A novel modified-curcumin 2.24 resolves inflammation by promoting M2 macrophage polarization

To assess resolving-like activity by a novel chemically-modified curcumin (CMC2.24) in a “two-hit” model of diabetes-associated periodontitis. Macrophages from rats were cultured in the presence/absence of either Lipopolysaccharide (LPS, 1st hit); or advanced-glycation-end products (AGE, 2nd hit); or both combined. CMC2.24 was added as treatment. The conditioned media were analyzed for MMP-9, cytokines (IL-1β, IL-6, TNF-α), resolvins (RvD1, RvE1, lipoxin A4), and soluble receptor for AGE (sRAGE). The phenotypes of M1/M2 macrophage were analyzed by flow cytometry. Both LPS/AGE-alone, and two-combined, dramatically increased the secretion of MMP-9 by macrophages. CMC2.24 “normalized” the elevated levels of MMP-9 under all conditions. Moreover, CMC2.24 significantly reduced the secretion of IL-1β and IL-6 with a fewer effects on TNF-α. Importantly, CMC2.24 increased RvD1 and sRAGE secretion by macrophages exposed to LPS/AGE; and both treatment groups exhibited increased M2 relative to M1 populations. Furthermore, scatter-diagram showed the macrophages gradually shifted from M1 towards M2 with CMC2.24-treated, whereas LPS/AGE-alone groups remained unchanged. CMC2.24 “normalized” cytokines and MMP-9, but also enhanced RvD1 and sRAGE in macrophages. Crucially, CMC2.24 appears to be a potent inhibitor of the pro-inflammatory M1 phenotype; and a promotor of the pro-resolving M2 phenotype, thus acting like a crucial “switch” to reduce inflammation.

Cell culture studies.Peritoneal Mφs.Mφs were obtained from peritoneal washes (PW) of 8 SD rats.Briefly, 15 mL of cold sterile PBS (pH7.4) containing 3 mM of ethylenediaminetetraacetic acid (EDTA) intraperitoneally (i.p.) was injected into the peritoneal cavity, and the peritoneum was gently massaged for 1 min.Approximately 12 mL/per rat of PW was collected and stored on ice (4 °C).These lavages were centrifuged (1000 rpm, 4 °C, 5 min) to separate supernatant cell-free peritoneal fractions (CFPFs) and resident mixed cells that were re-suspended in 25 mL/each sample of PBS/3 mM EDTA.The cell suspension was layered onto Lymphoprep™ (Accurate Chemical & Scientific Corporation, Westbury, NY, USA) at a ratio of 2 to 1-1.5 (v/v) and centrifuged at 1,800 rpm for 30 min at 25 °C.The resident peritoneal Mφs were isolated and purified from the mixed cells by density gradient centrifugation and adherence as described previously 38 .The peritoneal Mφs were cultured in macrophage serum-free media (SFM) (Thermo Fisher Scientific Inc., Boston, MA, USA) in 24-well plates.Each well contained 10 6 cells/mL, supplemented with 100 units/mL penicillin, and 100 μg/mL streptomycin, in a humidified atmosphere of 5% CO 2 and 95% air at 37 °C.
Blood-derived Mφs.Resident blood-derived Mφs were isolated from the heart of 8 SD rats by cardiac puncture.In brief, 10-15 mL of whole blood from each rat was drawn, then transferred to the heparin-coated tubes (anti-clotting).Mφs isolation was achieved by the same procedure as described above (Fig. 1a), then cultured in the Teflon Beakers (Savillex, MN, USA) to prevent cell attachment (no adherence) 39,40 .At time 0, Mφs were isolated and distributed into 6 groups (triplicates/group) to mimic the "two-hit" model of diabetes-associated periodontitis: in the absence (N), or presence of either LPS (L) or AGE (A).CMC2.24 was added to the cultures of different groups at a final concentration of 5 μM (N + 2.24, L + 2.24, and A + 2.24), then harvested at 0, 6, and 18 h for flow cytometry analysis.

