Curcumin attenuates proangiogenic and proinflammatory factors in human eutopic endometrial stromal cells through the NF‐κB signaling pathway

Abstract Endometriosis is a chronic gynecological inflammatory disorder in which immune system dysregulation is thought to play a role in its initiation and progression. Due to altered sex steroid receptor concentrations and other signaling defects, eutopic endometriotic tissues have an attenuated response to progesterone. This progesterone‐resistance contributes to lesion survival, proliferation, pain, and infertility. The current agency‐approved hormonal therapies, including synthetic progestins, GnRH agonists, and danazol are often of limited efficacy and counterproductive to fertility and cause systemic side effects due to suppression of endogenous steroid hormone levels. In the current study, we examined the effects of curcumin (CUR, diferuloylmethane), which has long been used as an anti‐inflammatory folk medicine in Asian countries for this condition. The basal levels of proinflammatory and proangiogenic chemokines and cytokines expression were higher in primary cultures of stromal cells derived from eutopic endometrium of endometriosis (EESC) subjects compared with normal endometrial stromal cells (NESC). The treatment of EESC and NESC with CUR significantly and dose‐dependently reduced chemokine and cytokine secretion over the time course. Notably, CUR treatment significantly decreased phosphorylation of the IKKα/β, NF‐κB, STAT3, and JNK signaling pathways under these experimental conditions. Taken together, our findings suggest that CUR has therapeutic potential to abrogate aberrant activation of chemokines and cytokines, and IKKα/β, NF‐κB, STAT3, and JNK signaling pathways to reduce inflammation associated with endometriosis.

Hormone-dependent immune system dysregulation through aberrant production of chemokines and cytokines may play a role in endometriosis initiation and progression (Klemmt, Carver, Kennedy, Koninckx, & Mardon, 2006;Meola et al., 2010;Mu et al., 2008;Reis et al., 2013;Ulukus et al., 2009). The symptoms of endometriosis contribute substantially to the burden of disease and add substantial cost to society through expensive surgical and medical therapies and reduced economic productivity (Nnoaham et al., 2011;Simoens et al., 2012;Simoens, Hummelshoj, & Dhooghe, 2007). A definitive diagnosis of endometriosis requires the surgery and the treatment which include pharmacological or operative approaches (Gordts, Puttemans, Gordts, & Brosens, 2015). Unfortunately, current hormonal therapies, including synthetic progestins, GnRH agonists, and danazol are often with limited efficacy and counterproductive to fertility, and they commonly cause systemic side effects (vasomotor symptoms and osteopenia) because of suppression of endogenous steroid hormone levels (Vercellini et al., 2014). Therefore, developing new ways of treating endometriosis using nonhormonal drugs has been a subject of intense investigation. Although many researchers have evaluated individual chemokines and cytokines in patients with endometriosis and matching controls, their exact role in the pathogenesis of the disease remains unsolved (Borrelli, Abrao, & Mechsner, 2014;Borrelli, Hediger et al., 2013;Khorram, Taylor, Ryan, Schall, & Landers, 1993;Li, Luo et al., 2012;Margari et al., 2013). Chemokines represent a family of small cytokines or proteins released by cells, especially lymphocytes, which are capable of inducing chemotaxis (directed movement through chemical gradients in the microenvironment) of nearby responsive cells as a means of directing cellular migration (Burney & Giudice, 2012;Minici et al., 2008;Reis et al., 2013;Vercellini et al., 2014). Inflammatory chemokines are released from a variety of cells in response to bacterial and viral infections and other pathogenic agents that act under both physiological conditions and pathological processes (Acker, Voss, & Timmerman, 1996;Borrelli et al., 2014;Borrelli, Hediger et al., 2013;Lira & Furtado, 2012).
Interestingly, recent studies have demonstrated that dietary supplements of CUR in combination with standard therapies may lead to the improvement of the regular medical treatment of endometriosis (Signorile, Viceconte, & Baldi, 2018). However, there are no detailed studies of chemokines and cytokines expression profiles in human endometrial stromal cells (ESCs) from normal women and those affected with endometriosis, particularly with respect to the effects of CUR on the secretion of these proteins. Therefore, our current experimental studies were designed to quantify and compare the secretion of chemokine and cytokine from normal endometrial stromal cells (NESC) with that from eutopic endometrium of endometriosis subjects (EESC). We also sought to analyze the ability of CUR to alter chemokines and cytokines secreted from these cells. Immunoblot studies were carried out under various experimental conditions to analyze the phosphorylation and total expression levels of selective inflammatory signaling molecules including inhibitor of nuclear factor κ-B kinase subunit α/β (IKKα/β), NF-κB, signal transducer and activator of transcription 3 (STAT3), and c-Jun N-terminal kinases (JNK), which are common proinflammatory signaling molecules and upregulated during inflammation (Hoesel, & Schmid, 2013;Huminiecki, Horbańczuk, & Atanasov, 2017;Israël, 2010).

