Selectively Targeting Breast Cancer Stem Cells by 8-Quinolinol and Niclosamide

Cancer maintenance, metastatic dissemination and drug resistance are sustained by cancer stem cells (CSCs). Triple negative breast cancer (TNBC) is the breast cancer subtype with the highest number of CSCs and the poorest prognosis. Here, we aimed to identify potential drugs targeting CSCs to be further employed in combination with standard chemotherapy in TNBC treatment. The anti-CSC efficacy of up to 17 small drugs was tested in TNBC cell lines using cell viability assays on differentiated cancer cells and CSCs. Then, the effect of 2 selected drugs (8-quinolinol -8Q- and niclosamide -NCS-) in the cancer stemness features were evaluated using mammosphere growth, cell invasion, migration and anchorage-independent growth assays. Changes in the expression of stemness genes after 8Q or NCS treatment were also evaluated. Moreover, the potential synergism of 8Q and NCS with PTX on CSC proliferation and stemness-related signaling pathways was evaluated using TNBC cell lines, CSC-reporter sublines, and CSC-enriched mammospheres. Finally, the efficacy of NCS in combination with PTX was analyzed in vivo using an orthotopic mouse model of MDA-MB-231 cells. Among all tested drug candidates, 8Q and NCS showed remarkable specific anti-CSC activity in terms of CSC viability, migration, invasion and anchorage independent growth reduction in vitro. Moreover, specific 8Q/PTX and NCS/PTX ratios at which both drugs displayed a synergistic effect in different TNBC cell lines were identified. The sole use of PTX increased the relative presence of CSCs in TNBC cells, whereas the combination of 8Q and NCS counteracted this pro-CSC activity of PTX while significantly reducing cell viability. In vivo, the combination of NCS with PTX reduced tumor growth and limited the dissemination of the disease by reducing circulating tumor cells and the incidence of lung metastasis. The combination of 8Q and NCS with PTX at established ratios inhibits both the proliferation of differentiated cancer cells and the viability of CSCs, paving the way for more efficacious TNBC treatments.


Introduction
Breast cancer is the most frequent cancer among women, accounting for 24.5% of all new female cancer cases in 2020 worldwide, and overall, it is the leading cause of cancer death in women [1]. The triple negative breast cancer (TNBC) subtype represents approximately 15% to 20% of breast cancers and comprises a heterogeneous group of

Identification of Candidates with Selective Anti-CSC Activity
Selected drugs were then subjected to proliferation assays in CSC and non-CSC subpopulations sorted from the MDA-MB-231-ALDH1A1:tdTomato reporter cell line [13]. Sorted cells were characterized by the expression of stemness markers (Supplementary Figure S2) and further subjected to conventional proliferation. No significant differences in treatment sensitivity were observed between CSCs and non-CSCs when treated with the reference drug PTX, thereby confirming that the chemotherapeutic agent did not show a selective cytotoxic effect against CSCs ( Figure 1A). Moreover, 8Q and NCS were the only drugs that exhibited significantly higher sensitivity in CSCs compared to non-CSCs ( Figure 1B). Indeed, IC 50 values for 8Q and NCS drugs were 1.33 and 3.66-fold lower (p < 0.05) in the CSC in comparison with the non-CSC subpopulation, respectively ( Figure 1A,B). The results for SAL also moved in the same direction, but differences among IC 50 values of CSCs and non-CSCs were not statistically significant. Conversely, NTC, VS, DSF, FLU and DFT drugs showed no differences in cell viability between CSCs and non-CSCs and were discarded. Figure S2) and further subjected to conventional proliferation. No significant differences in treatment sensitivity were observed between CSCs and non-CSCs when treated with the reference drug PTX, thereby confirming that the chemotherapeutic agent did not show a selective cytotoxic effect against CSCs ( Figure 1A). Moreover, 8Q and NCS were the only drugs that exhibited significantly higher sensitivity in CSCs compared to non-CSCs (Figure 1B). Indeed, IC50 values for 8Q and NCS drugs were 1.33 and 3.66-fold lower (p < 0.05) in the CSC in comparison with the non-CSC subpopulation, respectively ( Figure 1A,B). The results for SAL also moved in the same direction, but differences among IC50 values of CSCs and non-CSCs were not statistically significant. Conversely, NTC, VS, DSF, FLU and DFT drugs showed no differences in cell viability between CSCs and non-CSCs and were discarded. The effect of 8Q, NCS and SAL on CSCs was further studied in low attachment conditions, where mammosphere formation (MSF) is promoted. As shown in Figure 2A, the inhibition of mammosphere-forming efficiency increased in a dose-dependent manner in all three anti-CSC drugs, being higher in the case of 8Q and NCS (Figure 2A), which prevented MSF in 27.4% ± 6.7% and 13.8% ± 7.5%, respectively, when the highest concentration was tested ( Figure 2B). Accordingly, these were the compounds that showed the lowest IC50 values, 2.71 µM ± 0.14 for 8Q and 1.65 µM ± 0.79 for NCS. Significantly, the IC50 value of 8Q in low attachment conditions was lower than the IC50 value in conventional cultures (25.55 ± 0.151 µM, data in Table 1). On the other hand, the effect on MSF was remarkably lower for SAL (IC50 of 4.95 ± 0.83 µM) in comparison with 8Q and NCS, but significantly higher (p < 0.0001) than that produced by the anti-neoplastic reference drug The effect of 8Q, NCS and SAL on CSCs was further studied in low attachment conditions, where mammosphere formation (MSF) is promoted. As shown in Figure 2A, the inhibition of mammosphere-forming efficiency increased in a dose-dependent manner in all three anti-CSC drugs, being higher in the case of 8Q and NCS (Figure 2A), which prevented MSF in 27.4% ± 6.7% and 13.8% ± 7.5%, respectively, when the highest concentration was tested ( Figure 2B). Accordingly, these were the compounds that showed the lowest IC 50 values, 2.71 µM ± 0.14 for 8Q and 1.65 µM ± 0.79 for NCS. Significantly, the IC 50 value of 8Q in low attachment conditions was lower than the IC 50 value in conventional cultures (25.55 ± 0.151 µM, data in Table 1). On the other hand, the effect on MSF was remarkably lower for SAL (IC 50 of 4.95 ± 0.83 µM) in comparison with 8Q and NCS, but significantly higher (p < 0.0001) than that produced by the anti-neoplastic reference drug PTX (Figure 2A,B). Indeed, the higher dose tested for PTX (2 µM, twice the IC 50 value in conventional cultures) did not reach 50% mammosphere growth inhibition. These results pointed out that PTX was not as effective as 8Q, NCS or SAL in selectively eradicating CSC populations. Moreover, given that 8Q and NCS showed the highest efficacy in inhibiting MSF, these two drugs were chosen to continue with functional assays. PTX (Figure 2A,B). Indeed, the higher dose tested for PTX (2 µM, twice the IC50 value in conventional cultures) did not reach 50% mammosphere growth inhibition. These results pointed out that PTX was not as effective as 8Q, NCS or SAL in selectively eradicating CSC populations. Moreover, given that 8Q and NCS showed the highest efficacy in inhibiting MSF, these two drugs were chosen to continue with functional assays. Percentage of mammosphere after 7-day incubation with 2 µM PTX, 10 µM SAL and 8Q, and 15 µM NCS. Statistically significant results were obtained for all tested compounds when compared to PTX. Data are represented as the mean ± SEM of three independent experiments. **** p < 0.0001.

