrophy (16), atherosclerosis plaques (17), the liver in mouse models of nonalcoholic steatohepatitis (18), in the kidney after injury [reviewed in (19)], and in the central nervous system in models of autoimmune encephalomyelitis (20)). However, there is a lack of consensus regarding the molecular mechanisms that drive these phenotypes, and they may be condition-specific.

As an alternative animal model, the zebrafish provides distinct advantages when studying mechanisms of MR action, such as the ability to harbor ubiquitous knockouts of both GR and MR (21). This allows for a more comprehensive characterization of the roles of MR and GR at the systems level. Indeed, these ubiquitous knockouts have been used to discover the physiological significance of MR, specifically uncovering novel roles in metabolism (22, 23), behavior (21, 24, 25), and gene regulation (26). Additionally, the immune system is well-conserved between fish and mammals, and because of the optical transparency of embryos and larvae, fluorescently labelled leukocytes can be visualized in real time (27, 28). Using this model, the dynamics of leukocytes of the innate immune system (macrophage and neutrophils) in response to GCs have been well-established (27, 29). Additionally, the development of the innate immune system, including the production of myeloid cells (myelopoiesis), is well-characterized and commences with primitive myelopoiesis at 12 to 24 hours postfertilization (hpf), with the maturation of the adaptive immune system only occurring after 1 month of age (30, 31).

Here, we used zebrafish transgenic models to test the hypothesis that MR is a key modulator of the innate immune system during inflammation. Our results show a novel role of MR in early macrophage development and that distribution as a loss of MR causes a global reduction in the macrophage populations throughout the embryo. This was associated with a lack of macrophage responsivity to developmental distribution signals. Moreover, in wild-type (WT) larvae, where macrophage development was allowed to proceed normally, treatment with an MRA shows an anti-inflammatory effect, which is associated with a reduced responsiveness to the chemokine Ccl2, suggesting a pro-inflammatory role for MR under basal conditions. Taken together, we demonstrate that MR activity promotes the production, distribution, and function of macrophages, suggesting a fundamental role for this receptor in the function of the innate immune system.

Materials and Methods

Zebrafish Husbandry

Zebrafish were maintained in accordance with guidelines from the zebrafish model organism database (zfin.org) and in compliance with the directives of the animal welfare body at Leiden University. Briefly, fish were held in a recirculating system on a 14 hour:10 hour light:dark cycle (light on 08:00 hours; light off 22:00 hours). Water was maintained at 28 °C, 300 μS conductivity, and a pH of 7.5. Fish were fed twice daily with a diet consisting of Gemma Micro 500 (Skretting, Stavanger, Norway) in the morning and live artemia in the afternoon. Fertilization was performed by natural spawning at the beginning of the light period. Eggs were collected and embryos and larvae were reared from 0 to 5 days postfertilization in a 28.5 °C incubator in 10-cm Petri dishes (Sarstedt, Nümbrecht, Germany) at a density of 100 embryos/dish in E3 embryo media (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO₄ + 0.1 ppm methylene blue antifungal agent). Embryos and larvae were raised on a 14 hours:10 hours light:dark cycle, and 50% of the embryo media was refreshed daily. The following lines were used in this work: Tg(mpx:GFPⁱ¹¹⁴/mpeg1.1:mCherry-F^umsF001), Tg(mpeg1.1:mCherry-F^umsF001/tnfa:GFP-F^ump5), homozygous MR mutants (nr3c2^inr11/^inr11), and WT siblings (nr3c2^+/+).

Live Imaging

Zebrafish larvae from 50 to 72 hpf were imaged as previously described (27). Briefly, larvae were mounted in 2% low melting agarose containing 0.2 mg/mL MS-222 (Sigma-Aldrich, St. Louis, MO, USA) as a sedative/anesthetic in E3 medium. The imaging dish was then filled with E3 media with 0.2 mg/mL MS-222 to prevent dehydration. Confocal imaging was performed using a Leica Stellaris 5 (Leica Microsystems, Wetzlar, Germany). Imaging of the GFP signal was performed using the 488-nm laser and imaging of the mCherry signal was performed using the 561-nm laser. Analysis of imaging data was performed using ImageJ (FIJI).

Pharmacological Treatments, Tail fin Amputation, and Leukocyte Recruitment

To differentially activate GR and MR, embryos were treated with either an agonist (cortisol; 5 µg/mL [14 µM]) or an antagonist (GR: mifepristone [1.25 µM]; MR: eplerenone [1.25 µM]) 2 hours before tail fin amputation (all purchased from Sigma-Aldrich). At 2 days postfertilization, embryos were anesthetized with 0.168 g/L buffered MS222 (2:1 sodium bicarbonate; Sigma-Aldrich). Embryos were transferred to a 2% agarose-coated Petri dish and the tail fin fold was amputated just posterior to the notochord using a 1-mm sapphire blade (World Precision Instruments, Sarasota, FL, USA). Larvae were allowed to recover, in the appropriate aforementioned treatments, for 4 hours after amputation and were subsequently euthanized (using 0.4 g/L MS222) and fixed in 4% paraformaldehyde (PFA; Sigma-Aldrich) overnight at 4 °C. The fluorescently labeled macrophages and neutrophils were visualized using a LeicaMZ16FA fluorescence stereomicroscope supported by LAS 3.7 software. The numbers of macrophages and neutrophils were quantified by manual counting after blinding of the samples.

Fluorescence-activated Cell Sorting of Macrophages

Macrophages were sorted from Tg(mpeg1.1:mCherry-F^umsF001) embryos as previously described (27). Briefly, dissociation was performed with 150 embryos for each sample using Liberase TL (Roche, Rotkreuz, Switzerland). Samples were allowed to dissociate for 20 minutes, and the reaction was stopped with fetal calf serum (10% final concentration). Isolated cells were suspended in Dulbecco's PBS and filtered through a 40-µm cell strainer. Actinomycin D (Sigma-Aldrich, Netherlands) was added to each step (1 µg/mL final concentration) to inhibit transcription. Macrophages were sorted based on their red fluorescent signal using a fluorescence-activated cell sorting (FACS) Aria II cell sorter (BD Biosciences, San Jose, CA, USA). The sorted cells were collected in 1x PBS and stored at −80 °C until further analysis.

Quantitative Real-time PCR

Transcript levels of specific genes were measured by quantitative real-time PCR (qPCR). Total RNA was extracted from larvae using Trizol reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions and quantified using a Nanodrop (Thermo Fisher Scientific). RNA (1 μg) was treated with DNase I (Thermo Fisher Scientific) and converted to cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific), according to the manufacturer's protocols. Transcript levels were measured by qPCR using SYBR green (Sso Advanced, Bio-Rad) in duplicate using gene-specific primers as described previously (Table 1).

