Synthesis of Curcumin Derivatives and Analysis of Their Antitumor Effects in Triple Negative Breast Cancer (TNBC) Cell Lines

We analyzed antitumor effects of a series of curcumin analogues. Some of them were obtained by reaction of substitution involving the two phenolic OH groups of curcumin while the analogues with a substituent at C-4 was prepared following an original procedure that regards the condensation of benzenesulfenic acid onto the nucleophilic central carbon of the curcumin skeleton. We analyzed cytotoxic effects of such derivatives on two TNBC (triple negative breast cancer) cell lines, SUM 149 and MDA-MB-231, but only three of them showed an IC50 in a lower micromolar range with respect to curcumin. We also focused on these three derivatives that in both cell lines exhibited a higher or at least equivalent pro-apoptotic effect than curcumin. The analysis of molecular mechanisms of action of the curcumin derivatives under study has highlighted that they decreased NF-κB transcriptional factor activity, and consequently the expression of some NF-κB targets. Our data confirmed once again that curcumin may represent a very good lead compound to design analogues with higher antitumor capacities and able to overcome drug resistance with respect to conventional ones, even in tumors difficult to treat as TNBC.


Introduction
Cancer can be described as uncontrolled DNA replication and cell division, evasion from programmed cell death, breaking through normal tissue boundaries and invasion to new sites in the body. Actually, the most promising anti-cancer therapies combine agents with different molecular mechanisms such as specific drugs and chemo-or radiotherapies, resulting in a better efficacy and longer survival [1].
In this scenario, polyphenols are to be considered molecules with antitumor action because of their capacity to interfere with a large number of pathways that in the neoplastic cell are simultaneously deregulated. Their role in improving bioavailability of drugs has been often underlined, perfecting the response of cancer cell to different therapies [2]. Recently, researchers suggested that natural polyphenols might be also used to sensitize tumor cells to chemo-and radiotherapy by inhibiting pathways that lead to treatment resistance [3]. For this reason, these compounds are often defined as "privileged" structures [1], the implications of which in human health are vast, including many

Antiproliferative Activity
We analyzed cytotoxic effects of curcumin derivatives 1-5 on the two TNBC cell lines SUM 149 and MDA-MB-231, and also in a normal cell line, 1-7HB2. In Table 1

Antiproliferative Activity
We analyzed cytotoxic effects of curcumin derivatives 1-5 on the two TNBC cell lines SUM 149 and MDA-MB-231, and also in a normal cell line, 1-7HB2. In Table 1

Antiproliferative Activity
We analyzed cytotoxic effects of curcumin derivatives 1-5 on the two TNBC cell lines SUM 149 and MDA-MB-231, and also in a normal cell line, 1-7HB2. In Table 1 the IC 50 of Compounds 1-5 and curcumin are reported and only Compounds 1-3 showed an IC 50 lower than curcumin (Figures 2  and 3), with SUM 149 cells being the most sensitive to the action of these derivatives. This result led us to focus our studies on Compounds 1-3. interesting, because they can be used in different types of cancer. In line with the results obtained in the TNBC cells, also in the non-TNBC cells, Analogues 1-3 are more active than curcumin. Noteworthy the IC50 of Compounds 1 and 2 are lower in the multidrug resistant variant, HL60R, than the parental HL60; these results are worthy to further experimental analyses to verify if these analogues are capable to bypass the wide problem of acquired multidrug resistance. We investigated the mechanism by which compounds could exert any anticancer activity with the study of their pro-or antioxidant properties, due to the well-known role that oxidative stress plays in the progression of a number of cancers. Therefore, our investigation on the mechanism by  In order to verify if these analogues are characterized by a specificity of action towards the cancer model examined, we performed a cell growth assay in non-TNBC cells, in particular, in a cell line of acute promyelocytic leukemia HL60 and in its multidrug resistant variant (HL60: curcumin IC 50 = 47.5 µM ± 3.3, 1 IC 50 = 14.0 µM ± 1.7, 2 IC 50 = 17.0 µM ± 1.2, 3 IC 50 < 1 µM; HL60 R: curcumin IC 50 = 40.5 µM ± 2.1, 1 IC 50 = 7.0 µM ± 0.8, 2 IC 50 = 9.0 µM ± 2.8, 3 IC 50 = 6.7 µM ± 1.1). These results confirmed that Compounds 1-3 do not have specific effects on TNBC cell lines, on the contrary, they show a strong action also in the non-TNBC cancer model analyzed and for this are even more interesting, because they can be used in different types of cancer. In line with the results obtained in the TNBC cells, also in the non-TNBC cells, Analogues 1-3 are more active than curcumin. Noteworthy the IC 50 of Compounds 1 and 2 are lower in the multidrug resistant variant, HL60R, than the parental HL60; these results are worthy to further experimental analyses to verify if these analogues are capable to bypass the wide problem of acquired multidrug resistance.