| Human subjects and tissue acquisition
The current study was approved by the institutional review boards of the Emory and Morehouse Schools of Medicine, Atlanta. Primary ESCs were obtained from reproductive age women with regular menstrual cycles, and who had not received hormonal therapy for at least 3 months before laparoscopic surgery (Yu et al., 2014). Written Histological confirmation of ectopic glands and stroma was confirmed in all endometriosis cases.
The intracellular uptake of CUR in NESCs and cells derived from EESCs, and their survival status. Cells were cultured and treated with or without CUR (1, 5, 10, 20, and 40 µg/ml) for 24, 48, and 72 hr in DMEM/Ham's F-12 media with 5% exosome-depleted fetal bovine serum. (a) ESCs were fixed and stained with Hoechst 33248 to identify nuclei. Data represent the percentage of cells displaying morphological alteration of apoptosis based on quantification of nuclear morphologic changes. At least 250-300 cells were counted for each data point. The bar graph represents the mean ± SEM of results from three independent experiments. Significant (p ≤ 0.05) differences are represented with star "*" and compared to the parallel control group. (b) To assess if morphological changes occur in cells, live cell photographs were taken under an inverted epifluorescence microscope to image the green fluorescence signals for the CUR autofluorescence or the control (untreated) group alone along with phase contrast pictures at ×200 magnification at 48 hr posttreatment. Inset images are at a higher magnification, demonstrating CUR autofluorescence. Data are representative of three individual experiments (n = 3) from eutopic endometrial biopsies from three subjects with and three without evidence of endometriosis that were performed for each of the two patient groups. CUR: curcumin; DMEM: Dulbecco's modified Eagle's medium; EESC: eutopic endometrium of endometriosis subjects; ESCs: endometrial stromal cells; NESCs: normal human endometrial stromal cells [Color figure can be viewed at wileyonlinelibrary.com]

| CUR treatment of NESC and EESCs
ESC cultures were grown to 95-100% confluence in six-well plates (Fisher Scientific, Hampton, NH). Cells were treated with CUR (molecular weight = 368.41, purity = 99%; Sigma-Aldrich) at a concentration of 1, 5, 10, 20, and 40 μg/ml for 24, 48, and 72 hr. F I G U R E 2 CUR attenuated proinflammatory interleukin secretion from human NESCs and cells derived from EESCs, but not IL-10 or IL-12. Cells were treated with or without CUR (5 µg/ml or 10 µg/ml) for 24 and 48 hr in DMEM/Ham's F-12 media with 5% exosome-depleted fetal bovine serum. Basal levels of chemokines and cytokines were measured and analyzed in the supernatants using Bio-Plex Pro Human Cytokine, Chemokine, and Growth Factor Magnetic Bead-Based Assays, coupled with the Luminex 200 system. All bar graphs represent the mean ± SEM of results from three individual experiments (n = 3) from eutopic endometrial biopsies from three subjects with and three without evidence of endometriosis. The superscript "a" represents significant differences (p ≤ 0.05) in EESCs groups compared with respective NESCs groups at 24 and 48 hr. Star (*) represents significant differences (p ≤ 0.05) in EESCs groups treated with CUR compared with respective NESCs groups treated with CUR at 24 and 48 hr. CUR: curcumin; DMEM: Dulbecco's modified Eagle's medium; EESCs: eutopic endometrium of endometriosis subjects; NESCs: normal endometrial stromal cells T A B L E 2 Two-way ANOVA analysis of CUR dose and duration effects and dose-duration interaction on proinflammatory cytokines and chemokines secretion in human normal endometrial stromal cells (NESCs) and cells derived from eutopic endometrium of endometriosis subjects (EESCs) in vitro  Note. CCL11: chemokine eotaxin; FGF: fibroblast growth factors; G-CSF: granulocyte-colony stimulating factor; GM-CSF: granulocyte-macrophage colony stimulating factor; IFNγ: interferon γ; IL: interleukin; IP-10/CXCL10: interferon γ-induced protein 10; MCP-1/CCL2: monocyte chemotactic protein-1; MIP-1α/CCL3: macrophage inflammatory proteins 1α; PDGF: platelet-derived growth factor; TNF-α: tumor necrosis factor-α; VEGF: vascular permeability factor/vascular endothelial growth factor. F represents degrees of freedom numerator (Dfn) and degrees of freedom denominator (Dfd) for each group.

