EGCG-Derivative G28 Shows High Efficacy Inhibiting the Mammosphere-Forming Capacity of Sensitive and Resistant TNBC Models.

Recent studies showed that Fatty Acid Synthase (FASN), a lipogenic enzyme overexpressed in several carcinomas, plays an important role in drug resistance. Furthermore, the enrichment of Breast Cancer Stem Cell (BCSC) features has been found in breast tumors that progressed after chemotherapy. Hence, we used the triple negative breast cancer (TNBC) cell line MDA-MB-231 (231) to evaluate the FASN and BCSC population role in resistance acquisition to chemotherapy. For this reason, parental cell line (231) and its derivatives resistant to doxorubicin (231DXR) and paclitaxel (231PTR) were used. The Mammosphere-Forming Assay and aldehyde dehydrogenase (ALDH) enzyme activity assay showed an increase in BCSCs in the doxorubicin-resistant model. Moreover, the expression of some transcription factors involved in epithelial-mesenchymal transition (EMT), a process that confers BCSC characteristics, was upregulated after chemotherapy treatment. FASN inhibitors C75, (-)-Epigallocatechin 3-gallate (EGCG), and its synthetic derivatives G28, G56 and G37 were used to evaluate the effect of FASN inhibition on the BCSC-enriched population in our cell lines. G28 showed a noticeable antiproliferative effect in adherent conditions and, interestingly, a high mammosphere-forming inhibition capacity in all cell models. Our preliminary results highlight the importance of studying FASN inhibitors for the treatment of TNBC patients, especially those who progress after chemotherapy.


Introduction
Breast cancer is the most common cancer among women worldwide with approximately 2.1 million diagnoses estimated in 2018 according to the International Agency for Research on Cancer [1]. Triple negative breast cancer (TNBC) represents approximately 15-20% of patients with breast carcinomas and is characterized by the lack of expression of estrogen and progesterone receptors (ER/PR) and neither expression nor amplification of the HER2 oncogene [2]. The absence of these three biomarkers prevents the use of currently available targeted therapies for breast cancer [3,4], and leaves systemic cytotoxic chemotherapy as the sole treatment option [5]. Despite a good initial response to chemotherapy in neoadjuvant settings, only 30% of patients will exhibit a survival of more than five years following diagnosis [6][7][8]. Compared to other BC subtypes, TNBC presents the highest recurrence rate along with the shortest time, which is associated with a significantly poorer overall survival [2,7]. Two major molecular subtypes are represented in TNBC, basal-like (80%) and claudin-low (CL) or mesenchymal-like (ML; 20%), this last one being enriched in cells with stem features [9]. Current therapies failure and resistance may be explained in part by the presence of a small population of cells that display stem properties within breast tumors known as breast cancer stem cells (BCSCs). In vitro, BCSC-enriched populations display specific features: cell-surface specific marker pattern (CD44 high /CD24 low ), the ability to form mammospheres when growing in non-adherent conditions, aldehyde dehydrogenase 1 (ALDH1) enzyme activation, and enhanced resistance to chemotherapy [10]. Additionally, the induction of epithelial-mesenchymal transition (EMT) through E-cadherin expression decrease and upregulation of mesenchymal proteins, such as vimentin or N-cadherin, results in stem properties [11]. Some transcription factors involved in EMT regulation are slug, zeb1, zeb2, twist and snail [12,13].
Moreover, it has recently been demonstrated that the regulation of lipid metabolism promotes BCSCs and cancer chemoresistance [14]. Back in 1924, Warburg made evident metabolism deregulation in cancer cells [15,16], becoming many years later a hallmark of cancer [17]. Cell membranes are formed by long-chain fatty acids, being also important substrates for energy cell metabolism. The Fatty Acid Synthase (FASN) is the enzyme responsible for the de novo synthesis of palmitate, the most abundant fatty acid [18]. Several carcinomas such as breast, colon, lung, prostate, among others, overexpress FASN [19][20][21][22], suggesting it as a unique onco target. Blocking FASN activity causes in vitro and in vivo anticancer activity by inhibiting tumor progression [23][24][25][26][27][28], hindering angiogenesis [29,30], overcoming drug-resistance [31,32], and synergistically increasing the efficacy of chemotherapy [26,33,34]. A recent study showed that FASN was expressed in 92% of tumor tissue samples coming from a cohort of 100 TNBC patients and its association with positive node status made evident its role as a possible predictive biomarker in this aggressive BC subtype [35].
(−)-Epigallocatechin 3-gallate (EGCG) is a powerful antioxidant and the most abundant catechin in green tea. Its apoptotic effect leads to antiproliferative activity [36][37][38][39]. Although EGCG targets HER1-HER2, MAPK, and AKT signaling pathways among others, it has been described that its apoptosis-inducing effect occurs through FASN inhibition [28,40,41]. Several studies have demonstrated a weak effect of EGCG in 20 different human cancer stem cell populations when used alone but synergistically increased in combination with different anticancer drugs [42]. We have produced a battery of new polyphenolic derivatives structurally related to EGCG, from which G28, G56, and G37 proved to possess enhanced FASN inhibitory activity [43][44][45]. These compounds also showed cancer cell cytotoxicity in a set of human breast cancer cells. G28 displayed a potent tumor volume reduction in vivo with no weight loss or anorexia, the main side-effects of other FASN inhibitors like the cerulenin-derived compound C75 [28,41,43]. G28 also showed apoptosis induction in HER2+ resistant cell lines and tumor diminishment in HER2+ breast cancer xenografts [26,46].
In the present study, we evaluated FASN and BCSC characteristics, i.e., mammosphere-forming capacity and ALDH1 activity, in the acquisition of chemoresistance in the TNBC model MDA-MB-231 (231). Moreover, we used the natural FASN inhibitor EGCG and its synthetic derivatives G28, G56, and G37 in comparison to C75 (Figure 1) to target FASN through these BCSC features from these TNBC models resistant to doxorubicin (231DXR) and paclitaxel (231PTR), the most common drugs currently used in this BC subtype devoid of a validated targeted therapy.