Table 1.

Gene-specific primers

Gene	Forward (5′-3′)	Reverse (5′-3′)	Reference
ccl2	GTCTGGTGCTCTTCGCTTTC	TGCAGAGAAGATGCGTCGTA	(27)
ccr2	TGGCAACGCAAAGGCTTTCAGTGA	AGGTTTCCCGAAGGTGAAGT	(27)
cxcl11aa	ACTCAACATGGTGAAGCCAGTGCT	CTTCAGCGTGGCTATGACTTCCAT	(27)
cxc3.2	CTGGAGCTTTGTTCTCGCTGAATG	CACGATGACTAAGGAGATGATGAGCCC	(32)
Il1b	TGTGTGTTTGGGAATCTCCA	CTGATAAACCAACCGGGACA	(27)
tnfa	ACCAGGCCTTTTCTTCAGGT	TTTGCCTCCGTAGGATTCAG	(27)
actab2	CGAGCAGGAGATGGAACC	CAACGGAAACGCTCATTGC	(33)

Macrophage Proliferation and Cell Death

To identify newly formed macrophages, embryos were pulse-labelled with the thymidine analogue ethynyl deoxyuridine (EdU; Molecular Probes, Eugene, OR, USA) immediately after tail fin amputation to label all newly formed macrophages. To visualize macrophages, we used anti-Mfap4 (GeneTex; Irvine, CA, USA; Cat# GTX132692, RRID:AB_2886714), which has previously been used in zebrafish (34). Briefly, embryos were either treated immediately after amputation for 4 hours, or from 24 to 48 hpf with 400 µM EdU in a 48-well plate. To evaluate macrophage proliferation, embryos were sacrificed by an overdose of MS222 (0.4 mg/mL) and fixed in 4% PFA overnight under continuous shaking. Macrophages were detected using immunohistochemistry for microfibril-associated glycoprotein 4 (mfap4), a macrophage-specific marker (35). To evaluate macrophage cell death, embryos (48 hpf) were sacrificed by an overdose of MS222 (0.4 mg/mL) and fixed in 4% PFA overnight under continuous shaking. This was followed by immunohistochemistry for Mfap4. Cell death was subsequently evaluated using terminal deoxynucleotidyl transferase-mediated dUTP nick end labelling (TUNEL) using the In Situ Cell Death Detection Kit, TMR red (Roche Life Sciences, Zurich, Switzerland), as described previously (36).

Immunohistochemistry

After fixation samples were washed (1 × 5 minutes) in 1x PBS + 0.1% Triton X (PBS-T). Samples were then dehydrated in a methanol series (25%, 50%, 75%, 100%) and stored overnight at −20 °C. The next day, samples were rehydrated in the reverse methanol series and permeabilized with 100% acetone at −20 °C for 12 minutes, followed by treatment with 10 µg/mL proteinase K (Roche) for 20 minutes. Samples were washed 3x with 1x PBS-T and fixed with 4% PFA for 20 minutes. After fixation, samples were washed 5 × 5 minutes with 1x PBS-T and then blocked for 2 hours in blocking buffer (10% BSA, 10% goat serum, 1% DMSO, 0.1% Triton X, and 0.1% Tween-20; all reagents acquired from Sigma-Aldrich). Larvae were then incubated overnight at 4 °C overnight with primary Mfap4 antibody (1:300). The next day, samples were washed 5 × 5 minutes in PBS-T and incubated for 2 hours with secondary antibody (anti-rabbit, Alexa Fluor 488 (1:400; Thermo Fisher Scientific Cat# A-11001, RRID:AB_2534069). To visualize the EdU incorporation, a Click-iT reaction was performed according to the manufacturer's directions (Thermo Fisher Scientific). Briefly, the master mix was assembled, and larvae were incubated for 1 hour at room temperature. Larvae were then washed with 3 × 10 minutes with PBS-T. Finally, samples were stained with 4′,6-diamidino-2-phenylindole (DAPI; 12.5 µg/mL (1:400), 5 minutes; Molecular Probes). Mfap4 (Alexa Fluor 488), EdU (Alexa Fluor 555), and DAPI signals were visualized using a Leica TCS SP8 or Stellaris 5 confocal microscope.

Macrophage Recruitment to Hindbrain Ventricle

To determine chemokine responsivity, we either injected 1 nL of CCL2 protein (PeproTech; Thermo Fisher Scientific, UK, 100 nM) into the hindbrain ventricle of Tg(mpeg1.1:mCherry-F^umsF001) larvae at 48 hpf, or 1 nL of 1x PBS as a vehicle control. This protocol has previously been used in zebrafish (32). After 4 hours, larvae were fixed in 4% PFA. Samples were then visualized under a Leica TCS SP8 confocal microscope, and the number of macrophages that had migrated to the hindbrain ventricle was determined by going through the z-stack images.

Statistics

Statistical analyses were performed using GraphPad Prism 6. Data were log-transformed where necessary to meet the assumptions of normality and equal variance. Untransformed data are shown in all graphs. Unless otherwise stated, bars show means ± SEM of data pooled from three different experiments, with at least 3 biological replicates per experiment. Significance was accepted at P < .05 and different significance levels are indicated: *P < .05; **P < .01; ***P < .001; ****P < .0001.

Results

MR Deficiency Results in a Global Depletion of Macrophages

To visualize the impact of MR deficiency on leukocyte function in zebrafish embryos and larvae, we first created Tg(mpeg1.1:mCherry,mpx:eGFP);nr3c2^+/+ (WT) or Tg(mpeg1.1:mCherry,mpx:eGFP);nr3c2^{−inr11/inr11} (MRKO) lines. In these larvae, macrophages are observed by visualization of the mCherry signal (Fig. 1A; magenta) and neutrophils by the GFP signal (Fig. 1A; green). Quantification of whole-body macrophage counts (excluding the yolk) revealed that primitive myelopoiesis, which begins around 12 to 24 hpf, was initiated normally because there was no observable difference in total numbers of macrophages between WT and MRKO larvae whole body (Fig. 1B). However, by 48 hpf, MRKO larvae had one-fifth of the number of macrophages observed in WT fish (21.9 ± 5.8 vs 104 ± 4.8; P < .001). By 72 hpf, MRKO larvae had increased the number of macrophages approximately 9-fold (194 ± 46.4) compared with the number at 48 hpf, but this was still about half the number of macrophages that populated WT larvae at 72 hpf (405 ± 46.4; P < .01). At 96 hpf, MRKO larvae still had about half the number of macrophages (127 ± 46.4) compared with WT larvae at the same age (405 ± 46.4; P < .05). In contrast, the number of neutrophils was similar between WT and MRKO fish at all developmental time points studied (Fig. 1C).