Effects of CMC2.24 on pro-inflammatory cytokines.
As shown in Fig. 2a, IL-1β concentrations were significantly increased in both L and A-alone (p < 0.001; p < 0.005) challenged groups, and with a synergistic and significant increase in the L + A challenged group (p < 0.05), respectively, compared to the N group.IL-1β concentrations in the L + A group were also significantly greater than each "hit" (L or A) alone (p < 0.05; p < 0.05).When the L-alone was treated with 2 and 5 µM of CMC2.24, the higher concentration (5 µM) of CMC2.24 produced a significant reduction in IL-1β concentrations (p < 0.05), as well as a trend of reduction at the lower concentration (2 µM) of CMC2.24 (p > 0.05).In addition, when the A-alone was treated with 2 and 5 µM of CMC2.24, both concentrations of CMC2.24 treatment significantly reduced IL-1β concentrations (p < 0.05; p < 0.05) in an apparent dose-response manner.Impressively, when the "two-hit" (L + A) group was treated with 2 and 5 µM of CMC2.24, both concentrations of CMC2.24 significantly decreased IL-1β concentrations (p < 0.05; p < 0.05) and brought them back to normal levels in a similar dose-response relationship (Fig. 2a).
In addition, IL-6 concentrations were significantly increased in both L and A-alone (p < 0.05; p < 0.005) groups, also with a similar synergistic and significant increase in the L + A challenged group (p < 0.05), respectively, compared to the N group (Fig. 2b).IL-6 concentrations in the "two-hit" (L + A) group were significantly greater than each "hit"-alone (both p < 0.05).Moreover, the CMC2.24-treated group (5 µM) significantly reduced IL-6 concentrations (p < 0.05), compared to the L-alone.When the A-alone was treated with 2 and 5 µM of CMC2.24, both concentrations significantly reduced IL-6 concentrations (both p < 0.05) compared to the A-alone.Furthermore, when the L + A group was treated with 2 and 5 µM of CMC2.24, both concentrations significantly decreased IL-6 concentrations (both p < 0.05), again, in the apparent dose-response fashion, and brought them back to normal levels.
Next, we examined TNF-α concentrations in the same culture media.As shown in Fig. 2c, TNF-α levels were significantly increased in all three challenged groups: L and A-alone, as well as the L + A groups (all p < 0.001), respectively, compared to the N group.However, the combination of "two-hit" did not increase TNF-α concentrations over the already (apparently) maximum increase in this cytokine produced by each "hit" separately.Impressively, when the L-alone was treated with 2 and 5 µM of CMC2.24, both concentrations significantly reduced TNF-α concentrations (both p < 0.05) in a dose-response manner.However, when the A-alone was treated with 2 and 5 µM of CMC2.24,only the higher concentration (5 µM) of CMC2.24 appeared to produce a more significant reduction in TNF-α concentrations (p < 0.05), although a trend of reduction was seen in the lower concentration (2 µM) of CMC2.24-treated group (p > 0.05).Finally, when the L + A group was treated with 2 and 5 µM of CMC2.24, a lesser effect on the "two-hit" group was observed (p > 0.05), even though there was still a trend of reduction in TNF-α concentrations.

Effects of CMC2.24 on resolvins.
Regarding RvD 1 , as shown in Fig. 3a, the concentrations of this resolvin were significantly reduced to undetectable levels both in the L and A-alone groups (both p < 0.001), respectively, compared to the N group.When the L-alone was treated with 2 and 5 µM of CMC2.24, both concentrations significantly increased RvD 1 concentrations (both p < 0.001) in a dose-response manner, compared to the L-alone.It should also be noted that the higher concentration (5 µM) of CMC2.24 treatment, but not the lower concentration (2 µM) tremendously increased RvD 1 concentrations by eightfold above the L-alone challenged group and by 6-folds above normal (N) levels.Similarly, when the A-alone was treated with 2 and 5 µM of CMC2.24, the higher concentration (5 µM) of CMC2.24 significantly increased this resolvin by sixfold above the A-alone and 4-folds above N levels (both p < 0.05).However, these effects were less prominent in the "two-hit" group.
Concentrations of RvE 1 were not statistically different among the normal, the challenged groups (L-, A-alone, and L + A), and the CMC2.24-treatedgroups.In addition, concentrations of LxA 4 were undetectable in the rat peritoneal Mφs culture media (data not shown).
Effects of CMC2.24 on sRAGE.Although LPS-challenge did not reduce the levels of sRAGE, the 2 µM of CMC2.24-treated group significantly raised the levels of sRAGE (p < 0.05); and a trend towards increase of the levels of this molecule was seen in the 5 µM of CMC2.24-treated group as well, both compared to the L-alone (Fig. 3b).Besides, the concentrations of sRAGE were dramatically reduced in the A-alone challenged group (p < 0.005) compared to the N group.Both concentrations of CMC2.24-treated groups increased the sRAGE concentrations.But the higher concentration (5 µM) of CMC2.24 seemed to be more potent than the lower one (2 µM), to bring the levels of sRAGE back to normal (p < 0.05), compared to the A-alone.In addition, the L + A group had a trend toward reduction in sRAGE levels, compared to the N group.However, there were fewer effects of CMC2.24 on sRAGE levels in the "two-hit" group (p > 0.05).

