*2.3. Identifying Di*ff*erentially Expressed Protein Phosphorylation Sites Induced by Coelonin Treatment*

To identify differentially expressed signaling-associated phosphorylated proteins between coelonin-treated and untreated RAW264.7 cells induced by LPS, the expression levels of phospho-antibody specific proteins were compared. Of the 1318 antibodies analyzed in microarray experiments, a total of 32 different phosphorylation proteins showed downregulated expression using a fold ratio ≥2 as the cutoff criterion (Figure 4, Table 1). In addition, we performed a protein network analysis using REACTOME (http://reactome.ncpsb.org/) to identify major interactions, three major cellular processes were significantly enriched (*p* value < 0.05) as follows: (1) Immune system, (2) signal transduction, and (3) cell cycle. LPS is the component of the outer membrane of Gram-negative bacteria, and is one of the most well characterized pathogen-associated molecular patterns (PAMPs), which can be recognized by Toll-like receptor 4 (TLR4), and trigger innate responses [19], such as enhancing the secretion of cytokines and chemokines, and promoting macrophages migration and proliferation [20]. Therefore, it is evident that the protein network analysis results imply coelonin may block LPS induced RAW264.7 cell signal transduction, immune response, and proliferation.

**Figure 4.** Protein and phosphorylation altered after coelonin treatment. From a total of 1318 differentially phosphorylated proteins, 32 were selected as significantly downregulated by coelonin using a volcano plot analysis.

**Table 1.** The phosphorylation of 32 proteins were significantly downregulated by coelonin treatment (*p* < 0.05).


Moreover, we entered the ENTREZ Gene IDs of the 32 genes into the DAVID Bioinformatics Resources 6.8 database. As depicted in Figure 5, the numbers of changed genes in the PI3K/AKT signaling pathway, the mitogen-activated protein kinase (MAPK) signaling pathway, the neurotrophin signaling pathway, the Sphingolipid signaling pathway and Ras signaling pathway were ranked in the top 15. In addition, we found that within these genes, the maximum number of genes belonged to the PI3K/AKT signal pathway. It is noteworthy that the phosphorylation of both subunits p65 and p105/50 of transcription factor NF-κB in this signaling pathway was reduced, which was closely related to the regulation of inflammatory gene expression [21].

**Figure 5.** Number of genes in signaling pathways changed by coelonin pretreatment.

*2.4. Validation of Proteomic Findings–Coelonin Treatment Inhibits Inflammatory Cytokines Secretion by Blocking NF-*κ*B Activation*

NF-κB is the most important signaling molecule induced by LPS through TLR4. The phosphorylated NF-κB can translocate to the nucleus, interacts with the κB elements and cause numerous cytokines secretion such as IL-1β, MCP-1 and TNF-α et al. [21]. As observed in Table 1, both NF-κB p65 and p105/50 phosphorylation levels were significantly reduced by 2.5 μg/mL coelonin treatment. Therefore, we performed western blot and immunofluorescence assays by nucleus translocation of p65 for validation. As shown in Figure 6A, the western blot results indicated that LPS stimulation significantly increased p65 accumulation in the nucleus, but the coelonin dose dependently reduced the effect of LPS. Moreover, confocal microscopic analysis reconfirmed that LPS stimulation significantly induced p65 translocation from the cytoplasm to the nucleus, which was remarkably inhibited by 2 μM of ammonium pyrrolidine dithiocarbamate (APDC), an inhibitor of NF-κB, and 5.0 μg/mL of coelonin pre-treatment (Figure 6B). In addition to cytokines, numerous studies have shown that NF-κB also regulates the expression of iNOS and COX2 genes [22,23]. Furthermore, over expression of iNOS and COX2 can lead to inflammation, tissue damage and even tumorigenesis [24,25]. As observed in Figure 6A, 200 ng/mL LPS treatment for 24 h caused a remarkable increased expression of iNOS and COX2. Pre-treatment with coelonin dose dependently reversed LPS-induced iNOS and COX2 expression. These results were consistent with the microarray results and confirmed that coelonin exerts its anti-inflammatory effect by inhibiting NF-κB activity.