Figure 1.

A loss of MR results in global macrophage depletion throughout development. (A) Representative images of either Tg(mpeg:mCherry/mpx:eGFP);nr3c2^+/+ (WT) or Tg(mpeg:mCherry/mpx:eGFP);nr3c2^inr11/inr11 (MRKO), at 48 hours postfertilization (hpf). The macrophages are visualized by the mCherry signal (magenta) and the neutrophils by the GFP signal (green). Scale bar: 200 µm. (B) The number of mpeg+ macrophages in the whole body (excluding the yolk) at 24, 48, 72, and 96 hpf (n = 5-15). (C) The number of mpx+ neutrophils in the whole body (excluding the yolk) at 24, 48, 72, and 96 hpf (n = 5-15). (D) The number of macrophages (mpeg+ cells) in the caudal hemopoietic tissue (CHT) of larvae at 48 hours postfertilization (hpf) (n = 5-9). (E) The number of mpeg+ macrophages in the caudal hemopoietic tissue (CHT) region of WT, MRKO, and MRKO injected with a zfMR expression vector at 48 hpf (n = 9-12). (F) Representative images of the zebrafish CHT stained with TMR-Red (magenta; TUNEL, cell death), Alexa 488 (green; anti-Mfap4, macrophages) and a merged image of TUNEL and Mfap4 signals (white arrows denote colocalization). (G) The number of Mfap4-positive cells in the CHT. (H) The number of TUNEL-positive cells in the CHT. (I) The percentage of TUNEL-positive macrophages in the CHT. Bars show mean ± SEM of data pooled from 3 different experiments (B-E, G-I; each data point representing a single embryo/larva). Data were analyzed using a 2-way ANOVA (B-D; Holm-Sidak post hoc test), a 1-way ANOVA (E), or a t-test (G-I). Statistical significance is indicated by *P ≤ .05, **P ≤ .01, ***P ≤ .001, ****P ≤ .0001.

The difference in macrophage number were particularly evident in the caudal hemopoietic tissue (CHT) region (Fig. 1A [white boxes]; Fig. 1D). The CHT region is the embryonic precursor to the adult kidney in fish and is considered the primary source of leukocytes that are recruited at the site of inflammation. Indeed, reflective of whole-body macrophage numbers, at 48 hpf, the number of macrophages in the CHT of WT larvae (29.1 ± 2.4) was 12-fold greater compared with macrophage numbers in CHT of MRKO (2.3 ± 1.2; P < .0001). However, by 72 hpf, this difference was only 2-fold (WT 22.3 ± 1.2 vs MRKO 13.5 ± 2.4; P = .003), and there was no difference in macrophage numbers in the CHT at 96 hpf (Fig. 1D), suggesting this phenotype is specific to early developmental periods.

To confirm that this decrease in macrophage numbers was specific for a loss of MR, we examined whether recovery of MR levels might rescue the macrophage phenotype. For this purpose, we injected a zebrafish MR expression vector (zfMR) into the larvae at the 1-cell stage (Supplementary Fig. S1A (37)), and the number of macrophages in the CHT region was quantified in larvae at 48 hpf, the time point where the greatest reduction in macrophages was observed (Fig. 1D). At 48 hpf, MRKO embryos had less than half the population of CHT macrophages (17 ± 6.3) compared with WT embryos (47 ± 7.1; P < .01), and this phenotype was rescued by injection of a zfMR construct (33.6 ± 5.1; Fig. 1E).

To determine whether the decrease in macrophage number in the CHT of MRKO larvae was due to decreased proliferation and/or an increase in cell death in this region, we next examined macrophage proliferation via EdU incorporation over 24 hours (24-48 hpf; Supplementary Fig. S1B-S1E (37)), and cell death via terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) (Fig. 1F-1I). When assessing EdU and TUNEL incorporation, we used Mfap4 as a marker of macrophages. This is in contrast to the transgenic zebrafish lines that use the mpeg promoter to drive mCherry expression (Fig. 1A-1E). Using Mfap4 as a marker, we first confirmed a reduction in the number of macrophages detected in the CHT region in MRKO compared with WT larvae (Fig. 1G, Supplementary Fig. S1C (37)).

When we determined the number of EdU-positive macrophages in the CHT region, we observed that nearly all macrophages present at 48 hpf in both the WT and MRKO larvae had failed to incorporate EdU, indicating that they had been formed before the addition of EdU at 24 hpf (Supplementary Fig. S1B-S1E (37)). These data are supported by a lack of change at this time point in expression levels of mpeg1.1, a gene encoding a specific macrophage marker (Supplementary Fig. S1F (37)) (38), pu.1/spi1, encoding a key transcription factor involved in leukocyte production (Supplementary Fig. S1G (37)) (39), and irf8, encoding a transcription factor involved in macrophage production (Supplementary Fig. S1H (37)) (40).

Subsequently, we assessed the number of apoptotic cells at 48 hpf in the CHT using TUNEL staining. We found that the total number of TUNEL-positive cells was not different between WT and MRKO larvae (Fig. 1H), but that the percentage of TUNEL-positive macrophages in MRKO larvae (13 ± 1.7%) was ∼3-fold larger compared with the level of TUNEL-positive macrophages detected in WT larvae (4.5 ± 2.7%; P = .04; Supplementary Fig. S1I (37)). This suggests that macrophages lacking MR are more susceptible to cell death, which may contribute to the reduced number of macrophages observed in the CHT in these larvae.

A Loss of MR Disrupts Both Production and Distribution of Macrophages During Development

During development, macrophages first proliferate in the rostral blood island of the yolk (41), and indeed, at 48 hpf, EdU staining revealed little to no proliferation in the CHT region (Supplementary Fig. S1E-SG (37)), suggesting that these macrophages had originated elsewhere. Therefore, we next examined whether macrophage production and distribution from the yolk was impacted by a loss of MR. Using FACS-isolated macrophages, we analyzed the transcript abundance of colony-stimulating factor 1 receptor (csf1ra), a key receptor essential for macrophage distribution during development. Indeed, there was approximately a 10-fold decrease in transcript abundance of csf1ra in MRKO macrophages compared to WT macrophages (Fig. 2A).

Figure 2.

A loss of MR impacts macrophage production and distribution during development. (A) Transcript abundance of csf1ra in FACS-sorted macrophages from WT and MRKO larvae at 48 hours postfertilization (hpf), determined by qPCR (n = 3). (B) Number of mpeg+ macrophages in the yolk at 30, 50, and 72 hpf in WT and MRKO embryos/larvae at different stages of development (n = 3-5). (C) Representative images of mpeg+ macrophage (magenta) populations within the yolk of Tg(mpeg:mCherry);nr3c2^+/+ (WT) or Tg(mpeg:mCherry);nr3c2^inr11/inr11 (MRKO) larvae at 30, 54, and 72 hpf. Scale bar: 100 µm. Bars show mean ± SEM of data pooled from 3 different experiments (B; each data point representing a single embryo/larva). Data were analyzed using a 2-way ANOVA (Holm-Sidak post hoc test). Statistical significance is indicated by *P ≤ .05, **P ≤ .01.