| Assessment of chemokines and cytokines in secretion media
To determine the effects of CUR treatment on NESC and EESC, cytokine and chemokine levels were measured in conditioned media.

| Intracellular uptake of CUR in normal and EESCs
We used our well-established cell culture model of endometriosis to understand the differential chemokine and cytokine secretory capacity of the cells. Given that the bioavailability of natural CUR is low (Lee et al., 2013;Shen & Ji, 2012), therefore, we first determined the optimum concentration and its intracellular uptake in ESCs (Figure 1). ESCs were grown to 95-100% confluence and treated with different doses (1, 5, 10, 20, and 40 μg/ml) of CUR for 24, 48, and 72 hr. As shown in Figure 1a,  These results further suggest that EESCs are significantly more resistant to cell death compare to NESCs (Dmowski, Gebel, & Braun, 1998). Thus, based on these results we selected 5 and 10 μg/ml dose for all other experimental studies.
In addition, we determined cell morphology under various experimental conditions. Phase contrast photomicrograph pictures ( Figure 1b) showed that both NESC and EESC have classical mesenchymal characteristics with spindle shaped morphology and oval or round nuclei when grown in exosome free low serum media.
As previously reported, under basal conditions there were no significant apparent morphological differences observed between NESC and EESC (Yu et al., 2014). After treatment with CUR for 48 hr, a dose-dependent increase in green autofluorescence was noted, confirming that CUR was absorbed intracellularly.

| Differential secretion of chemokines and cytokines in NESCs versus EESCs
As shown in Figures 2, 3 and Table 2, most chemokines and cytokines were secreted in significantly (p ≤ 0.05) higher concentrations by EESC compared with NESC at 24 and 48 hr. Some proteins, for example, VEGF, MIP-1β, and IFNy were at or below the limit of detectability in media from NESC at 24 hr; and IL-17 was completely absent in media from NESC at 24 and 48 hr.
IL-2, IL-5, IL-9, GM-CSF, and PDGF were not detected in culture media from either EESC or NESC at 24 and 48 hr (not shown).
Consistent with previous reports, several chemokines and cytokines were highly overexpressed in EESC (e.g., IL-6, IL-8, IP-10, G-CSF, MCP-1, and RANTES were orders of magnitude higher than other chemokine and cytokines in EESC). By contrast, under basal conditions, IL-10 and IL-12 expression were not different between EESC and NESC.
Therefore, we evaluated the expression and phosphorylation of STAT3 and JNK in normal and endometriotic ESCs (Figure 5a,b and Table 3). The levels of phosphorylated STAT3 and JNK were significantly (p ≤ 0.05) higher in EESCs compared with NESCs at 24 and 48 hr. Interestingly, CUR treatment significantly inhibited (p ≤ 0.05) the phosphorylation of STAT3 and JNK in a dose-and time-dependent manner in EESCs.
Moreover, CUR treatment also significantly (p ≤ 0.05) decreased the overall expression of JNK (Figure 5a,b and Table 3).

