POMC Gene-Derived Peptides Activate Melanocortin Type 3 Receptor on Murine Macrophages, Suppress Cytokine Release, and Inhibit Neutrophil Migration in Acute Experimental Inflammation¹

More than 10 years ago the existence of an immunological network between pro-inflammatory cytokines and pituitary neuropeptides was identified as one of the most effective ways to control the host inflammatory response (1). Cytokines such as IL-1 release into the circulation adrenocorticotrophic hormone (ACTH),³ from the anterior pituitary, which stimulates secretion of corticosteroids (cortisol in human, corticosterone in rodents) with consequent down-regulation of the inflammatory response (2). The possibility that ACTH might be produced locally in extra-pituitary tissues has been suggested by several independent studies, which have detected either ACTH immunoreactivity or the product of the pro-opiomelanocortin (POMC) gene in peripheral organs and cells (3, 4, 5). It is unclear whether the POMC gene is expressed peripherally to release corticosteroids from the adrenal gland, or POMC-derived products produce their effects locally.

Treatment of patients with ACTH is a well-known but seldom used strategy for the clinical management of gouty arthritis. Besides being efficacious in patients who do not tolerate nonsteroidal anti-inflammatory drugs or colchicine (6), systemic treatment with ACTH was found to possess therapeutic efficacy over and above that attained with corticosteroids (7), strongly suggesting the existence of a mechanism of action distinct from that achieved by direct stimulation of the adrenal gland. Based on these clinical observations, and using a recently characterized murine model of experimental gout (8), we have sought to characterize the antimigratory profile of some POMC gene products. Only a single study has so far addressed the potential anti-inflammatory activity of these peptides, with the finding that ACTH_1–39 and a nonsteroidogenic fragment, ACTH_4–10, inhibited PGE₁ generation and edema formation in rat skin (9).

We have studied the effects of the heptapeptide ACTH_4–10, ACTH_1–39, α-MSH, and β-MSH in this model and observed novel pharmacological actions of the peptides, including inhibition of neutrophil (PMN) migration and chemokine generation. We have identified the peritoneal macrophage (Mφ) as the principal target cell and the melanocortin type 3 receptor (MC3-R) as the receptor responsible for transducing the observed effects.

Male Swiss Albino mice (20–22 g body weight) were purchased from Banton & Kingsman (T.O. strain; Hull, Humberside) and maintained on a standard chow pellet diet with tap water ad libitum using a 12-h light/dark cycle. Animals were used 3–4 days after arrival. Animal work was performed according to Home Office regulations (Guidance on the Operation of Animals, Scientific Procedures Act, 1986).

Mice received i.v. injections of 20 ng (4.4 pmol) ACTH, 100 μg (104 nmol) ACTH_4–10, or 30 μg (11.3 nmol) β-MSH 2 h before blood collection by cardiac puncture following terminal anesthesia. CCS and differential leukocyte counts in plasma aliquots were determined by RIA and light microscopy, respectively, as previously described (10).

Crystal-induced PMN recruitment was produced using a technique recently reported by our group (8). Briefly, mice were treated i.p. with 3 mg of monosodium urate (MSU) crystals in 0.5 ml PBS, and peritoneal cavities washed at different time points with 3 ml PBS containing 3 mM EDTA and 25 U/ml of heparin. Aliquots of the lavage fluids were then stained with Turk’s solution (0.01% crystal violet in 3% acetic acid), and differential cell counts were assessed using a Neubauer haemocytometer and a light microscope. Mononuclear cells and PMN were easily identified by their different morphology and nuclear staining. Data are reported as 10⁶ PMN per mouse. Lavage fluids were then centrifuged at 400 × g for 10 min, and supernatants were stored at −20°C for biochemical determinations (see below).

