**Combined E**ff**ects of Eicosapentaenoic Acid and Adipocyte Renin–Angiotensin System Inhibition on Breast Cancer Cell Inflammation and Migration**

**Fahmida Rasha 1,2, Chanaka Kahathuduwa 2,3, Latha Ramalingam 1,2, Arelys Hernandez 1, Hanna Moussa 2,4 and Naima Moustaid-Moussa 1,2,\***


Received: 30 November 2019; Accepted: 11 January 2020; Published: 16 January 2020

**Abstract:** Obesity is a major risk factor for breast cancer (BC). Obesity-related metabolic alterations such as inflammation and overactivation of the adipose renin–angiotensin system (RAS) may contribute to the progression of BC. Clinically used antihypertensive drugs such as angiotensin-converting enzyme inhibitors (ACE-I) and dietary bioactive components such as eicosapentaenoic acid (EPA) are known for their anti-inflammatory and adipose RAS blocking properties. However, whether EPA enhances the protective effects of ACE-I in lessening adipocyte inflammation on BC cells has not been studied. We hypothesized that combined EPA and ACE-I would attenuate BC cell inflammation and migration possibly via adipose RAS inhibition. To test our hypothesis, we examined the (i) direct effects of an ACE-I (captopril (CAP)) or EPA, individually and combined, on MCF-7 and MDA-MB-231 human BC cells, and the (ii) effects of conditioned medium (CM) from human adipocytes pretreated with the abovementioned agents on BC cells. We demonstrated that CM from adipocytes pretreated with EPA with or without captopril (but not direct treatments of BC cells) significantly reduced proinflammatory cytokines expression in both BC cell lines. Additionally, cell migration was reduced in MDA-MB-231 cells in response to both direct and CM-mediated CAP and/or EPA treatments. In summary, our study provides a significant insight into added benefits of combining anti-inflammatory EPA and antihypertensive ACE-I to attenuate the effects of adipocytes on breast cancer cell migration and inflammation.

**Keywords:** obesity; breast cancer; renin–angiotensin system; eicosapentaenoic acid; adipocyte inflammation

#### **1. Introduction**

Breast cancer (BC) is the most common type of cancer among U.S. women and has a lifetime risk of more than 12% [1]. Obesity is one of the major modifiable risk factors for BC, especially in postmenopausal women, and is associated with poor cancer outcomes and survival in patients with BC [2,3]. Moreover, as inflammation is an important underlying basis for both diseases, it is critical to understand its involvement in mechanism(s) of obesity-related BC [4]. In obesity, adipose tissue expands and becomes hypertrophic and hypoxic, with invasion of proinflammatory M1-type macrophages and exaggerated secretion of protumor adipocytokines [5]. Various proinflammatory adipocytokines such as interleukin (IL)-6, IL-8, and monocyte chemoattractant protein-1 (MCP-1) can promote the activation and transcription of many protumor signaling pathways such as nuclear factor kappa B (NF-κB) and signal transducer and activator of transcription 3 (STAT3), resulting in increased migratory and invasive capacities in the adjacent cancer cells [6]. Additionally, cancer-associated adipocytes form a proinflammatory microenvironment via induced macrophage invasion and thus synergistically activate STAT3 and NF-κB pathways [7], which in turn leads to apoptosis inhibition and cancer cell proliferation [8]. Moreover, aberrant STAT3 signaling also upregulates angiogenic vascular endothelial growth factor (VEGF), leading to higher tumor angiogenesis [9]. Hence, obesity-associated proinflammation increases cancer cell progression via their interaction with adjacent adipocytes [6]. Furthermore, obese adipocytes also overexpress metabolic markers such as fatty acid synthase (FASN), which is associated with adipocyte hypertrophy and increased proinflammatory adipocytokine secretion in obesity [10,11]. FASN provides an alternative energy source to proliferating cancer cells in the form of free fatty acids helping their growth and proliferation [12]. This, in turn, facilitates the formation of a toxic tumor microenvironment (TME) in the breast, which contributes to obesity-associated BC progression [13].

The renin–angiotensin system (RAS) is classically known to regulate fluid balance and blood pressure. Angiotensinogen (Agt) is the main precursor protein for the bioactive hormone angiotensin II (Ang II). Agt is first cleaved by renin into Ang I, which is converted into Ang II by the angiotensin-converting enzyme (ACE). Ang II mediates its effect via binding to two major G-protein-coupled receptors, namely, type 1 receptor (AT1R) and type 2 receptor (AT2R) [14,15]. Adipose RAS activation has multiple roles in obesity, such as increasing inflammatory, prothrombotic, and angiogenic markers [16] that can also promote tumor growth, proliferation, and invasion in the breast [17,18]. In addition, Ang II/AT1R signaling triggers production and infiltration of tumor-associated macrophages (TAM) in different tumor models, while RAS inhibitors can restrain tumor growth and TAM response [19]. Thus, RAS is a potential key player in adipocyte–breast-cancer-cell interactions. RAS inhibitors such as angiotensin-converting enzyme inhibitors (ACE-I) and AT1R blockers (ARB) have recently gained interest as possible anticancer agents in both clinical and preclinical models of BC [17,18,20,21]. Some ACE-I and ARBs were found to reduce BC risk in a time-dependent manner in BC patients [21].

RAS can also be downregulated by anti-inflammatory dietary bioactive compounds such as omega-3 polyunsaturated fatty acids (n-3 PUFA) [22], which we have also shown to reduce Agt secretion from adipocytes [23]. The dietary n-3 PUFAs eicosapentaenoic acid (EPA)-20:5n-3 and docosahexaenoic acid (DHA)-22:5n-3 have potent anti-inflammatory properties that include reducing obesity-associated systemic and adipose tissue inflammation, thus alleviating TME formation in adipocyte–BC-cell crosstalk [24–27]. Therefore, given the promising anti-inflammatory and anticancer properties of ACE-I and EPA, we hypothesized that ACE-I (captopril (CAP)) and EPA alone or in combination would attenuate BC cell inflammation and migration, in part through inhibition of adipocyte RAS. We examined both direct and adipose CM-mediated effects of captopril and EPA and their combination on cultured BC cell genes and proteins as well as cell migration using wound healing analyses.

#### **2. Results**

#### *2.1. E*ff*ect of Captopril and EPA on Markers of Fatty Acid Synthesis and Inflammation in BC Cells and Role of Human Adipocyte-Conditioned Media (CM)*

MDA-MB-231 and MCF-7 cells were treated with 100 μm of EPA or 100 μm of CAP with or without EPA either directly or using CM from CAP, EPA, or both pretreated adipocytes for 48 h. Direct treatments with EPA, CAP, or CAP + EPA did not alter most measured markers of cancer cell growth and inflammation in MDA-MB-231 cells, with the exception of IL-8, which was significantly reduced by EPA and CAP + EPA treatments for 48 h in MDA-MB-231 cells (*p* < 0.05) compared with control (CT), while CAP alone had no effect. However, EPA and CAP + EPA had comparable effects, indicating no additional effect of direct CAP + EPA combination on BC cell inflammatory markers

(Figure 1E). Treatment of MDA-MB-231 cells with human adipocyte CM significantly increased all tested markers of cell growth and inflammation after 48 h, as shown by increased mRNA levels of FASN, STAT3, NF-κB, IL-6, and IL-8 compared with control (Figure 1A–E; also see Tables S1–S7) (*p* < 0.05). However, treatment of MDA-MB-231 cells with CM from adipocytes pretreated with EPA, CAP, and their combination significantly reduced the mRNA content of all measured markers of BC cell growth and inflammation compared with treatment with CM derived from untreated adipocytes (Figure 1A–E) (*p* < 0.05). However, no changes were observed in FASN, STAT3, NF-κB, and IL-8 mRNA transcription levels in MDA-MB-231 BC cells treated with CAP + EPA pretreated adipocyte CM, compared with CAP-CM or EPA-CM (Figure 1A–C,E) (*p* < 0.05). Interestingly, CAP + EPA pretreated CM reduced IL-6 mRNA levels to a greater extent in MDA-MB-231 cells compared with CAP-CM and/or EPA-CM treatments, indicating potential additive anti-inflammatory effects of CAP and EPA combination (Figure 1D) (*p* < 0.05). Exploratory factorial regression analyses performed to examine the interactions of CAP and EPA when administered as a combination resulted in significant negative regression coefficients for CM-CAP and CM-EPA factors but significant positive CM-CAP × EPA interaction for mRNA levels of all measured markers of MDA-MB-231 cell growth and inflammation (Tables S1–S5). This result suggests that CAP and EPA may act via a common pathway in reducing mRNA expression in CM-treated MDA-MB-231 cells.