Q and NCS Affect Specific Stem Cell-Like Features of Breast CSC Subpopulation
We next examined whether 8Q and NCS impacted other distinctive features of CSCs, including migration, anchorage-independent growth (neoplastic transformation) and cell invasion ability, using the MDA-MB-231 cell line ( Figure 3). Regarding migration, significantly lower wound closure was observed in cells treated with anti-CSC drugs compared to the controls ( Figure 3A,B). At 24 h, the open wound detected for the control samples was almost minimal (4.45% ± 1.12%), while in cells treated with 8Q and NCS, it was 33.9% ± 1.8% and 24.9% ± 1.7%, respectively ( Figure 3A). Of note, treatment with PTX also showed a high inhibitory effect on cell migration ( Figure 3A,B). Interestingly, anchorageindependent growth of MDA-MB-231 cells in soft agar was significantly inhibited by treatments with 8Q and NCS ( Figure 3C). At the highest tested concentrations of 8Q (50 µM) and NCS (15 µM), the capacity of malignant cells to form colonies was significantly reduced to 5.40% ± 1.23% and 11.48% ± 1.11% of the growth observed in the non-treated cells, respectively. Conversely, PTX treatment was not able to reduce the anchorage-independent growth below 50%, thereby demonstrating that 8Q and NCS had superior anti-CSC activity ( Figure 3C). Furthermore, we explored the effects of 8Q and NCS on the invasiveness of stem celllike cells. The results showed that 8Q significantly reduced the ability of cells to cross the Matrigel-covered 8-µm filter, leading to almost complete inhibition of cell invasion ability ( Figure 3D). In the case of NCS, no inhibitory effect was observed in the cell invasion assays. PTX also showed some efficacy in reducing cell invasion, but it was significantly lower than that obtained with 8Q ( Figure 3D). Altogether, the results demonstrated that 8Q and NCS had strong anti-CSC activity in MDA-MB-231 cells by inhibiting crucial stemness features of CSCs.

Q and NCS Affect Specific Stem Cell-Like Features of Breast CSC Subpopulation
We next examined whether 8Q and NCS impacted other distinctive features of CSCs, including migration, anchorage-independent growth (neoplastic transformation) and cell invasion ability, using the MDA-MB-231 cell line ( Figure 3). Regarding migration, significantly lower wound closure was observed in cells treated with anti-CSC drugs compared to the controls ( Figure 3A,B). At 24 h, the open wound detected for the control samples was almost minimal (4.45% ± 1.12%), while in cells treated with 8Q and NCS, it was 33.9% ± 1.8% and 24.9% ± 1.7%, respectively ( Figure 3A). Of note, treatment with PTX also showed a high inhibitory effect on cell migration ( Figure 3A,B). Interestingly, anchorage-independent growth of MDA-MB-231 cells in soft agar was significantly inhibited by treatments with 8Q and NCS ( Figure 3C). At the highest tested concentrations of 8Q (50 µM) and NCS (15 µM), the capacity of malignant cells to form colonies was significantly reduced to 5.40% ± 1.23% and 11.48% ± 1.11% of the growth observed in the non-treated cells, respectively. Conversely, PTX treatment was not able to reduce the anchorage-independent growth below 50%, thereby demonstrating that 8Q and NCS had superior anti-CSC activity ( Figure 3C). Furthermore, we explored the effects of 8Q and NCS on the invasiveness of stem cell-like cells. The results showed that 8Q significantly reduced the ability of cells to cross the Matrigel-covered 8-µm filter, leading to almost complete inhibition of cell invasion ability ( Figure 3D). In the case of NCS, no inhibitory effect was observed in the cell invasion assays. PTX also showed some efficacy in reducing cell invasion, but it was significantly lower than that obtained with 8Q ( Figure 3D). Altogether, the results demonstrated that 8Q and NCS had strong anti-CSC activity in MDA-MB-231 cells by inhibiting crucial stemness features of CSCs. All statistical analyses were performed compared to the control non-treated cells (** p < 0.01, *** p < 0.001, **** p < 0.0001).

Q and NCS Display a Synergic Anti-Proliferation Effect When Combined with PTX at Specific Ratios in Different TNBC Cell Lines
Since anti-CSC drugs are not meant to be used solely in cancer treatment, the combination of 8Q and NCS in combination with PTX was evaluated on MDA-MB-231, HCC-1806 and MDA-MB-468 cell lines by MTT assays. Depending on the drug doses, a combination can yield a synergistic, additive or antagonistic effect; thus, the combination index (CI) for each drug interaction was estimated for different drug ratios ( Figure 4). To better determine the potential synergism of the drugs, assays were conducted by fixing each drug at its IC50 value and varying the other one. For the PTX-8Q combination, an enhanced cytotoxic effect was observed at multiple PTX:8Q ratios, of which 1:12.5 was the one that resulted in the highest synergistic effect with the lowest CI value (CI = 0.06) ( Figure 4A,B).

Q and NCS Display a Synergic Anti-Proliferation Effect When Combined with PTX at Specific Ratios in Different TNBC Cell Lines
Since anti-CSC drugs are not meant to be used solely in cancer treatment, the combination of 8Q and NCS in combination with PTX was evaluated on MDA-MB-231, HCC-1806 and MDA-MB-468 cell lines by MTT assays. Depending on the drug doses, a combination can yield a synergistic, additive or antagonistic effect; thus, the combination index (CI) for each drug interaction was estimated for different drug ratios ( Figure 4). To better determine the potential synergism of the drugs, assays were conducted by fixing each drug at its IC 50 value and varying the other one. For the PTX-8Q combination, an enhanced cytotoxic effect was observed at multiple PTX:8Q ratios, of which 1:12.5 was the one that resulted in the highest synergistic effect with the lowest CI value (CI = 0.06) ( Figure 4A,B). For the PTX-NCS combination, the greatest drug synergy was observed at a 1:2 ratio of PTX:NCS with a CI value close to 0, thus indicating a strong positive synergism between both drugs ( Figure 4C,D). Accordingly, these ratios were used in later studies on MDA-MB-231 cells.

