Preclinical Evaluation of Fatty Acid Synthase and EGFR Inhibition in Triple-Negative Breast Cancer

Breast cancer is the most prevalent and deadly cancer type in Western women (1). Triple-negative breast cancer (TNBC) represents 15% to 20% of the patients diagnosed with breast carcinomas and is characterized by the lack of expression of estrogen and progesterone receptors (ER/PR) and no amplification of the HER2 oncogene (2). TNBC runs an aggressive course and has a poor prognosis. As TNBC lacks a validated targeted therapy, patients are treated mainly with cytotoxic chemotherapy (anthracyclines and taxanes). Even though patients with TNBC show a good initial response to chemotherapy, the recurrence rate within 5 years following diagnosis is about 30% (3, 4).

Perou and colleagues defined five major molecular subtypes within breast cancer: luminal A, luminal B, HER2-enriched, basal-like, and claudin-low or mesenchymal-like (5–7). The two major subtypes comprised in the TNBC are basal-like (∼50%) followed by the mesenchymal-like (∼20%; ref. 8). Other molecular subtypes are also represented within TNBC subtype because TNBC is such a heterogeneous disease. However, between 50% and 70% of TNBC have been shown to express the EGF receptor (EGFR; refs. 7, 9), and its expression has been associated with poor prognosis (10). While EGFR inhibition has been considered a promising approach for TNBC, minimum benefit has been observed in the clinical settings alone or in combination with chemotherapy (11, 12).

Alteration of metabolism is one of the hallmarks of cancer, providing advantages to cancer cells in the tumor environment (13, 14). Fatty acid synthase (FASN), the major enzyme capable of de novo synthesis of fatty acids, has been described to be overexpressed in several carcinomas, such as breast, prostate, and colon among others (15). This enzyme is not expressed in normal tissue, except in lipogenic tissues where its expression is tightly regulated by diet. In cancer, however, cells overexpress FASN enzyme to synthesize almost all their fatty acids de novo (14, 16). We and others have shown that blocking FASN activity has anticancer activity via the apoptotic pathway in vitro and in vivo (17–19), by disrupting lipid membrane synthesis and modification of protein palmitoylation and deregulating oncogenic signaling pathways (20, 21).

The purpose of this study was to evaluate FASN as a target or co-target in sensitive and chemoresistant TNBC. First, we evaluated FASN expression in a panel of TNBC cell lines and in TNBC primary tumor samples. In vitro, C75 and EGCG (FASN inhibitors) were tested for antitumor efficacy in combination with chemotherapy and with cetuximab (an EGFR inhibitor). In vivo orthoxenografts of two different sensitive and resistant TNBC models were performed to test the antitumor effect of EGCG in combination with cetuximab. FASN, EGFR, and downstream protein expression were studied to determine cellular signaling pathways changes resulting from the different mono and dual treatments.

All TNBC cells were obtained from ATCC. MDA-MB-231, MDA-MB-468, and MDA-MB-157 were routinely grown in DMEM (Gibco). HCC1806, Du4475, and BT549 were maintained with RPMI. Cell culture media were supplemented with 10% FBS (HyClone Laboratories), 1% l-glutamine (Gibco), 1% sodium pyruvate (Gibco), and 50 U/mL penicillin/streptomycin (Linus). RPMI was additionally supplemented with 0.023 IU/mL insulin for BT549 culture. Cell lines were kept at 37°C and 5% CO₂ atmosphere. Doxorubicin-resistant cells MDA-MB-231 (231DXR) and HCC1806 (HCCDXR) were developed using a stepwise selection method. Increasing doses of doxorubicin (TEDEC-Meiji Farma) were performed until their corresponding IC₅₀ was reached. Briefly, initially cells were treated with a concentration of doxorubicin of 0, 1× IC₅₀. After 48 hours, the 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 doxorubicin concentration for 48 hours. The stepwise selection method was subsequently performed until the final concentration of doxorubicin achieved the parental IC₅₀. It took around 6 months for each cell line. Resistance was confirmed by cell viability assay.

