Reversal of Epithelial–Mesenchymal Transition by Natural Anti-Inflammatory and Pro-Resolving Lipids

Epithelial mesenchymal transition (EMT) is a key process in the progression of malignant cancer. Therefore, blocking the EMT can be a critical fast track for the development of anticancer drugs. In this paper, we update recent research output of EMT and we explore suppression of EMT by natural anti-inflammatory compounds and pro-resolving lipids.


Introduction
The epithelial-mesenchymal transition (EMT) is defined as a phenomenon that epithelial cells transform into mesenchymal cells [1]. EMT plays a key role in cancer progression and fibrosis. Many researchers and pharmaceutical companies have tried to develop novel EMT blockers due to its importance in such diseases [2][3][4].
Inflammation is one of ten in cancer hallmark [5]. It is a critical factor of tumor microenvironment affecting EMT. Several reviews have emphasized role of inflammation in EMT [6]. However, there are a few reviews that deal with the blocking of EMT by natural anti-inflammatory compounds and pro-resolving lipids.
In this review, I have dealt with ingredients derived from natural products that were not covered in the 2018 review on EMT [6]. I have also added the story of inflammasomes, which play an essential role in the early steps of inflammation, and how they are involved in EMT. Naturally derived compounds that control these inflammasome-related molecules in the EMT have been discussed. About the resolution of inflammation, the newly discovered pro-resolving lipids including RvTs are added and the receptors specifically acting on pro-resolving lipids have also been discussed. Understanding the action of natural anti-inflammatory compounds and pro-resolving lipids with anti-EMT activities might provide a new armory to suppress the progression of cancer.

EMT in Cancers
EMT shows the reduced expression of epithelial markers including E-cadherin and keratins and the increased expression of mesenchymal marker proteins such as vimentin and N-cadherin via actions of transcription factors including as SNAIL1 and ZEB1 (Figure 1) [6]. The molecular mechanism of EMT process has been well explained in the report (references in it) [1]. In this part, we will briefly update the concept that reflects recent achievements for EMT.
EMT occurs by a various mediators from tumor microenvironments via receptor through signal transduction. EMT-related transcription factors blocks the expression of epithelial cell-marker genes and evoke mesenchymal-marker genes. E-cadherin, keratin, ZO-1, miR-34, and miR-200 belong to the epithelial markers, and N-cadherin, vimentin, fibronectin, SNAIL, ZEB1, TWIST, Brachyury, Foxq1, Runx2, GATA, and SOX belong to the mesenchymal markers. P-cadherin is the marker of partial EMT. The dot triangle indicates the relative ratio of p-cadherin expression. Modified from Lee's report [6].

Upgrade of Inflammation Concept: From Initiation (alpha) to Resolution (omega)
Inflammation is intrinsically a protective process via microcirculation. Local or systemic inflammatory reactions delete the causing stimuli and reboot repair and healing processes of tissue [40].
Acute inflammation has two phases: initiation (alpha) and resolution (omega). Inflammation starts by the soluble inflammatory mediators such as complement, cytokines including chemokines, free radicals, vasoactive amines, and eicosanoids (including prostaglandins) by adjacent cells of the infected or injured part in the body [6,40,41].
Inflammasome is a multiprotein oligomer responsible for the activation of inflammatory responses and consists of NLRP protein such as NLRP3, ASC, and procaspase-1 [42,43]. The inflammasome can promote the maturation and secretion of interleukin 1β (IL-1β) and IL-18 [44].
Inflammasomes are involved in the EMT of cancer or other epithelial cells. For example, knockdown of NLRP3 alleviates high glucose or TGF-β1-induced EMT in human renal tubular cells [45]. NLRP3 regulates cellular proliferation and metastasis via EMT and the PTEN/AKT signaling pathway [46]. NLRP3 inhibition can attenuate silica-induced EMT in human bronchial epithelial cells [47]. NLRP3 also participates in the regulation of EMT in bleomycin-induced pulmonary fibrosis [48]. Uric acid can activate NLRP3 inflammasome in the EMT in the kidney of rats [49]. NLRP3 appears to be important for EMT since inflammasome-independent NLRP3 is enough to EMT in colon cancer cells [50].