**Figure 1.** Eicosapentaenoic acid (EPA) and captopril (CAP) (angiotensin-converting enzyme inhibitors; ACE-I) effects on mRNA expression in MDA-MB-231 cells. MDA-MB-231 cells were treated with 100 μm of CAP with or without 100 μm of EPA for 48 h. Human mesenchymal stem cells (HMSCs) were differentiated into adipocytes and treated with 100 μm of CAP with or without 100 μm of EPA for 24 h. Conditioned media (CM) was collected and transferred to breast cancer (BC) cells for 48 h. Cells were harvested and changes in mRNA levels of fatty acid synthase (FASN) (**A**), signal transducer and activator of transcription 3 (STAT3) (**B**), nuclear factor kappa B (NF-κB) (**C**), interleukin (IL)-6 (**D**), and IL-8 (**E**) were measured (*p* < 0.05; N = 3; three replicates under each treatment group; bars with different letters (a, b, c) indicate significance).

On the other hand, CM from human adipocytes significantly increased markers of cell growth and inflammation in MCF-7 cells after 48 h, as shown by increased mRNA levels of FASN, STAT3, NF-κB, and IL-8 compared with CT (*p* < 0.05), while CM from adipocytes pretreated with EPA, CAP, and CAP + EPA significantly reduced the abovementioned markers of cell growth and inflammation after 48 h compared with CM-control (Figure 2A–C,E; Tables S8–S14) (*p* < 0.05). However, no changes in the mRNA levels of the respective markers were observed between CAP and EPA treated groups with or without CAP–EPA combination. Additionally, direct treatments with EPA and CAP + EPA significantly reduced MCF-7 BC cell inflammation, as demonstrated by significantly lower IL-6 and IL-8 mRNA transcription levels, while direct treatments with CAP reduced only IL-6 mRNA levels after 48 h compared with control in MCF-7 cells (Figure 2D,E) (*p* < 0.05). However, the changes were not significant between EPA and CAP + EPA treated groups, indicating no additional effects of CAP–EPA combination in MCF-7 cells compared with EPA alone or CAP alone.

**Figure 2.** EPA and captopril (ACE-I) effects on mRNA expression in MCF-7 cells. MCF-7 cells were treated with 100 μm of CAP with or without 100 μm of EPA for 48 h. HMSCs were differentiated into adipocytes and treated with 100 μm of CAP with or without 100 μm of EPA for 24 h. CM was collected and transferred to BC cells for 48 h. Cells were harvested and mRNA level changes of FASN (**A**), STAT3 (**B**), NF-κB (**C**), IL-6 (**D**), and IL-8 (**E**) were measured (*p* < 0.05; N = 3; three replicates under each treatment group; bars with different letters (a, b, c) indicate significance).

Next, we measured proinflammatory IL-6 and IL-8 protein levels secreted by both MDA-MB-231 and MCF-7 cells after 48 h of EPA and CAP treatments. Medium was collected from both cell lines after direct and adipocyte CM (with or without EPA, CAP, or combined EPA and CAP) treatments for 48 h. Consistent with the above-described results (Figure 3A,B), no changes were observed in secreted IL-6 and IL-8 levels in MDA-MB-231 cells with direct treatments of EPA, CAP, CAP + EPA for 48 h. In addition, increased IL-6 and IL-8 secretion in MDA-MB-231 cells were observed in response to adipocyte CM treatments compared with CT and direct CAP ± EPA treatments (*p* < 0.05). Moreover, significant reductions in both IL-6 and IL-8 protein levels were identified in MDA-MB-231 cells in response to CM-EPA and CM-CAP treatments (with or without EPA) compared with CM-control (*p* < 0.05) (Figure 3A,B). However, the changes were not significant between CM-EPA and CM-CAP groups as well as compared to CT and direct CAP ± EPA treatments after 48 h (*p* < 0.05). Additionally, both IL-6 and IL-8 protein levels were undetectable (\* nd) without CM treatments in MCF-7 cells, as reported earlier by others [28] (Figure 3C,D). Interestingly, adipocyte CM with CAP + EPA significantly reduced IL-6 secretion in MCF-7 cells compared with CM-control treatment, but no changes were observed compared to CM-EPA or CM-CAP individual treatments (Figure 3C) (*p* < 0.05). However, no significant changes were observed in secreted IL-8 levels by MCF-7 cells when treated with adipocyte CM-EPA or CM-CAP (with or without EPA) (Figure 3D). Therefore, these results suggest possible inflammatory-reducing benefits of CAP–EPA combination in triple-negative breast cancer (TNBC) MDA-MB-231 cells but not in MCF-7 cells, indicating the potential role of n-3 PUFAs and antihypertensive ACE inhibitors in attenuating the tumor-promoting proinflammatory effects of adipocytes on breast cancer cells.

**Figure 3.** Adipose-CM with ACE-I with or without EPA reduced proinflammatory interleukins secretion in BC cells. Mature human adipocytes were pretreated with 100 μm of CAP with or without 100 μm of EPA for 24 h followed by CM collection and transferred to MDA-MB-231 (**A**,**B**) and MCF-7 (**C**,**D**) cells for 48 h. Medium was collected, and secreted IL-6 and IL-8 levels were measured in response to direct (no adipocyte CM) or CM-mediated CAP and EPA effects (*p* < 0.05; N = 3; and in each experiment, three replicates were used for each treatment group; bars with different letters (a, b) indicate significance;); nd: not detectable; ns: not significant).

#### *2.2. Combined E*ff*ect of Captopril and EPA on Breast Cancer Cell Migration in Response to Treatment with Adipocyte CM, Measured by a Wound Healing Assay*

Since inflammation is associated with BC cell migration, we performed in vitro wound healing assays to dissect the effects of both direct and human adipocyte-mediated effects of CAP ± EPA treatments in MDA-MB-231 cells. First, MDA-MB-231 cells were treated directly with CAP with or without EPA and images of wound closure were taken at various times up to 48 h (Figure 4A,B; also

see Tables S15 and S16). No significant effects of direct CAP treatment were found in MDA-MB-231 cells at any of the time points tested (*p* < 0.05) (Figure 4A,B). Intriguingly, a significantly higher percent wound area (or lower percent wound healing) was observed in response to direct EPA and CAP + EPA treatments at 4, 8, 12, 24, and 36 h time points when compared with bovine serum albumin (BSA)-control (*p* < 0.05), indicating reduced wound healing cell migration with EPA and CAP + EPA. However, the reductions in wound healing were not significant between EPA and CAP + EPA treatments at the respective time points (*p* < 0.05), indicating no additional effect of CAP when combined with EPA on MDA-MB-231 cell wound-healing capacity (Figure 4A,B).

**Figure 4.** Direct effect of captopril and EPA on cell migration in MDA-MB-231 cells. Representative wound healing images at 0, 4, 8, 12, 24, 36, and 48 h. Wounds were inflicted with a 200 uL pipette tip on MDA-MB-231 (**A**,**B**) cells in confluent monolayers. Graphical representation of migrating MDA-MB-231 (**A**,**B**) cells in response to CAP with or without EPA treatments of BC cells. Cells treated with regular Dulbecco's Modified Eagle's Medium (DMEM) with 1% bovine serum albumin (BSA) conjugation were used as controls (BSA). Bars represent mean of the percent wound area in MDA-MB-231 (**A**,**B**) cells in the two independent experiments ± SEM. *p* < 0.05 (comparison between 0 and 48 h; N = 2 combined experiments, in which each had three replicates per treatment group). Different letters (a, b) indicate significance.

Next, we identified the adipocyte-CM-mediated effects of CAP ± EPA treatments in MDA-MB-231 cells. Mature human adipocytes were pretreated with CAP with or without EPA for 24 h and medium was transferred to MDA-MB-231 cells, followed by capturing time-dependent images, as described above (Figure 5A,B). Cell migration, as denoted by the changes in percent wound area due to closure, was significantly increased in response to CM-mediated EPA, CAP, and CAP + EPA treatments compared with CM-control in MDA-MB-231 cells at all time points tested, such as 4, 8, 12, 24, 36, and 48 h (*p* < 0.05) (Figure 5A,B), whereas CM-CAP + EPA effects were not significantly different from

CM-EPA or CM-CAP, indicating no additional effect of CAP and EPA combination on MDA-MB-231 cell migration (Figure 5A,B). When time was modeled as a continuous variable, the wound area with CM treatment decreased by 2.13% per hour. This rate of reduction of wound area was hindered by treatment of adipocyte CM with each of EPA, CAP, and their combination as evidenced by significant positive interactions (Table S16). The subsequent factorial model revealed a significant positive effect of EPA and CAP but a significant negative interaction of CAP and EPA; these findings suggest that EPA and CAP may act via a common pathway in slowing down cell MDA-MB-231 cell migration (Tables S17 and S18). We did not perform wound healing assays in MCF-7 cells since the inflammatory markers were expressed and secreted at a lower level in MCF-7 compared with the MDA-MB-231 cell line we used. In addition, we observed greater changes in these markers, such as IL-6 in MDA-MB-231 cells, in response to both direct and CM-mediated treatments of CAP and EPA. These results further confirm our previous observation that CAP and EPA effects may be mediated by a common pathway to reduce expression of mRNA biomarkers of MDA-MB-231 cell proliferation.