#### *2.5. Colonin May Partially Inhibit the Activation of NF-*κ*B through PTEN*/*AKT Pathway*

It is well know that LPS can activate NF-κB through the Toll-like receptor 4/myeloid differentiation factor 88/IL-1 receptor associated kinase/TNF receptor associated factor 6/TGF beta-Activated Kinase 1/inhibitor of nuclear factor-κB kinase (TLR4/MyD88/IRAK/TRAF6/TAK1/IKKs) pathway [26] and now, several lines of evidence suggest that LPS can also activate NF-κB through the TLR4/MyD88/PI3K/AKT/IKKs pathway [27,28]. However, the PI3K/AKT pathway is negatively regulated by the PTEN [29,30]. Previous research has shown that the down regulation of PTEN can activate NF-κB activity by increasing p65 nucleus translocation in mouse mesangial cells and bovine alveolar macrophages, contrary, activate PTEN would reverse the effect [29,31]. The microarray results indicated that coelonin treatment significantly down-regulated the phosphorylation of PTEN at Ser380/Thr382/Thr383 (see Table 1), which implies that coelonin may inactivate NF-κB by restoring the activity of PTEN, as phosphorylation of PTEN will make it inactivated [30]. Thus, western blotting was carried out to verify this presumption. As show in Figure 6A, LPS significantly increased the phosphorylation of PTEN, AKT and inhibitor of NF-κB (IκBa), which was dose-dependently reduced by coelonin pre-treatment (Figure 6A). In order to further confirm the possible inhibition activity of coelonin against LPS-induced NF-κB activation mediated by PTEN/AKT pathway, RAW264.7 cells

were pre-treated with PTEN inhibitor SF1670. As show in Figure 7A, pre-treatment with SF1670 significantly increased LPS-induced AKT phosphorylation, which could not be downregulated by additional coelonin. Meanwhile, SF1670 also dramatically promoted LPS-induced secretion of IL-1β, IL-6 and TNF-α, and significantly reduced but could not completely block the inhibitory activity of coelonin. This result further indicates that PTEN did participate in PI3K/AKT/NF-κB activation pathway. However, it is note worthy that most levels of IL-1β and IL-6 were still significantly inhibited by coelonin co-treated with SF1670, suggesting that more critical pathways need to be identified besides the PTEN/AKT pathway.

**Figure 6.** (**A**) Coelonin inhibits LPS induced NF-κB activation in macrophage. RAW264.7 cells were pretreated with coelonin for 1 h and then treated with 200 ng/mL LPS for 30 min. Then, the nuclear protein was extracted, the content of p65 in the nucleus was determined, the total cellular protein was extracted, and the levels of phosphorylated PTEN, AKT and IκBα were determined. For iNOS and COX2 detection, RAW264.7 cells were pretreated with coelonin for 1 h, then treated with 200 ng/mL LPS for 24 h, then total cellular protein was extracted and analyzed. (**B**) Confocal microscopy analysis of p65 nucleus translocation. RAW264.7 cells were incubated with solvent (a–c) or 200 ng/mL LPS for 1 h in the absence (d–f) or presence 2 μM of the NF-κB inhibitor APDC (g–i) or 5 μg/mL of coelonin (j–l).

**Figure 7.** (**A**) Western blotting result of pAKT. RAW264.7 cells were incubated with solvent or 200 ng/mL LPS for 30 min in the absence or presence 5 μg/mL of coelonin or 2 μM of the PTEN inhibitor SF1670 or 2 μM of SF1670 combined with 5 μg/mL of coelonin. Then total cellular protein was extracted and analyzed by a simple western immunoblotting technique on a Peggy Sue system. (**B**–**D**) Effect of PTEN inhibitor SF1670 on anti-inflammatory activity of coelonin. RAW264.7 cells were incubated with solvent or 200 ng/mL LPS for 12 h in the absence or presence 2 μM of the PTEN inhibitor SF1670 or 5 μg/mL of coelonin or 2 μM of SF1670 combined with 5 μg/mL of coelonin. Culture supernatants was collected for IL-1β, IL-6 and TNF-α detection by CBA method. Data are expressed as mean ± SD (*n* = 6). \*\* *p* < 0.01.