Pro-and Antioxidant Activity
We investigated the mechanism by which compounds could exert any anticancer activity with the study of their pro-or antioxidant properties, due to the well-known role that oxidative stress plays in the progression of a number of cancers. Therefore, our investigation on the mechanism by which Compounds 1-3 could exert any anticancer activity started with the study of their pro-or antioxidant properties.
In order to assess whether Analogues 1-3 could behave as pro-oxidant substances, the two TNBC cells were treated with the antioxidant N-acetyl-L-cysteine (NAC), which is a well-known scavenger of reactive oxygen intermediates, at 2mM for 1h, before exposure to curcumin and Compounds 1-3 at the corresponding IC 50 values. As shown in Table 2, both for SUM 149 cells (A) and MDA-MB-231 cells (B), the addition of NAC reduced the cytotoxic activity of Compounds 1 and 2 as well as of curcumin, whereas the same results were not obtained for Compound 3. In our study, the analysis carried out with DPPH (2,2-diphenyl-1-picrylhydrazyl) reduction assay indicated that Analogues 1 and 2 do not possess antioxidant activity, since the efficient dose (ED 50 ) of these was not identified (Table 3). From a chemical point-of-view, these results are in line with previous observations showing that the protection of the phenolic groups causes the complete abolition of the antioxidant activity [43]. From a biological point of view, they indicated that the mechanism of antitumor activity against TNBC can involve an at least partially pro-oxidant effect for Compounds 1 and 2. Surprisingly, substitution in the active methylene site of Compound 3 showed a scavenging activity lower then curcumin [43], reaching an ED 50 at a concentration of 19.2 µM (Table 3).

Pro-Apoptotic Activity
In the two TNBC cell lines, Compounds 1-3 showed increased or at least equivalent cell death induction than curcumin. The cells were treated for 24 h with curcumin and its derivatives at concentrations close to the corresponding IC 50 values. The flow cytometry analysis with propidium iodide revealed that in SUM 149 cells only Compound 1 determined a significant block in a pre-G 0 -G 1 position, while in MDA-MB-231 cells the order of potency in inducing cell death was Compound 1 > Compound 3 > curcumin (Figures 4 and 5). Anyway, the results appeared to be in agreement with the cytotoxicity data.

NF-κB Inhibition
Since curcumin is an inhibitor of the nuclear activation of NF-κB, we analyzed the capacity of curcumin derivatives 1-3 to inhibit NF-κB DNA-binding activity by TransAM assays. Each cell line was treated for 8 h and 24 h with curcumin and its derivatives, at the relative IC50 values. In the SUM 149 cell line we observed a high reduction of NF-κB DNA-binding activity for Compounds 1-3, in particular, Compound 3 exhibited a stronger decrease than curcumin, only after 24 h of treatment ( Figure 6). In MDA-MB-231 cells Compounds 1 and 3 were NF-κB inhibitors but only Compound 1 showed a greater effect than curcumin, after 8 h of treatment ( Figure 7).

NF-κB Inhibition
Since curcumin is an inhibitor of the nuclear activation of NF-κB, we analyzed the capacity of curcumin derivatives 1-3 to inhibit NF-κB DNA-binding activity by TransAM assays. Each cell line was treated for 8 h and 24 h with curcumin and its derivatives, at the relative IC50 values. In the SUM 149 cell line we observed a high reduction of NF-κB DNA-binding activity for Compounds 1-3, in particular, Compound 3 exhibited a stronger decrease than curcumin, only after 24 h of treatment ( Figure 6). In MDA-MB-231 cells Compounds 1 and 3 were NF-κB inhibitors but only Compound 1 showed a greater effect than curcumin, after 8 h of treatment ( Figure 7).