**Figure 5.** CM-mediated effect of captopril and EPA on cell migration in MDA-MB-231 cells. Representative wound healing images at 0, 4, 8, 12, 24, 36, and 48 h. Wounds were inflicted with a 200 uL pipette tip on MDA-MB-231 (**A**,**B**) cells in confluent monolayers. Graphical representation of migrating MDA-MB-231 (**A**,**B**) cells in response to CM-EPA and CM-CAP with or without EPA treatments of BC cells. Cells treated with human adipose conditioned medium were used as controls (CM). Bars represent mean of the percent wound area in MDA-MB-231 (**A**,**B**) cells in the two independent experiments ± SEM. *p* < 0.05 (comparison between 0 and 48 h; N = combined experiments, in which each had three replicates per treatment group). Different letters (a, b) indicate significance.

#### **3. Discussion**

The aim of the current study was to provide new insights into adipocyte–BC interaction through inhibition of RAS in adipocytes. We previously showed protective effects of n-3 PUFAs in adipocyte–BC cell interactions [27]. Moreover, we demonstrated that n-3 PUFAs reduce Agt secretion from adipocytes and reduce both systemic and adipose tissue inflammation [29]. Hence, we proposed to determine whether an n-3 PUFA (EPA), individually or in combination with ACE-I, would reduce BC cell inflammation and migration directly on BC cells or via adipose RAS inhibition. In this study, we demonstrated reduced BC cell inflammation and motility in response to CM-derived EPA, CAP, and CAP + EPA treatments while no changes were observed among these treatments, indicating that the combination of CAP and EPA had no synergistic effects in attenuating adipocyte RAS effects on BC cells.

The doses of EPA and CAP that we used were within or lower than ranges used in human studies, making our findings translatable to humans. For omega-3 fatty acids, doses ranging from 1 to 6 g of EPA per day have been used in various human cancer studies and are associated with both improved cardiovascular and cancer outcomes [30,31]. Researchers reported that a ≥150 μg/mL (~150 μm) plasma EPA level was linked to benefits in preventing cardiovascular disease (CVD) outcomes [32]. In addition, plasma EPA concentrations in healthy subjects taking 2–4 g/day of EPA ethyl ester for 28 days increased up to a mean of 366 μg/mL, without any adverse events [33]. Although it is difficult to translate in vitro doses into actual human intake, the above mentioned reports clearly indicate that our 100 μm of EPA dose was lower than doses that can be reached in human plasma levels with prescribed fish oil or EPA. Moreover, other cell studies used doses of fatty acids up to 5 times higher than our dose [34,35]. For captopril, it has been used safely at doses up to 450 mg/day in humans [36]**,** which may lead to plasma doses likely higher than the 100 μm dose we used for cell treatments. Captopril was given at 150 mg/day to lung cancer patients [37] and lisinopril (another ACE-I) given as 10 mg tablets to breast cancer patients to prevent cardiotoxicity [38]. We were unable to find data regarding this drug's plasma level concentration in cancer patients since ACE-I is only prescribed as antihypertensive medication for cancer patients and not as an anticancer agent. Hence, the suitable dose for preventing breast cancer is not known. However, earlier research reported plasma captopril levels in normal versus chronic renal failure subjects as 364 and 347 μm, respectively, after 1 h, but it was undetectable in blood after 6h[39]. In addition, we previously performed a cell viability (MTT) assay with both 100 μm of EPA and 100 μm of captopril and found no toxic effects against breast cancer cell viability.

Adipose tissue is a major component of the breast tissue and is actively involved in forming TME; thus, adipose tissue through its inflammatory and lipogenic activities may increase the risk of BC development [3,40]. Human adipose CM is a source of various growth factors, several adipocyte secretory hormones, cytokines, and metabolites [41], which are immensely responsible for inducing expression of markers associated with breast cancer cell signaling, motility, and metastases [42,43]. Consistent with this, we have found increased inflammatory, IL-6, IL-8, NF-κB, and STAT3 mRNA expression as well as induced expression of the lipogenic enzyme FASN in both MCF-7 and MDA-MB-231 cells in response to human adipocyte CM. Overexpression of FASN plays an important role in tumorigenesis and is thus a vital target of obesity-mediated BC therapy [44]. Additionally, selective FASN inhibitors are potent antitumor agents showing antiproliferative and apoptosis inhibition effects in in vitro and in vivo models of receptor-positive (ER/HER2+) BC [45].

Moreover, two metabolically different BC cell lines, MCF-7 and MDA-MB-231, were used in the current study to identify the presence of any metabolic responses of these cell lines to n-3 PUFA or ACE-I. While the observed changes in mRNA levels between two cell lines were almost identical, their cytokine secretory profiles varied in response to direct or CM treatments. Both IL-6 and IL-8 levels were below detection in response to direct EPA/CAP treatments in MCF-7 cells, which is in agreement with other studies investigating the basal protein secretion levels by different BC cell lines [46,47]. Furthermore, we observed higher IL-6/IL-8 secretion by MDA-MB-231 cells compared with MCF-7 cells, which is consistent with previous studies indicating a stronger secretory profile in MDA-MB-231

TNBC cells compared with its less aggressive counterpart MCF-7 cells [48]. Another possible reason for the differential proinflammatory profile of MCF-7 versus MDA-MB-231 could be the intrinsically reduced transcriptional/post-transcriptional and/or translational/post-translational IL-6/IL-8 profile of MCF-7 cells in vitro compared with MDA-MB-231 cells [48].

Moreover, IL-6 activates STAT3 and NF-κB by (i) autocrine and paracrine signaling [8] and (ii) via synergistic crosstalk between the pathways in BC cells [7], resulting in cancer cell proliferation and apoptosis evasion [8]. By contrast, EPA is a natural anti-inflammatory bioactive compound previously reported to reduce IL-6 production, in part by inhibiting NF-κB activation [49] and through its downstream anti-inflammatory lipid mediators such as E-resolvins [50]. Consistent with this, we have also shown reduced mRNA levels of IL-6/STAT3/NF-κB in response to EPA-CM treatments. On the other hand, EPA also reduces both inflammation and Agt secretion from adipocytes both in vivo and in vitro [22,29]. Hence, when combining EPA with ACE-I (captopril), we observed reduced IL-6 inflammation in both MDA-MB-231 and MCF-7 BC cells as indicated by lower IL-6 mRNA and protein levels, respectively. Interestingly, the effect of EPA and ACE-I combination was more pronounced in reducing TNBC cell inflammation as shown by reduced IL-6 levels in MDA-MB-231 cells, which could be attributed to the more aggressive, inflammatory, and metastatic nature of MDA-MB-231 cells compared with MCF-7 cells [51]. Similarly, EPA and EPA-derived ethanol amines were more potent in reducing cell proliferation, migration, and invasion of MDA-MB-231 cells compared with non-invasive MCF-7 cells in vitro, which is in agreement with our present study [52]. We did not observe any synergistic effects of ACE-I and EPA combination in BC cell proliferation and migration, possibly due to their competitive anti-inflammatory in vitro mechanisms of action via EGFR and/or NF-κB (Rel-A) signaling pathways [19,53]. Moreover, many preclinical studies reported antihypertensive and anti-inflammatory roles of n-3 PUFAs respectively via ACE-2 upregulation, Ang II downregulation, and partly via GPCR-mediated NF-κB modulation [22,54,55]. Importantly, to our knowledge, we are the first to report the potential anti-inflammatory roles of ACE-I and EPA combination in targeting adipocyte and BC cell interactions. Despite this knowledge, potential mechanisms of ACE-I and EPA interactions in lessening the effects of adipocytes on BC cells remain unknown. Clinical studies are also lacking to understand the crosstalk between RAS inhibitors and n-3 PUFAs or other anti-inflammatory bioactive compounds to improve obesity-mediated cancer outcomes.