FASN Expression in MDA-MB-231 Derived Chemoresistant Cell Lines
FASN activity has demonstrated to play an important role in drug resistance through new phospholipid synthesis for membrane renovation and plasticity. It also decreases ceramide levels, inhibiting apoptosis via PARP activation [32,[47][48][49][50].
To assess the role of FASN in chemoresistance acquisition in TNBC, we developed MDA-MB-231 (231) cells resistant to doxorubicin (231DXR) [34] and paclitaxel (231PTR) (Supplementary Figure  1). It has been described that doxorubicin-resistant cell lines become sensitive through the inhibition of FASN [34,51]. Therefore, we studied how FASN protein levels were modified after drug treatment of sensitive and chemoresistant TNBC cells. Our results showed that 231DXR FASN levels experienced a 2-fold increase after 24 h of doxorubicin treatment (Figure 2A), while such effect was not observed in parental cells. On the other hand, paclitaxel did not show any effect on FASN protein levels neither in 231 nor in 231PTR ( Figure 2B). PARP cleavage, a marker for apoptosis, is increased in parental cell lines compared to its resistant counterparts for both drugs, making evident their chemoresistance ( Figure 2).

FASN Expression in MDA-MB-231 Derived Chemoresistant Cell Lines
FASN activity has demonstrated to play an important role in drug resistance through new phospholipid synthesis for membrane renovation and plasticity. It also decreases ceramide levels, inhibiting apoptosis via PARP activation [32,[47][48][49][50].
To assess the role of FASN in chemoresistance acquisition in TNBC, we developed MDA-MB-231 (231) cells resistant to doxorubicin (231DXR) [34] and paclitaxel (231PTR) (Supplementary Figure S1). It has been described that doxorubicin-resistant cell lines become sensitive through the inhibition of FASN [34,51]. Therefore, we studied how FASN protein levels were modified after drug treatment of sensitive and chemoresistant TNBC cells. Our results showed that 231DXR FASN levels experienced a 2-fold increase after 24 h of doxorubicin treatment (Figure 2A), while such effect was not observed in parental cells. On the other hand, paclitaxel did not show any effect on FASN protein levels neither in 231 nor in 231PTR ( Figure 2B). PARP cleavage, a marker for apoptosis, is increased in parental cell lines compared to its resistant counterparts for both drugs, making evident their chemoresistance ( Figure 2).