#### *2.6. Coelonin Treatment Leads to G1 Cell Cycle Arrest through PTEN*

The PI3K-Akt pathway has been shown be involved in a variety of cellular processes, including inflammation response [27], cell survival and proliferation [32]. Many reports indicate that the PI3K/AKT pathway plays a pivotal role in regulating cell cycle progression through phosphorylation and degradation of cell cycle regulator p27Kip1 [33,34]. P27Kip1 can interact with cyclin-dependent kinase 2 (CDK2) and cyclinE to prevent cell entry into the S phase, and over-expression p27Kip1 would promote cell G1 phase cell cycle arrest [35]. In many cancer cells, such as cervical cancer cells, oral squamous cell carcinoma cells, and prostatic carcinoma cells, the expression protein of p27Kip1 was significantly reduced, the invasion and migration ability was enhanced, and the expression of p27Kip1 was negatively correlated with survival [33,36]. We noticed that the microarray results showed that coelonin significantly down regulated the phosphorylation of p27kip1, and interestingly, a previous study indicated that the up-regulation of PTEN prevents p27Kip1 phosphorylation and proteolysis [37]. Hence, we speculate that coelonin may inhibit p27Kip1 phosphorylation and degradation mediated by PTEN. Western blot confirmed that 5 μg/mL coelonin treatment significantly recovered the p27Kip1 level, which was even higher than the un-treatment group, while LPS stimulation significantly reduced the p27Kip1 level, and PTEN inhibitor SF1670 almost completely abrogated the effect of coelonin (see Figure 8A). Theoretically, high levels of p27Kip1 can block cells in the G1 phase cell-cycle, so we used flow cytometry to verify whether coelonin could induce G1 phase cell cycle arrest in RAW264.7 cells. As observed in Figure 8B, coelonin pre-treatment remarkably induced G1 phase cell cycle arrest and SF1670 completely inhibited the effect of coelonin. These data suggest that coelonin inhibits p27Kip1

degradation in a PTEN-dependent manner, thereby exerting its G1 phase cell cycle arrest effect in RAW264.7 cells.

**Figure 8.** (**A**) Western blotting result of p27kip1. RAW264.7 cells were incubated with solvent or 200 ng/mL LPS for 12 h in the absence or presence 5 μg/mL of coelonin or 2 μM of the PTEN inhibitor SF1670 or 2 μM of SF1670 combined with 5 μg/mL of coelonin. Then total cellular protein was extracted and analysed by a simple western immunoblotting technique on a Peggy Sue system. (**B**) Coelonin induce G1 cell cycle arrest through PTEN analysed by flow cytometry. RAW264.7 cells were incubated with solvent or 200 ng/mL LPS for 12 h in the absence or presence 5 μg/mL of coelonin or 2 μM of the PTEN inhibitor SF1670 or 2 μM of SF1670 combined with 5 μg/mL of coelonin. Cells were harvested, fixed and stained with PI, and cell cycle distributions were analyzed on a BD Accuri™ C6 flow cytometer in triplicate.

#### **3. Discussion**

To date, over 200 phenanthrene compounds have been isolated and identified, most of which come from the Orchidaceae family [38]. In *Bletilla striata*, a member of Orchidaceae, more than 30 phenanthrenes have been isolated from its pseudobulbs and fibrous roots, most of which showed antitumor and antimicrobial activities [39,40]. In addition, numerous studies have shown that phenanthrene and dihydrophenanthrene derivatives with remarkable anti-inflammation activity [38,41,42]. In this study, coelonin, a compound with a strong anti-inflammatory activity, was isolated and identified from the pseudobulbs of *Bletilla striat* under the guidance of biological activity screening. Few studies have reported its anti-inflammatory activity, and our studies show that coelonin can dose-dependently inhibit LPS-induced expression and secretion of IL-1β, IL-6 and TNF-α in RAW264.7 cells. Furthermore, we found that *Bletilla striata* has a high content of coelonin (0.020%–0.301% in different samples) [43]. Therefore, coelonin is probably one of the main anti-inflammatory active components of *Bletilla striata*.

In order to better elucidate the potential anti-inflammatory molecular mechanism of coelonin, PEX100 protein microarrays containing 1318 antibodies were used. The results indicated that 32 different phosphorylated proteins were significantly downregulated by pre-treatment with coelonin, which were closely related to the response of LPS-induced signal transduction, immune response and cell proliferation. Additionally, most of these genes were focused on the PI3K/AKT signal pathway, and we