NF-κB Inhibition
Since curcumin is an inhibitor of the nuclear activation of NF-κB, we analyzed the capacity of curcumin derivatives 1-3 to inhibit NF-κB DNA-binding activity by TransAM assays. Each cell line was treated for 8 h and 24 h with curcumin and its derivatives, at the relative IC 50 values. In the SUM 149 cell line we observed a high reduction of NF-κB DNA-binding activity for Compounds 1-3, in particular, Compound 3 exhibited a stronger decrease than curcumin, only after 24 h of treatment ( Figure 6). In MDA-MB-231 cells Compounds 1 and 3 were NF-κB inhibitors but only Compound 1 showed a greater effect than curcumin, after 8 h of treatment ( Figure 7).  In order to verify if the three analogues alter the expression of some targets of NF-κB, like curcumin, western blot analysis was used. The two cell lines were treated with curcumin and derivatives 1-3 in the same conditions. In line with the results of NF-κB DNA-binding activity inhibition, only Compounds 1 and 3 caused a decrease of expression of some targets in the same cell lines. In particular, in SUM 149 cells only Compound 3 caused a decreased expression of Survivin (a 55% of reduction respect to control) and Bcl-2 (a 48% of reduction respect to control), while in MDA-MB-231 cells only Compound 1 caused a reduction of IAP1 expression (a 32% of reduction with respect to the control) (Figures 8 and 9).  In order to verify if the three analogues alter the expression of some targets of NF-κB, like curcumin, western blot analysis was used. The two cell lines were treated with curcumin and derivatives 1-3 in the same conditions. In line with the results of NF-κB DNA-binding activity inhibition, only Compounds 1 and 3 caused a decrease of expression of some targets in the same cell lines. In particular, in SUM 149 cells only Compound 3 caused a decreased expression of Survivin (a 55% of reduction respect to control) and Bcl-2 (a 48% of reduction respect to control), while in MDA-MB-231 cells only Compound 1 caused a reduction of IAP1 expression (a 32% of reduction with respect to the control) (Figures 8 and 9). In order to verify if the three analogues alter the expression of some targets of NF-κB, like curcumin, western blot analysis was used. The two cell lines were treated with curcumin and derivatives 1-3 in the same conditions. In line with the results of NF-κB DNA-binding activity inhibition, only Compounds 1 and 3 caused a decrease of expression of some targets in the same cell lines. In particular, in SUM 149 cells only Compound 3 caused a decreased expression of Survivin (a 55% of reduction respect to control) and Bcl-2 (a 48% of reduction respect to control), while in MDA-MB-231 cells only Compound 1 caused a reduction of IAP1 expression (a 32% of reduction with respect to the control) (Figures 8  and 9).

Conclusions
In summary, the synthesis of three new curcumin derivatives has been described, one of which has been prepared by an original procedure via sulfenic acid condensation. Five curcumin derivatives have been analyzed in a study on TNBC cell lines, of which Compounds 1-3 showed cytotoxic and pro-apoptotic activity towards the cellular models studied. Among the three compounds, 1 and 3 seem to be able to inhibit the activation of NF-κB and the expression of some of its targets in a stronger way than curcumin and in a mode which is peculiar with respect to the two TNBC cell lines examined. This difference could be due to some differences, both phenotypic and molecular, between the two cell lines studied. In fact, SUM 149 is a model of inflammatory breast cancer, a basal-like subtype, while MDA-MB-231 is claudin-low subtype. Other authors observe different responses to drugs between these two cell lines though both are TNBC models [44,45].

Conclusions
In summary, the synthesis of three new curcumin derivatives has been described, one of which has been prepared by an original procedure via sulfenic acid condensation. Five curcumin derivatives have been analyzed in a study on TNBC cell lines, of which Compounds 1-3 showed cytotoxic and pro-apoptotic activity towards the cellular models studied. Among the three compounds, 1 and 3 seem to be able to inhibit the activation of NF-κB and the expression of some of its targets in a stronger way than curcumin and in a mode which is peculiar with respect to the two TNBC cell lines examined. This difference could be due to some differences, both phenotypic and molecular, between the two cell lines studied. In fact, SUM 149 is a model of inflammatory breast cancer, a basal-like subtype, while MDA-MB-231 is claudin-low subtype. Other authors observe different responses to drugs between these two cell lines though both are TNBC models [44,45].