MB-231 cells.
Combination assays of 8Q and NCS with PTX in HCC-1806 and MDA-MB-468 cell lines also showed improved cytotoxicity when combined at specific ratios (see Supplementary Figures S4 and S5). In HCC-1806 cells, the greatest synergy was obtained at 1:5 (with a CI close to 0) and 1:0.4 (CI = 0.17) ratios for the PTX-8Q and PTX-NCS treatments, respectively. In MDA-MB-468 cells, it was obtained at 1:1250 (CI = 0.02) and 1:1000 (CI = 0.12) drug ratios. The differences in the established optimal ratios between these cell lines are likely a reflection of differential molecular and intracellular signaling changes, as well as differential drug sensitivity to treatments tested, demonstrating the cell-type-specific response to combined therapy.  Combination assays of 8Q and NCS with PTX in HCC-1806 and MDA-MB-468 cell lines also showed improved cytotoxicity when combined at specific ratios (see Supplementary  Figures S4 and S5). In HCC-1806 cells, the greatest synergy was obtained at 1:5 (with a CI close to 0) and 1:0.4 (CI = 0.17) ratios for the PTX-8Q and PTX-NCS treatments, respectively. In MDA-MB-468 cells, it was obtained at 1:1250 (CI = 0.02) and 1:1000 (CI = 0.12) drug ratios. The differences in the established optimal ratios between these cell lines are likely a reflection of differential molecular and intracellular signaling changes, as well as differential drug sensitivity to treatments tested, demonstrating the cell-type-specific response to combined therapy.