EMT Inducers from Chronic Inflammatory Tumor Microenvironments
The tumor microenvironment mainly influences the progression of cancers via secretion of various factors that cause EMT [54,55]. Cancer-related chronic inflammation is described as a chaotic state where both pro-inflammatory and anti-inflammatory signals are present to permit tumor growth and immune evasion [6]. Besides, the tumor microenvironment contributes to the cancer heterogeneity. Therefore, the tumor microenvironment has been regarded as a promising target for the cure of cancer. Thus, Vanneman and Dranoff have demonstrated a novel way of curing cancer by re-educating the tumor microenvironment [56].
Here, we will briefly introduce EMT inducers from chronic inflammatory tumor microenvironment ( Figure 3).
IL-6 serves as either a pro-inflammatory or anti-inflammatory cytokine [6]. The mean serum concentration of IL-6 was observed as 31.7 pg/mL in patients with breast cancer and 3.3 pg/mL in the normal cohort [6,60]. IL-6 induces EMT of human breast cancer cells [61]. IL-6/STAT3-induced expression of lncTCF7 can promote EMT of liver cancer cells [62]. The IL-6 pathway induces EMT in biliary tract cancer via cross-talking to the SMAD4 in the TGF-1 pathway [63].
IL-8 is a chemokine mainly secreted by macrophages [64]. The mean serum concentration of IL-8 was found as 40.1 pg/mL in patients with breast cancer and 5.3 pg/mL in the normal group [60]. IL-8 expression is highly increased in TGF-1-induced EMT in colon carcinoma and nasopharyngeal carcinoma [65]. IL-8 is also involved in mast cell-induced EMT of human lung and thyroid cancer cells [66,67]. JAK2/STAT3/Snail pathway is involved in the IL-8-induced EMT of HCC cells [68]. Brachyury-induced EMT of the tumor is mediated by IL-8/IL-8R signaling pathway [69].
High-mobility group box 1 (HMGB1) is a nuclear DNA-binding protein and released to the outside from macrophages, NK cells, dendritic cells, necrotic cells, and apoptotic cells according to infection, injury, and inflammation [74]. The mean serum HMGB1 level was 4.64 ng/mL in patients with malignant breast cancer, which was remarkably higher than in patients with benign breast cancer (1.32 ng/mL) or in healthy subjects (1.36 ng/mL) [75]. HMGB1 (2 µg/mL) induces EMT of colorectal and prostate cancer cells via the RAGE/NF-κB pathway [76,77].
IL-10 is a potent anti-inflammatory cytokine that suppresses T cell/macrophage cytokine synthesis and blocks their antigen-presenting capacity [78]. In vitro generated M1-and M2-macrophages both can induce EMT of pancreatic cancer cells via the IL-10 signaling pathway [79].
PGE2 is biosynthesized from arachidonic acids ( Figure 2). The mean serum level of PGE2 was 6.324 pg/mL in patients with brain cancer and 1.677 pg/mL in the compared normal cohort [85]. PGE2 acts through G protein-coupled receptors such as EP1-EP4 [86]. PGE2 (2-10 µg/mL) works in an autocrine or paracrine manner, leading to stimulation of EMT through the expression of SNAIL and ZEB1 [87]. In turn, SNAIL induces blocking of degradation of PGE2 by repressing prostaglandin dehydrogenase, generating a positive loop that promotes cancer progression [88]. PGE2 (5 µM) enhances invasion of HCC cells via EP1-mediated expression of YB-1, which induces TGF-1-induced EMT by AKT activation [89]. PGE2 (0.01-1 µM) inhibits fibroblast chemotaxis but stimulates chemotaxis of epithelial cells in the airway [90]. PGE2 s inhibitory action against TGF-1-induced EMT seems to be via the EP2 pathway since EP2 agonist can block TGF-1-induced EMT [91]. As such, the effect of PGE2 on EMT varies depending on concentration and in cell types.
Leukotriene B4 (LTB4) can be produced via the 5-lipoxygenase pathway. Significantly higher levels of LTB4 have been found in the whole blood of lung cancer patients than those in the control group (44.1 vs. 17.9 pg/mL) [92]. LTB4 acts via two distinct GPCR called BLT1 and BLT2. BLT2 is involved in keratin phosphorylation and perinuclear reorganization, which is a prelude of EMT [93]. Accordingly, BLT2 is involved in the ras-promoted TGF-1-induced EMT [94]. As expected, LTB4 (100 nM) can induce EMT leading to vimentin expression through the BLT2/ERK2 activation [95].
Cysteinyl leukotrienes (CysLTs) include LTC4, LTD4, and LTE4 [96]. Mean serum levels of LTD4 found in HCC patients and healthy groups were 174.95 and 10.75 pg/mL, respectively [97]. Actions of CysLTs are mediated via GPCRs such as CysLT1 and CysLT2 [98]. LTD4 (100 nM) can suppress E-cadherin expression in cancer cells through enhanced translocation of -catenin to the nucleus while montelukast (0.1 mM) suppresses eosinophil-induced EMT in bronchial epithelial cells [99]. Recently, exosomes and cells from ascites in lung cancer patients can convert LTC4 to LTD4 to promote cancer cell migration and invasion via CysLT1 [100].

Reversal of EMT by Anti-inflammatory and Pro-Resolving Natural Compounds
The 2018 review by us did not cover the regulation of EMT by natural products [6]. In this part, we describe anti-inflammatory natural compounds and pro-resolving lipids that can prohibit EMT stimulated by mediators released in the tumor microenvironment.