Inflammation is positively associated with BC cell motility and wound healing, although their mechanistic inter-relationship requires further studies [56]. Adipose CM contains various cytokines, including IL-6 and IL-8, which could be responsible for increased BC cell motility and wound healing, as reported by others [57]. We found a similar induced wound-healing capacity in MDA-MB-231 cells in response to human adipose CM. On the contrary, given the anti-inflammatory and antiproliferative effects of EPA on MDA-MB-231 cell motility [27] and the proposed similar effects for ACE-I, we combined them to identify their synergistic or additive effects in MDA-MB-231 TNBC. Surprisingly, both direct and CM-mediated EPA and ACE-I combination were ineffective in reducing TNBC cell motility compared with either EPA or ACE-I alone. This is in disagreement with Krusche et al., who reported the synergistic inhibitory effect of ACE-I and a Chinese herb (artesunate) on human umbilical vein endothelial cells (HUVEC) motility [58]. This might be due to the competitiveness between EPA and captopril to target common inflammatory mechanisms of action such as NF-κB downregulation and/or inactivation [59]. Taken together, despite some protective anti-inflammatory effects of EPA and ACE-I combination, such as attenuating IL-6, the underlying mechanisms require further studies in both in vitro and in vivo models of obesity-induced breast cancer.

#### **4. Materials and Methods**

#### *4.1. Cell Culture Experiments*

Human breast cancer cells such as MDA-MB-231 (HTB-26; Lot: 700792) TNBC and ER/PR positive MCF-7 (HTB-22; Lot: 64125078) BC cells were purchased from the American Type Culture collection (ATCC) (Manassas, VA, USA). MDA-MB-231 and MCF-7 cells were seeded at a density of 50,000 cells per well and maintained under standard culture conditions at 37 ◦C in a humid atmosphere of 5% CO2 in Dulbecco's Modified Eagle's Medium (DMEM). The medium was supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, Waltham, MA, USA) and antibiotics (50 μg/mL penicillin, 50 μg/mL streptomycin, and 100 ug/mL neomycin) (Thermo Fisher Scientific, Waltham, MA, USA). Bone-marrow-derived human mesenchymal stem cells (HMSCs) (PT-2501; Lot: 0000483199) were purchased from Lonza (Allendale, NJ, USA). These cells were differentiated into mature human adipocytes according to the optimized published protocol [60] and maintained under standard culture conditions at 37 ◦C and 5% CO2 in DMEM: Nutrient Mixture F-12 (DMEM/F-12) (Thermo Fisher Scientific, Waltham, MA USA). Following differentiation, cells were transferred to regular DMEM with 10% FBS for 24 h and CM experiments were performed. EPA (Nu-Check Prep, Waterville, MN, USA) and/or captopril (Sigma-Aldrich, St. Louis, MO, USA) was conjugated with 1% BSA (Sigma-Aldrich, St. Louis, MO, USA) for 2 h in a shaking water bath at 37 ◦C prior to cell treatment. Since fatty acids are found in the circulation as attached to albumin, many studies used BSA as a complex medium for EPA and other fatty acids in in vitro culture experiments [61]. For consistency, all treatments were performed in a BSA complexed medium.

#### *4.2. Treatment with ACE Inhibitor, Captopril, and Eicosapentaenoic Acid for Conditioned Medium Experiments*

MDA-MB-231 and MCF-7 cells were starved with 1% BSA for 2 h prior to CAP ± EPA experiments. The BC cells were treated directly with the ACE inhibitor CAP (100 μm) with or without EPA (100 μm) combination for 48 h to identify their individual/combined effect in modulating BC cell metabolism (Figure S1). The time and dose of our current treatments (CAP and EPA) were based on our previous work using EPA alone [27] or CAP alone (manuscript under review) against adipocyte–BC-cell interactions. CM experiments were performed between differentiated human adipocytes (HMSCs) and BC cells. Prior to the treatments, differentiated mature adipocytes were starved with 1% BSA for 2 h. Following starvation, mature adipocytes were treated with 100 μm of CAP ± 100 μm of EPA complexed with BSA, and BSA alone was used as the control. Conditioned media from differentiated adipocytes were collected after 24 h and centrifuged at 10,000× *g* for 10 min, then filter-sterilized to remove any cell debris. MDA-MB-231 and MCF-7 cells were seeded in six-well plates and then at 70% confluency exposed to CM from human adipocytes for 48 h. BC cells and media were collected and stored at −80 ◦C for further analyses (Figure S1).

#### *4.3. Enzyme-Linked Immunosorbent Assay (ELISA)*

Quantification of secreted IL-6 and IL-8 levels were determined by ELISA (R&D Systems, Minneapolis, MN, USA) according to the manufacturer's protocol.

#### *4.4. RNA Isolation and Real-Time Quatitative Polymerase Chain Reaction (RT-qPCR)*

RNA was purified using the Quick RNA mini kit (Zymo Research, Irvine, CA, USA) according to the established manufacturer's protocol followed by cDNA synthesis. cDNA was reverse-transcribed using Maxima reverse transcriptase (Thermo Fisher Scientific, Waltham, MA, USA). mRNA transcription levels were assessed with RT-qPCR using the Sybr green master mix (Thermo Fisher Scientific, Waltham, MA, USA). All genes were normalized to two housekeeping genes (18S ribosomal RNA and TBP (TATA box binding protein)). We used both 18S and TBP as our reference genes for normalization of RT-qPCR data, both were normalized to each other, and we did not see any regulatory effects in response to our proposed treatment conditions.

Sequences of the primers (Sigma Aldrich, St. Louis, MO, USA) are listed as follows (forward, reverse):

IL-6 (5- -AGACAGCCACTCACCTCTTCAG-3- , 5- -TTTCTGCCAGTGCCTCTTTGC-3- ), IL-8 (5- -AGGACAAGAGCCAGGAAGAA-3- , 5- -GGGTGGAAAGGTTTGGAGTATG-3- ), NF-κB (5- -ATGGCTTCTATGAGGCTGAG-3- , 5- -GTTGTTGTTGGTCTGGATGC-3- ), STAT3 (5- -AGAAGGACATCAGCGGTAAGA-3- , 5- -GGATAGAGATAGACCAGTGGAGAC-3- ), FASN (5- -TCGTGGGCTACAGCATGGT-3- , 5- -GCCCTCTGAAGTCGAAGAAGAA-3X), 18S (5- -CTACCACATCCAAGGAAGCA-3- , 5'-TTTTTCGTCACTACCTCCCCG-3- ), and TBP (5- -ATGGTGGTGTTGTGAGAAGATG-3- , 5- -CAGATAGCAGCACGGTATGAG-3- ).

#### *4.5. Wound Healing Assay*

MDA-MB-231 cells were seeded (50,000 cells per well) in six-well plates. At 95% confluence, cells were washed with phosphate-buffered saline (PBS) (Thermo Fisher Scientific, Waltham, MA, USA) and starved with 0.5% FBS in DMEM overnight to inhibit cell proliferation, thereby ensuring wound closure was due to cell migration only. Using a sterilized 200 μL pipette tip, two straight scratches were made through the cell's monolayer, stimulating wound. Cells were first rinsed once gently with PBS, then subjected to treatments described in the results, such as CAP with or without EPA. Images were taken at 0, 4, 8, 12, 24, 36, and 48 h using the EVOS Cell Imaging System (Thermo Fisher Scientific, Waltham, MA, USA) and data were analyzed using the in-house Cellular Growth Analyzer for Windows software (version 2.2). The in-house analyzer program was used to change the original image into black (the scratch) and the cell area into white (inside cells). The detection parameter assigned in the program was the pixel size, and the program checked along the *x* direction and later along the *y* direction, moving from left to right and counting the number of black pixels it came across. This stopped when it encountered a white pixel. This worked on the basis that the wound area (black) was the largest gap visible in the scratch assay.

#### *4.6. Statistical Analysis*

Data obtained from the experiments were normalized and analyzed using IBM SPSS (version 20, 2015) and R statistical software (version 3.5.3), and graphs were made using GraphPad Prism (Version 8.0). Statistical significance between groups (control vs. treatments) was tested using one-way ANOVA and subsequent pairwise post hoc comparisons adjusted using Tukey's correction. Differences were considered statistically significant for a Tukey-corrected *p*-value < 0.05. All experiments were performed at least two times and results are expressed as mean ± SEM of three independent biological replicates. Additional exploratory factorial regression analyses were performed to examine the interactions between EPA and CAP on all measured biomarkers.

For the wound healing assays, the data were analyzed separately for BSA and CM-treated conditions using two-way ANOVA models. The two-way ANOVA models were constructed to examine for the main effects of treatment (i.e., EPA, CAP, and combination), time modeled as a categorical variable, and their interaction. To determine the effect of each treatment on each biomarker per unit time, additional exploratory linear regression analyses were performed to modeling time as a continuous variable and the treatment condition as a categorical variable and their interaction. The interactions of EPA and CAP on temporal trends of biomarkers were examined in factorial regression analyses.

#### **5. Conclusions**

In conclusion, our study reported significant protective roles of EPA with and without ACE-I combination in attenuating adipocyte-induced proinflammation and cancer cell migration. To our knowledge, our study is the first to report inflammatory IL-6 downregulation effects of captopril and EPA combination in both MDA-MB-231 and MCF-7 BC cells. One acknowledged limitation of our study is that we only established a proof of principle for protective effects of CAP–EPA combination against adipocyte–BC-cell interaction, possibly via modulation of inflammation. This provided us a basis for future more in-depth mechanistic studies to further dissect effects of adipocytes on cancer cell proliferation and motility in response to CAP and EPA treatments in the tumor microenvironment.

