Resolution-phase macrophages possess a unique inflammatory phenotype that is controlled by cAMP

Abstract

Neutralizing injurious stimuli, proinflammatory mediator catabolism, and polymorphonuclear leukocyte (PMN) clearance are determinants of inflammatory resolution. To this, we recently added innate-type lymphocyte repopulation as being central for restoring postinflammation tissue homeostasis with a role in controlling innate immune-mediated responses to secondary infection. However, although macrophages dominate resolution, their phenotype and role in restoring tissue physiology once inflammation abates are unknown. Therefore, we isolated macrophages from the resolving phase of acute inflammation and found that compared with classically activated proinflammatory M1 cells, resolution-phase macrophages (rMs) possess weaker bactericidal properties and express an alternatively activated phenotype but with elevated markers of M1 cells including inducible cyclooxygenase (COX 2) and nitric oxide synthase (iNOS). This phenotype is controlled by cAMP, which, when inhibited, transforms rM to M1 cells. Conversely, elevating cAMP in M1 cells transforms them to rMs, with implications for cAMP in the resolution of systemic inflammation. It transpires that although rMs are dispensable for clearing PMNs during self-limiting inflammation, they are essential for signaling postresolution lymphocyte repopulation via COX 2 lipids. Thus, rM macrophages are neither classically nor alternatively activated but a hybrid of both, with a role in mediating postresolution innate-lymphocyte repopulation and restoring tissue homeostasis.

Figure 1

Cell profile and associated cytokine and chemokine release in resolving inflammation. We established 2 models of inflammation—one where a low dose of zymosan (0.1 mg, □) was injected intraperitoneally causing a transient acute inflammatory response that resolved, as well as another model where 10 mg zymosan was injected (■) triggering a more aggressive inflammation. (A) Gr1(LY6G)-positive granulocytes and (B) F4/80-labeled macrophages were enumerated in each model over time by hemocytometer and FACS analysis. (C) Levels of typical proinflammatory cytokines were quantified in the cell-free inflammatory exudates at several time points during and after resolution in both models as well as (D) in plasma 2 weeks after resolution to determine whether inflammation resolved or became systemic. n = 6 to 8 mice per group with data expressed as mean plus or minus SEM.

Figure 2

Unique phenotype of resolution-phase macrophages is preserved in response to inflammatory stimuli ex vivo. Macrophages were isolated from a 72-hour peritonitis that either resolved (0.1 mg zymosan or rM macrophages, □) or progressed to systemic inflammation (10 mg zymosan or M1 macrophages, ■) and incubated for 6 hours ex vivo to determine profiles and levels of (A) cytokines and chemokines indicative of established M1 and M2 macrophage phenotype. Having established that rM macrophages secrete predominantly anti-inflammatory IL-10 and comparatively fewer proinflammatory mediators, expression of other inflammatory markers was determined including (B) iNOS as well as (C) COX 2 and COX 2–derived PGD₂ and intracellular markers of M2 phenotype including (D) mannose receptor as well as (E) cAMP. (F,G) Finally, we determined that exposure of these resolution-phase macrophages to a range of inflammatory stimuli did not alter their phenotype, which remained robust upon exposure to TLR ligands. n = 6-8 mice per group. *P ≤ .05, as determined by ANOVA, followed by Bonferroni t test or 2-tailed Student t test, with data expressed as mean plus or minus SEM.

Figure 3

Resolution-phase macrophage phenotype is determined by cAMP. Macrophages (0.5 × 10⁶) previously classified as either resolution phase (rM) or M1 were treated with modulators of cAMP with or without LPS for 20 hours in triplicate. Using (A,B) TNFα and (C,D) IL-10 as markers of macrophage inflammatory status, it was determined that cAMP elevation in M1 cells triggered IL-10 while also inhibiting TNFα. Changes in (E) COX 2 and (F) iNOS expression in response to cAMP were determined by Western blot analysis using 3 μg M1 and rM cell lysates in each lane. LPS-stimulated J774 macrophage extracts were used as positive controls. n = 6 to 8 mice per group. *P ≤ .05, as determined by ANOVA, followed by Bonferroni t test with data expressed as mean plus or minus SEM.