Combination of 8Q or NCS with PTX Increases the Anti-CSC Efficacy of the Drugs
To evaluate whether combination therapy of PTX-8Q and PTX-NCS could offer therapeutic advantages, drug ratios with the best CI values were evaluated in the fluorescent tdTomato CSC models. Changes in the CSC subpopulation were monitored by flow cytometry and stem cell gene expression analysis ( Figure 5 for the MDA-MB-231 cell line and Supplementary Figure S6 for MDA-MB-468 and HCC-1806). As expected, following PTX exposure, the relative abundance of CSCs (tdTomato+ cells) significantly increased (p < 0.0001) in a dose-dependent manner in MDA-MB-231 cells ( Figure 5A). Conversely, upon incubation with 8Q and NCS anti-CSC drugs, the relative amount of MDA-MB-231-tdTomato+ cells remarkably decreased (p < 0.0001 for both drugs) when compared to PTX treatment. Of note, low doses of 8Q achieved a significant reduction in the MDA-MB-231 CSC subpopulation, while at higher doses, the effect on CSCs was only preventive. In the case of NCS treatment, such a decrease in CSC-tdTomato+ was dose dependent. Interestingly, combined treatments of 8Q and NCS with PTX significantly abrogated the relative increase of MDA-MB-231-tdTomato+ cells induced by PTX, as well as significantly enhanced the anti-CSC effect of 8Q and NCS individual treatments (p = 0.0201 and p = 0.003 for lower doses of PTX-8Q and PTX-NCS, respectively, while for higher doses was p < 0.0001 for both combinations; compared to individual anti-CSC therapy). Values below than, equal to or above 1 indicate reduction, maintenance or increase of CSC-tdTomato+ cells within population, respectively. Data are represented as the mean ± SEM of three independent experiments. Statistic t-test analyses were performed to compare drug combinations with single anti-CSC drug treatments at the corresponding equivalent drug dose (* p < 0.05, *** p < 0.001, **** p < 0.0001). (B) Changes in the stem cell gene expression profile determined by quantitative RT-PCR. Results are expressed as normalized relative quantities (NRQ) and referred to as control conditions. Concentrations tested for drugs were 0.5 and 1 µM for PTX, 6.25 and 12.5 µM for 8Q, 1 and 2 µM for NCS and the corresponding combined ratios.
The effect of the combination of 8Q and NCS with PTX on the CSC subpopulation was further confirmed by analyzing their effect on mammosphere viability (MSV) and the ability to form new ones (mammosphere-forming efficiency, MSF) in all three TNBC cell lines ( Figure 6 and Supplementary Figure S6). Consistent with previous results, the ability of MDA-MB-231 stem-like cells to grow as mammospheres was significantly reduced when treated with 8Q and NCS (p < 0.0001 in both cases). This reduction was significantly greater when drugs were used in combination (p = 0.0014 and p = 0.0101 for combinations Values below than, equal to or above 1 indicate reduction, maintenance or increase of CSC-tdTomato+ cells within population, respectively. Data are represented as the mean ± SEM of three independent experiments. Statistic t-test analyses were performed to compare drug combinations with single anti-CSC drug treatments at the corresponding equivalent drug dose (* p < 0.05, *** p < 0.001, **** p < 0.0001). (B) Changes in the stem cell gene expression profile determined by quantitative RT-PCR. Results are expressed as normalized relative quantities (NRQ) and referred to as control conditions. Concentrations tested for drugs were 0.5 and 1 µM for PTX, 6.25 and 12.5 µM for 8Q, 1 and 2 µM for NCS and the corresponding combined ratios.
These results were further confirmed in the MDA-MB-468 and HCC-1806 TNBC cell lines (Supplementary Figure S6A,C). Upon incubation of TNBC cell lines with 8Q and NCS anti-CSC drugs, the relative presence of tdTomato+ cells was significantly reduced (p < 0.0001; for both drugs in both cell lines) when compared to PTX treatment. In the case of NCS treatment, such a decrease was more evident and dose-dependent in HCC-1806 and MDA-MB-468 cell lines than in MDA-MB-231 (Supplementary Figure S5A,C). Furthermore, the anti-CSC effects of 8Q and NCS were significantly enhanced when drugs were administered in combination with PTX rather than as an individual therapy. Interestingly, the extent of the reduction was similar to that observed in MDA-MB-231 cells, moving from values at 1. These results were further confirmed by gene expression analysis in all three TNBC cell lines. In MDA-MB-231 cells, we found a remarkable relative increase in the expression of all stem cell genes analyzed, including ALDH1A1, ALOX5, CMKLR1, ABCg2, Notch4, Nanog and OCT4 after PTX individual treatment ( Figure 5B). Meanwhile, the 8Q and NCS treatments, either alone or in combination with PTX, substantially downregulated the expression of stem cell genes. In particular, 8Q single treatment led to a strong decrease in ALDH1A1, ALOX5 and ABCG2 mRNA levels, while for the other genes, the effect obtained was slighter. In the case of NCS treatment, such a decrease was greater in ALDH1A1, ALOX5 and Nanog mRNA, and indeed, was stronger than when treated with 8Q. The combination of both drugs with PTX remarkably downregulated the expression of almost all stem cell-like markers analyzed, such as ALDH1A1, ALOX5, Notch4 and Nanog, thus indicating preferential and efficient effects against CSCs when 8Q and NCS were used in combination with PTX ( Figure 5B). These results were further confirmed in the HCC-1806 and MDA-MB-468 cell lines (Supplementary Figure S6B,D). Importantly, in both cell lines, a prominent increase in stem cell gene expression was also obtained after PTX treatment. Such an increase was stronger in Notch4 and OCT4 mRNA than in the MDA-MB-231 cell line. Moreover, 8Q and NCS, either alone or in combination with PTX, remarkably downregulated stem cell gene expression in both cell lines, overcoming and, in some cases, only matching the effect of individual treatments in almost all genes analyzed.
The effect of the combination of 8Q and NCS with PTX on the CSC subpopulation was further confirmed by analyzing their effect on mammosphere viability (MSV) and the ability to form new ones (mammosphere-forming efficiency, MSF) in all three TNBC cell lines ( Figure 6 and Supplementary Figure S6). Consistent with previous results, the ability of MDA-MB-231 stem-like cells to grow as mammospheres was significantly reduced when treated with 8Q and NCS (p < 0.0001 in both cases). This reduction was significantly greater when drugs were used in combination (p = 0.0014 and p = 0.0101 for combinations of PTX with 8Q and NCS, respectively) than in single treatments ( Figure 6A). The anti-neoplastic drug PTX showed a limited impact on mammosphere growth, since the PTX dose did not succeed in reducing to 50% the MSF (75.9% ± 4.7%). Similar results were obtained when MSV was analyzed in MDA-MB-231 ( Figure 6B), where the effect of PTX was limited (85.5% ± 4.7%), but the use of 8Q and NCS induced a relevant loss of mammosphere viability, especially when combined with PTX (p = 0.0005 and p = 0.0186 for combinations with 8Q and NCS, respectively). The described results were confirmed in the HCC-1806 and MDA-MB-468 TNBC cell lines ( Figure S7). In these cells, PTX showed no impact on MFA or MSV (65.9% ± 4.3% and 83.3% ± 2.6% in MSF for HCC-1806 and MDA-MB-468 cells, respectively, while in MSV, it was 92.4% ± 3.9% and 86.7% ± 5.1%, respectively). On the contrary, the ability of stem-like tumor cells to grow as mammospheres from both TNBC cell lines was significantly reduced when treated with 8Q and NCS (p < 0.0001 for both drugs). Such reduction was significantly greater when drugs were used in combination (p < 0.0001 and p = 0.0031 for combinations of PTX with 8Q and NCS in HCC-1806 cells, respectively; p < 0.0001 and p = 0.0043 for MDA-MB-468 cells, respectively) than as single treatments ( Figure S7A,C). Similar findings were observed when MSV was analyzed ( Figure S6B,D). The use of 8Q and NCS induced a relevant loss of mammosphere viability, especially when combined with PTX (p = 0.0069 and p = 0.0014 for combinations with 8Q and NCS in HCC-1806, respectively; p = 0.004 and p = 0.0046 for MDA-MB-468 cells, respectively). Altogether, NCS showed a superior effect in comparison to 8Q treatment, either alone or in combination with PTX, regarding mammosphere formation and growth.  Results are represented as the mean ± SEM of three independent experiments and referred to as the non-treated control condition. Statistical t-test analysis was performed comparing combination therapy with single anti-CSC treatments at the corresponding equivalent drug dose (* p < 0.05, ** p < 0.01, *** p < 0.001).
2.6. The Combination of 8Q or NCS with PTX Inhibits NF-κB and Wnt/β-Catenin Signaling Pathways 8Q and NCS have been described as inhibitors of nuclear factor-κB (NF-κB) and Wnt/β-catenin signaling pathways, respectively. Both pathways are known to be crucial in maintaining the stem phenotype [24,25]. In order to elucidate whether the synergism observed on cell viability between 8Q and NCS and the PTX was sustained by alterations in these two pathways, MDA-MB-231 cells were treated with 8Q, NCS and PTX for 24 h, and signaling protein levels were assessed by Western blot. The results pointed out that a significant reduction of phospho-NF-κB (p-NF-κB) subunit p65 was observed for 8Q concentrations as low as 12.5 µM, while the total expression of NF-κB subunit p65 was not changed, thus confirming the role of the 8Q drug as a selective inhibitor of the NF-κB signaling pathway (Supplementary Figure S8A,C). As expected, PTX treatment showed no efficacy in inhibiting the NF-κB signaling pathway. Moreover, a marked reduction of the p-NF-kB p65 subunit was also observed in cells treated with the combination compared to non-treated (control) or PTX-treated cells, and the inhibition of NF-kB p65 phosphorylation with the PTX+8Q combination was as efficient as with the 8Q alone ( Figure  7). 2.6. The Combination of 8Q or NCS with PTX Inhibits NF-κB and Wnt/β-Catenin Signaling Pathways 8Q and NCS have been described as inhibitors of nuclear factor-κB (NF-κB) and Wnt/β-catenin signaling pathways, respectively. Both pathways are known to be crucial in maintaining the stem phenotype [24,25]. In order to elucidate whether the synergism observed on cell viability between 8Q and NCS and the PTX was sustained by alterations in these two pathways, MDA-MB-231 cells were treated with 8Q, NCS and PTX for 24 h, and signaling protein levels were assessed by Western blot. The results pointed out that a significant reduction of phospho-NF-κB (p-NF-κB) subunit p65 was observed for 8Q concentrations as low as 12.5 µM, while the total expression of NF-κB subunit p65 was not changed, thus confirming the role of the 8Q drug as a selective inhibitor of the NF-κB signaling pathway (Supplementary Figure S8A,C). As expected, PTX treatment showed no efficacy in inhibiting the NF-κB signaling pathway. Moreover, a marked reduction of the p-NF-kB p65 subunit was also observed in cells treated with the combination compared to non-treated (control) or PTX-treated cells, and the inhibition of NF-kB p65 phosphorylation with the PTX+8Q combination was as efficient as with the 8Q alone (Figure 7). The results are expressed as normalized protein levels, referred to β-actin expression, represented as mean ± SEM of three independent experiments. Statistic t-test analysis of 8Q and combined therapy was done in comparison with control non-treated cells (black asterisks, ** p < 0.01) as well as with PTX treatment (blue asterisks, * p < 0.05 and ** p < 0.01).
As for the NCS, MDA-MB-231 cells treated with the drug showed a significant dosedependent reduction of β-catenin protein level as well as an increase of phosphorylated GSK3-β (p-GSK3-β) protein (while total expression of GSK3-β was not changed) ( Figure 8 and Supplementary Figure S8B,D). Conversely, PTX treatment showed no efficacy in inhibiting the Wnt signaling pathway (Figure 8). Importantly, the Western blot analyses confirmed a significant increase of p-GSK3-β and reduction of β-Catenin proteins in TNBC cells when treated with combined therapy (Figure 8). The results are expressed as normalized protein levels, referred to β-actin expression, represented as mean ± SEM of three independent experiments. Statistic t-test analysis of 8Q and combined therapy was done in comparison with control non-treated cells (black asterisks, ** p < 0.01) as well as with PTX treatment (blue asterisks, * p < 0.05 and ** p < 0.01).
As for the NCS, MDA-MB-231 cells treated with the drug showed a significant dosedependent reduction of β-catenin protein level as well as an increase of phosphorylated GSK3-β (p-GSK3-β) protein (while total expression of GSK3-β was not changed) ( Figure 8 and Supplementary Figure S8B,D). Conversely, PTX treatment showed no efficacy in inhibiting the Wnt signaling pathway (Figure 8). Importantly, the Western blot analyses confirmed a significant increase of p-GSK3-β and reduction of β-Catenin proteins in TNBC cells when treated with combined therapy (Figure 8). Representative β-Catenin and GSK3-β protein levels after drug treatments. The β-actin protein expression level was used as a loading control. (B-D) Graphs represent the quantification band intensity signal referred to β-actin expression, represented as mean ± SEM of three independent experiments. Statistic t-test analysis of NCS and combined therapy was performed in comparison with control non-treated cells (black asterisks, ** p < 0.01 and **** p < 0.0001) as well as with PTX treatment (blue asterisks, * p < 0.05 and ** p < 0.01).