BCSC-Enriched Population in Sensitive and Resistant Cell Lines
Despite the fact that BCSCs are a very low-represented population within tumors, selection or enrichment of these cells after chemotherapy treatment can result in a tumor relapse due to their intrinsic chemoresistance and their tumorigenic ability [52][53][54]. Here, we aim at the evaluation of the BCSC population in our sensitive and resistant models by means of two of these techniques, i.e., mammospheres-forming capacity and ALDH1 activity. The influence of doxorubicin and paclitaxel treatment was also evaluated.
To determine whether there was an enrichment of this population in our resistant models, we used the Mammosphere-Forming Assay (MFA; Figure 3A) in 231 parental cell line and in the chemoresistant derivatives 231DXR and 231PTR ( Figure 3B). 231 and 231DXR displayed similar Mammosphere-Forming Index (MFI), comparable to those of 231 and 231PTR. Likewise, no morphological differences were found among the three cell models (data not shown). Since one of the BCSC population characteristics is its inherent chemoresistance [55], we carried out doxorubicin and paclitaxel treatments in adherent and mammosphere cell culture, in both sensitive and resistant  (70 nM) or paclitaxel (5 nM) at 48 h), and treatment duration were performed equally in both approaches. The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was carried out in adherent conditions to determine cell viability, whereas, in non-adherent culture, the Mammosphere-Forming Inhibition (MFIn) was used. The BCSC-enriched population showed inherent doxorubicin resistance when compared to adherent conditions for both sensitive and resistant cell lines ( Figure 3C). In adherent settings, the inhibition of viability ranged from 89.81% ± 1.65 (231) to 86.7% ± 1.85 (231DXR). Nevertheless, MFIn values exhibited lower inhibition in both cell lines treated with doxorubicin. Furthermore, in non-adherent culture, a doxorubicin cytotoxic effect was significantly higher in parental cells (60.66% ± 1.17) compared to the resistant model (51.29% ± 0.58, p < 0.01). Similar results were found with paclitaxel treatment ( Figure 3D). The reduction in viability in adherent condition was significantly higher in sensitive cells (90.47% ± 1.40) compared with paclitaxel-resistant model (72.61% ± 0.94) with p < 0.001. Additionally, BCSC-enriched population in mammospheres assay showed paclitaxel intrinsic resistance, with MFIn values of 47.82% ± 3.17 (231) and 31.24% ± 5.59 (231PTR). We next measured the ALDH1 activity using the ALDEFLUOR TM assay. Obtained data revealed a significant increase of ALDH1 activity in the resistant counterpart 231DXR (26.52% ± 1.81) compared to parental (10.81% ± 5.74; p < 0.05) ( Figure 3E). Interestingly, ALDH-bright (ALDH br ) cells from both sensitive and resistant cell lines were also increased, with no significance, after 12 and 24 h of doxorubicin treatment, reaching values of 42.15% ± 10.53 and 38.10% ± 6.71, respectively. 231DXR cells exhibited a non-significant increase of ALDH1 activity in comparison with 231 cells after 12 h treatment (36.87% ± 3.56 vs. 17.28% ± 5.68), showing similar levels at 24 h after treatment.