The peptides reported in Table I were used in this study. ACTH_4–10 (10–200 μg ranging from 10 to 208 nmol), α-MSH (3–30 μg ranging from 1.8 to 16.2 nmol), and β-MSH (1–30 μg ranging from 0.38 to 11.3 nmol) were administered s.c. 30 min before MSU crystals. The tetrapeptide HFRW (ACTH_6–9) was administered s.c. at a dose of 80 μg (10⁴ nmol) per mouse, equimolar to 100 μg ACTH_4–10. For experiments in vitro, full-length ACTH_1–39 was used at concentrations up to 100 ng/ml (22 nM). Different batches of these peptides were obtained from Sigma (Poole, Dorset, U.K.). A scrambled ACTH_4–10 peptide (sequence MGREWFH) was prepared by solid phase step-wise synthesis at The Advance Biotechnology Centre (The Charing Cross and Westminster Medical School, London, U.K.). Purity of this peptide was more than 90% as assessed by HPLC and capillary electrophoresis (data supplied by the manufacturer).

The MC3/4-R receptor agonist MTII (Ac-Nle⁴-c[Asp⁵,d-Phe⁷,Lys¹⁰]NH₂ ACTH_4–10) and MC3/4-R antagonist SHU9119 (Ac-Nle⁴-c[Asp⁵,d-²Nal⁷,Lys¹⁰]NH₂ ACTH_4–10) (11) were purchased from Bachem (Saffron Walden, Essex, U.K.) and Phoenix Pharmaceuticals (Mountain View, CA), respectively. MTII was given s.c. at 10 μg (9.6 nmol) per mouse, whereas the antagonist SHU9119 was administered i.p. (3–10 μg corresponding to 2.7–9 nmol) 30 min before MSU crystals. The compound S110, a generous gift of Prof. E. T. Wei (University of California, Berkeley), also called dynorphin A_[6–12] (p-methoxybenzoyl-Arg-Arg-Ile-Arg-Pro-Lys-d-Leu-NH₂), has been recently reported to show antagonistic activity at MC-R (12) and was given i.p. at the dose of 10 μg (9.33 nmol).

Murine KC, MIP-2, and TNF-α levels in the peritoneal lavage fluids were determined using commercially available ELISA kits purchased from R&D Systems (Abingdon, U.K.), whereas the murine IL-1β ELISA Cytoscreen was from BioSource International (Canarillo, CA). In brief, lavage fluids (50 μl) were assayed for each cytokine and compared with a standard curve constructed with 0–1 ng/ml of the standard cytokine. The ELISAs showed negligible (<1%) cross-reactivity with several other murine cytokines and chemokines (data supplied by the manufacturer).

Peritoneal cells (5 × 10⁶; >80% Mφ) were collected from untreated mice by lavage and incubated in RPMI 1640 medium supplemented with 2% FCS and different concentrations of peptides in a total volume of 1 ml at 37°C for 15 min. Cells were then diluted to 1 × 10⁶/ml in Kreb’s solution before the addition of 10 μl of the reagent Fc Oxyburst Red (Molecular Probes, Eugene, OR). Uptake of Fc oxyburst Red Complexes by the peritoneal Mφ population was monitored in real time by use of a FACScan (Becton Dickinson, Oxford, U.K.), which not only allowed the identification of the Mφ population by forward and side scatter characteristics, but also the quantification of the fluorescence acquired in the FL-3 channel during the 200 s of reaction. Cumulative changes in fluorescence at constant time intervals were then constructed and the area under the curve measured (13).

An enriched population (>95% pure) of peritoneal Mφ was prepared by 2-h adherence at 37°C in 5% CO₂/95% O₂ atmosphere in RPMI 1640 + 10% FCS and 1% strep-pen (Sigma). Nonadherent cells were then washed off, and adherent cells (>95% Mφ) were incubated with the inhibitory peptides for 15 min in RPMI 1640 medium. Cells were then stimulated with 1 mg/ml MSU crystals (a concentration chosen from preliminary experiments), and the cell-free supernatants were collected 2 h later. KC levels were measured by ELISA as described above.