Previous work has established essential roles for colony-stimulating factor (Csf) receptors in the distribution and differentiation of macrophages during zebrafish development, in particular exit from the yolk after initial propagation in the rostral blood island (42). Therefore, we quantified the number of yolk macrophages to determine if the macrophages lacking MR were sequestered there. We hypothesized that if distribution and not production was impacted, we would observe equal numbers of macrophages in the yolk. However, at 30 hpf, there was a 2-fold reduction in the number of macrophages in the yolk of MRKO larvae compared with WT counterparts (P > .01), suggesting a reduction in macrophage proliferation and/or differentiation at this stage. In contrast, at 50 and 72 hpf, equal numbers of yolk macrophages were observed between WT and MRKO larvae (Fig. 2B). Imaging revealed that although WT macrophages exit the yolk between 50 and 72 hpf, macrophages in the MRKO larvae remained primarily in the yolk (Fig. 2C).

Macrophages Lacking MR Retain Their Migratory Ability in Response to Local Inflammation

To further investigate the migratory capacity of macrophages in MRKO larvae, we used the tail fin amputation model and examined the migration of macrophages toward the wound site at 4 and 24 hours postamputation (hpa; Fig. 3A and 3B). The number of macrophages that were recruited to the wound site was similar between WT and MRKO fish at 4 hpa (WT, 10 ± 1.0 vs MRKO, 14.6 ± 1.3) (Fig. 3C). This could suggest a wounding-induced increase in macrophage number, a process known as emergency myelopoiesis; therefore, we also quantified macrophages that populated the CHT region. However, at 4 hpa, the number of macrophages in the CHT region of nonamputated (sham) MRKO embryos (6.4 ± 1.8) was 5-fold less compared with sham WT embryos (28.8 ± 2.4), and this remained unchanged with amputation (4 hpa; Fig. 3D). This suggested that the migrated macrophages may have been sourced from increased mobilization of macrophages from elsewhere in the body of the MRKO larvae.

Figure 3.

Macrophages lacking MR are still responsive to an inflammatory stimulus. (A) Schematic diagram of tail fin amputation. Distinct areas of quantification are indicated. (B) Representative images of sham and amputated (AMP) Tg(mpeg:mCherry/mpx:eGFP);nr3c2^inr11/inr11 (MRKO) at 72 hours postfertilization (hpf); 24 hours postamputation (hpa). In the sham larvae, macrophages are sequestered in the rostral blood island (white horizontal arrow) and in AMP larvae macrophages have migrated toward the CHT and wound site (white vertical arrow). (C) Number of mpeg+ macrophages in the caudal hematopoietic tissue (CHT) of 52 hpf sham and AMP larvae (4 hpa; n = 10-11)). (D) Number of mpeg+ macrophages that have migrated toward the wound site (200 µm from point of amputation) at 4 hpa (n = 10-14). (E) Number of mpeg+ macrophages in the CHT of 72 hpf sham and AMP larvae (24 hpa; n = 9-12)). (F) Number of mpeg+ macrophages that have migrated toward the wound site at 24 hpa (n = 8-10). (G) Number of mpx+ neutrophils in the CHT of 52 hpf sham and AMP larvae (4 hpa; n = 8-9)). (H) Number pf mpx+ neutrophils that have migrated toward the wound site at 4 hpa (n = 7-9). (I) Number of mpx+ neutrophils in the CHT of 72 hpf sham and AMP larvae (24 hpa; n = 6-11)). (J) Number of mpx+ neutrophils that have migrated toward the wound site at 24 hpa (n = 5-8). Bars show mean ± SEM of data pooled from 3 different experiments (each data point representing a single larva). Data were analyzed using a 2-way ANOVA (D, F, H, J; Holm-Sidak post hoc test), or a t-test (C, E, G, I); Statistical significance is indicated by *P ≤ .05, **P ≤ .01, ***P ≤ .001.

At 24 hpa, the number of macrophages localized at the site of amputation was still similar between WT and MRKO larvae (WT, 12.25 ± 3.0 vs MRKO, 11.7 ± 3.1) (Fig. 3E). When we examined the number of macrophages in the CHT region at 24 hpa (Fig. 3F), amputated WT larvae showed double the number of macrophages compared with age-matched WT sham larvae, indicative of emergency myelopoiesis in WT larvae at this later time point. Similarly, amputated MRKO larvae also demonstrated emergency myelopoiesis, reflected by an approximately 4-fold increase in the number of macrophages that populated the CHT compared with sham controls; however, their macrophage numbers were still lower than those in WT larvae at this time point (Fig. 3F).

We next examined the number of neutrophils at the wound site in WT and MRKO larvae at 4 hpa. The MRKO larvae showed double the number of neutrophils localized at the wound site at this time point (MRKO, 24 ± 1.8 vs WT, 14.9 ± 1.5; P < .001) (Fig. 3G). Although the sham MRKO embryos showed a reduced number of neutrophils in the CHT region compared with sham WT embryos (WT, 34 ± 2.3 vs MRKO, 24 ± 1.0), they showed similar neutrophil numbers as sham WT embryos by 4 hpa (WT, 23.5 ± 2.0 vs MRKO, 23.7 ± 1.3) (Fig. 3H). By 24 hpa the MRKO larvae experienced a marked reduction in the number of neutrophils localized at the wound compared with the number of neutrophils observed in the WT larvae at 24 hpa (WT, 10.4 ± 1.5 vs MRKO, 2.5 ± 0.6; Fig. 3I). However, there was no difference in the number of neutrophils in the CHT either between the sham or amputated WT and MRKO groups (Fig. 3J).

Production of New Macrophages is Reduced in MRKO Larvae, and Macrophages Migrating Toward the Wound Exhibit a Less Inflammatory Phenotype

To confirm that MRKO macrophages were migrating toward the wound, and not the result of rapid, de novo differentiation and/or proliferation, we labeled newly formed macrophages with the thymidine nucleoside analogue, EdU, and used an anti-Mfap4 antibody to label the endogenous macrophage population. We first confirmed that Mfap4+ macrophages constitute a similar population as the mCherry-labeled macrophages in the Tg(mpeg1.1:mCherry) line (Fig. 4A and 4B). The number of Mfap4+ macrophages that migrated toward the wound by 4 hpa was similar between WT and MRKO (Fig. 4B), although the absolute number of Mfap4+ macrophages was lower compared with the number of migrated macrophages observed in the Tg(mpeg1.1:mCherry) line (Fig. 3C). When we examined the percentage of new macrophages in this migrated population, we discovered that only 5% of the macrophages in the MRKO embryos were EdU-positive, whereas 25% of the macrophages in WT embryos were EdU-positive (Fig. 4C). These results indicate that the number of newly produced cells in the migrated macrophage population is actually reduced in the MRKO larvae.