Additionally, we did not see any differences between individual and combined ACE-I and EPA treatments except for IL-6. This indicates the need for more in-depth mechanistic studies possibly via knockdown of inflammatory signaling pathways in adipocytes and BC cells such as RAS and IL-6/STAT3/NF-κB pathways. Moreover, comparison of the efficacy of other ACE-I such as lisinopril, ARBs such as telmisartan, losartan, and/or n-3 PUFAs such as DHA or fish oil is necessary to better understand how these affect adipocyte–BC-cell interactions. Hence, our present study provides a novel significant validation for future mechanistic studies to combine diet and antihypertensive medication as a potential therapeutic approach for obesity-associated BC. More research is warranted to identify the possible mechanism of action and synergy of this unique combination of ACE-I and EPA or other n-3 PUFAs in obesity and breast cancer.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/12/1/220/s1, Figure S1: Experimental design; Table S1: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MDA-MB-231 cell growth and inflammation; Table S2: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MDA-MB-231 cell growth and inflammation; Table S3: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MDA-MB-231 cell growth and inflammation; Table S4: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MDA-MB-231 cell growth and inflammation; Table S5: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MDA-MB-231 cell growth and inflammation; Table S6: Exploratory factorial regression analyses to examine CAP and EPA interactions of protein markers associated with MDA-MB-231 cell inflammation; Table S7: Exploratory factorial regression analyses to examine CAP and EPA interactions of protein markers associated with MDA-MB-231 cell inflammation; Table S8: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MCF-7 cell growth and inflammation; Table S9: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MCF-7 cell growth and inflammation; Table S10: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MCF-7 cell growth and inflammation; Table S11: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MCF-7 cell growth and inflammation; Table S12: Exploratory factorial regression analyses to examine CAP and EPA interactions of mRNA markers associated with MCF-7 cell growth and inflammation; Table S13: Exploratory factorial regression analyses to examine CAP and EPA interactions of protein markers associated with MCF-7 cell inflammation; Table S14: Exploratory factorial regression analyses to examine CAP and EPA interactions of protein markers associated with MCF-7 cell inflammation; Table S15: Exploratory time-series regression analyses to examine CAP and EPA interactions in MDA-MB-231 cell migration; Table S16: Exploratory time-series regression analyses to examine CAP and EPA interactions in MDA-MB-231 cell migration; Table S17: Exploratory time-series factorial regression analyses to examine CAP and EPA interactions in MDA-MB-231 cell migration; Table S18: Exploratory time-series factorial regression analyses to examine CAP and EPA interactions in MDA-MB-231 cell migration.

**Author Contributions:** N.M.-M. conceived this project, and N.M.-M. and F.R. designed the experiments with input from and discussions with L.R. and C.K.; F.R. conducted most experiments with assistance from A.H.; H.M. developed the analysis software for the wound healing assays. F.R. and H.M. analyzed wound healing experimental data with input from N.M.-M., C.K. and L.R.; C.K. assisted with statistical analyses. All coauthors were involved in discussions of this work during its execution, reviewed the paper, and provided feedback. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Texas Tech University (TTU) Transdisciplinary Research Academy (N.M.-M.), TTU startup funds (N.M.-M.), and the Obesity Research Institute. Internal funds are used to cover the costs to publish in open access.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Isobolographic Analysis Demonstrates the Additive and Synergistic E**ff**ects of Gemcitabine Combined with Fucoidan in Uterine Sarcomas and Carcinosarcoma Cells**

### **Marcin Bobi ´nski 1,\*, Karolina Okła 1, Jarogniew Łuszczki 2, Wiesława Bednarek 1, Anna Wawruszak 3, Gema Moreno-Bueno 4, Magdalena Dmoszy ´nska-Graniczka 3, Rafał Tarkowski 1,**† **and Jan Kotarski 1,**†


Received: 29 October 2019; Accepted: 30 December 2019; Published: 31 December 2019

**Abstract:** Background: Uterine sarcomas and carcinosarcoma are associated with unfavorable prognosis. The regimens that are used in chemotherapy are associated with high incidence of side effects and usually do not significantly increase patients' survival rates. In this study we investigated the activity and interactions between gemcitabine and fucoidan, the natural compound known for its anti-tumor properties, in human sarcomas and carcinosarcoma cell models. Methods: SK-UT-1, SK-UT1-B (carcinosarcoma), MES-SA (leiomyosarcoma), and ESS-1 (endometrial stromal sarcoma) cell lines were used for the experiments. Cells were incubated in the presence of gemcitabine, fucoidan, and mixtures, after the incubation the MTT tests were performed. In order to assess the interactions between tested compounds isobolographic analysis was performed. Additional assessments of apoptosis and cell cycle were done. Results: Additive effect of combined treatment with gemcitabine and fucoidan was observed in ESS-1 and SK-UT-1 cell line. Although the supra-additive (synergistic) effect noticed in SK-UT-1B cell line. It was not possible to determine the interactions of fucoidan and gemcitabine in MES-SA cell line due to insufficient response to treatment. Addition of fucoidan to gemcitabine enhances its proapoptotic activity, what was observed especially in ESS-1 and SK-UT-1B cell lines. The arrest of cell cycle induced by mixture of gemcitabine and fucoidan, superior comparing gemcitabine alone was observed in SK-UT-1B. Conclusions: Obtained data showed that a combination of fucoidan and gemcitabine in uterine endometrial stromal sarcoma and carcinosarcoma cell lines has additive or even synergistic effect in decreasing cell viability. Furthermore, this drug combination induces apoptosis and arrest of cell cycle. The resistance of uterine leiomyosarcoma cell line, justifies searching for other drugs combinations to improve therapy efficacy.

**Keywords:** uterine sarcoma; gemcitabine; fucoidan; isobolography

#### **1. Introduction**

Uterine sarcomas are a group of malignancies consist of various types of tumors arising from mesenchymal tissue. The most common type of uterine sarcomas is leiomyosarcoma (about 60% of cases), less common types are endometrial stromal sarcoma (low and high grade), liposarcoma, rhabdomyosarcoma, and many other rare types [1]. The incidence of uterine sarcomas remains low respectively to epithelial malignancies (3–7% of all uterine malignancies) [2]. The diagnosis of uterine sarcoma is associated with bad prognosis and low rate of response to chemotherapy. Carciosarcoma is a type of tumor consisting of both mesenchymal and epithelial cells, that nowadays is considered as a type of endometrial cancer. These tumors are associated with bad prognosis and, similarly to uterine sarcomas, the results of its systemic treatment remain unsatisfactory. Systemic treatment of uterine mesenchymal tumors is based on cytostatic agents such as doxorubicine, gemcitabine, docetaxel, dacrbazine, and ifosfamde [3].

The exceptions from the above characteristics are endometrial stromal sarcomas, associated with indolent clinical behavior and favorable prognosis. The majority of these tumors express estrogen receptors and usually respond to hormonal treatment [2].

The retrospective trial comparing gemcitabine plus docetaxel, ifosfamide plus cisplatin, doxorubicin plus ifosfamide, ifosfamide alone, topotecan alone, and observation only in stage I and II leiomyosarcoma revealed no significant differences between groups in recurrence rate [4]. Due to the rarity of uterine sarcomas, the number of completed, randomized, prospective trials is limited. The response rates among patients suffering from uterine sarcomas, in prospective trials assessing the efficacy of various combinations of gemcitabine, docetaxel, and bevacizumab were between 25% and 35.8% [5–7].

The results obtained in systemic therapy of sarcomas are unsatisfactory. Most authors strongly recommend conducting clinical and pre-clinical trials in this field. The situation described above led to a search for new agents to be active against uterine sarcomas. Fucoidan is sulphated polysaccharide derived from brown seaweed, that recently gained attention due to its biological activities. It is known to affect multiple pathways in cancer cells including PI3K/AKT, MAPK, PTEN, VEGF, and caspases. Its effectiveness was proven in various models including lymphoma, leukemia, prostate cancer, breast cancer, hepatic cancer [8]. Recently we reported the anti-proliferative and pro-apoptotic activity of fucoidan in monotherapy among uterine sarcoma and carcinosarcoma cell lines. Simultaneously we confirmed previous observations that fucoidan do not affect normal (benign) human cells. Additionally, fucoidan is widely using as a dietary supplement. The characteristics listed above allow it to be considered as a safe product [8–10].