Mφ (1 × 10⁵) were seeded into 96-well plates as above and incubated with serum-free RPMI 1640 medium containing 1 mM isobutylmethylxanthine and different concentrations of ACTH or ACTH_4–10. The effect of the direct adenyl cyclase stimulator forskolin (3 μM) was also tested. In selected wells, the antagonist SHU9119 was added in the presence or absence of ACTH or ACTH_4–10. After 30 min at 37°C, supernatants were removed and the cells washed and lysed. cAMP levels in the lysates were determined with a commercially available enzyme immunoassay (Amersham, Little Chalfont, Buckinghamshire, U.K.) using a standard curve constructed with 0–3.0 pmol cAMP.

Peritoneal Mφ (5 × 10⁶) enriched by 2 h adherence at 37°C in 24-well plates were lysed in 1 ml of Trizol (Life Technologies, Paisley, U.K.), and RNA was isolated according to the manufacturer’s protocol. Briefly, RNA was extracted with chloroform and isopropanol, precipitated with ethanol, and the pellet resuspended in diethyl pyrocarbonate-treated water. The yield and purity of the RNA was then estimated spectrophotometrically at 260 nm and 280 nm. Total RNA (3 μg) was used for the generation of cDNA using the T-Primed First-Strand kit (Pharmacia Biosystems Europe; St Albans, U.K.). PCR amplification reactions were then performed on aliquots of the cDNA. All PCR reactions were performed using PCR beads (Pharmacia) in a final volume of 25 μl using a Hybaid OmniGene thermal cycler (Middlesex, U.K.). The murine MC-R primer sequences were as follows: MC1-R, 5′-GTC-CAG-TCT-CTG-CTT-CCT-GG-3′ and 5′-TCT- TCA-GGA-GCC-TGT-GGT-CT-3′ (forward and reverse), which amplified a fragment 825 bp in length; MC3-R, 5′-GCC-TGT-CTT-CTG-TTT- CTC-CG-3′ and 5′-GCC-GTG-TAG-CAG-ATG-CAG-TA-3′ (forward and reverse) which amplified a fragment 820 bp in length; MC4-R, 5′-ATC-CAT-TTG-CAG-CTT-GCT-TT-3′ and 5′-ATG-AGA-CAT-GAA- GCA-CAG-ACG-C-3′ (forward and reverse) which amplified a fragment 445 bp in length; MC5-R, 5′ATG-AAC-TCC-TCC-TCC-ACC-CT-3′ and 5′-GCA-GTA-GAC-GTT-CTG-AGG-GC-3′ (forward and reverse) which amplified a fragment 810 bp in length The cycling parameters were as follows: initial denaturation for 3 min at 94°C, followed by 30 cycles of denaturation (94°C for 45 s), annealing (60°C for 30 s), extension (72°C for 1 min), and a final extension of 72°C for 10 min. Primers for murine GAPDH (14) were also used as positive controls. Amplification products were visualized by ethidium bromide fluorescence in agarose gels. Images were inverted using the Graphic Converter software (version 2.1; Lemke Software, Peine, Germany) running on a Macintosh Performa 6200.

Data are reported as mean ± SE of n distinct observations. Statistical differences were calculated on original data by ANOVA followed by Bonferroni test for intergroup comparisons (15), or by unpaired Student’s t test (two-tailed) when only two groups were compared. A threshold value of p < 0.05 was taken as significant.

Intraperitoneal administration of MSU crystals produced an intense and long-lasting accumulation of PMN, with maximal influx in the 6–24 h time period postinjection (Fig. 1,a). The peak of cell influx was preceded by a transient release of KC in the lavage fluids, which was maximal at the 2-h time point (Fig. 1 b).