Conversion of Mφs from M1 to M2 phenotype.
LPS-challenged Mφs.At 0, 6, and 18 h, the ratio of M2/M1 was indicated in Fig. 4a.With 6 to 18-h treatment of CMC2.24, this ratio was dramatically increased in the L + 2.24 group, and was significantly higher than the L-alone group (p < 0.005).Interestingly, the ratio of M2/ M1 in N + 2.24 was also higher at these time points compared to the untreated (N and L-alone) groups.Thus, treatment with CMC2.24 appeared to "prime" the normal monocyte/macrophage to develop to the M2 phenotype.In the absence of CMC2.24, the ratio of M2/M1 in the L-alone and N groups was maintained at low levels at each time point (Fig. 4a).Consistent with the ratio change of M2/M1, the percentage of M1 population was dramatically decreased in the L + 2.24 group at 18 h, compared to the L-alone (p < 0.005) (Fig. 4b).However, the percentage of M2 population was significantly increased in the CMC2.24-treated group at 6 h, compared to the L-alone (p < 0.005) (Fig. 4c).
These observations were further confirmed by the scatter diagrams of flow cytometry analysis (Fig. 4d).In the L + 2.24 group, we observed a population of Mφs that was gradually "shifting" from M1 towards M2 phenotype during the time.This "shifting" population was indicated in the CMC2.24-treatedMφs (L + 2.24; N + 2.24).However, the population of Mφs remained unchangeable in the N and L-alone groups.