| DISCUSSION
In the current study, we performed a systematic assessment of chemokine and cytokine secretion and confirmed that many of these autacoids are differentially expressed by stromal cells derived from EESC subjects, relative to women without the disease. Interestingly, CUR treatment renders normalization of these proteins, in many cases to the basal secretion levels observed in NESCs. It is wellestablished that eutopic endometrial cells function differently in women with endometriosis compared with a normal endometrium in disease-free women (Burney et al., 2007). These cells are resistant to apoptosis and have other selective advantages for survival outside the uterine cavity, which lead to their implantation and invasion of the peritoneum and other ectopic sites (Dmowski et al., 1998). The detailed identification of molecular differences in the eutopic endometrium of women with endometriosis is an important step toward understanding the pathogenesis of this condition and developing effective strategies for the treatment of its associated infertility and pain. Therefore, we hypothesized that an increase in chemokines, cytokines, and/or, growth factors produced in eutopic endometrial tissue from women with endometriosis may contribute to increases in angiogenesis and proliferation.
Our results indicate that EESCs have an increased basal production of almost all the selected proinflammatory and proangiogenic chemokines and cytokines (except IL-10) and that they can promote a chronic inflammatory environment within the pelvis of these women (Vercellini et al., 2014). Also, a large body of evidence indicates that TNF-α, IL-1β, IFN-γ, IL-6, IL-8, eotaxin, and RANTES are involved in recruitment and activation of macrophages, neutrophils, eosinophils, basophils, monocytes, and NK-cell to the sites of endometriosis, thus promoting inflammatory changes and enhance angiogenesis through increase production of VEGF (Reis et al., 2013).
Several hormonal treatments and analgesics are available to endometriosis patients suffering pain (Vercellini et al., 2014). The current medical strategies for endometriosis management involve inhibition of ovulation, abolition of menstruation, and achievement of a stable steroid hormone milieu (Vercellini et al., 2014). Creation of hypoestrogenic (GnRH agonists), hyperandrogenic (danazol, gestrinone), or hyperprogestogenic (oral contraceptives, progestins) environments result in the suppression of endometrial and endometriosis cell proliferation. However, serious side effects (vasomotor symptoms, mood instability, and negative calcium balance) and unfavorable changes in serum cholesterol lipoprotein distribution (HDL levels decrease and LDL levels increase) are associated with these therapies. Thus, we chose to evaluate the effects of CUR, a natural, medicinal Asian herb, on proinflammatory and proangiogenic chemokine and cytokine secretion in EESCs and NESCs. Our findings reveal that CUR is a potent inhibitor of proinflammatory and proangiogenic chemokine and cytokine secretion from these cells.
By contrast, IL-10 and IL-12, which themselves have anti-inflammatory properties, were upregulated by CUR, particularly in EESCs.
Interestingly, the biological actions of these two ILs include inactivation of macrophages and inhibition of proinflammatory and proangiogenic cytokines and chemokines.
F I G U R E 4 Effects of CUR on phosphorylation and total expression of IKKα, IKKβ, and NF-κB proteins in human NESCs and cells derived from EESCs subjects. Cells were treated with or without curcumin (CUR, 5 µg/ml or 10 µg/ml) for 24 and 48 hr in DMEM/Ham's F-12 media with 5% exosome-depleted fetal bovine serum. Total protein was isolated, followed by equal amounts of protein (25 μg) from each sample separated by one-dimensional gel electrophoresis and analyzed for phospho-IKKα, phospho-IKKβ, and phospho-NF-κB; and total IKKα, IKKβ, and NF-κB protein.