As shown in Tables II and III, the in vivo administration of ACTH_4–10 or β-MSH did not affect circulating CCS levels and PMN counts: this is in contrast to ACTH which produced a marked increase in both parameters. However, treatment of mice with ACTH_4–10 inhibited both the cellular and the humoral response measured in our model of experimental gouty arthritis. Fig. 2,a illustrates the ability of ACTH_4–10 and related molecules to attenuate the 6-h PMN influx into the mouse peritoneal cavity. Approximate ED₄₀ values of 3, 4.5, and 50 nmol per mouse s.c. could be calculated for β-MSH, α-MSH, and ACTH_4–10, respectively (Fig. 2,b). Of the three peptides tested, only ACTH_4–10 was able to produce a full dose-response curve (with a maximal inhibition of 60% at the highest dose tested), whereas both β-MSH and α-MSH produced a bell-shaped curve. Similarly, ACTH_4–10 inhibited the release of KC as measured both at 2 and 6 h post-MSU crystal injection, whereas the α-MSH and β-MSH were effective only at the latter time point (Fig. 2,c). Importantly, s.c. treatment of mice with the scrambled ACTH_4–10 peptide did not modify either PMN accumulation or KC release (Fig. 2, a and c).

Table IV reports that TNF-α, IL-1β, and MIP-2 could be detected both in the 2-h and 6-h exudates. Treatment of mice with ACTH_4–10 significantly inhibited MSU crystal-induced release of IL-1β, but not that of MIP-2 and TNF-α.

ACTH, ACTH_4–10, and β-MSH inhibited Mφ phagocytosis as measured by flow cytometry. Again, whereas a linear concentration-response curve was observed with both ACTH and ACTH_4–10 (the latter peptide being almost 1000 times less potent than the parent molecule) (Fig. 3,a), a bell-shaped curve was obtained with β-MSH (Fig. 3,b). ACTH_4–10, α-MSH, and β-MSH were also tested on KC release from adherent Mφ: Fig. 3 c shows that a significant inhibitory effect (∼60%) was observed when cells were incubated with ACTH_4–10 or α-MSH.

The ability of ACTH_4–10 to attenuate MSU crystal-induced PMN recruitment and KC production were prevented by treatment of mice with the MC3/4-R antagonist SHU9119 (11). Both parameters were essentially unaltered by SHU9119 treatment alone, whereas the MC3/4-R antagonist reduced the inhibitory action of ACTH_4–10 in a dose-dependent manner, with total inhibition of the effect seen at the dose of 10 μg i.p. (Fig. 4). The nonselective MC-R antagonist S110 (12) did not affect ACTH_4–10 inhibition of PMN influx and KC release.

Similarly, ACTH_4–10 inhibition of KC release from Mφ activated in vitro was abolished by coincubation with 10 μg/ml SHU9119 (Fig. 5). The MC3/4-R antagonist also attenuated the inhibition exerted by ACTH (from 70% to 10%) and α-MSH (from 45% to 2%). In addition, the MC3/4-R agonist MTII was also able to suppress MSU crystal-induced KC production, giving 56 ± 7% of inhibition (n = 4; p < 0.05).

Both ACTH and ACTH_4–10 were able to stimulate cAMP accumulation in peritoneal Mφ (Fig. 6,a). The MC3/4-R agonist MTII, tested at a single concentration of 10 μg/ml, also stimulated cAMP formation to a degree similar to that attained with maximal concentrations of ACTH or ACTH_4–10. Conversely, 10 μg/ml SHU9119 abrogated the stimulating effect of both 100 ng/ml ACTH or 100 μg/ml ACTH_4–10 (Fig. 6 a).

To discover whether either MC3-R and/or MC4-R were expressed in murine Mφ, RT-PCR analysis was performed. Primers for murine MC1-R, MC3-R, MC4-R, and MC5-R were designed and validated using mouse genomic DNA preparations and GAPDH mRNA as a positive control (Fig. 6,b). When RNA extracted from peritoneal Mφ was used, only the MC3-R band was detected (Fig. 6 b).

In the present study we have investigated the antimigratory action of POMC gene derived peptides containing the core ACTH_4–10 sequence. The mechanism of action we propose is that activation of MC3-R on mouse peritoneal macrophages by POMC gene-derived peptides reduces the release of pro-inflammatory cytokines and subsequent recruitment of PMN. On this basis, we suggest that agonism at MC3-R may be a novel way to control the cellular component of the host inflammatory response.