The Combination of PTX with NCS Enhances Abrogates the In Vivo Growth of MDA-MB-231 Tumors
As shown in previous in vitro studies, lower doses of NCS are needed to obtain more significant anti-CSC activity than 8Q. Therefore, we selected NCS to further test the effect of its combination with PTX in vivo. In detail, NOD/SCID mice with orthotopic MDA-MB-231 tumors were administered with the vehicle, PTX or the combination of PTX with NCS. The treatment schedule followed was established in accordance with previous tests performed in order to reduce side effects without compromising therapeutic efficacy. Tumor growth during the treatment period and the number of circulating tumor cells and lung metastasis at the end-point were evaluated. As for tumor growth, the combination of PTX with NCS had the strongest effect on reducing MDA-MB-231 tumor growth, resulting in smaller tumor volumes ( Figure 9A) and weights ( Figure 9B). These significant differences were already visible on day 9 and increased throughout the experiment ( Figure 9A). More specifically, analysis of excised tumors showed that combined treatment with PTX-NCS reduced both tumor volumes and weights 3.1-fold after co-treatment. In addition, the mice β-Catenin, total-and phosphorylated-GSK3-β protein levels expression levels after PTX and NCS treatment, both alone and in combination (1:2 ratio), were determined. (A-C) Representative β-Catenin and GSK3-β protein levels after drug treatments. The β-actin protein expression level was used as a loading control. (B-D) Graphs represent the quantification band intensity signal referred to β-actin expression, represented as mean ± SEM of three independent experiments. Statistic t-test analysis of NCS and combined therapy was performed in comparison with control non-treated cells (black asterisks, ** p < 0.01, *** p < 0.001 and **** p < 0.0001) as well as with PTX treatment (blue asterisks, * p < 0.05 and ** p < 0.01).

The Combination of PTX with NCS Enhances Abrogates the In Vivo Growth of MDA-MB-231 Tumors
As shown in previous in vitro studies, lower doses of NCS are needed to obtain more significant anti-CSC activity than 8Q. Therefore, we selected NCS to further test the effect of its combination with PTX in vivo. In detail, NOD/SCID mice with orthotopic MDA-MB-231 tumors were administered with the vehicle, PTX or the combination of PTX with NCS. The treatment schedule followed was established in accordance with previous tests performed in order to reduce side effects without compromising therapeutic efficacy. Tumor growth during the treatment period and the number of circulating tumor cells and lung metastasis at the end-point were evaluated. As for tumor growth, the combination of PTX with NCS had the strongest effect on reducing MDA-MB-231 tumor growth, resulting in smaller tumor volumes ( Figure 9A) and weights ( Figure 9B). These significant differences were already visible on day 9 and increased throughout the experiment ( Figure 9A). More specifically, analysis of excised tumors showed that combined treatment with PTX-NCS reduced both tumor volumes and weights 3.1-fold after co-treatment. In addition, the mice seemed to tolerate treatments without overt severe signs of toxicity or excessive loss of body weight (Supplementary Figure S9). seemed to tolerate treatments without overt severe signs of toxicity or excessive loss of body weight (Supplementary Figure S9). The results are represented as the mean ± SEM in both graphs (animals/group ≥ 6). Statistic t-test analyses of combined treatment were performed in comparison with vehicle (black) as well as with PTX single treatment (blue). In all cases, * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

The Combination of PTX with NCS Reduces the Presence of Circulating Tumor Cells (CTCs) and the Presence of Lung Metastasis In Vivo
The CTCs in the mice bloodstream at the treatment endpoint (day 24) were analyzed in order to evaluate the effect of the combination in restraining metastatic spread. Although the effect of PTX in primary tumors was clear, PTX-only treatment showed no significant effect on reducing the CTC population when compared to the vehicle ( Figure  10A). Meanwhile, flow cytometry analysis revealed that animals treated with the PTX-NCS combination effectively eliminated the CTCs in comparison to single PTX treatment or vehicle (p = 0.0070 and p < 0.0124, respectively), reducing 2.27-and 2.10-fold, respectively, the CTC content ( Figure 10A). Additionally, the presence of lung metastasis was evaluated by ex vivo BLI of the lungs. An example of the BLI images acquired from the excised lungs of all three study groups at endpoint is shown in Figure 10B All animals treated with vehicle (6/6) or PTX-only (9/9) showed positive BLI signals in the lungs (100% incidence), while this incidence dropped down to 67% (6/9 animals) in the case of animals treated with the combination. Moreover, when lung BLI intensity was analyzed, the combination treatment showed a significant decrease in the BLI signal when compared to the single PTX treatment (p = 0.0492) or the vehicle (p = 0.0087), reducing 8.56-and 42.83-fold lung BLI intensity, respectively ( Figure 10B). The results are represented as the mean ± SEM in both graphs (animals/group ≥ 6). Statistic t-test analyses of combined treatment were performed in comparison with vehicle (black) as well as with PTX single treatment (blue). In all cases, * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.

The Combination of PTX with NCS Reduces the Presence of Circulating Tumor Cells (CTCs) and the Presence of Lung Metastasis In Vivo
The CTCs in the mice bloodstream at the treatment endpoint (day 24) were analyzed in order to evaluate the effect of the combination in restraining metastatic spread. Although the effect of PTX in primary tumors was clear, PTX-only treatment showed no significant effect on reducing the CTC population when compared to the vehicle ( Figure 10A). Meanwhile, flow cytometry analysis revealed that animals treated with the PTX-NCS combination effectively eliminated the CTCs in comparison to single PTX treatment or vehicle (p = 0.0070 and p < 0.0124, respectively), reducing 2.27-and 2.10-fold, respectively, the CTC content ( Figure 10A). Additionally, the presence of lung metastasis was evaluated by ex vivo BLI of the lungs. An example of the BLI images acquired from the excised lungs of all three study groups at endpoint is shown in Figure 10B All animals treated with vehicle (6/6) or PTX-only (9/9) showed positive BLI signals in the lungs (100% incidence), while this incidence dropped down to 67% (6/9 animals) in the case of animals treated with the combination. Moreover, when lung BLI intensity was analyzed, the combination treatment showed a significant decrease in the BLI signal when compared to the single PTX treatment (p = 0.0492) or the vehicle (p = 0.0087), reducing 8.56-and 42.83-fold lung BLI intensity, respectively ( Figure 10B). A statistic t-test analysis was performed to compare the results between the study groups. (B) Quantification of BLI signal intensity from mice lungs of all three study groups. Results are expressed as the mean ± SEM of the BLI signal (ph/s) per gram of lung tissue (n ≥ 6). Statistical Mann-Whitney analysis was performed comparing the results between the study groups. Representative BLI images of excised mice lungs from all three study groups at endpoint necropsy are also shown. Bioluminescent signals were analyzed by ex vivo imaging using the IVIS Spectrum and further quantified using Living Image ® 4.5.2 software. Statistical significance is shown as ns p > 0.05, * p < 0.05, and ** p < 0.01.