BCSC-Enriched Population in Sensitive and Resistant Cell Lines
Despite the fact that BCSCs are a very low-represented population within tumors, selection or enrichment of these cells after chemotherapy treatment can result in a tumor relapse due to their Regarding an ALDH br population in paclitaxel-resistant 231 cell line, no significant differences were found between sensitive and resistant cell models ( Figure 3F). Paclitaxel treatment decreased ALDH br population after 24 h of treatment in sensitive cells (p < 0.01), whereas ALDH br population in 231PTR cells remained similar during treatment.
We next measured the ALDH1 activity using the ALDEFLUOR TM assay. Obtained data revealed a significant increase of ALDH1 activity in the resistant counterpart 231DXR (26.52% ± 1.81) compared to parental (10.81% ± 5.74; p < 0.05) (Figures 3E). Interestingly, ALDH-bright (ALDH br ) cells from both sensitive and resistant cell lines were also increased, with no significance, after 12 and 24 h of doxorubicin treatment, reaching values of 42.15% ± 10.53 and 38.10% ± 6.71, respectively. 231DXR cells exhibited a non-significant increase of ALDH1 activity in comparison with 231 cells after 12 h treatment (36.87% ± 3.56 vs. 17.28% ± 5.68), showing similar levels at 24 h after treatment.
Regarding an ALDH br population in paclitaxel-resistant 231 cell line, no significant differences were found between sensitive and resistant cell models ( Figure 3F). Paclitaxel treatment decreased ALDH br population after 24 h of treatment in sensitive cells (p < 0.01), whereas ALDH br population in 231PTR cells remained similar during treatment. In the same conditions, ALDH activity quantification (%) for (E) 231DXR and (F) 231PTR compared to 231. Results are expressed as mean ± SEM. * (p < 0.05), ** (p < 0.01) and *** (p < 0.001) indicate levels of statistically significance or n.s. when no significance was found.

Evaluation of EMT in Drug-Resistant TNBC Models
Cell plasticity is a key process in the development of drug resistance [56]. It has been described that induction of EMT can lead to cell dedifferentiation, acquiring BCSC features such as for five days. In the same conditions, ALDH activity quantification (%) for (E) 231DXR and (F) 231PTR compared to 231. Results are expressed as mean ± SEM. * (p < 0.05), ** (p < 0.01) and *** (p < 0.001) indicate levels of statistically significance or n.s. when no significance was found.

Evaluation of EMT in Drug-Resistant TNBC Models
Cell plasticity is a key process in the development of drug resistance [56]. It has been described that induction of EMT can lead to cell dedifferentiation, acquiring BCSC features such as chemoresistance [11,39]. We wondered if the development of this phenotype was due to EMT process triggered by chemotherapy treatment. To that purpose, RNA expression of the EMT-related transcription factors snail, slug, twist, zeb1 and zeb2 was evaluated after doxorubicin or paclitaxel treatment. The basal levels of snail were 2-fold higher in 231DXR compared to 231 (p < 0.05). When doxorubicin was added, interestingly, a significant increase of snail was observed in the 231 cell line, while such a strong effect was not observed in 231DXR ( Figure 4A). Vimentin gene expression levels were analyzed since it is an intermediate filament protein related to the mesenchymal phenotype. After 12 h of doxorubicin treatment, vimentin expression was significantly increased in both cell lines with a higher growth in 231DXR ( Figure 4A).
Concerning paclitaxel resistant cells, slug transcription factor was upregulated in 231PTR compared with sensitive 231 cells (p < 0.05) but opposite results were observed in vimentin levels at the same time points ( Figure 4B). while such a strong effect was not observed in 231DXR ( Figure 4A). Vimentin gene expression levels were analyzed since it is an intermediate filament protein related to the mesenchymal phenotype. After 12 h of doxorubicin treatment, vimentin expression was significantly increased in both cell lines with a higher growth in 231DXR ( Figure 4A).
Concerning paclitaxel resistant cells, slug transcription factor was upregulated in 231PTR compared with sensitive 231 cells (p < 0.05) but opposite results were observed in vimentin levels at the same time points ( Figure 4B).

C75, EGCG and Its Derivatives G28, G56, and G37 Effect in BCSC-Enriched Populations
MFIn was performed in 231, 231DXR, and 231PTR cell lines in order to study the impact of FASN inhibition in the BCSC-enriched population.

C75, EGCG and Its Derivatives G28, G56, and G37 Effect in BCSC-Enriched Populations
MFIn was performed in 231, 231DXR, and 231PTR cell lines in order to study the impact of FASN inhibition in the BCSC-enriched population.
were analyzed since it is an intermediate filament protein related to the mesenchymal phenotype. After 12 h of doxorubicin treatment, vimentin expression was significantly increased in both cell lines with a higher growth in 231DXR ( Figure 4A).
Concerning paclitaxel resistant cells, slug transcription factor was upregulated in 231PTR compared with sensitive 231 cells (p < 0.05) but opposite results were observed in vimentin levels at the same time points ( Figure 4B).