Parental and resistant TNBC cells were synchronized by starvation in serum-deprived medium (0.5% FBS) for 24 hours. Cells were lysed in ice-cold lysis buffer (Cell Signaling Technology, Inc.) with 100 μg/mL phenylmethylsulfonylfluoride (PMSF) by vortexing every 5 minutes for 30 minutes. Frozen tumors were first ground (Dounce) in ice-cold lysis buffer and then vortexed as described above. Equal amounts of protein were heated in LDS Sample Buffer with Sample Reducing Agent (Invitrogen) for 10 minutes at 70°C, electrophoresed on SDS-PAGE, and transferred onto nitrocellulose membranes. Blots were incubated for 1 hour in blocking buffer [5% powdered-skim milk in PBS 0.05% Tween (PBS-T)] and incubated overnight at 4°C with the appropriate primary antibodies diluted in blocking buffer. Specific horseradish peroxidase (HRP)-conjugated secondary antibodies were incubated for 1 hour at room temperature. The immune complexes were detected using a chemiluminescent HRP substrate [Super Signal West Femto (Thermo Scientific Inc.) or Immobilon Western (Millipore)]. β-Actin (Santa Cruz Biotechnology Inc.) was used as a control of protein loading. Western blot analyses were repeated at least three times and representative results are shown.

Parental and resistant cells were plated in 96-well plates at a cell density of 4 × 10³ cells per well in their corresponding growth medium. After 24 hours, growth medium was removed, and 100 μL of fresh medium containing the corresponding concentration of doxorubicin (TEDEC-Meiji Farma), EGFR inhibitor [cetuximab (erbitux, Merck)], or FASN inhibitors [EGCG (Sigma) or C75 (Sigma)] were added to each well. For drug combination experiments, cells were treated with a fixed FASN inhibitor concentration in combination with a serial of concentrations of (i) doxorubicin for 48 hours or (ii) cetuximab for 4 days. Same treatments were assessed in monotherapy. Following treatment, cell viability was measured using the standard colorimetric MTT assay as previously described (19). Combinatorial effects were evaluated using the Interaction index (Ix) = ∑(% CPI drug alone)/% CPI combination. Ix < 1, synergism; Ix = 1, additivism; and Ix > 1, antagonism.

Cells were PBS washed, and then 1 mL of Qiazol (Qiagen) was added. Total RNA was isolated using RNeasy Mini Kit (Qiagen) following the instructions provided by the manufacturer. RNA was reverse-transcribed into cDNA using High Capacity cDNA Archive Kit (Applied Biosystems). Gene expression levels of EGFR and FASN were assessed using LightCycler 480 Real-time PCR System (Roche) with LightCycler 480 SYBR Green I Master (Roche). Primers were designed as follows: FASN [forward (Fw), 5′-CAGGCACACACGATGGAC-3′; reverse (Rv), 5′-CGGAGTGAATCTGGGTTGAT-3′], EGFR (Fw, 5′- CATGTCGATGGACTTCCAGA -3′; Rv 5′- GGGACAGCTTGGATCACACT -3′), and β-actin (Fw, 5′- ATTGGCAATGAGCGGTTC -3′; Rv, 5′- CGTGGATGCCACAGGACT -3′). RT-PCR analyses were performed at least three times and each gene was run in triplicate. β-Actin was used for normalization of EGFR and FASN expression levels.

FASN and EGFR tumor expression levels were retrospectively evaluated in paraffin-embedded core biopsies of 29 patients with TNBC diagnosed in the Hospital Josep Trueta of Girona (Girona, Italy) between 2007 and 2012. Tissue slides were previously treated with the PT link (DAKO) solution at high pH as a deparaffination and antigen retrieval steps. Immunohistochemical (IHC) staining was performed using an anti–fatty acid synthase polyclonal antibody (dilution 1:100, Enzo Life Sciences) and an anti-EGFR monoclonal antibody (dilution 1:100, Cell Signaling) with the detection kit EnVision (DAKO) in the AutostainerPlus Link (DAKO). FASN expression was graded from 0 to 2+: 0, no staining; 1+, moderate staining; 2+, high staining. EGFR expression was classified as positive when more than 5% of the cells showed specific membrane staining. Analysis was carried out by a board-certified pathologist. Written informed consent was obtained from all patients before the study was started. This study was approved by the ethical board of the Dr. Josep Trueta University Hospital.