Figure 4.

The generation of new macrophages is reduced in MRKO larvae, and macrophages migrating toward the wound exhibit a less inflammatory phenotype. (A) Representative images of macrophage proliferation at 48 hours postfertilization (hpf), 4 hours postamputation (hpa), where the nuclei are stained with DAPI (blue), new cells have incorporated the thymidine nucleoside analogue EdU (red), and macrophages are visualized by immunohistochemistry using a Mfap4 antibody (green). White dotted lines denote the location of amputation. (B) Number of Mfap4+ macrophages localized to the wound site at 4 hpa (n = 7). (C) The percentage of new (EdU+) macrophages localized to the wound site at 4 hpa (n = 7). (D and E) Transcript abundance of macrophage-specific chemokines cxcl11aa (D) and ccl2 (E), determined by qPCR analysis of RNA isolated from larvae at 4 hpa (n = 4). (F) Circularity index of macrophages that had migrated towards the wound at 2 hpa. (G) Transcript abundance of il1b, cxcr3.2, and ccr2 in FACS-sorted macrophages determined by qPCR (n = 3). (H) Number of mpeg+ macrophages that have migrated to the hindbrain 4 hours after CCL2 injection at 48 hpf (n = 6-9). (I) Representative images of WT and MRKO Tg(mpeg:mCherry) at 48 hpf, injected with either PBS or recombinant human CCL2 in the hindbrain ventricle. Bars show mean ± SEM of data pooled from 3 different experiments (B,C,H; each data point representing a single larva) or mean ± SEM of 3 individual experiments. Data were analyzed using a t-test (B, C, G) or a 2-way ANOVA (D, E, H; Holm-Sidak post hoc test). Statistical significance is indicated by *P ≤ .05, **P ≤ .01, ***P ≤ .001.

Because we used a ubiquitous knockout, the MRKO larvae also lack MR in key tissues involved in the wounding response. To study whether the chemokine release from the inflamed tissue of the wounded tail fin fold is altered as a result of MR deficiency, we measured the expression levels of two genes encoding key chemokines (Cxcl11aa and Ccl2) involved in macrophage migration toward the wound. On amputation, both cxcl11aa and ccl2 mRNA levels were increased in WT larvae, but there was no change in the MRKO group on amputation (Fig. 4D and 4E).

We next analyzed how the loss of MR influenced the differentiation of the macrophages toward a pro-inflammatory (M1) phenotype. For this purpose, we first studied the morphology of the macrophages because this has previously been shown to be an indicator of differentiation, with pro-inflammatory M1 macrophages displaying a higher level of circularity compared with anti-inflammatory M2 macrophages (27). At 2 hpa, 30% of WT macrophages had a circularity index of 0.6 (Figure 4F), which is consistent with what has been reported previously (27). In contrast, macrophages from the MRKO larvae showed a reduced circularity index (30% of the macrophage population had a circularity index of 0.4). To confirm that MRKO macrophages exhibited a decreased differentiation toward an M1 phenotype, we measured the expression levels of the pro-inflammatory gene il1b in FACS-sorted macrophages. MRKO macrophages showed a reduced expression of the gene encoding this cytokine (Fig. 4G), in line with a reduced differentiation of macrophages to an M1 phenotype in MRKO larvae. Further analysis of MRKO macrophages also revealed a higher transcript abundance of genes for 2 key receptors involved in macrophage chemotaxis: cxcr3.2 (encoding the receptor of Cxcl11aa; P < .01) and ccr2 (encoding the receptor of Ccl2; P < .01) (Fig. 4G).

These results suggest increased responsivity of MRKO macrophages to chemokines such as Ccl2. To study this responsivity, we injected recombinant human CCL2 into the hindbrain of the zebrafish, a location that under basal conditions hardly contains any macrophages at this time point. Ccl2 injection doubled the number of macrophages present in the hindbrain in WT larvae at 4 hours after injection (WT [PBS]: 3 ± 0.4 vs WT [CCL2]: 7.2 ± 1.6; P < .01; Fig. 4H and 4I). The response to CCL2 injection was nearly 2-fold larger in MRKO larvae compared with WT (WT [CCL2]: 7.2 ± 1.6 vs MRKO: 11.3 ± 2.0; P < .01). This increased sensitivity of macrophages to chemoattractants such as Ccl2 may explain the intact migration toward the wound site that we observed in the MRKO larvae, despite the reduced expression levels of cxcl11aa and ccl2 in the MRKO macrophages.

Treatment With the MR Antagonist Eplerenone Phenocopies the Macrophage Depletion Observed in MRKO Larvae

To further confirm the specificity of MR during myelopoiesis, we used the MRA, eplerenone, to block MR activation. Indeed, chronic eplerenone treatment from 2 hpf onward, at either 1.25 µM or 12.5 µM, phenocopied the global reduction in macrophage observed in the MRKO larvae (Fig. 5A). Similar to what we had observed in the MRKO larvae, chronic eplerenone treatment (1.25 µM) did not affect macrophage recruitment at the wound site after tail fin amputation (Fig. 5B).

Figure 5.

Pharmacological blockade of MR results in a reduced macrophage migration toward a wound. (A) Number of mpeg+ macrophages in the whole body of zebrafish at 52 hpf after incubation with either 1.25 μM or 12.5 μM eplerenone for 50 hours (n = 8-12). (B) Number of mpeg+ macrophages that have migrated toward the wound site (200 µm from point of amputation) at 4 hpa in zebrafish exposed to either vehicle or the MR antagonist eplerenone (1.25 μM) for 50 hours (n = 16). (C) Number of macrophages that have migrated toward the wound site, at 4 hpa, in zebrafish exposed to either eplerenone (1.25 μM) or the GR antagonist mifepristone (RU486 (1.25 μM) (n = 12-21). (D) Representative images of Tg(mpeg:mCherry) zebrafish at 48 hpf, treated with either vehicle, eplerenone (1.25 µM) or mifepristone (1.25 µM) for 6 hours. Bars show mean ± SEM of data pooled from three different experiments (A-C; each data point representing a single larva). Data were analyzed using a 1-way ANOVA (A,C) or a t-test (B). Statistical significance is indicated by *P ≤ .05.