Discussion
The identification of the CSC population has undoubtedly provided a crucial breakthrough in the understanding of cancer biology. These cells have self-renewal capacity and differentiation potential and contribute to multiple tumor malignancies, such as recurrence, metastasis, heterogeneity, multidrug resistance, and radiation resistance [26,27]. Accordingly, CSCs are considered a promising therapeutic target for developing efficient therapies for breast cancer treatment. Such therapies are particularly relevant for TNBC patients due to the lack of specific treatments and the high content of CSC populations in TNBC tumors [9]. Currently, conventional chemotherapy is a standard treatment for TNBC, but it spares CSC populations, which cause tumor recurrence and progression [28]. In recent decades, multiple CSC-targeting strategies have shown therapeutic potential for TNBC in multiple preclinical studies [29], and although some of these strategies are currently being evaluated in clinical trials, such as the AKT inhibitor MK-2206 [30] and the chemokine receptor I/II inhibitor reparixin [31], no therapy specifically targeted against CSCs has already been approved for TNBC treatment. Therefore, the identification of novel therapeutic anti-CSC agents is urgently needed to develop effective therapeutics for this cancer subtype. In this work, we focus on the validation of potential anti-CSC drugs that could inhibit CSC proliferation and downregulate signaling pathways, keeping the stemness phenotype in cancer cells. For this purpose, a battery of 17 compounds was initially identified, all clinically approved drugs with previously described anti-CSC activity  6). A statistic t-test analysis was performed to compare the results between the study groups. (B) Quantification of BLI signal intensity from mice lungs of all three study groups. Results are expressed as the mean ± SEM of the BLI signal (ph/s) per gram of lung tissue (n ≥ 6). Statistical Mann-Whitney analysis was performed comparing the results between the study groups. Representative BLI images of excised mice lungs from all three study groups at endpoint necropsy are also shown. Bioluminescent signals were analyzed by ex vivo imaging using the IVIS Spectrum and further quantified using Living Image ® 4.5.2 software. Statistical significance is shown as ns p > 0.05, * p < 0.05, and ** p < 0.01.

Discussion
The identification of the CSC population has undoubtedly provided a crucial breakthrough in the understanding of cancer biology. These cells have self-renewal capacity and differentiation potential and contribute to multiple tumor malignancies, such as recurrence, metastasis, heterogeneity, multidrug resistance, and radiation resistance [26,27]. Accordingly, CSCs are considered a promising therapeutic target for developing efficient therapies for breast cancer treatment. Such therapies are particularly relevant for TNBC patients due to the lack of specific treatments and the high content of CSC populations in TNBC tumors [9]. Currently, conventional chemotherapy is a standard treatment for TNBC, but it spares CSC populations, which cause tumor recurrence and progression [28]. In recent decades, multiple CSC-targeting strategies have shown therapeutic potential for TNBC in multiple preclinical studies [29], and although some of these strategies are currently being evaluated in clinical trials, such as the AKT inhibitor MK-2206 [30] and the chemokine receptor I/II inhibitor reparixin [31], no therapy specifically targeted against CSCs has already been approved for TNBC treatment. Therefore, the identification of novel therapeutic anti-CSC agents is urgently needed to develop effective therapeutics for this cancer subtype. In this work, we focus on the validation of potential anti-CSC drugs that could inhibit CSC proliferation and downregulate signaling pathways, keeping the stemness phenotype in cancer cells. For this purpose, a battery of 17 compounds was initially identified, all clinically approved drugs with previously described anti-CSC activity (Supplementary Table S1). Of note, repurposing approved drugs is a cost-effective strategy that can bypass the time-consuming stages of drug development and thus facilitate rapid clinical translation [32].
From the initial drug set, 8Q and NCS compounds emerged, showing remarkably antiproliferative activity on the CSC subpopulation, either in attachment or in low attachment culture conditions. It is important to note that we did not select the compounds according to their potency but to their selectivity. SAL, for instance, had lower IC 50 values against CSC than 8Q, but the effect of the latter was more specific to CSC. In addition, we further demonstrated that 8Q and NCS efficiently inhibited different stemness properties, i.e., migration, invasion and neoplastic transformation. These results are consistent with the works published by Wang et al. [25] and Zhou et al. [24], in which 8Q and NCS were identified as having preferential activity against breast cancer sphere cells, showing their effect through the downregulation of stem pathways and inducing CSC apoptosis. In agreement with previously published studies [33,34], the reference drug PTX was not capable of selectively targeting the CSC subpopulation, either in attachment or in low attachment conditions.
Traditional chemotherapies have failed to remove CSCs and therefore, they have proved ineffective in preventing tumor recurrence. In this scenario, anti-CSC drug administration in combination with traditional chemotherapy may ameliorate clinical outcomes and long-term patient survival. Furthermore, combination therapy is a novel strategy widely used in clinical studies for cancer treatment. It pursues the achievement of a high synergistic therapeutic efficacy at lower drug doses, namely, increasing the treatment efficacy while reducing non desired side-effects [35]. In this scenario, we hypothesized that the use of specific anti-CSC drugs in combination with current anti-cancer reference agents may become a real therapeutic alternative to improve the treatment of TNBC. Following this hypothesis, the synergistic effect of both anti-CSC drugs 8Q and NCS with the anticancer drug PTX was explored through combination assays in vitro in different TNBC cell lines. Our data provided strong evidence that both anti-CSC drugs displayed synergistic anti-proliferation activity with PTX when combined at specific ratios. It is worth noting that the synergistic ratios were drug and cell line dependent, probably due to the molecular and phenotypic heterogeneity among the different cell lines. In this regard, such differences evidence the complexity and heterogeneity of cancer, especially in the case of TNBC. There are many ongoing efforts to understand this inherent variability and resistant nature of cancer, as well as to characterize the molecular differences between tumors, with the aim of developing specialized treatments for each specific subtype of cancer, mainly based on the measurement and manipulation of key patient genetic and omics data (transcriptomics, metabolomics, proteomics, etc.) [36]. The understanding and application of these data as tools in clinical trial design and treatment selection have steered the field of cancer treatment toward the concept of precision and personalized medicine, in which therapy selection is tailored to each individual [36,37]. An example of such a personalized approach is the well-known drug trastuzumab, which was approved years ago for the treatment of HER2 receptor positive breast cancer [38], but also the promising use of the poly-ADP ribose polymerase inhibitor olaparib in the treatment of BRCA-mutant ovarian cancer [39]. Accordingly, clinical implementation of drug combination approaches should be considered in a personalized context for the individual patient, ensuring maximum performance and getting the most out of combined therapy in each case. Several studies have already demonstrated the safety, feasibility and importance of designing precision oncology trials that emphasize personalized, individually tailored combination therapies rather than scripted monotherapies. The targeting of a larger fraction of identified molecular alterations has been correlated with significantly improved disease control rates, as well as with longer progression-free and overall survival rates, compared to targeting just one driver mutation, according to a study published by Jason K. Sicklick et al. [40]. In this work, the researchers demonstrated that the use of multi-drug therapies helped improve outcomes among patients with therapy-resistant cancers, indicating that combination drug treatments could improve precision medicine for cancer care. It is clear that integrating a PPM perspective into cancer research and tumor treatment could result in major improvements in fighting cancer, especially due to its complexity and interpatient variability.
Our findings in combination studies showed how two drugs could move from acting synergistically to being antagonistic depending solely on their relative doses. In this regard, optimization of the customized drug ratio is vital prior to further treatment implementation. Interestingly, in our hands, the combination of 8Q with NCS abrogated the relative increase in CSCs induced by PTX. In TNBC cells, 8Q and NCS inhibited NF-κB and Wnt/βcatenin signaling pathways. Both pathways have been reported to be overexpressed in CSC subpopulations, promoting proliferation and resistance to therapy and preserving the undifferentiated state of stem cells [8,10]. An enhanced inhibition of both pathways was obtained when cells were treated with established ratios of PTX with 8Q and NCS, indicating that combination therapy resulted in the suppression of bulk tumor cell proliferation but also enhanced the sensitivity of chemo-resistant cells. Based on these results, we moved forward in testing the efficacy of PTX-NCS treatment in vivo in an orthotopic TNBC mice model. A number of studies have evaluated the potential of NCS as an anti-cancer therapy in various cancer types in vivo, including TNBC [41][42][43][44] and stated the ability of NCS to overcome cancer chemoresistance when combined with PTX [45][46][47] and other anti-cancer agents [48,49]. However, the synergistic effect of combined treatment in CSC subpopulations has not yet been clearly elucidated. In the present study, we showed that the combined treatment of PTX and NCS effectively arrested MDA-MB-231 tumor growth in the TNBC cancer xenograft model and significantly reduced the metastatic ability of aggressive tumor cells, indicating that combined therapy not only improved the therapeutic efficacy of treatments in differentiated tumor cells but also in the CSC subpopulation. Such synergism in MDA-MB-231 cells is mediated via Wnt/β-catenin pathway, and probably executed through the inhibition of Bcl-2 protein [50,51]. Importantly, combination therapy successfully decreased the number of CTC in blood and, more importantly, prevented the generation of lung metastasis, results not observed following treatment with the PTX drug alone.
Based on our findings, treatment with PTX combined with NCS may offer an effective therapeutic approach to improving the prognosis of TNBC by simultaneously targeting both bulk differentiated cancer cells and the minor population of CSCs. Personalized treatment with combination therapies may yield novel chemotherapy strategies in the future.