Discussion
In this study, we demonstrated the current view on the pathogenesis of diabetes-associated periodontitis, which includes a series of dysregulated host inflammatory/collagenolytic responses.HMT is a novel therapeutic strategy to resolve uncontrolled and unresolved inflammation, which could be an adjunct to conventional treatment of diabetes-associated periodontitis.
A NON-antimicrobial formulation of doxycycline (NAD), as the first generation of HMT agents developed by our group, has been and still is the only governmentally-approved agent for clinically treating periodontitis in humans 52 .A second generation of HMT agents, namely chemically-modified tetracyclines (CMTs, especially CMT-3), has also shown evidence of safety and efficacy in cell and tissue culture, in vivo (rat models), and in clinical studies [51][52][53][54] .Many recent studies indicate that curcumin can modulate macrophage polarization through its anti-inflammatory effects 55,56 .As the latest generation of HMT agents, CMC2.24 is a phenylaminocarbonyl curcumin which has a triketonic structure at its C-4 position, resulting in higher bioactivity, better solubility, and greater zinc-binding capability than natural curcumin, as well as no evidence of toxicity even at high doses 36,57,58 .The novelty of CMC2.24 has also been tested in different models and conditions as we described previously [28][29][30][32][33][34] . For exaple, we examined the signaling and transcriptional factors in two in vivo models of diseases (diabetes and periodontitis) 28,29,31,32 .The results demonstrated that CMC2.24 significantly reduced TLR-2 and p38 MAPK expressions involved in the inflammatory signaling cascade in a dog model of natural periodontitis; and inhibited NF-κB activation and inflammatory bone loss in murine models of LPS-induced experimental periodontitis and diabetes-associated natural periodontitis.
More importantly, our group also has compared the effects of the oral administration of natural curcumin and a chemically-modified curcumin (CMC2.24) on osteoclast-mediated bone resorption, apoptosis, and inflammation in a rat model of experimental periodontitis 59 .We found that CMC2.24 and curcumin inhibit inflammation by different mechanisms, CMC2.24 was capable of reducing alveolar bone resorption in the LPS-induced model of periodontitis.In addition, we did an in vitro study regarding the inhibitory concentration at 50% (IC50) of CMC2.24 and natural curcumin.It demonstrated that CMC2.24 exhibited inhibitory IC50 values in vitro, ranging from 2-8 μM against two collagenases (MMP-8 and MMP-13), two gelatinases (MMP-2 and MMP-9), MMP-3, MMP-7 and MMP-12, which were much lower than curcumin, ranging from 3 to 52 μM 57 .Therefore, CMC2.24 exhibited stronger potency and efficacy than natural curcumin on MMPs inhibition.
For these reasons, currently, we focused on the multiple mechanisms of efficacy of CMC2.24.This ex-vivo study provided insights that CMC2.24 can restore the balance between the pro-inflammatory and the antiinflammatory/pro-resolving responses, (a) by decreasing the production of pro-inflammatory mediators (IL-1β, IL-6, and TNF-α), as well as the collagenolytic MMP (MMP-9); and (b), by also dramatically increasing RvD 1 levels, as well as sRAGE levels.Many studies showed that the cytokines, like IL-1β, TNF-α, and IL-6 were involved in MMPs upregulation 60 , and connective tissue breakdown, including bone resorption 17 .Therefore, our results indicate that CMC2.24 is a potent regulator with additional mechanisms to resolve inflammation and restore homeostasis.The pleiotropic mechanisms of this novel compound on resolving inflammation have been summarized in Fig. 6.
RvD 1 is a well-characterized D-series resolvin that is generated in situ from DHA by two sequential lipoxygenation steps 61 .Many researchers have found that RvD 1 is important in (a) ameliorating LPS-induced inflammatory diseases 62 ; (b) improving insulin sensitivity by resolving the chronic inflammation that is associated with diabetes and obesity 63 ; and importantly, (c) activating the anti-inflammatory response, partially by the conversion of Mφs from classically-activated (M1 phenotype) to alternatively activated (M2 phenotype) cells 14 .This "shift" accelerates the resolution of persistent inflammation at the cellular level.And our study identified that RvD 1 , not the other resolvins or lipoxins such as RvE 1 and LxA 4 , is the main target for the resolvin-like activity of CMC2.24.This did provide the additional mechanisms that help explain its potent efficacy in the resolution of tissue-destructive effects of chronic inflammation.
Besides, the role of AGE-RAGE interaction in promoting chronic inflammatory diseases has been increasingly recognized.AGE is formed when reducing sugars nonenzymatically bind to proteins or lipids, a process that is enhanced by hyperglycemic and hyperlipidemic environments of numerous metabolic disorders including diabetes and its complications (e.g., periodontitis) 9 .sRAGE, a decoy receptor for AGE, is associated with diabetes when a lower level of this soluble receptor which present in the circulation.In this study, AGE-challenged Mφs showed a decrease in the levels of sRAGE.But CMC2.24-treated Mφs significantly increased the levels of sRAGE and brought them back to normal.This indicated an additional mechanism that CMC2.24 resolve inflammation through AGE-RAGE/sRAGE signaling.www.nature.com/scientificreports/Moreover, another mechanism of CMC2.24's pleiotropic efficacy is indicated by the conversion of the M1 pro-inflammatory phenotype to the M2 pro-resolving phenotype.In terms of both phenotype and function, Mφs have remarkable heterogeneity 64 , reflected in their specialization and also the marked changes in their activity and gene expression when facing challenges in the microenvironment.Our study demonstrated that the CMC2.24-treatedMφs significantly increased the ratio of M2/M1, and enhanced the conversion of Mφs from the M1 to M2 phenotype.This effect appeared to be attributed primarily due to the suppression of the M1 phenotype population; and due to the early promotion of the M2 phenotype population.As we observed, at 18 h, the challenge with either LPS or AGE and treatment with CMC2.24 can dramatically suppress the M1 phenotype as indicated by a lower expression of the M1 cell surface marker, CD38, compared to the non-treated or challenge-only groups.In contrast, the M2 phenotype population under the same conditions was markedly increased at 6 h after CMC2.24 therapy.This early effect can be lasting for 18 h as made evident by a high expression of the M2 cell-surface marker, CD163.The scatter diagram confirmed our hypothesis that CMC2.24 produced a "shifting" of Mφs from the M1 towards the M2 phenotype using flow cytometry (Figs.4d, 5d).Therefore, CMC2.24 appears to be a highly potent inhibitor of the M1 phenotype (probably related to the inhibition of the M1 macrophage differentiation or maturation, or both); and a promoter of the pro-resolving M2 macrophage, thus acting like a "switch".

Conclusion
Taken together, these results continue to expand the pleiotropic mechanisms of the novel chemically-modified curcumin 2.24, by (a) inhibiting inflammatory and tissue-destructive mediators including cytokines and MMPs; (b) upregulating resolvin-enhancing (RvD 1 ) activity; and (c) promoting the expression of sRAGE; as well as (d) contributing to macrophage polarization from the M1 towards M2 phenotype (Fig. 6).Therefore, CMC2.24 can be a promising drug in treating diabetes-associated periodontitis.And the challenge for the future will be to demonstrate the safety and efficacy of this pleiotropic compound for the treatment of human chronic inflammatory diseases including diabetes, periodontitis, and others.Figure 6.A scheme to summarize the pleiotropic mechanisms of the novel chemically-modified curcumin 2.24 on resolving inflammation in a "two-hit" model.This "two-hit" model demonstrates a bi-directional manner between diabetes (or other systemic diseases) and periodontitis.Treatment with CMC2.24 can inhibit inflammatory and tissue-destructive mediators including cytokines and MMPs; upregulate resolvin-enhancing (RvD1) activity; and promote the expression of sRAGE; as well as contribute to macrophage polarization from the M1 towards M2 phenotype.