Obtained results led us to question the activity of fucoidan in combination with standard chemotherapy among uterine sarcomas. Gemcitabine was selected to be investigated together with fucoidan. Gemcitabine is commonly used anti-tumor drug belonging to the group of antimetabolites. It express antimetabolic effect by interruption of DNA synthesis. It is widely used as an option of standard approach for the systemic therapy in most of uterine sarcomas subtypes [3,11,12]. Furthermore, it is a comparator in many ongoing clinical trials enrolling uterine sarcoma patients [12]. Combinations of gemcitabine with novel agents including olaratumab, nivolumab, and pazopanib are under investigation in phases II–III of clinical trials [13].

This study was aimed to test if the combination of gemcitabine and fucoidan will have better therapeutic effect in uterine sarcomas and carcinosarcomas cell lines than the compounds applied alone and to assess the type of interactions between them in concomitant treatment.

#### **2. Materials and Methods**

#### *2.1. Reagents*

Gemcitabine (100 mg/mL, 0.38 mMol/mL) was purchased from Accord Healthcare (UK), and fucoidan (*Undaria pinnatifida*) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Fucoidan was diluted in respective complete culture medium at a concentration of 10 mg/mL just before use.

The Roswell Park Memorial Institute 1640 (RPMI-1640), Eagle's Minimum Essential Medium (MEM), McCoy's 5a Medium Modified, fetal bovine serum (FBS), trypsin-EDTA were purchased from PAN-Biotech (Aidenbach, Germany), penicillin-streptomycin and MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) were obtained from Sigma-Aldrich (St. Louis, MO, USA). The Cell Proliferation ELISA System assay kit was purchased from Roche (Molecular Biochemicals, Manneihm, Germany). PE Active Caspase-3 Apoptosis Kit and Propidium iodide utilizing the PI/RNase Staining Buffer were obtained from Becton Dickinson Biosciences (San Jose, CA, USA).

#### *2.2. Cell Lines and Cultures*

Carcinosarcoma cell lines (SK-UT-1, SK-UT1-B), uterine leiomyosarcoma cell line (MES-SA), and endometrial stromal sarcoma cell line (ESS-1) were obtained from the Laboratorio de Investigación Traslacional (MD Anderson Cancer Center, Madrid). The selection of cell lines was done in order to assess the differences in response to tested compounds among wide range of uterine sarcomas. Each cell line used for experiments is a model of particular tumor what makes obtained data more reproducible and comparable with other research.

The cells were cultured in MEM (SK-UT-1, SK-UT1-B), McCoy's 5a Medium Modified (MES-SA), RPMI-1640 (ESS-1), containing 10% (SK-UT-1, SK-UT-1B, MES-SA) or 20% (ESS-1) FBS and 1% penicillin-streptomycin at 37 ◦C in a humidified 5% CO2 atmosphere. Cells from the 4th to 9th passage were used for all experiments.

**Cell line: SKUT-1 \* SKUT-1B \* MES-SA ESS-1 Organism:** Homo sapiens, human Homo sapiens, human Homo sapiens, human Homo sapiens, human **Tissue:** uterus uterus/endometrium uterus uterus **Culture properties:** adherent adherent adherent adherent **Disease:** grade III, mesodermal tumor (mixed); consistent with leiomyosarcoma grade III, mesodermal tumor (mixed); consistent with leiomyosarcoma grade III, recurrent, uterine leiomyosarcoma [14] endometrial stromal sarcoma **Age:** 75 years 75 years 56 years 76 years **Gender:** female female female female

Detailed characteristic of cell lines is presented in Table 1.

**Table 1.** The characteristics of used cell lines [14,15].

**Ethnicity:** Caucasian Caucasian Caucasian Caucasian \* SKUT-1 and SKUT-1B cell lines were derived from the same patient from different sites of tumor SKUT-1 line is a model of sarcomatous part of the tumor (forms spindle cell sarcomas), SKUT-1B is a model of carcinomatous part of the tumor (forms well differentiated adenocarcinomas).

#### *2.3. Cell Viability Assay*

Cells were platted on 96-well microplates SK-UT-1, SK-UT-1B, MES-SA, and ESS-1 (3 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/mL) and cells were incubated in the presence of gemcitabine (0.1–200 ng/mL), fucoidan (0.01–5 mg/mL) and mixtures of both compounds for 96 h. The maximum concentration of gemcitabine achieved in human serum was 26.79 ± 10.06 μg/mL, reported by Wang et al. so the concentrations used

in our experiments were much lower comparing to ones available in vivo [16]. The safe concentrations of fucoidan in human serum are still under investigations and no conclusive data are available up to date. Afterwards, the cells were incubated for 3 h with the MTT. During the time MTT was metabolized by living cells to purple formazan crystals, which were later solubilized in SDS buffer (10% SDS in 0.01 N HCl) overnight. Separate experiments were performed in triplicate. The optical density of the product was measured at 570 nm with the use of an ELX-800 plate reader (Bio-Tek, Instruments, Winooski, VT, USA) and analyzed using Gen5 software (Bio-Tek, Instruments, Winooski, VT, USA).

#### *2.4. Assessment of Apoptosis*

Cell lines were treated with tested compounds and mixtures for 48 h, as a control cells without treatment were used. Afterwards cells were harvested, fixed and permeabilized using the Cytofix. All of experiments were performed according to the manufacturer's instructions of PE Active Caspase-3 Apoptosis Kit (BD Biosciences, San Jose, CA, USA). Labeled cells were analyzed by flow cytometer FACSCalibur (Becton Dickinson, Franklin Lakes, NJ, USA), operating with CellQuest software (BD Biosciences, San Jose, CA, USA) to quantitatively assess the caspase-3 activity. The methodology of apoptosis assessment were described in detail elsewhere [9].

#### *2.5. Cell Cycle Analysis*

Cell lines were treated with tested compound and mixture for 48 h (as a control we used cells without exposure to tested compounds) and then fixed in 80% ethanol at −20 ◦C for 24 h. The experiment was conducted by utilizing PI/RNase Staining Buffer (BD Biosciences, USA) according to the manufacturer's instructions. Cell cycle analysis was performed using flow cytometry (FACSCalibur (Becton Dickinson, USA). Acquisition rate was at least 60 events/sec in low acquisition mode and at least 10,000 events were measured. Methodology of cell cycle analysis was presented in detail previously [10].

#### *2.6. Isobolographic and Statistical Analysis*

In order to determine the inhibition rate of cell viability, measured by MTT assay, per dose of gemcitabine and fucoidan, log-probit linear regression analysis was performed according to method described by Litchfield and Wilcoxon [17,18]. Median inhibitory concentrations (IC50) for gemcitabine and fucoidan in ESS-1, SK-UT-1, and SK-UT-1B cell lines were calculated according to method previously described [19]. Due to lack of cells response, the IC50 was not achievable for fucoidan in MES-SA cell line [10]. Parallelism between dose–response curves for gemcitabine and fucoidan in ESS-1, SK-UT-1, and SK-UT-1B cell lines was confirmed by log-probit method, as it was described in detail previously [18]. Next, isobolographic analysis of interactions between drugs for the combination of gemcitabine and fucoidan in tested cell lines were performed according to the method presented by i.a. Tallarida et al. [20]. The median additive inhibitory concentrations (IC50 add) for two-drug mixtures were theoretically calculated according to the method described elsewhere [20]. The calculated values were used for performing MTT tests on ESS-1, SK-UT-1, and SK-UT-1B cell lines—the assessment of experimentally derived IC50 mix values for tested drug combinations in a fixed 1:1 ratio. The particular drug concentrations (gemcitabine and fucoidan) in the mixture were calculated by multiplying IC50 mix values accordingly to proportions in additive mixture. Detailed description of isobolographic method was introduced by Tallarida, Grabovsky and Luszczki [20–22]. The results of MTT test were analyzed by one-way ANOVA test, Tukey's Multiple Comparison Post-test using GraphPad Prism 5.0 (GraphPad Softwere Inc., San Diego, CA, USA). The *p* < 0.05 was considered as statistically significant.

#### **3. Results**

#### *3.1. Cell Viability Assay*

Anti-proliferative effects of gemcitabine on tested cell lines is presented on Figure 1. Experimentally determined IC50 values for gemcitabine in SK-UT-1, SK-UT-1B, ESS-1, andMES-SA cell lines, were 31.173, 25.243, 13.875, and 72.482 ng/mL respectively.

As we previously reported fucoidan significantly affects SK-UT-1, SK-UT-1B, and ESS-1 cell lines, meanwhile MES-SA cells seem to be resistant for this agent. IC50 was 0.966, 3.348, and 0.848 mg/mL respectively, it was not possible to determine IC50 for fucoidan in MES-SA cell line due to insufficient response to treatment [9]. The IC50 values are summarized in Supplementary Table S1.