C75, EGCG and Its Derivatives G28, G56, and G37 Effect in BCSC-Enriched Populations
MFIn was performed in 231, 231DXR, and 231PTR cell lines in order to study the impact of FASN inhibition in the BCSC-enriched population.

Discussion
The two major molecular subtypes within TNBCs are the CL and ML. The last one accounts for 20% of TNBCs (the second in incidence after basal-like in TNBC [9]), and exhibits an enrichment of the EMT features, showing BCSC properties, and having a poor prognosis [9,57]. Furthermore, the enrichment of CSC features has been found in tumors that progressed after therapy [52,53]. In the present study, an ML subtype cell line MDA-MB-231 (231) and its resistant derivatives 231DXR and 231PTR were used to evaluate the BCSC population in chemotherapy resistance acquirement.
BCSCs are a low-represented population within tumors, primary and established cell lines; therefore, different methodologies have been settled to study the BCSC-enriched population with these cells in vitro. The 231 cell line is known to form vague mammospheres that can be passaged for several generations that can be passed on for several generations [58][59][60][61]. While no differences were found regarding the index formation between the three cell lines assayed, significant differences were observed using this methodology under drug treatment. Both 231DXR and 231PTR showed increased ability to form mammospheres under drug treatment compared to 231 showing both, drug resistance and anchorage-independent growth. The ALDEFLUOR TM assay indicated that ALDH br cell population was significantly increased in the 231DXR cell line, whereas ALDH br 231PTR cell percentage remained similar to that in the parental. Interestingly, the number of ALDH br cells significantly increased in the parental cell line after doxorubicin treatment. Taken together, these results suggest that 231DXR cell line shows an increase in BCSC features, and that these features can be induced after doxorubicin treatment. On the other hand, despite the fact that we did not observe an obvious enrichment in BCSC characteristics in 231PTR compared to parental cells, the intrinsic BCSC features of this ML cell line were maintained.
Cell plasticity plays a key role in tumor progression, and the plastic CSC concept defines that bidirectional conversions between non-CSCs and CSCs may exist. Thus, the activation of the EMT process may lead to the CSC phenotype [11,38,39]. 231DXR and 231PTR cell lines showed higher resistance to their respective chemotherapeutic agent in comparison with 231 in adherent or suspension condition. Additionally, and as expected, doxorubicin and paclitaxel showed more cytotoxic effects in adherent conditions compared to non-adherent ones. Interestingly, the ALDH1 activity was enhanced after doxorubicin treatment in both sensitive and resistant 231DXR cell models. This phenotype is associated with chemoresistance in patients [62]. Remarkably, basal expression of snail, an EMT inducer directly associated with tumor development and relapse [63][64][65], significantly increased in 231DXR compared to sensitive cells. Regarding paclitaxel resistance, no differences were observed in the ALDH br population. The EMT-related transcription factor slug, linked to cancer progression and BCSC activity [66], was upregulated in 231PTR cells compared to 231 cell model with no changes in vimentin levels.
Although doxorubicin counts with undeniable advantages as an antitumor drug, it has been shown to promote stemness in murine cell lines [67][68][69]. Recent evidence showed that doxorubicin treatment increased stem cell-related signaling pathways, explaining the inherent chemoresistance of MDA-MB-231 (compared to another ML cell line, Hs578T) [70]. Previous studies showed the ability of paclitaxel to increase the BCSC subpopulation in 231 TNBC cell lines [71]; however, non-obvious enrichment was observed in our hands, but a maintenance of the intrinsic stem features of this ML subtype. FASN overexpression confers many advantages to tumor cells such as the ability to preserve a high proliferation rate, and it also plays a key role in drug resistance acquisition [32,47,49]. We have proven that FASN inhibition overcame doxorubicin resistance in chemoresistant cell lines [34], and FASN protein levels significantly increased in the 231DXR cell model when treated with doxorubicin. It has been previously described in breast cancer that FASN protein levels increase as a mechanism to become resistant [72]. Gonzalez-Guerrico and coworkers showed that the inhibition of FASN led to the conversion of the ML phenotype to a non-malignant one by downregulating EMT markers in a breast cancer model [73]. Additionally, the FASN inhibitor metformin blocked lipogenesis, leading to a switch from basal-to luminal-like sphere morphologies [73]. Likewise, resveratrol, another FASN inhibitor, precluded the growth of the CSC population both in vitro and in vivo using a breast cancer xenograft model [74]. FASN expression has also been shown to be important in stemness preservation in glioma [75]. In this work, we evaluated the effect of the inhibition of FASN in an CSC population through the ability to hinder mammosphere-formation capacity compared to the cytotoxic effect in adherent culture. No differences were found using C75 in adherent or suspension conditions. G28 displayed a noticeably better inhibitory effect in non-adherent conditions in both sensitive and resistant cell lines. Interestingly, this outcome was not found when using any of the chemotherapeutical agents (doxorubicin or paclitaxel). These results set up the basis for further investigation using FASN inhibitors in combination with chemotherapy to target BCSCs as well as the bulk of the tumor to improve the outcome of patients with TNBC.
In conclusion, the results presented in this work suggest that chemoresistance of 231DXR and 231PTR (ML cell models) is not only a static drug-selected state, but also an acquired reversible phenotype that can be triggered during treatment. 231DXR presented a larger BCSC population compared to the doxorubicin sensitive 231 cell line, which was also enhanced with the presence of the chemotherapeutical agent. This might be explained by the activation of EMT induced through snail. Despite the preliminary nature of the obtained results, they provide a rationale to suppress FASN as a potential strategy for a highly proliferative neoplasia such as TNBC.