Conclusions
In summary, the synthesis of three new curcumin derivatives has been described, one of which has been prepared by an original procedure via sulfenic acid condensation. Five curcumin derivatives have been analyzed in a study on TNBC cell lines, of which Compounds 1-3 showed cytotoxic and pro-apoptotic activity towards the cellular models studied. Among the three compounds, 1 and 3 seem to be able to inhibit the activation of NF-κB and the expression of some of its targets in a stronger way than curcumin and in a mode which is peculiar with respect to the two TNBC cell lines examined. This difference could be due to some differences, both phenotypic and molecular, between the two cell lines studied. In fact, SUM 149 is a model of inflammatory breast cancer, a basal-like subtype, while MDA-MB-231 is claudin-low subtype. Other authors observe different responses to drugs between these two cell lines though both are TNBC models [44,45].
We thought that SUM 149 and MDA-MB-231 cells could represent the enormous clinical heterogeneity which is found in patients. In fact, although both cell lines belong to the triple negative type, they represent a good study model to compare the response to drugs just for their different characteristics.
Even though Compound 2 has an IC 50 inferior than curcumin in both cell lines, on NF-κB it has comparable or lower effects to those of curcumin (respectively in SUM 149 cells and in MDA-MB-231 cells). Its lower efficacy on the inhibition of the targets could depend precisely on the effect on NF-κB that was more limited compared to Analogues 1 and 3. This result would not seem to depend on the chemical structure of Compound 2, rather than on the peculiarity of the mechanism of action of the different analogues on the inhibition of the transcription factor. Of all three compounds, Compound 3 appears to be the most promising, given that in both cell lines it shows strong inhibitory effects of cell proliferation and pro-apoptotic as well as a downregulation of NF-κB activity. For these biological effects, this compound appears to be worthy for further analysis, as in vivo assays, to demonstrate the real efficacy as a new anticancer agent.
Taken together, our data corroborate potentiality of curcumin as a lead compound in cancer diseases. In particular, in this work we highlighted Analogues 1, 2 and 3 were effective molecules towards triple negative cancer cell lines. Their cytotoxic and pro-apoptotic capacities with ability to interfere with NF-κB pathway can contribute to paving the way to their use as new anticancer agents.

Cell Lines
Dr. Elda Tagliabue (Molecular Targeting Unit, Department of Experimental Oncology and Molecular Medicine, Fondazione Institute of Hospitalization and Scientific Care, National Cancer Institute, Milan, Italy) kindly provided us the human breast cancer cell lines: MDA-MB-231(ATCC: HTB-26-Rockville, MD, USA) and SUM 149 (SUM149PT-Asterand Bioscience Detroit, MI). The first was cultured in RPMI-1640 and the second was cultured in DMEM/F-12 supplemented with insulin (5 μg/mL). The cells were authenticated using the short tandem repeat profiling method in their Institute. Prof. Giulio Ghersi (STEBICEF Department, University of Palermo, Italy) kindly provided us the non-tumorigenic cell line 1-7HB2 (ECACC 10081201-Cancer Research Technology, London, UK) that was cultured in DMEM low glucose supplemented with hydrocortisone (5 μg/ml) and insulin (10 μg/mL). HL60, obtained from ATCC® (CCL-240, Rockville, MD, USA), and its variant HL60R, obtained by exposure to gradually increasing concentrations of doxorubicin, were cultured in RPMI-1640. All media were supplemented with 10% heat-inactivated fetal calf serum, 2

Cell Lines
Dr. Elda Tagliabue (Molecular Targeting Unit, Department of Experimental Oncology and Molecular Medicine, Fondazione Institute of Hospitalization and Scientific Care, National Cancer Institute, Milan, Italy) kindly provided us the human breast cancer cell lines: MDA-MB-231(ATCC: HTB-26-Rockville, MD, USA) and SUM 149 (SUM149PT-Asterand Bioscience Detroit, MI). The first was cultured in RPMI-1640 and the second was cultured in DMEM/F-12 supplemented with insulin (5 μg/mL). The cells were authenticated using the short tandem repeat profiling method in their Institute. Prof. Giulio Ghersi (STEBICEF Department, University of Palermo, Italy) kindly provided us the non-tumorigenic cell line 1-7HB2 (ECACC 10081201-Cancer Research Technology, London, UK) that was cultured in DMEM low glucose supplemented with hydrocortisone (5 μg/ml) and insulin (10 μg/mL). HL60, obtained from ATCC® (CCL-240, Rockville, MD, USA), and its variant HL60R, obtained by exposure to gradually increasing concentrations of doxorubicin, were cultured in RPMI-1640. All media were supplemented with 10% heat-inactivated fetal calf serum, 2