Pharmacological Agents
The compounds used in this thesis were selected based on their reported activity against CSC, their presence in clinical trials and their commercial availability. First, a literature screening was carried out in PubMed using the terms 'breast cancer stem cells,' 'targeting' and 'triple negative breast cancer.' The objective was to find compounds with specific activity against CSC to further demonstrate their activity in our CSC models. Seventeen compounds were finally selected. Molecular structures, as well as references to previous works, have been detailed in the Supplementary Material (Table S1) [13], were cultured in complete medium supplemented with the selection antibiotic blasticidin at 1 µg/mL (Gibco, Thermo Fisher Scientific, Madrid, Spain). Of note, the stemness nature of tdTomato-expressing cells from all three-CSC models has already been fully confirmed in vitro by increased expression of stemness markers, mammosphere formation and in vivo tumorigenic capacity [13]. All cell lines were incubated in a humidified incubator at 37 • C containing 5% CO 2 .

Wound Healing Assay
The effect of drugs on the migratory potential of MDA-MB-231 cells was studied by recording with a microscope (FSX100 microscope, Olympus Life Science) the closure of a scratch made in confluent monolayers [54]. Wound area was measured at 0, 8 and 24 h by image analysis with ImageJ software [55]. To exclude the effect of cell proliferation in wound healing, treatment doses (1 µM PTX, 25 µM 8Q and 0.5 µM NCS) corresponded to the maximum dose, keeping cell viability above 80% after 24 h incubation.

Colony Formation Assay (Anchorage-Independent Growth) in Soft Agar
Anchorage-independent growth was assessed using CytoSelect™ Cell Transformation Assay Kit (Cell Biolabs Inc., San Diego, CA, USA). A semisolid agar medium was prepared according to the manufacturer's instructions. Selected drugs were added, and the plates were incubated for 6-8 days at 37 • C in a 5% CO 2 incubator. Following treatment, colonies formed were observed under an optical microscope and viable transformed cells were quantified by measuring cell viability by MTT at 570 nm.

Matrigel Cell Invasion Assay
Invasion experiments were performed using a conventional 24-well plate with cell culture inserts containing 8 µM pore size filters (Falcon TM , Fisher Scientific, Madrid, Spain) and coating the filters with the basement membrane Matrigel (1 mg/mL, BD Biosciences, Franklin Lakes, NJ, USA). After Matrigel solidification, MDA-MB-231 cells (10,000 cells/300 µL, with previous 24 h of FBS starvation) were added into the upper chamber of each insert together with the selected treatments. The chemo-attractant (complete medium with 10% FBS) was placed in the lower chamber of each well. After 24 h incubation at 37 • C in a 5% CO 2 incubator, the upper surface of the filter was wiped with a cottontipped applicator to remove non-invading cells. Cells that had invaded through the filter pores and attached to the under surface of the filter were fixed with 100% methanol and stained with 0.4% crystal violet (Sigma-Aldrich, St. Louis, MO, USA) solution for 15 min. The membranes were mounted on glass slides, and cells from 10 random microscopic fields (20× magnifications) were counted using ImageJ software [56].

Fluorescence-Activated Cell-Sorting Enrichment
A cell sorting assay was performed to select and expand CSCs that expressed tdTomato fluorophore under the ALDH1A1 promoter [12,13]. Briefly, TNBC cells stably transfected with pLenti6-ALDH1A1-tdTomato plasmid (MDA-MB-231-, HCC1806-and MDA-MB-468-ALDH1A1:tdTomato cells) were harvested and suspended in PBS with 10% FBS and 10 µg/mL DAPI and filtered using 30 µm sterile filters (CellTrics ® , Sysmex Europe GmbH, Norderstedt, Germany) before subjecting to cell sorting with the BD FACSAria TM flow cytometer (High speed FACSAria digital cell sorter, BD Biosciences, Franklin Lakes, NJ, USA). A yellow-green laser of 561 nm was used for the tdTomato detection and a violet laser of 405 nm was used for DAPI detection. Thereafter, enriched cells were seeded for the correct expansion of both cell subpopulations. Importantly, the stem-like phenotype of flow-sorted CSC tdTomato+ cells was confirmed before performing further experiments by the increased expression of stemness markers compared to tdTomato-cells. An example of the stem cell-like gene expression profile of enriched CSC and non-CSC subpopulations from the MDA-MB-231 fluorescent model is shown in Supplementary Figure S2. Similarly, the in vivo tumorigenicity of tdTomato+ cells derived from the HCC1806 cell line is shown in Figure S3.