After studying the effect of MR deficiency on macrophages during the development of the immune system, we next wanted to determine whether a loss of MR had the same effect on macrophages after they had differentiated regularly in the presence of MR. Therefore, we next allowed myelopoiesis to proceed normally in WT larvae until 48 hpf (definitive hematopoiesis (31)). At this time point, endogenous cortisol levels cannot be increased because of hyporesponsiveness of the hypothalamus-pituitary-interrenal axis (21, 43, 44). Indeed, there was no effect of amputation on endogenous cortisol levels at this stage of development (Supplementary Fig. S2A (37)). To study the acute effects of MR deficiency on differentiated macrophages, eplerenone was added 2 hours before the tail fin amputation. In addition, the GR antagonist (GRA) mifepristone was added to inactivate the GR for comparison. After the tail fin amputation, we allowed the fish to recover for 4 hours and then assessed the number of macrophages that had migrated toward the wound. Neither mifepristone nor eplerenone altered endogenous cortisol levels (Supplementary Fig. S2B (37)), and the antagonistic effect of these drugs was validated using gene readouts for MR- and GR-specific target genes (Supplementary Fig. S2C and S2D (37)). Treatment with eplerenone reduced the number of macrophages that migrated toward the wound by half (9.0 ± 1.1; P < .05; Fig. 5C) compared with the number of macrophages in vehicle-treated larvae (14.6 ± 1.1), whereas mifepristone did not alter the number of macrophages localized to the wound site (16.3 ± 1.1). This was surprising because both chronic eplerenone treatment (Fig. 5B) and the MRKO model (Fig. 3D) displayed no difference in macrophage localization at the wound site. This suggests that the effect of MR antagonism on differentiated macrophages is distinct from the effects observed after chronic MR inactivation.

The Eplerenone-induced Reduction in Macrophage Migration Toward Inflammation is Associated With Decreased Macrophage Responsivity

To determine whether the lack of macrophage migration in eplerenone-treated macrophages was due to a loss of responsivity at the level of the macrophage or to reduced signalling in the wounded tissue, we first measured expression levels of ccl2 and cxcl11aa. Amputation increased the transcript levels of both ccl2 (Fig. 6A) and cxcl11aa (Fig. 6B). However, larvae treated with eplerenone before amputation experienced a 4-fold increase in the transcript abundance of ccl2 compared with that of amputated vehicle-treated larvae (Fig. 6A). In the presence of mifepristone, the increase in ccl2 expression after amputation was similar to that of vehicle-treated larvae.

Figure 6.

The anti-inflammatory effect of eplerenone is due to altered responsivity at the level of the macrophage. (A and B) Transcript abundance (determined by qPCR) of genes encoding the macrophage-specific chemo-attractants ccl2 (A) or cxcl11aa (B) in sham and AMP zebrafish larvae exposed to either eplerenone (1.25 µM) or mifepristone (1.25 µM) (n = 4). (C) Transcript abundance of the gene encoding the pro-inflammatory cytokine interleukin 1β (il1b) and the chemokine receptors ccr2 and cxcr3.2 in FACS-sorted macrophages (n.d., not detectable) (n = 3). (D) Number of mpeg+ macrophages that have migrated to the hindbrain 4 hours after Ccl2 injection at 48 hpf (n = 8-13). (E) Representative images of Tg(mpeg:mCherry) at 48 hpf, treated with either vehicle or eplerenone (1.25 µM) and injected with either PBS or 100 nM of recombinant Ccl2. Bars show the mean ± SEM of 3 individual experiments. Data were analyzed using a 1-way ANOVA (A-C) or a 2-way ANOVA (D). Statistical significance is indicated by *P ≤ .05, ****P ≤ .0001.

The high levels of ccl2 and cxcl11aa expression did not agree with the reduction in macrophages at the wound site (Fig. 5C and 5D). Therefore, we postulated that it was due to a lack of responsivity at the level of the macrophage. Indeed, mRNA levels of ccr2, the gene encoding the receptor for Ccl2, were not detectable in macrophages isolated from eplerenone-treated larvae (Fig. 6C). Interestingly, the mRNA levels of cxcr3.2, the gene encoding the Cxcl11aa receptor, was unchanged (Fig. 6C), suggesting that a loss of MR specifically affects the Ccl2-Ccr2 signalling pathway (Fig. 6C). To confirm the deficiency in Ccl2-Ccr2 signalling, we injected recombinant human CCL2 into the hindbrain of the zebrafish. CCL2 injection promoted macrophage migration to the hindbrain in vehicle-treated larvae, but this response was abolished in the eplerenone-treated group (Fig. 6D and 6E).

Further validation was done using the expression of the tnfa gene as a readout. These experiments showed that the eplerenone-induced increase in the pro-inflammatory genes ccl2 and cxcl11aa was also observed for tnfa (Supplementary Fig. S3A and S3B (37)) and that this increase in tnfa expression was not due to increased expression of this gene in macrophages (Supplementary Fig. S3C-S3F (37)). Additionally, il1b transcript levels were shown to be similarly increased in macrophages of the amputated larvae after vehicle and eplerenone treatment (Fig. 6C).

Eplerenone Abolishes Cortisol-induced Changes in Macrophage Recruitment During Inflammation

To determine the role of MR in cortisol-induced changes in leukocyte trafficking (27), we first examined how increasing concentrations of cortisol would impact the migration of both macrophages (Fig. 7A) and neutrophils (Supplementary Fig. S4A (37)). When larvae were treated with 10 µg/mL (28 µM) of cortisol, the number of macrophages was reduced by ∼30% (vehicle, 15.8 ± 1.5 vs cortisol, 10.6 ± 0.7; P < .01). There was no effect of lower doses of cortisol on the number of macrophages that migrated toward the wound (Supplementary Fig. S3A (37)). The effect of cortisol on the migration of neutrophils was similar, with a 50% reduction (vehicle, 17.5 ± 1.9 vs cortisol, 9.3 ± 1.4) in the number of neutrophils that migrated toward the wound in the presence of 10 µg/mL of cortisol (Supplementary Fig. S4A (37)). There was no effect of lower cortisol doses on the number of migrated neutrophils.

Figure 7.