This study was prompted by the observation that treatment with ACTH possesses a unique therapeutic profile in the management of human gouty arthritis which suggest the existence of mechanism(s) of action besides glucocorticoid release from the adrenal glands (6, 7). The heptapeptide ACTH_4–10 is known to lack any glucocorticoid stimulating action (9), and we confirmed that ACTH_4–10 did not stimulate plasma CCS levels and consequent neutrophilia in our current model. We have previously reported that ACTH-induced neutrophilia is totally dependent on CCS release (10).

Intraperitoneal injection into the mouse of MSU crystals produced an intense and long-lasting accumulation of PMN into the peritoneal cavity, which we recently characterized in terms of the role of resident cells and a requirement for adhesion molecules (8). We now show that a panel of chemotactic cytokines and chemokines is also released during this inflammatory reaction, and in our study we chose to monitor the CXC chemokines KC and MIP-2 (16), and the pro-inflammatory cytokines TNF-α and IL-1β. All these mediators have been shown to be produced in other models of experimental (17, 18) and human (19, 20) gouty arthritis.

Systemic administration to mice of ACTH_4–10 inhibited MSU crystal-induced PMN accumulation, and this was associated with a reduction in KC and IL-1β release in the inflammatory exudates. In this set of experiments the effects of α-MSH and β-MSH (which contain the ACTH_4–10 sequence) were also shown to suppress PMN influx and KC release in vivo. Whereas this is not surprising for α-MSH (for a recent review, see 21), we believe this is the first time that an anti-inflammatory activity has been described for β-MSH. Importantly, β-MSH was also unable to modify circulating CCS and PMN levels. As for other in vitro and in vivo experimental systems (22, 23), bell-shaped dose responses were constructed for α-MSH and β–MSH. Further studies are required to investigate whether a catabolic pathway may become activated at higher doses of the melanocortins. The effect of ACTH_4–10 was validated further by using as a control a scrambled sequence: this peptide was totally ineffective on the two parameters under observation in this model of experimental gout.

The next step was to identify the cellular target(s) responsible for the observed inhibitory actions of ACTH_4–10. In view of the relatively large body of literature which relates POMC gene-derived peptides and ACTH to the Mφ (24, 25, 26, 27, 28), we tested the hypothesis that the resident peritoneal Mφ could be targeted by these peptides. Mouse peritoneal Mφ have been shown to be deactivated by ACTH, such that IFN-γ-mediated tumoricidal activity (23) and latex beads phagocytosis (24) were blocked by micromolar concentrations of this hormone. Our preliminary experiments indicated that POMC gene-derived peptides were inactive in experiments where PMN or mast cell activation was measured (data not shown). Both ACTH and ACTH_4–10 inhibited Mφ phagocytosis as assessed by FACS analysis. Full concentration-response curves could be constructed for the two peptides, and the shorter fragment was almost 1000 times less potent than the parent molecule. β-MSH was also effective in this assay, again producing a bell-shaped curve.

More relevant to inflammation itself, the KC release from Mφ stimulated in vitro with MSU crystals was also determined. ACTH_4–10 also inhibited this parameter of Mφ activation, whereas no such effect was found with β-MSH. The reason for this discrepancy is at the moment obscure. α-MSH, used as an internal positive control in view of its reported effect on Mφ activation (21), was also found to inhibit KC release. These data complement a previous report which showed α-MSH to inhibit KC mRNA expression in the mouse liver during endotoxin-induced inflammation (30), and adds KC to the list of Mφ-derived mediators whose release is affected by this melanocortin (21).