Cell Culture and Development of Doxorubicin-and Paclitaxel-Resistant TNBC Cells
MDA-MB-231 cell line was obtained from the American Type Culture Collection (ATCC) and was routinely grown in DMEM (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% FBS (HyClone Laboratories, GE Healthcare, Chicago, IL, USA), 1% L-glutamine (Gibco), and 1% sodium pyruvate (Gibco), 50 U/mL Pen/Strep (Linus). Cells were kept at 37 • C and 5% CO 2 atmosphere. Doxorubicin-resistant cells MDA-MB-231 (231DXR) and paclitaxel-resistant cells (231PTR) were developed using a stepwise selection method. Increasing doses of doxorubicin (TEDEC-Meiji Farma, Alcalá de Henares, Spain) and paclitaxel (Accord Healthcare Ltd., Thaltej, India) were added until their corresponding IC 50 was reached. Briefly, cells were initially treated with a concentration of 0.1xIC 50 of each drug. After 48 h, treatment medium was replaced for fresh medium. When the cells were capable of growing and reached appropriate confluence, they were treated with double the previous concentration for 48 h. The stepwise selection method was subsequently performed until the final concentration of drug was that of the parental IC 50 . It took around six months for each cell line. Resistance was confirmed by cell viability assay.

Western Blot Analysis of Cell Lysates
Cell lines were treated during 12 or 24 h with half the concentration of doxorubicin that inhibited 50% (0.5xIC 50 ) of the MDA-MB-231 parental cells' viability previously determined with MTT assay after 48 h [34]. A final concentration of 3xIC 50 of paclitaxel was used (Supplementary Figure S1). After treatment, parental and resistant TNBC cells were lysed in ice-cold lysis buffer (Cell Signaling Technology Inc., Danvers, MA, USA) with 100 µg/mL phenylmethylsulfonyl fluoride (PMSF) by vortexing every 5 min for 30 min. Equal amounts of protein were heated in an lithium dodecyl sulfate (LDS) Sample Buffer with Sample Reducing Agent (Invitrogen, Carlsbad, CA, USA) for 10 min at 70 • C, electrophoresed on SDS-polyacrylamide gel (SDS-PAGE), and transferred onto nitrocellulose membranes. Blots were incubated for 1 h in blocking buffer (5% powdered-skim milk in Phosphate-buffered saline 0.05% Tween (PBS-T)) and incubated overnight at 4 • C with the appropriate antibodies diluted in blocking buffer (Table 1). Specific horseradish peroxidase (HRP)-conjugated secondary antibodies were incubated for 1 h at room temperature. The immune complexes were detected using a chemiluminescent HPR substrate (Super Signal West Femto (Thermo Scientific Inc.) or Immobilon Western (Merck Millipore, Burlington, MA, USA)). β-actin (Santa Cruz Biotechnology Inc., Dallas, TX, USA) was used as a control of protein loading. Western blot analyses were repeated at least four times and representative results are shown. Quantification analyses of Western blot data were performed using a computer-assisted densitometer. The results are expressed as ratio of protein levels vs. β-actin levels.