Reversal of EMT by Anti-Inflammatory Natural Compounds
Diverse natural products can inhibit EMT of cancer cells or epithelial cells. We focused on EMT inhibitors found from natural compounds that can inhibit the EMT of cancer cells caused by EMT inducers from the chronic inflammatory tumor microenvironment due to limited space ( Figure 4, Table 1).   Arctigenin from Asteraceae has anti-inflammatory effects [101]. Arctigenin (12-50 µM) can suppress TGF-induced EMT of human lung cancer cells, thus blocking invasion [102]. Arctigenin (0.25-1 µM) can inhibit the expression of MCP-1 and subsequent EMT induced by ROS-dependent ERK/NF-κB pathway of renal tubular epithelial cells [103].
Baicalin found in Scutellaria baicalensis and Scutellaria lateriflora has well-known anti-inflammatory effects [104]. Baicalin (2 µM) can suppress TGF-1-mediated EMT in MCF10A cells by reducing the expression of slug [105]. Baicalin (12.5-25 µM) can inhibit the expression of TGF-1-induced EMT-related transcription factors in osteosarcoma cells and inhibit the aggressive metastasis of breast cancer by blocking EMT via inhibiting the activation of -catenin [106,107] Berberine found in Berberis can reduce the secretion of IL-1 and TNF-α [108,109]. Berberine (5-20 µM) can reverse EMT in uterine cancer, leading to suppression of cancer metastasis [110]. Berberine can inhibit the metastatic ability of prostate cancer cells by suppressing EMT-associated genes [111]. Berberine (50 µM) can make nasopharyngeal carcinoma cells sensitive to radiation through EMT inhibition [112].
Betanin from beets can decrease the production of superoxide anion and cytokines TNF-α and IL-1 [113,114]. Betanin (25-50 µM) can inhibit high glucose-induced EMT of renal proximal tubular cells [115]. However, activity of betanin for EMT of cancer cells has not been reported yet.
Brusatol from the seeds of Brucea sumatrana can inhibit the response of cultured beta-cells to pro-inflammatory cytokines in vitro [116]. Brusatol (2 µM) can inhibit the EMT of pancreatic cancer cells [117].
Celastrol (0.1-1 µM), from the root extracts of Tripterygium wilfordii and Celastrus regelii, can suppress experimental autoimmune encephalomyelitis [129]. Celastrol can inhibit the expression of snail and increased the expression of E-cadherin in the lung cancer cells [130]. Many studies have reported the effect of celastrol on EMT of diseases other than cancer (please refer the ref 128 by Kashyap et al.) [131].
Gallic acid is a trihydroxybenzoic acid found in gallnuts, sumac, witch hazel, and tea leaves [162]. For various anti-inflammatory effects of gallic acid, please refer to the 2017 review [163]. Gallic acid-coated sliver nanoparticle (50 µg/mL) can alter the expression of radiation-induced EMT in NSCLC [164]. Black tea polyphenols (10-40 µM) can also reverse EMT and inhibit invasion of human oral cancer cells [165].
Gedunin, one of the main chemical compounds in the neem tree, can protect TLR-mediated inflammation by suppression of inflammasome activation and cytokine production [169,170]. Gedunin (15 µM) suppresses EMT of pancreatic cancer by inhibiting sonic hedgehog signaling pathway [171].
Genistein first isolated from Genista tinctoria can suppress psoriasis-related inflammation through a STAT3/NF-κB-dependent mechanism in keratinocytes [172,173]. For more detailed information about genistein's anti-inflammatory action or anti-EMT, please refer to a previous review by Spagnulo et al. and Lee et al. [174,175]. Genistein (200 µM) can induce apoptosis of colon cancer cells by reversal of EMT via a notch1/NF-κB/slug/E-cadherin pathway [176]. miR-223 inhibitor and genistein (20 µM) can synergistically reverse in EMT of gemcitabine-resistant pancreatic cancer cells [177].
Glycyrrhizin (50-200 µM) from the roots of Glycyrrhiza glabra (Licorice) has anti-inflammatory and antiviral activities. It is also a novel pharmacological inhibitor of HMGB1 [192]. Glycyrrhizin can attenuate the EMT of prostate cancer cells by suppressing HMGB1-involved signaling pathway [192].
Honokiol from the Magnolia possesses anti-inflammatory activity by blocking downstream signaling of MEKK-1 in NF-κB activation pathway [193,194]. It (30 µM) can also inhibit EMT-mediated migration of human NSCLC cells in vitro by targeting c-FLIP and EMT of breast cancer cells by targeting STAT3/Zeb1/E-cadherin axis [195,196]. Honokiol (20 µM) can inhibit the metastasis of renal cancer cells by blocking EMT through regulating miR-141/ZEB2 pathway [197]. It (5-20 µM) can also inhibit the invasion of U87MG human glioblastoma cell via regulation of EMT [198].