Five MC-R have to date been identified and cloned. MC1-R binds α-MSH and ACTH preferentially, whereas the MC2-R is highly selective for ACTH and is expressed predominantly in the adrenal gland (31, 32). Human and murine MC3-R, MC4-R, and MC5-R (33, 34, 35) bind to all these peptides with varying affinities (31, 32). The lack of selective drugs has hampered a full pharmacological characterization of these receptors so far. Fan et al. (11) have recently described two compounds, MTII and SHU9119, obtained by cyclization and amino acid substitution of the ACTH_4–10 sequence, as a potent MC3/4-R agonist and antagonist, respectively. A recent study has shown that dynorphin-A (6, 7, 8, 9, 10, 11, 12), or S110, acts as an nonselective antagonist to MC-R (12). When tested in our experimental systems, S110 did not affect ACTH_4–10 inhibition of PMN migration and KC release in vivo or from activated Mφ in vitro (S.J.G., unpublished data), but the MC3/4-R antagonist SHU9119 was highly effective. When used at 10 μg i.p. (a dose already used in vivo in the mouse) (11), which is 10 times lower than ACTH_4–10 on a molar basis), SHU9119 inhibited the in vivo effects of the peptide in the MSU crystal peritonitis, whereas a partial inhibition was seen at the lowest dose of 3 μg/mouse. Importantly, SHU9119 reversed the inhibitory effect of ACTH, ACTH_4–10 and α-MSH on KC release from activated Mφ. In the latter assay we also tested the selective agonist MTII, finding a reduction in MSU crystal-induced release of the CXC chemokine. These data are strongly suggestive of an MC3-R and/or MC4-R as the molecular target(s) for the biologically activities of ACTH_4–10 and related peptides described here.

These data do not allow us to pin-point which of these two MC-R was actually responsible for the observed effects: in fact, the core sequence (ACTH_4–10) binds MC3-R and MC4-R with almost equal affinity (36). As discussed above, the efficacy of SHU9119 also does not allow receptor discrimination (11, 37). To unravel this aspect, we moved to PCR analysis of the mRNA expressed in Mφ. The specific product for the MC3-R, but not for the MC4-R (and indeed the MC1-R or MC5-R), could be found in resting mouse peritoneal Mφ. Thioglycollate injection before Mφ collection did not change this profile of MC-R expression (S.J.G., unpublished observation). In addition, MC2-R mRNA was also absent in basal or elicited Mφ (data not shown). The specificity of MC3-R primers was verified by using the murine genomic DNA preparation, and also by testing mouse brain tissue (data not shown and 38).

Finally, since MC3-R activation leads to intracellular accumulation of cAMP (34), we tested formation of this second messenger in mouse peritoneal Mφ. Following cell incubation with ACTH or ACTH_4–10, intracellular cAMP accumulated in a concentration-dependent manner. Again ACTH was 1000 times more potent than the heptapeptide on a molar basis but, as in the case of Mφ phagocytosis, ACTH_4–10 was able to produce a degree of inhibition similar to that attained with the parent molecule. Both ACTH and ACTH_4–10, but not forskolin, induced cAMP formation in Mφ and was blocked by SHU9119. Overall these data indicate that MC3-R is not only expressed in murine Mφ but is fully functional such that cAMP formation occurs after agonist activation. These observations link well with the known ability of ACTH to bind mouse leukocytes and cause intracellular accumulation of cAMP (39).

There is a resurgence of interest in the peripheral expression of the POMC gene (40); for instance, both rodent splenocytes (27, 41) and human leukocytes (42, 43) express POMC gene products. Basal expression of the POMC gene has been detected in rat Mφ (40), and the POMC gene product seems to be normally processed to produce immunoreactive ACTH (4). We report here that peptides containing the ACTH_4–10 sequence suppress PMN accumulation in acute inflammation in a CCS-independent manner. These data identify MC3-R as the molecular target for these peptides, and together with other studies (4, 40, 41, 42, 43), may suggest the existence of a novel anti-inflammatory loop based on ACTH and MC3-R that may operate to down-regulate the acute inflammatory response or the acute phases of chronic inflammation.

We thank Prof. Eddie Wei (University of California, Berkeley) for the supply of S110 and for valuable advice. We also thank Drs. R. de Médicis and A. Lussier (University of Sherbrooke, Sherbrooke, Canada) for the supply of MSU crystals.