Cell Viability Assays
To elucidate inhibitory effects of chemotherapy drugs and FASN inhibitors, 231 and resistant counterparts were seeded in 96-well microplates at a cell density of 4 × 10 3 cells/well in their corresponding growth medium. After 24 h, culture medium was removed and 100 µL of fresh medium with increasing doses of paclitaxel, or FASN inhibitors (EGCG (Sigma-Aldrich, St. Louis, MO, USA; dissolved in PBS 5% DMSO), C75 (Sigma), EGCG-derivatives G28, G56 or G37 (dissolved in DMSO)) were added to each well. The synthesis of FASN inhibitors G28, G56, and G37 was performed as previously described [45]. Following 48 h treatment, a colorimetric MTT assay was used to measure cell viability as previously described [27]. Graph plots were performed in order to elucidate paclitaxel resistance and inhibitory concentrations.
On the other hand, cell viability inhibition experiments were performed seeding 5000 cells in adherent 6-well cell culture microplates with DMEM. Cells were incubated for 24 h to allow cell attachment and then chemotherapy drugs (doxorubicin or paclitaxel) or FASN inhibitors were added at a unique specific concentration (IC 30 of parental cells calculated at 48 h) for five days. Finally, cell viability was also measured using a colorimetric MTT assay.

Mammosphere-Forming Assay
The mammosphere-forming procedure was used to evaluate an CSC population, following previously described protocols [55,56]. After cell detachment, 2000 cells were seeded into a non-adherent 6-well cell culture microplate. Cells were cultured with DMEM/F12 medium supplemented with B27, hEGF and hFGF (20 ng/mL), 1% L-glutamine, and 1% sodium pyruvate. Microplates were incubated for 5 days and mammospheres bigger than 50 µm were counted using an inverted optical microscope. Different parameters were calculated using the formulas described above: ( Formula:

Quantitative Real-Time PCR Analysis
Cells were treated with a concentration of doxorubicin or paclitaxel equal to the 0.5xIC50 and 3xIC50, respectively, obtained after 48 h of treatment. After 12 or 24 h treatment, cells were washed with PBS, and then Qiazol (Qiagen, Hilden, Germany) was added. RNeasy mini kit (Qiagen) was used to isolate total RNA according to manufacturer instructions. RNA was reverse-transcribed into complementary DNA (cDNA) using High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, USA). Different gene expression levels were determined using LightCycler® 480 Real-time PCR System (Roche, Basel, Switzerland) with LightCycler® 480 SYBR Green I Master (Roche), following manufacturer instructions. Primers used are shown in Table 2. RT-PCR analyses were performed at least three times and each gene was run in triplicate. Gene expression levels were quantified using the standard formula 2^CT and normalized to the housekeeping gene β-actin (2^CT).