performed verification on the phosphorylation levels of PTEN, NF-κB and p27Kip1. NF-κB is considered as a central regulator of LPS-induced pro-inflammatory response in macrophage activation [44], which is usually formed as p65:p50 heterodimer. In addition, PEX100 microarray results indicated that the phosphorylation levels of both p65 and p50 were significantly downregulated by coelonin pre-treatment (Table 1), which was further confirmed by western blot and confocal microscopic analysis on p65 nucleus translocation (Figure 6). Studies have shown that numerous dihydrophenanthrene derivatives exert anti-inflammatory effects by inhibiting NF-κB pathway [38,41,42], which was significantly correlated with the presence of phenolic hydroxyl groups [45]. Interestingly, coelonin has two phenolic hydroxyl groups, which is in accordance with its anti-inflammatory effect. However, few reports have studied the upstream targets that regulate the activity of NF-κB by dihydrophenanthrene derivatives. According to the results of the PEX100 microarray, we propose that coelonin may play an anti-inflammatory role by inhibiting the activity of NF-κB through the PTEN/AKT pathway. However, some studies have indicated that the PI3K/AKT pathway negatively regulates LPS induced NF-κB activation [46]. In contrast, other studies have shown that the PI3K/AKT pathway does positively regulate LPS-induced gene expression [27]. We found that LPS stimulation could dramatically induce AKT phosphorylation, IκBa degradation and NF-κB nucleus translocation (Figure 6), and the PI3K inhibitor LY294002 significantly inhibited LPS inducted IL-1β, IL-6 and TNF-α expression in RAW264.7 cells (see Supplementary Figure S5), which is consistent with the findings that the PI3K/AKT pathway is required for LPS induction of gene expression in RAW264.7 cells. Meanwhile, PTEN inhibitor SF1670 significantly increased LPS-induced AKT phosphorylation and cytokines secretion (Figure 7), which is consistent with the report of Zhao et al. [47]. These results indirectly supported that both PI3K/AKT and PTEN/AKT pathways are involved in LPS inducted NF-κB activation. However, the phosphorylation level of Akt did not decrease after co-treatment with PTEN inhibitor SF1670 and coelonin, but coelonin still could significantly down-regulate the expression of inflammatory cytokines induced by LPS. This contradictory result not only indicated that coelonin has other pathways to exert its anti-inflammation activity, but also indicated that AKT pathway has complex mechanisms in the regulation of inflammation. In fact, the exact mechanism of activation of NF-κB by AKT pathway remains controversial. Some reports indicated that activation AKT lead to IKK-dependent IκBα degradation and nucleus translocation of NF-κB, while others shown that AKT-dependent activation of NF-κB by stimulating the transactivation potential of the p65 subunit, rather than inducing IκBα degradation [48]. Obviously, although coelonin could inhibit IκBα phosphorylation and degradation in a dose-dependent manner, it is not clear whether coelonin plays a role through the AKT pathway or TLR4/MyD88/TRAF6/TAK1 pathway, the canonical pathway of NF-κB activation (Figure 9). Therefore, it is difficult to draw a definite conclusion only through the intervention of inhibitors, and more scientific and reasonable experiments need to be designed to elucidate the alternative mechanism of coelonin inactivating NF-κB.

In addition, our study also detected the degradation of p27Kip1 after the LPS challenge, which is a downstream target of PTEN/AKT. This effect was dramatically inhibited by coelonin pre-treatment. However, the PTEN inhibitor SF1670 completely abrogated the effect of coelonin (Figure 8). We noticed that LPS treatment also induced RAW264.7 cells G1 phase arrest (Figure 8B), which contradicts a previous report [20], but is consistent with report [49]. The exact molecular mechanism underlying these contradictory results need to be further studied, but it is clear that LPS-induced G1 phase arrest in RAW264.7 cells did not occur through the up-regulation of p27Kip1. However, LPS treatment did promote p27Kip1 degradation (Figure 8A), and which may be dependent on PTEN inactivation (Figure 9). While coelonin could block the LPS-induced degradation of p27Kip1 by restoring PTEN activity (Figure 9). Studies have shown that alveolar macrophage proliferation may play an important role in the formation of multinucleated giant cells and granuloma induced by silica or asbestos [50]. These results suggest that, in addition to inhibiting the secretion of inflammatory factors by macrophages, inhibiting the proliferation of macrophages may also play a role in alleviating silica or asbestos-induced lung pathological changes to an extent.

**Figure 9.** Proposed mechanism of coelonin inhibiting LPS-stimulated activation of NF-κB and inducing G1 phase cell-cycle arrest in RAW264.7 cells. Coelonin inhibits the expression of inflammatory cytokines IL-1β, IL-6 and TNF-α in RAW264.7 cells treated by LPS may partially through the PTEN/AKT pathway. However, it induces G1 cell cycle arrest of RAW264.7 cells by inhibiting the degradation of p27Kip1 in a PTEN-dependent manner. The black arrow section has been verified, while the red arrow section needs further validation. The question mark indicated that coelonin may inactivate NF-κB by inhibiting TLR4/MyD88/TRAF6/TAK1 signal pathway, the canonical pathway of NF-κB activation.