(a) Representative western blot analysis of protein for phospho-and total IKKα, IKKβ, and NF-κB levels in NESCs and EESCs treated with or without CUR. α Tubulin was used as an internal constitutive control. (b) Bar diagrams represent the densitometric analyses of protein in WBs of three independent experiments (n = 3) as mean ± SEM that were performed for each individual group. The bar graphs represent the ratios of phospho-IKKα, phospho-IKKβ, and phospho-NF-κB protein levels normalized to total IKKα, IKKβ, and NF-κB, respectively, and the ratios of total IKKα, IKKβ, and NF-κB protein levels, normalized to α tubulin. The superscript "a" represents significant differences (p ≤ 0.05) in EESCs groups compared with respective NESCs groups at 24 and 48 hr. Star (*) represents significant differences ( (Hoesel, & Schmid, 2013;Huminiecki et al., 2017;Israël, 2010). IKKα and IKKβ are part of a multiprotein complex involved in F I G U R E 5 Effects of CUR on phosphorylation and total expression of STAT3 and JNK proteins in human NESCs and cells derived from EESCs subjects. Cells were treated with or without curcumin (CUR, 5 µg/ml or 10 µg/ml) for 24 and 48 hr in DMEM/Ham's F-12 media with 5% exosome-depleted fetal bovine serum. Total protein was isolated, followed by equal amounts of protein (25 μg) from each sample separated by one-dimensional gel electrophoresis and analyzed for phospho-STAT3 and phospho-JNK; and total STAT3 and JNK protein. (a) Representative WBs analysis of protein for phospho-and total STAT3 and JNK levels in NESCs and EESCs treated with or without CUR. α Tubulin was used as an internal control. (b) Bar diagrams represent the densitometric analyses of specific protein bands in WBs of three independent experiments (n = 3) as mean ± SEM that were performed for each individual group. The bar graphs represent the ratios of phospho-STAT3 and phospho-JNK protein levels normalized to total STAT3 and JNK, respectively, and the ratios of total STAT3 and JNK protein levels normalized to α tubulin. The superscript "a" represents significant differences (p ≤ 0.05) in EESCs groups compared with respective NESCs groups at 24 and 48 hr. Star (*) represents significant differences (p ≤ 0.05) in EESCs groups treated with CUR compared with respective NESCs groups treated with CUR at 24 and 48 hr. CUR: curcumin; DMEM: Dulbecco's modified Eagle's medium; EESCs: eutopic endometrium of endometriosis; JNK: c-Jun N-terminal kinases; NESCs: normal endometrial stromal cells; STAT3: signal transducer and activator of transcription 3; WBs: western blots mediating transcription of multiple chemokine and cytokine genes through Ikβ (Hoesel, & Schmid, 2013;Huminiecki et al., 2017;Israël, 2010). JNK is a member of the mitogen-activated protein kinase family and cytokine/chemokine-dependent phosphorylation of JNK modifies the activity of numerous proteins that reside or act in the mitochondria or nucleus. Downstream molecular targets of JNK regulate several important cellular functions including cell growth, differentiation, and survival. Similarly, in response to cytokines, chemokines, and growth factors, STAT3 is phosphorylated by receptor-associated Janus kinases (JAK), form homodimers or heterodimers, and translocate to the cell nucleus where they act as transcription activators and promote cell proliferation and differentiation (Hoesel, & Schmid, 2013;Huminiecki et al., 2017;Israël, 2010). Our results further demonstrated that CUR treatment of EESCs completely inhibited or eliminated phosphorylated forms of JNK and STAT3, along with IKKα, IKKβ, and NF-κB. Thus, our results are consistent with reports showing that CUR has strong antiproliferative, anti-inflammatory, and antiangiogenic properties (Beevers, & Huang, 2011;Lee et al., 2013;Shen, & Ji, 2012). Moreover, CUR has no negative effects on serum cholesterol and lipoproteins and does not promote the development of a pseudopregnant endocrine state (Beevers, & Huang, 2011;Lee et al., 2013;Shen, & Ji, 2012).
Together, CUR's anti-inflammatory, antiproliferative, and proapoptotic effects may offer a well-tolerated alternative to standard, approved antiendometriosis medications. Nonhormonal therapeutics of plant origin may be especially suitable for young women with severe symptoms of endometriosis-associated pain who require a protracted duration of therapy ( Figure 6). Additional experiments are underway to elucidate CUR's precise anti-inflammatory and potential fertility-preserving effects in in vivo models of endometriosis. Our findings support the investigation of novel, natural, and safe therapeutics for future endometriosis treatment.

CONFLICTS OF INTEREST
The authors declare that there are no conflicts of interest.