MR modulates the cortisol-induced reduction in macrophage migration toward a wound site. (A) Number of mpeg+ macrophages that have migrated toward the wound site (200 µm from amputation site), at 4 hours after amputation, in larvae exposed to different levels of cortisol ranging from doses that reflect physiological stress levels (100 ng/mL) to a pharmacological dose (10 μg/mL) (n = 11-30). (B) The number of mpeg+ macrophages that have migrated toward the wound site in the presence of cortisol (10 µg/mL), in combination with either eplerenone (1.25 µM) or mifepristone (1.25 µM) (n = 10-28). (C and D) Transcript abundance of genes encoding the macrophage-specific chemoattractants ccl2 (C) or cxcl11aa (D), determined by qPCR analysis of RNA isolated from zebrafish larvae exposed to cortisol (10 µg/mL) in combination with either eplerenone (1.25 µM) or mifepristone (1.25 µM) (n = 5-6). Bars show mean ± SEM of data pooled from three different experiments (A, B; each data point representing a single larva) or mean ± SEM of 3 individual experiments (C, D). Data were analyzed using a 1-way ANOVA. Statistical significance is indicated by *P ≤ .05, **P ≤ .01, ***P ≤ .001.

Subsequently, we wanted to determine whether this effect of cortisol was MR- and/or GR-mediated. Therefore, 30 minutes before cortisol treatment, 1.25 µM of either the GRA mifepristone or the MRA eplerenone was added. With the addition of cortisol, the number of macrophages was reduced by ∼30% (10.8 ± 0.9), and this effect was abolished in the presence of both mifepristone (15.5 ± 1.2) and eplerenone (14 ± 1.0), suggesting that both MR and GR are involved in this effect (Fig. 7B). Finally, we assessed whether MR accomplished this by modulating the levels of chemokines important for macrophage recruitment. However, neither treatment with cortisol nor with eplerenone or mifepristone altered the transcript abundance of ccl2 (Fig. 7C) or cxcl11aa (Fig. 7D) after amputation.

Discussion

Currently, little is known regarding the immunomodulatory role of MR, but recent studies and clinical data point to key roles for this receptor in modulating inflammation (5, 11). To determine the extent of MR signalling during inflammation, we established a transgenic zebrafish model to examine the impact that a ubiquitous MR knockout had on leukocyte dynamics. Surprisingly, we discovered that a loss of MR was associated with a global reduction in the number of macrophages because of an increase in cell death and reduced proliferation and distribution during early myelopoiesis, suggesting that MR plays an immune-protective role during development. Although reduced macrophage migration was associated with a loss of csf1ra expression, a receptor essential to macrophage distribution from the yolk in zebrafish (42, 45), migration of macrophages toward a site of inflammation was intact. This was likely because of increased levels of the chemokine receptors Ccr2 and Cxcr3.2 in MRKO macrophages. In contrast, when macrophages were allowed to develop normally and were exposed to an MRA, there was a reduction in macrophage migration toward an inflamed site, which could be explained by a reduced macrophage ccr2 expression. Taken together, our data reveal key roles for MR in modulating macrophage development and function, and that the mechanisms of action are context-specific.

The reduced number of macrophages in the MR knockout zebrafish is a novel observation in any model organism, and we subsequently wanted to delineate whether the lack of macrophages was due to reduced production or to restricted migration from the sites where they are produced. Myelopoiesis is a highly conserved process and is well-defined in both mammals and zebrafish. In zebrafish, primitive macrophages develop in the rostral blood island between the 3 and 5 somite stage (∼12 hpf) with eventual migration from the yolk toward circulation or the head (41). This is followed by definitive hematopoiesis, in which macrophages and neutrophils are produced in the posterior blood island (later the caudal hematopoietic tissue [CHT]) with functional macrophages and neutrophils detectable throughout the zebrafish body by 28 to 32 hpf (41). The molecular mechanisms that drive macrophage differentiation are also well characterized. Both macrophages and neutrophils are derived from a common myeloid progenitor, and their initial development is dependent on the transcription factor Pu.1, whereas the transcription factor interferon regulatory factor 8 (irf8) is essential for further differentiation of macrophages. In our study, there was no difference in EdU incorporation at 48 hpf, and transcript abundances of pu.1 and irf8 were comparable between WT and MRKO fish, suggesting that production of macrophages was not appreciably impacted.

In contrast, the increase in the percentage of TUNEL-positive macrophages in the CHT region of MRKO larvae suggests that MR signalling may play an immunoprotective, antiapoptotic role during macrophage development. Although the role of MR signalling in mediating immune cell death is largely unknown, MyMRKO mice show decreased accumulation of apoptotic cells in atherosclerotic lesions. However, this was attributed to increased efferocytosis, rather than modulation of apoptosis (17). The antiapoptotic role of MR has also been observed in neurons, where MR signalling plays a neuroprotective role (reviewed in (46)). In other cells types, such as those of the cardiac and renal systems, MR signalling is known to promote apoptosis (47, 48), suggesting the effects of MR on cell death may be tissue-specific. Taken together, the evidence presented here suggests an antiapoptotic role for MR in macrophages during early development.

Further examination also revealed changes in the migration of macrophages from the yolk, suggesting that the promotion of the macrophage population during development by MR is multifactorial. Macrophage migration from the yolk has been shown to be dependent on the colony-stimulating factor receptors and their ligands, most notably interleukin 34 (45). Zebrafish csf1r-knockout mutants have a persistent global macrophage deficiency (49, 50) and lack microglia (42, 45). Macrophages isolated from MRKO larvae had reduced expression of csf1ra, which could further explain the reduced macrophage number. These results indicate that MR activity promotes macrophage differentiation and distribution by enhancing the transcriptional activity of genes encoding key factors such as Csf1ra. Although we can speculate that the reduction in macrophage number in MRKO fish is likely developmental, the function of the innate immune system in adult animals was not evaluated here and future work should characterize the impact of MR on adult innate and adaptive immune systems.

When we assessed the migration of the macrophages to a site of inflammation in the MR-deficient larvae, we found no effect of MR knockout on their migratory capacity. This was unexpected given the lack of macrophages in the periphery and CHT of MRKO larvae, which are the primary areas from where macrophages migrate (51). We next assessed whether MR was simply producing more macrophages via emergency myelopoiesis, but we observed a marked reduction in the number of EdU-positive macrophages at the wound site. This suggests that there were fewer newly formed macrophages in MRKO compared with WT larvae. Although this lack of proliferation is consistent with what has been described in the MyMRKO mice (15), these results failed to clarify why we see macrophages being recruited toward the wound in MRKO larvae. As an alternative hypothesis, we postulated that macrophages lacking MR are more sensitive to the chemokine gradient from the wound. Indeed, in the macrophages of MRKO larvae, there was a marked increase in the expression of genes encoding receptors responsible for chemotaxis toward the wound (ccr2, cxcr3.2). Interestingly, the expression levels of genes encoding chemokines (cxcl11aa and ccl2) were not increased after amputation in the MRKO larvae. It has previously been established that transcript abundances of these chemokines increase after amputation in WT larvae and are representative of the chemokine gradient established on a localized wound (27), and indeed the WT larvae demonstrated a robust response after amputation. Taken together, these results suggest that the intact migration of macrophages toward a site of inflammation in MRKO larvae is due to an increased responsivity of these cells to inflammatory chemokine signals, which likely compensates for the lower macrophage numbers and chemokine gradient that are observed in the MR-deficient larvae.