Aldefluor Assay
Aldefluor TM kit (STEMCELL Technologies) was used to determine ALDH enzyme activity, following company guidelines. Briefly, ALDH reagent (BAAA) freely diffuses into viable cells and ALDH enzyme catalyzes its conversion to BAA, which is negatively charged and retained into cells. Intracellular BAA increases fluorescence which can be analyzed through flow cytometry. As seen before for 12 and 24 h treatments, 0.5xIC50 and 3xIC50 (calculate at 48 h on parental cells) for doxorubicin and paclitaxel, respectively, were used. After treatment, cells were detached and 2 × 10 5 cells per sample were collected. After washing with PBS, 500 µL of the kit buffer was added to each condition. Afterwards, 2.5 µL of the ALDH reagent was added and 250 µL of the suspension was immediately transferred to a new Eppendorf with 2.5 µL of the ALDH inhibitor N,Ndiethylaminobenzaldehyde (DEAB), in order to consider background fluorescence (Supplementary Figure 3). All samples were incubated at 37 C for 40 min. Cells were then centrifuged and washed and samples were prepared to be analyzed through cytometry (FACSCalibur II, BD Bioscience). Images were obtained with FlowJo software (Flow Jo LLC, Ashland, OR, USA.

Conclusions
The study of the BCSC-enriched population in an ML TNBC cell line (MDA-MB-231) and its doxorubicin-resistant derivative (231DXR) showed increased BCSC features in both sensitive and resistant cell models after chemotherapy treatment. We have previously recognized FASN as a

Quantitative Real-Time PCR Analysis
Cells were treated with a concentration of doxorubicin or paclitaxel equal to the 0.5xIC 50 and 3xIC 50 , respectively, obtained after 48 h of treatment. After 12 or 24 h treatment, cells were washed with PBS, and then Qiazol (Qiagen, Hilden, Germany) was added. RNeasy mini kit (Qiagen) was used to isolate total RNA according to manufacturer instructions. RNA was reverse-transcribed into complementary DNA (cDNA) using High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, USA). Different gene expression levels were determined using LightCycler ® 480 Real-time PCR System (Roche, Basel, Switzerland) with LightCycler ® 480 SYBR Green I Master (Roche), following manufacturer instructions. Primers used are shown in Table 2. RT-PCR analyses were performed at least three times and each gene was run in triplicate. Gene expression levels were quantified using the standard formula 2ˆ∆CT and normalized to the housekeeping gene β-actin (2ˆ∆∆CT).

Aldefluor Assay
Aldefluor TM kit (STEMCELL Technologies) was used to determine ALDH enzyme activity, following company guidelines. Briefly, ALDH reagent (BAAA) freely diffuses into viable cells and ALDH enzyme catalyzes its conversion to BAA, which is negatively charged and retained into cells. Intracellular BAA increases fluorescence which can be analyzed through flow cytometry. As seen before for 12 and 24 h treatments, 0.5xIC 50 and 3xIC 50 (calculate at 48 h on parental cells) for doxorubicin and paclitaxel, respectively, were used. After treatment, cells were detached and 2 × 10 5 cells per sample were collected. After washing with PBS, 500 µL of the kit buffer was added to each condition. Afterwards, 2.5 µL of the ALDH reagent was added and 250 µL of the suspension was immediately transferred to a new Eppendorf with 2.5 µL of the ALDH inhibitor N,N-diethylaminobenzaldehyde (DEAB), in order to consider background fluorescence (Supplementary Figure S3). All samples were incubated at 37 • C for 40 min. Cells were then centrifuged and washed and samples were prepared to be analyzed through cytometry (FACSCalibur II, BD Bioscience). Images were obtained with FlowJo software (Flow Jo LLC, Ashland, OR, USA.

Conclusions
The study of the BCSC-enriched population in an ML TNBC cell line (MDA-MB-231) and its doxorubicin-resistant derivative (231DXR) showed increased BCSC features in both sensitive and resistant cell models after chemotherapy treatment. We have previously recognized FASN as a possible co-target to resensitize chemoresistant cells to doxorubicin [34]. Moreover, its implication in drug resistance acquisition in breast cancer has already been identified [72]. Results derived from the present paper showed that the newly developed FASN inhibitor EGCG-derived G28 had a strong inhibitory effect on the mammosphere-formation capacity (a stem feature) in sensitive and both drug-resistant models, 231DXR and 231PTR. However, these outcomes are still in initial states and more approaches should be performed. According to recent published research, FASN would not only play an important role in a highly proliferative neoplasia such as TNBC, but also in drug resistance acquisition and in sustaining malignancy in cancer [73][74][75], being a promising (co)target for patients with TNBC that progress to current treatments.