Polyphyllin I, a component in the Rhizoma of Paris, can improve collagen-induced arthritis by blocking the inflammation response in macrophages through the NF-κB Pathway [253]. Polyphyllin I (0.3 µM) can overcome EMT-associated resistance to erlotinib in lung cancer cells via IL-6/STAT3 pathway inhibition [254].
Withaferin A from the Solanaceae family attenuates bleomycin-induced scleroderma by targeting FoxO3a and NF-κB signaling [276]. Withaferin A (0.5 µM) can inhibit the EMT of NSCLC cells [277]. It (2 µM) can also inhibit the EMT of MCF10A cells and suppress vimentin expression in breast tumors [278]. Extracts of root in Withania somnifera can suppress mammary EMT and cancer metastasis [279].
MaR1 (0.1-10 ng/mice), a docosahexaenoic acid-derived pro-resolution lipid, can protect skin from inflammation and oxidative stress caused by UVB irradiation [302]. Receptor for MaR1 is yet unknown. The effect of MaR1 on EMT of cancer cells has not been reported yet except one study has shown that incubating DHA with A549 lung cancer cells can produce MaR1 (1.58 ng/mL) and PD1 (1.67 ng/mL) [303]. MaR1 (10 nM) can inhibit TGF-1-induced proliferation, migration, and differentiation in human lung fibroblasts [304].
PD1 has anti-inflammatory and survival effects on neuronal diseases such as Alzheimer's disease and retinal degenerations [305]. PDX (1-100 nM), one of PD1 derivatives, can suppress bleomycin-induced lung fibrosis through blocking EMT [6,306]. However, it is hard to find reports about the effect of PDs on the EMT of cancer cells [6,306]. However, it is hard to find reports about the effect of PDs on the EMT of cancer cells.
Pro-resolving lipid classes are mainly originated from ω-3 fatty acids which constitute a group of essential fats that humans cannot synthesize endogenously [307]. Several pro-resolving lipids can be obtained by total synthesis. Studies on derivatives are in progress.
It is possible to produce pro-resolving lipids by biological methods. EPA and DHA are the major long chain ω-3 fatty acids in the diet. Algae are the major producers of EPA and DHA in the ecosystem. Therefore, fish that consumes algae contains a lot of EPA and DHA [307]. Accordingly, algae or fish might be a starting point for isolating or producing pro-resolving lipids in industrial scale. Brain cells of rainbow trout (Oncorhynchus mykiss) can produce novel DHA-derived Rvs and PDs (Table 2) [308]. However, baking reduces proportions of PG, Rv, and hydroxy-fatty acid in farm-raised Atlantic salmon (Salmo salar; Table 2) [309].
Infectious organisms can produce pro-resolving lipids to control host inflammation. Thus, supraphysiological levels of LxA4 are generated during infection by Toxoplasma gondii, which in turn reduces IL-12 production by dendritic cells, thus dampening Th1-type cell-mediated immune responses (Table 2) [310].
Candida albicans can modulate host defense by biosynthesizing the pro-resolving lipid RvE1 [311]. C. albicans can biosynthesize nanogram quantities of RvE1 from EPA without collaboration of other cellular partners. It can also biosynthesize PDs (Table 2) [311]. Trypanosoma cruzi is a protozoan parasite that causes Chagas disease and produces the RvD1, RvD5, and RvE2 (Table 2) [312]. These reports suggested that algae, fish, and some infectious organisms might be applied to produce pro-resolving lipids.

Perspectives
Several natural compounds have anti-inflammatory activities and/or anti-EMT activities. However, few reports have explained their anti-EMT activity by anti-inflammatory or pro-resolving mechanisms. Therefore, examining anti-EMT activities of natural compounds based on their anti-inflammatory or pro-resolving activities and assuring their anti-EMT activities in vivo might be important in the future. Especially, it might be a reasonable way to study the anti-EMT activity of natural products through interaction with the tumor microenvironment [255].
The flux of inflammatory or pro-resolving lipids from arachidonic acid related pathway ( Figure 2) could change if one pathway is blocked by natural anti-inflammatory or pro-resolving compounds. Therefore, prevention of the production of these lipids might influence levels of other inflammatory and pro-resolving lipids. Thus, further studies on effects of anti-inflammatory compounds or pro-resolving lipids on EMT of cancer might require entire profiles of lipid metabolites affecting the EMT.

Conclusions
Expansion of studies about other pro-resolving lipids and nature-derived pro-resolving compounds to inhibit EMT are needed. To do this, various types of pro-resolving lipids and new compounds from natural sources should be procured and diffused to researcher without barrier. Thus, collaboration of natural chemistry researchers with other fields is also required.