The similar macrophage migration between MRKO and WT larvae is distinct from what has been reported in studies of mammalian models, in which a loss of MR prevents macrophage migration toward sites of inflammation (11). To determine if this may be due to a developmental effect, we subsequently used short-term treatment of WT larvae with the MRA eplerenone to investigate the role of MR in macrophages that had undergone normal development. This treatment resulted in a reduced migration of the macrophages to a site of inflammation, even though inflammation induced higher expression levels of chemokine genes (ccl2, cxcl1aa) in eplerenone-treated larvae compared with the vehicle-treated controls. Analysis of eplerenone-treated macrophages demonstrated that transcript abundance of ccr2 was below detection, which may explain the decreased macrophage migration, despite the increased chemokine levels. The MR-mediated regulation of ccr2 expression in macrophages appears to be a highly conserved effect. MyMRKO mouse models of Duchenne muscular dystrophy had reduced macrophage infiltration in the skeletal muscle and those macrophages that had infiltrated were negative for Ccr2 (16). A reduction in Ccr2 signalling was also observed in peritoneal macrophages of MyMRKO mice after femoral artery injury (52). To confirm whether the reduction in macrophage migration was due to a lack of responsivity to chemokines, we injected the Ccl2 protein, a Ccr2 ligand, and a known macrophage chemoattractant in both fish (32) and mammals (16), into the hindbrain of zebrafish. Indeed, eplerenone treatment reduced the number of migrating macrophages toward the injected Ccl2, confirming the role of the Ccr2-Ccl2 axis in macrophage migration. This suggests that modulation of the Ccr2-Ccl2 signalling pathway by MR in macrophages is conserved between fish and mammals.

Treatment with the MRA eplerenone additionally uncovered a dual role for MR in inflammation, under both resting and stressed conditions. As described previously, when MR was blocked under basal conditions by eplerenone, there was a reduction in the number of macrophages localized toward the wound, suggesting a permissive effect of this receptor on macrophage action. However, eplerenone also abolished the cortisol-induced reduction in migration (27), suggesting that MR is also necessary for the cortisol-induced reduction in macrophage migration, thereby revealing a suppressive effect of this receptor under this condition. Interestingly, a similar inhibition of the cortisol-induced reduction in macrophage migration was observed on treatment with the GRA mifepristone. This dual role of MR is not unexpected because it is well-established that the function of MR under basal levels differs from its role under high cortisol levels, leading to increased GR activation. This may result in increased MR-GR heterodimerization, which may underly different interactions of MR with hormone response elements (53). Our data also show the functional consequences of the combined MR-GR-induced macrophage migration because treatment with either the MRA or GRA abolished cortisol-induced reductions in tissue regeneration (Supplementary Fig. S5 (37)). In humans, it has been shown that the GC-induced delay in wound reepithelialization can be prevented with the application of an MRA (54), which was attributed to reduced inflammation at the wound site (55).

Overall, our results show important roles for MR in macrophage development and action (Fig. 8). First, we demonstrate that a loss of MR is associated with a reduced macrophage population during key developmental stages, which suggests that MR is required for the proper development of the macrophage cell population. Besides the effect of MR signalling on the size of the macrophage population, we also observed increased macrophage responsivity toward an inflammatory stimulus in MR-deficient individuals, suggesting MR plays a crucial immunomodulatory role during development. Second, we determined that MR blockade with an MRA reduced inflammation-induced macrophage migration, indicating that MR activation enhances macrophage responsiveness under basal conditions. However, we also established that MR, in combination with GR, is necessary for the cortisol-induced reduction in macrophage migration. The conservation of the molecular mechanisms involved (Ccr2-Ccl2), from fish to mammals, also points toward the importance of MR in modulating macrophage responsivity to inflammatory signals. Therefore, although GR signalling has predominated in our existing paradigm of GC-induced immunomodulation, the evidence presented here highlights the importance of MR activation in the innate immune system, and that this may have consequences in the development of new strategies for stress-related immune disorders and disease susceptibility.

Figure 8.

Summary schematic of the different roles of mineralocorticoid receptor (MR) in modulating the development and function of macrophages in zebrafish. A reduction in MR signalling in zebrafish, either through genetic perturbations or pharmacological antagonism, leads to altered macrophage responsivity to developmental and inflammatory signals, showing that MR is crucial for macrophage development and function. (A) During a key developmental period (0-120 hours postfertilization [hpf]), larvae lacking MR (MRKO) experience a global reduction in whole-body macrophage numbers. This is associated with a lack of production in the yolk and distribution out of the yolk, and these macrophages show reduced transcript abundance of the colony-stimulating factor receptor 1 (csfr1a) gene, which encodes a receptor critical for early macrophage migration out of the yolk. In addition, these macrophages are more susceptible to apoptotic cell death. (B) MRKO macrophages show increased responsivity to inflammatory signals (eg, after tail wounding) associated with increased levels of C-C chemokine receptor 2 (ccr2) gene expression. (C) When WT larvae are treated with the MR antagonist eplerenone between 2 and 120 hpf, the global reduction in macrophage number that is observed in MRKO larvae (panel A) is phenocopied. (D) Eplerenone treatment over 6 hours (short-term), results in a reduced responsivity of macrophages to inflammatory signals, which is associated with a reduction in ccr2 expression in macrophages. This is in sharp contrast to the increased responsivity that is observed in MRKO larvae (panel B). Image created using Biorender.com.

Abbreviations

CHT

caudal hemopoietic tissue

Csf

colony-stimulating factor

CVD

cardiovascular disease

DAPI

4′,6-diamidino-2-phenylindole

EdU

ethynyl deoxyuridine

FACS

fluorescence-activated cell sorting

GC

glucocorticoid

GR

glucocorticoid receptor

GRA

glucocorticoid receptor antagonist

hpa

hours postamputation

hpf

hours postfertilization

mfap4

microfibril-associated glycoprotein 4

MR

mineralocorticoid receptor

MRA

mineralocorticoid receptor antagonist

MRKO

Tg(mpeg:mCherry/mpx:eGFP);nr3c2^inr11/inr11

PFA

paraformaldehyde

PBS-T

PBS + 0.1% Triton X

qPCR

quantitative real-time PCR

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labelling

WT

wild-type

Funding

Marie Curie Fellowship to E.F. (MeRGeR; Grant agreement ID: 891367).

Disclosures

The authors have nothing to disclose.

Data Availability

The datasets generated during the current study are available from the corresponding author on reasonable request.