A

B

**Figure 1.** The influence of gemcitabine on the proliferation of carcinosarcoma cell lines (SK-UT-1 (**A**), SK-UT1-B (**B**)), endometrial stromal sarcoma cell line (ESS-1 (**C**)) and uterine leiomyosarcoma cell line (MES-SA (**D**)). The cells were treated with the gemcitabine at various concentrations for 96 h. (\*\* *p* < 0.01, \*\*\* *p* < 0.001 were considered as statistically significant).

#### *3.2. Isobolographic Anaysis*

Additive effect of the combined treatment with gemcitabine and fucoidan was observed in ESS-1 and SK-UT-1 cell lines. Although the supra-additive (synergistic) effect was noticed in SK-UT-1B cell line.

The details of results obtained in isobolographic analysis are presented on Figures 2–4.

In Figures 2–4 the median inhibitory concentrations (IC50) for gemcitabine (GEM) and fucoidan (FUK) are plotted on the X- and Y-axes, respectively. The solid lines on both axes reflect the S.E.M. for the IC50 values for the studied drugs, when administered alone. The lower and upper isoboles of additivity represent the curves connecting the IC50 values for GEM and FUK administered alone. The dotted line illustrates the fixed-ratio of 1:1 for the combination of GEM with FUK. The points A' and A" depict the theoretically calculated IC50 add values for both, lower and upper isoboles of additivity. The point M reflects the experimentally-derived IC50 mix value for total dose of the mixture expressed as proportions of GEM and FUK that produced a 50% anti-proliferative effect (50% isobole) in the cancer

cell line (SK-UT-1, SK-UT-1B, and ESS-1, respectively for Figures 2–4) measured in vitro by the MTT assay. On the graph, the S.E.M. values are presented as horizontal and vertical error bars for every IC50 value. Type I isobolographic analysis of interactions are presented in Supplementary Table S2. The effect of combined fucoidan and gemcitabine on the proliferation of tested cell lines were presented of Figure 5.

**Figure 2.** Isobologram showing interaction between gemcitabine (GEM) and fucoidan (FUK) with respect to their anti-proliferative effects in the cancer cell line (SK-UT-1) measured in vitro by the MTT assay. The experimentally-derived IC50 mix value is placed within the area of additivity and indicates additive interaction between GEM and FUK in this cancer cell line.

**Figure 3.** Isobologram showing interaction between gemcitabine (GEM) and fucoidan (FUK) with respect to their anti-proliferative effects in the cancer cell line (SK-UT-1B) measured in vitro by the MTT assay. Because the experimentally-derived IC50 mix value is placed significantly below the point A', the interaction between GEM and FUK for the cancer cell line SK-UT-1B is supra-additive (synergistic). \* *p* < 0.05 vs. the respective IC50 add values.

**Figure 4.** Isobologram showing interaction between gemcitabine (GEM) and fucoidan (FUK) with respect to their anti-proliferative effects in the cancer cell line (ESS-1) measured in vitro by the MTT assay. Although the experimentally-derived IC50 mix value is placed below, but near to the point A', the interaction between GEM and FUK in this cancer cell line is additive.

B

A

68

**Figure 5.** *Cont.*

**Figure 5.** The influence of combined fucoidan and gemcitabine on the proliferation of carcinosarcoma cell lines (SK-UT-1 (**A**), SK-UT-1B (**B**)) and endometrial stromal sarcoma cell line (ESS-1 (**C**)). The cells were treated with fucoidan and gemcitabine at various concentrations for 96 h. (\*\* *p* < 0.01, \*\*\* *p* < 0.001 were considered as statistically significant). The values on axis X represent the multiplicity of calculated IC50. The combinations of gemcitabine and fucoidan were mixed 1:1 before added to cells.

#### *3.3. Assessment of Apoptosis*

C

The impact of gemcitabine and its combination with fucoidan on apoptosis, measured as a number of cells with activated caspase 9, is presented on Figure 6A–C. Gemcitabine was used in concentration of IC50. Results obtained in isobolographic analysis were used to select appropriate concentrations of mixture for apoptosis assessment. For SK-UT-1 and SK-UT-1B concentrations of 0.5 IC50 for both agents were used. Due to very strong effect observed in ESS-1 cell line concentration 0.05 IC50 was selected for experiments.

**Figure 6.** *Cont.*

**Figure 6.** Effects of gemcitabine (in IC50 concentration) and mixture of fucoidan and gemcitabine (in concentrations of 0.5 IC50 for SK-UT-1 and SK-UT-1B and 0,05 IC50 for ESS-1) on caspase-3 activation in SK-UT-1 (**A**), SK-UT-1B (**B**), and ESS-1 (**C**) cells. Results are expressed as mean ± SD of three separate experiments (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 versus the control, one-way ANOVA test).

The strongest induction of apoptosis by both gemcitabine and mixture gemcitabine + fucoidan was observed in ESS-1 cell line. This effect was also observed in SK-UT-1B cells. Although the induction of apoptosis in SK-UT-1 cell line of both single agent and combination was very weak. Statistical significance was observed between single agent and combination but the differences comparing to control were not significant.

#### *3.4. Cell Cycle*

The results of cell cycle analysis of cells treated with gemcitabine and its mixture with fucoidan are presented in Figure 7. Statistical significance of differences between each phase of cell cycle in treatments and control are listed in Supplementary Table S3. Tested compounds were used in concentration as in apoptosis assessments.

**Figure 7.** Effects of gemcitabine (in IC50 concentration) and mixture of fucoidan and gemcitabine (in concentrations of 0.5 IC50 for SK-UT-1 and SK-UT-1B and 0.05 IC50 for ESS-1) on cell cycle progression in SK-UT-1 (**A**), SK-UT1-B (**B**), and ESS-1 (**C**) cell lines. The results are presented as mean ± SD from three separate experiments.

The most significant impact of concomitant treatment with gemcitabine and fucoidan on cell cycle arrest (measured as percent of cells in phases pre-G1 and G0/G1) was detected in the SK-UT-1 B cell line (comparing to both control and gemcitabine treatment). In the ESS-1 cell line significant differences were noted between cells treated with gemcitabine and control, with no such differences between gemcitabine and mixture. The differences in number of cells in each phase of cell cycle in SK-UT-1 cells exposed to investigated agents were very small in spite of statistical significance observed in particular phases.

#### **4. Discussion**

This paper for our knowledge is the first report of interactions between fucoidan and gemcitabine in any model. The results obtained in our experiments show that both gemcitabine and fucoidan significantly affect cell viability in all tested cell lines (with exception of fucoidan in MES-SA cell line). The problem of MES-SA cell line resistance to fucoidan have been discussed previously [10]. In present study we observed also the worst response of MES-SA to gemcitabine among all of tested cell lines. Interestingly we used regular MES-SA line, but not its multidrug resistant variant MES-SA/Dx5. The mechanism of relative resistance of MES-SA cell line to fucoidan and gemcitabine is not known and requires further investigations. For our best knowledge the phenomenon of spontaneous MES-SA resistance to gemcitabine have not been published yet. Resistance to gemcitabine was artificially induced by genetic modifications (transfection) in gene coding deoxynucleoside kinase in MES-SA, although such modifications were not performed in the cell line we used [23]. Occurrence of some spontaneous mutations in deoxynucleoside kinase gene at least in part of cells could possibly explain our observation. The occurrence of serious side effects combined with limited activity makes systemic therapy of uterine sarcomas doubtful [13]. The most common unwanted results of chemotherapy are hematologic toxicities (leucopoenia, neutropenia, thrombocytopenia), fever, and concomitant infectious, that appear in the majority of treated patients [24]. Side effects of drugs are mostly dose-dependent, so it could be limited by the reduction of dosage, on the other hand, reduction of dose would decrease the therapeutic effect.

In multidrug regimens, in assumption, lower doses of multiple drugs may cause fewer side effects (or its lower intensity) without decreasing effectiveness or even increasing it [13]. But in clinical practice such approaches are commonly associated with higher rate of adverse events.

The rarity and heterogeneity of uterine sarcomas limit the possibilities of conducting clinical trials, which may be considered as a cause of the poor results of its treatment up to date. Even in preclinical studies number of research aimed to test new drugs combinations in this group of tumors is limited.

Coley et al. tested the combination of seliciclib (cyclin-dependent kinases (CDK)-inhibitor) and paclitaxel among selection of uterine sarcomas cell lines. Using isobolography, they observed synergism between both drugs. Interestingly, differences in activity depending on the sequence of each drug administration were reported. Although the study is interesting, seliciclib is not registered for the treatment of uterine sarcomas. Furthermore, paclitaxel in monotherapy is not recommended among this type of malignancies [25].

In the present study we propose a combination of standard chemotherapy with a natural compound that is known to be safe; its intake is not associated with serious adverse events [26]. Both substances were administered at the same time in order to avoid a "drug sequence effect" observed by Coley at al. that could influence the results. Isobolographic analysis showed additive interaction between gemcitabine and fucoidan in SK-UT-1 and ESS-1 cell lines and even supra-additive (synergy) in SK-UT-1B. No antagonism between tested substances was observed.