Cell Lines
Dr. Elda Tagliabue (Molecular Targeting Unit, Department of Experimental Oncology and Molecular Medicine, Fondazione Institute of Hospitalization and Scientific Care, National Cancer Institute, Milan, Italy) kindly provided us the human breast cancer cell lines: MDA-MB-231(ATCC: HTB-26-Rockville, MD, USA) and SUM 149 (SUM149PT-Asterand Bioscience Detroit, MI). The first was cultured in RPMI-1640 and the second was cultured in DMEM/F-12 supplemented with insulin (5 µg/mL). The cells were authenticated using the short tandem repeat profiling method in their Institute. Prof. Giulio Ghersi (STEBICEF Department, University of Palermo, Italy) kindly provided us the non-tumorigenic cell line 1-7HB2 (ECACC 10081201-Cancer Research Technology, London, UK) that was cultured in DMEM low glucose supplemented with hydrocortisone (5 µg/mL) and insulin (10 µg/mL). HL60, obtained from ATCC ® (CCL-240, Rockville, MD, USA), and its variant HL60R, obtained by exposure to gradually increasing concentrations of doxorubicin, were cultured in RPMI-1640. All media were supplemented with 10% heat-inactivated fetal calf serum, 2 mM L-glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin (all reagents were from EuroClone S.p.A., Milan, Italy; GE Healthcare Life Sciences, Logan, UT, USA). All cell lines were cultured in a humidified atmosphere at 37 • C in 5% CO 2 . Cells with a narrow range of passage number (4 ± 6) were tested for Mycoplasma contamination and used for all experiments. After obtaining the cells, the first passage carried out was assigned passage number 1 [35].

Cell Growth Aassays
The cells were seeded at 2 × 10 4 cells/well onto 96-well plates and incubated at 37 • C overnight; at time 0, the medium was replaced with fresh complete medium supplemented of compounds at the indicated concentrations. Following 72 h of treatment, 16 µL of a commercial solution obtained from Promega Corporation (Madison, WI, USA) containing 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxy methoxyphenyl)-2-(4-sulphophenyl)-2H-tetrazolium (MTS) and phenazine ethosulfate were added. After a incubation in a humidified atmosphere at 37 • C in 5% CO 2 , the bioreduction of MTS dye was evaluated by measuring the absorbance of each well at 490 nm, using a microplate absorbance reader (iMark Microplate Reader; Bio-Rad Laboratories, Inc., Hercules, CA, USA). Cell growth inhibition was expressed as a percentage (mean ± SE) of the absorbance of the control cells [35].

Anti-and Pro-oxidant Activity
The antioxidant activity was evaluated by the DPPH stable radical method. Trolox (6-hydroxy-2,5,7,8-tetramethyl-chroman-2-carboxylic acid) curve was used as the positive control. 100 µL of sample was added to aliquots of a solution made up with DPPH (4.8 mg) in MeOH (200 mL), and the mixture was incubated in the dark, for 1 h at room temperature. The absorbance was measured using a UV-VIS spectrophotometer, at 517 nm. The results were plotted as the percentage of absorbance disappearance at 517 nm [(1-A/A 0 ) × 100] against the amount of sample divided by the initial concentration of DPPH; lower absorbance values of reaction mixture indicate higher free radical scavenging activity. ED 50 corresponds to micrograms of fraction able to consume half the amount of free radical divided by micromoles of initial DPPH. The results were expressed as antiradical capacity (ARC), which is the inverse of ED 50 . Each point was acquired in triplicate.
Pro-oxidant effects were examined by cell counting, adding N-acetyl-L-cysteine (NAC), an antioxidant molecule, 1h before compounds. Results were expressed as mean ± standard error (SE) of at least three different experiments performed in duplicate. All the chemicals were supplied by Sigma Aldrich Srl, Milan, Italy [46].

Evaluation of Cell Death by Flow Cytometry
Cells were washed twice with ice-cold PBS and then resuspended in a hypotonic fluorochrome solution containing propidium iodide (PI) 50 µg/mL in 0.1% sodium citrate plus 0.03% (v/v) Nonidet P-40, at 1×10 6 / mL . After incubation (1 h) in this solution, the samples were filtered through nylon cloth, 40 µm mesh, and their fluorescence was analyzed using a FACSCanto instrument (Becton Dickinson, Montain View, CA, USA). The data were analyzed with BD FACSDiva software v.6.1.2. (Becton Dickinson). Cell death was determined by evaluating the percentage of events accumulated in the preG 0 -G 1 position [13].

NF-κB Activation
The DNA-binding capacity of NF-κB (p65 subunit) was defined in the nuclear extracts of SUM 149 and MDA-MB-231 cells using the TransAM NF-κB and Nuclear Extract kits (Active Motif, Carlsbad, CA, USA). The determination of binding capacity was based on a 96-well plate, on which an oligonucleotide containing the NF-κB consensus binding site was fixed. By use of an antibody directed against an epitope on p65, it may revealed NF-κB bound to the oligonucleotide. After addition of a horseradish peroxidase-conjugated secondary antibody, a sensitive colorimetric readout was quantified by densitometry (iMark Microplate Reader; Bio-Rad Laboratories, Inc.). The control of specificity of the assay carried out according to the indications of manufacturer's protocol. The results were expressed as arbitrary units: one unit indicated the DNA binding capacity exerted by 2.5 µg whole cell