Figure 4

Resolution-phase macrophages possess reduced bactericidal properties compared with M1 cells. (A) Macrophages (0.5 × 10⁶) were isolated from a 72-hour peritonitis that either resolved (0.1 mg zymosan or rM macrophages, □) or progressed to systemic inflammation (10 mg zymosan or M1 macrophages, ■) and incubated with S aureus. One hour later, media were taken and plated on LB-agar plates and bacterial colonies counted 24 hours later as a measure of macrophage ability to kill bacteria. Thus, the higher CFU reflect reduced bacterial clearance. In the first instance, macrophages from resolving inflammation (rM) had a lower ability to kill S aureus compared with M1 cells. However, upon incubation with the cAMP inhibitor, rp-cAMP, rM cells experienced enhanced bactericidal properties. Conversely, M1 cells, whose inherent ability to kill bacterial was greater than that of rM cells, were reduced by elevating cAMP. (B) Theophylline was injected to animals bearing an ongoing inflammation (10 mg zymosan) at 72 hours followed by group B streptococcus with plasma sampled 3 hours later to determine colony-forming units along with (C) animal survival over time at concentrations of drug that (D) elevated cAMP. n = 8 mice per group. *P ≤ .05; **P ≤ .01; and ***P ≤ .001, as determined by ANOVA, followed by Bonferroni t test with data expressed as mean plus or minus SEM.

Figure 5

Inflammatory environment dictates macrophage phenotype. rM cells from a 72-hour zymosan (0.1 mg)–induced peritonitis were labeled with PKH26-PCL^red and adoptively transferred at 72 hours to mice bearing a 10 mg zymosan (nonresolving inflammation) peritonitis and found not to trigger resolution as defined by little alterations in (A) total as well as individual cell numbers (PMNs, MΦs, and innate-type lymphocytes, not shown) determined 24 hours later. However, adoptively transferred rMs (identified by being PKH26-PCL^red) were isolated from the nonresolving milieu after 24 hours using FACSAria and were found to synthesize more (B) TNFα than native rM cells with little change in (C) IL-10, indicating that despite the robustness of ex vivo stimulation reported in Figure 2F,G, the inflammatory environment of whole animal systems dictates cellular phenotypes. n = 6 mice per group. * indicates P ≤ .05, and **P ≤ .01, as determined by ANOVA, followed by Bonferroni t test with data expressed as mean plus or minus SEM.

Figure 6

rM mediates postresolution homeostasis. Mice were injected intravenously with clodronate-filled liposomes followed by zymosan (0.1 mg). Inflammation was first determined 4 hours later showing a significant increase over controls of (A) PMNs at onset followed by a confirmed reduction in (B) macrophages during resolution (48 hours) concomitant with a surprising reduction in (C) PMN numbers during this phase, suggesting that when phagocytosing macrophage are limited in numbers, PMNs may be efficiently cleared by parenchymal/stromal cells. However, (D) T- and B-lymphocyte numbers were also significantly reduced, revealing that rM cells are central to recruiting postresolution innate-type lymphocytes, which we have shown to be critical for postinflammation susceptibility to superinfection and restoration of tissue homeostasis. (E) We identified that an rM COX 2–derived prostanoid mediated these effects as NS-398 at 10 mg/kg dosed orally during resolution (48 and 60 hours) impaired CD3 cell repopulation at 72 hours. Moreover, (F) adoptively transferring PKH26-PCL^red rM along with equal numbers of PKH2-PCL^green M1 cells to a murine peritoneal cavity revealed that after 24 hours rM cells remained in the peritoneum, whereas M1 displayed a greater clearance rate, suggesting that rM cells possess a greater propensity to remain in the peritoneum to elicit their prohomeostatic functions, with the dotted line denoting the original numbers of MΦs injected. (G) FACS dot plot to prove the successful transfer of differentially labeled MΦs and illustrating the greater number of PKH26-PCL^red MΦs (bottom right quadrant) compared with fewer PKH2-PCL^green MΦs (top left quadrant). *P ≤ .05 and **, P ≤ .01, determined by Bonferroni t test with data expressed as mean plus or minus SEM.