Post-Treatment CSC Quantification
The direct effect of the drugs on the CSC population was studied by flow cytometry in ALDH1A-tdTomato CSC models. For each cell line, tdTomato+ and tdTomato-cell subpopulations corresponding to the CSC and non-CSC cells, respectively, were maintained at a 1:1 ratio. Cells were seeded on 6-well plates at a density of 200,000 cells per well. After 24 h, the cells were incubated with the selected treatments (individual and in combination) for 72 h. Thereafter, the medium was removed, and the cultures were refed with complete medium to allow cellular recovery in the absence of drug/s for an additional 48 h. Changes in the amount of tdTomato+ were evaluated using the BD LSRFortessa™ Cytometer (BD Biosciences, Franklin Lakes, NJ, USA) and subsequently analyzed using FCS express 7 Flow cytometry software (De novo, Pasadena, CA, USA) to calculate the increase or decrease in the tdTomato+ population.

Quantitative PCR Analysis
Quantitative RT (real time)-PCR was performed to evaluate the relative expression of CSC stemness markers, such as ALDH1A1, ABCG2, ALOX5, CMKLR1, SOX2, Notch4, POU5F1/Oct-4 and Nanog, after treatment with the selected drugs. Briefly, total RNA was extracted using an RNeasy Mini Kit (Qiagen, Hilden, Germany) and 1 µg retrotranscribed to cDNA (Applied Biosystems, San Francisco, CA, USA). The generated cDNA was quantified using SYBR Green chemistry on a 7500 real-time PCR system (Thermo Fisher Scientific, Madrid, Spain). Two housekeeping (GAPDH and β-Actin) genes served as normalization controls for the analysis of the expression levels of the genes of interest using qbase+ software, version 2.0 (Biogazelle, Gent, Belgium). The cycle threshold values generated were used to calculate the fold change in the expression of the gene of interest. The primer sequences were the same as those previously described [13].

In Vivo Therapeutic Efficacy
Animal care and in vivo procedures were handled in accordance with the protocols for the Care and Use of Laboratory Animals of the Vall d'Hebron University Hospital Animal Facility. In addition, the Animal Experimentation Ethical Committee at the institution approved the experimental procedures. (ref. CEA-OH/9467/2). All in vivo studies were performed by the ICTS "NANBIOSIS" of CIBER-BBN's In vivo Experimental Platform for Functional Validation & Preclinical Research area (CIBBIM-Nanomedicine, Barcelona, Spain). Briefly, 1·10 6 MDA-MB-231.Fluc-GFP cells were suspended in a 1:1 mixture of culture medium and Matrigel (BD Biosciences, Franklin Lakes, NJ, USA) and injected into the mammary fat pad as previously described [57] in six-week-old female NOD/SCID mice (NOD.CB-17-Prkdcscid/Rj), obtained from Janvier Laboratories (France). When tumor volumes reached 75-80 mm 3 , mice were randomized into 3 groups (n = 6-9/each group). Paclitaxel at 10 mg/kg was administered intravenously by tail vein injection 3 times per week in the first week and two times/week the remaining 2 weeks. Niclosamide at 10 mg/kg was given by intraperitoneal injection 5 times per week for 3 weeks. A vehicle of NCS (DMSO: Cremophor ® EL, 1:1) was administered intraperitoneally 5 times/week. Tumor growth was monitored twice a week by conventional caliper measurements D × d 2 /2, where D is the major diameter and d is the minor diameter. Upon reaching the endpoint, mice were treated with their corresponding treatment and 1 h after administration were euthanized, and blood was collected by cardiac puncture for the analysis of CTC content by flow cytometry.

Detection of Lung Metastasis
To measure the extent of lung metastasis ex vivo, bioluminescence imaging (BLI) was performed using the IVIS Spectrum (PerkinElmer, Waltham, MA, USA) after administering 150 mg/kg of D-luciferin (Thermo Fisher Scientific, Madrid, Spain) to mice. Afterwards, tumors and tissues (lungs, kidney, spleen and liver) were harvested and weighted. Immediately after necropsy, lung tissues were placed individually into separate wells containing 300 µg/mL of D-luciferin, and imaged and quantified using Living Image ® 4.5.2 software (PerkinElmer, Waltham, MA, USA).

Detection of Circulating Tumor Cells from Mice Blood Samples
For isolation of circulating tumor cells (CTCs), blood samples were drawn from each animal by cardiac puncture and further processed using the following protocol [16]. First, the collected samples were subjected to several cycles of erythrocyte lysis using a lysis buffer, which consisted of a mixture of 90% of 0.16 M NH 4 Cl and 10% of 0.17 M Tris (pH 7.65). A total of 5 mL of lysis buffer was added to each tube and samples were incubated for 5 min at 37 • C. Following incubation, the samples were then centrifuged (500 g for 10 min at 4 • C). These steps were repeated until a white cell pellet was obtained, thus indicating the proper lysis of red blood cells. As inoculated tumor cells expressed the GFP fluorescent marker, the resulting pelleted cells were resuspended in a cytometry buffer and examined by flow cytometry. All samples were processed and analyzed using the same parameters. Subsequently, the number of GFP+ cells/mL blood collected was calculated and further normalized by tumor weight from excised tumors at the endpoint.

Statistical Analysis
In vitro experiments were repeated at least 3 times, each involving at least two technical replicates. The in vivo experiment was conducted once with ≥6 animals per group. Data were plotted as the mean value ± SEM (standard error of the mean) and analyzed using GraphPad Prism 6 software. One-way ANOVA analysis, unpaired two-tailed student's t-test or equivalent non-parametric tests were used to investigate the differences between tested compounds and controls. Differences were considered statistically significant when the p-value was equal to or below 0.05 (*), 0.01 (**), 0.001 (***) and 0.0001 (****).

Conclusions
In conclusion, 8Q and NCS anti-CSC drugs displayed synergistic anti-proliferation activity with PTX when combined at specific ratios in TNBC cells. The simultaneous use of PTX with 8Q or NCS combined therapy also increased their anti-CSC potential effect in vitro by inhibiting the inhibition of the NF-κB and Wnt/β-catenin signaling pathways, resulting in the reduction of the mammosphere growth and viability of TNBC cells. Furthermore, the combined therapy of PTX with NCS induced the regression of triple-negative breast tumors and reduced metastasis generation in vivo. These data support the combination of PTX and NCS as a potential treatment for advanced TNBC.