Cell viability was decreased equally or more than the sum of effects of single agents. If these observations are confirmed in animal models, clinical trials of fucoidan as an addition to treatment with gemcitabine may allow a decrease of its dose. Fucoidan is nowadays used as dietary supplement, and seaweeds containing it are widely used in Asian cuisine [26]. There are many researches aimed at checking its activity in various types of cancer. Some of them regard a combination of fucoidan and

other agents (including cytostatics), and the results of most of them are promising. In our study we assessed the type of interactions between fucoidan and a cytostatic agent, using an isobolographic method, and calculated IC50 for the mixture. Significantly lower doses of therapeutic compounds may be associated with lower risk of adverse events. Furthermore, it may solve another problem that was widely discussed previously— the bioavailability of fucoidan. That is currently still under investigation both in animal models as well as in clinical trials—the results of these studies are awaited [27–29]. IC50 values for fucoidan obtained in combination with gemcitabine are much lower than 1 mg/mL: Such concentrations seem to be achievable in vivo. The doubtful concentrations were around 5 mg/mL. Although a few studies in this field were performed, precise determination of available serum concentrations among human are still under investigation [29,30].

Quite similar results were obtained by Mathew et al. who assessed IC50 for various types of fucoidan from 0.3 to 1.3 mg/mL (depending on plant that it was derived from). In this study they also performed analysis of fucoidans' impact on CYP450 and COMT (catechol-O-methyltransferase) pathways, concluding that it is limited, which might be considered as a confirmation of its safety [31].

Our findings are comparable with other studies assessing activity of fucoidan and cytotoxic drugs. Burney et al. assessed activity of 2 types of fucoidan in combination with paclitaxel and tamoxifen on mouse models of human breast or ovarian cancer. They concluded that interaction between tested compounds were additive or supra-additive (depending on combination and model used) [32].

Similar observations were taken by Zhang et al. who tested combinations of fucoidan with cisplatin, tamoxifen, and paclitaxel in breast cancer cell lines. They concluded that fucoidan "enhances the anti-cancer activity" of tested cytostatic drugs. Furthermore, they observed the induction of apoptosis and cell cycle arrest among cells treated with fucoidan what corresponds with our previous findings [9,33].

Observed additive and supra-additive interactions between various fucoidans and chemotherapeutic drugs require explanation. The mechanism of fucoidans' activity is not fully investigated. As it was mentioned above there are various cellular pathways that are affected by this compound. Its multipotential activity is, on the one hand, promising, and does not allow the overcoming of single pathway change, but on the other it is very difficult to understand completely. One of possible mechanism of synergistic interaction with chemotherapeutics is the ability to down-regulate expression of Bcl-xL and Mcl-1, known as anti-apoptotic proteins [34]. The addition of fucoidan to gemcitabine enhances its ability to induce apoptosis. This effect is especially noticeable in ESS-1 cell line where, even if mixture was added in concentration 0.05 of IC50, almost 80% of cells enter apoptosis and the difference between gemcitabine in monotherapy and the mixture reaches almost 20%. Weaker effect was observed in SK-UT-1B cells. In the SK-UT-1 cell line almost no impact of gemcitabine and fucoidan to apoptosis was noted, even if agents were used in concentration of 0.5 of IC50. Abudabbus et al. compared the ability of fucoidan in combination with cisplatin, doxorubicin and taxol, to induce apoptosis and to arrest the cell cycle among benign and malignant breast cells [35]. They reported a strong impact to cell cycle arrest and induction of apoptosis in malignant cells but no significant effect in normal cells. A comparable effect of fucoidan in monotherapy was described by Arumugam et al. in hepatoblastoma cell line [36]. So, our results confirm proapoptotic features of fucoidan in selected models. Interestingly, observations taken in experiments assessing apoptosis were confirmed also in cell cycle analysis. The strongest effect of combination of gemcitabine and fucoidan was observed in ESS-1 line. Moderate and weak changes in number of cells in particular cell cycle phases were observed in SK-UT-1 and SK-UT-1B cell lines respectively.

Gemcitabine is known to affect the cell cycle in various cell models and in vivo; this feature is the consequence of its mechanism of action by DNA damaging [37]. The process of gemcitabine-induced cell cycle arrest can by interrupted by activity of DNA repairing mechanisms such as Chk1 [38]. Such a mechanism might be responsible for the limited changes in cells distribution among cell cycle phases observed in SK-UT-1 and SK-UT-1B treated with gemcitabine.

Park et al. reported the effect of cell cycle arrest by fucoidan treatment in bladder cancer cells. They also revealed the its association with down-regulation of cyclin D1, cyclin E, and cyclin-dependent-kinases (Cdks) in a concentration-dependent manner, without any change in Cdk inhibitors (p21, p27) [39]. Interestingly Han et al. experimented on a colon cancer model and observed cell cycle arrest associated with an increased expression of p21 [40]. Different mechanisms of cell cycle arrest induced by fucoidan in different models, combined with potential activity of DNA repairing factors affecting activity of gemcitabine, can explain differences among tested cell lines, we observed. So far, no studies assessing the impact of combination of gemcitabine and fucoidan have been published.

The results and survey of literature indicate that testing multicompound regimens gives promising results and further research in the field could reveal a combination that could be beneficial for patients suffering from uterine sarcomas.

The study is not without limitations and should be considered carefully. The main limitation of the study is a cancer model that was used, 2D cell cultures, which do not allow investigation of the tumor microenvironment and may not fully reflect the response of tumor cells in vivo. Although promising, obtained results require clinical studies to be considered in practice.

The results obtained in our study indicate 3 potential directions for further investigations:


#### **5. Conclusions**

Obtained data showed additive and supra-additive effect of fucoidan combined with gemcitabine in uterine endometrial stromal sarcoma (ESS-1) and carcinosarcoma (SK-UT-1, SK-UT-1B) cell lines, what confirms it has better or at least equal performance comparing to sum of effects of monotherapies.

The addition of fucoidan to gemcitabine enhances proapoptotic effect of gemcitabine in endometrial stromal sarcoma cells (ESS-1) but not in carcinosarcoma (SK-UT-1, SK-UT-1B) cell lines.

Gemcitabine in monotherapy do not induce cell cycle arrest in carcinosarcoma (SK-UT-1, SK-UT-1B) cell lines. Although the addition of fucoidan to gemcitabine induces it in model of carcinomatous part of carcinosarcoma (SK-UT-1B).

Differences in response to applied treatment among tested cell lines (in cell viability apoptosis and cell cycle distribution) can be explained by the multipotential and not-fully-investigated activity of fucoidan as well as differences in cellular mechanisms (such as DNA repairing) in selected models.

The relative resistance of uterine leiomyosarcoma cell line (MES-SA) to applied drugs combination justify searching for other therapeutic regimens to improve therapy efficacy.

Taking into consideration the disappointingly low effectiveness of systemic therapy among these types of cancer, a combination of gemcitabine and fucoidan seems to be a promising alternative, having the potential to increase effectiveness and safety of the treatment.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/12/1/107/s1, Table S1: Anti-proliferative effects of gemcitabine (GEM) and fucoidan (FUK) administered singly in cancer cell lines ESS-1, SKUT-1, SKUT-1B and MES-SA, as measured in vitro by the MTT assay, Table S2: Type I isobolographic analysis of interactions (for non-parallel dose-response effects) between gemcitabine (GEM) and fucoidan (FUK) in three cancer cell lines ESS-1, SKUT-1 and SKUT-1B, as measured in vitro by the MTT assay, Table S3: Statistical significance of differences in percents of cells in each cell cycle phases among cells treated with gemcitabine or mixture (gemcitabine and fucoidan) and control (no treatment). The results were analyzed by one-way ANOVA

test, Tukey's Multiple Comparison Post-test, *p* < 0.05 was considered as statistically significant. The results of experiment are presented on Figure 7.

**Author Contributions:** Conceptualization, M.B., W.B. and J.K.; methodology, K.O. and M.D.-G.; validation, A.W., M.D.-G.; formal analysis, J.K.; investigation, K.O., A.W. and M.D.-G.; resources, W.B. and G.M.-B.; data curation, J.Ł.; writing—original draft preparation, M.B.; writing—review and editing, K.O. and R.T.; supervision, J.K.; project administration, M.B.; funding acquisition, W.B., J.K. and R.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research was founded by Medical University of Lublin (grants No. DS 120, DS 121 and MNmd129).

**Acknowledgments:** The research was performed in collaboration under umbrella of ENITEC (European Network of Individual Treatment in Endometrial Cancer), the research group established by ESGO (European Society of Gynaecological Oncology).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