TNBC cell lines HCC1806 (HCC) and MDA-MB-231 (231) and their doxorubicin-resistant derivatives 231DXR and HCCDXR were orthotopically implanted (2 × 10⁶ cells in 25 μL Matrigel) into both inguinal cleared mammary fat pads of NRG (NOD-Rag1<null> IL2rg<null>) mice (The Jackson Laboratory). When tumors reached 50 mm³, animals were randomized into four different treatment groups. Each group received intraperitoneal injection of control (vehicle alone, 3 d/week), 30 mg/kg EGCG 3 d/week, 0.5 mg/mice cetuximab 1 d/week, or combination of EGCG + cetuximab. Tumor xenografts were measured with calipers, and tumor volumes were determined using the formula: [π/6 × (v₁ × v₂ × v₂)], where v₁ represents the largest tumor diameter and v₂ the smallest one. Body weight was registered every 2 days. At the end of the experiment or when tumors reached 1 cm in diameter, animals were weighed and then euthanized using CO₂ inhalation. Tumors were stored at −80°C. Apoptosis in control and treated tumors was analyzed by Western blotting (PARP). In vivo mouse experiments were performed in Dr. A. Welm's laboratory at the Huntsman Cancer Institute (University of Utah, Salt Lake City, UT) and at the Oklahoma Medial Research Foundation (ORMF, Oklahoma City, OK). All protocols and experiments involving animals, including care and handling, were conducted in accordance with guidelines on animal care and use established by institutional guidelines. All mice were maintained in a specific pathogen-free facility with controlled light/dark cycle, temperature, and humidity. All surgery was performed under inhaled isoflurane anesthesia; buprenorphine was administered as analgesic to mice after surgery. All efforts were made to minimize pain and distress.

Data were analyzed by the Student t test when comparing two groups or ANOVA using a Bonferroni post hoc test when comparing more than 2 groups. Nonparametric analysis by Kruskal–Wallis was used when data did not follow normal distribution. Statistical significant levels were P < 0.05 (denoted as *), P < 0.01 (denoted as **), and P < 0.001 (denoted as ***). P value is shown in results when significance is reached (P < 0.05). All data are means ± SEM. All observations were confirmed by at least three independent experiments.

We screened several TNBC cell lines for FASN, EGFR, and downstream protein expression and activation status (Fig. 1A). Molecular classification of TNBC cells was assessed on the basis of Lehman and colleagues (7). Interestingly, FASN was detected in all TNBC cells examined and showed higher protein expression levels in the basal-like than in the mesenchymal-like molecular subtype. Basal-like cell lines Du4475 and MDA-MB-468 showed the higher mRNA levels, whereas the other cell lines showed lower and similar FASN mRNA levels (Supplementary Fig. S1A). Activation of the EGFR (as assessed by phosphorylation) was also observed in almost all cell lines. The EGFR mRNA and protein expression showed that MDA-MB-468 had the highest levels and Du4475 the lowest (Supplementary Fig. S1B).

To further investigate the role of FASN as a target in TNBC, a cell proliferation inhibition (CPI) assay was performed using C75 and EGCG as FASN inhibitors (Fig. 1B). The most sensitive cell lines to C75 were the basal-like HCC1806 (IC₅₀: 29.4 ± 1.4 μmol/L) and MDA-MB-468 (IC₅₀: 31.38 ± 2.5 μmol/L). Higher values of IC₅₀ were obtained in the mesenchymal-like subgroup for the lines MDA-MB-231(IC₅₀: 46.6 ± 2.2 μmol/L) and MDA-MB-157 (IC₅₀: 58.3 ± 2.75 μmol/L). EGCG showed IC₅₀ values ranging from 149 ± 6.7 mol/L in MDA-MB-231 to 222 ± 6.7 mol/L in BT549.

As preclinical models of acquired resistance to chemotherapy, we developed two TNBC cell lines resistant to doxorubicin. MDA-MB-231 (231) from the mesenchymal-like subgroup and the basal-like cell line HCC1806 (HCC) were chosen as two different models for FASN expression levels. 231 doxorubicin-resistant (231DXR) and HCC doxorubicin-resistant (HCCDXR) cells were developed in our laboratory by dose-increasing treatments of doxorubicin as described in Materials and Methods. 231 and HCC were significantly more sensitive to doxorubicin for doses ranging from 10 to 350 nmol/L compared with the resistant models 231DXR and HCCDXR, respectively (Supplementary Fig. S2). As shown in Fig. 1C, 231DXR cells showed no apparent changes in either total protein or activation level in any of the proteins we examined related to the EGFR pathway. HCCDXR, however, showed increased expression levels of total EGFR and the downstream proteins p-AKT and p-ERK1/2. On the other hand, p-EGFR levels decreased compared with the parental, consistent with previous findings showing doxorubicin directly decreases EGF expression (22). FASN showed similar levels between the parental and resistant cells. Gene expression for FASN and EGFR proteins did not show any significant differences in 231 or HCC versus 231DXR and HCCDXR, respectively (Supplementary Fig. S3). Interestingly, no changes in sensitivity to FASN inhibitors alone (C75 or EGCG) were observed between parental and doxorubicin-resistant cell lines (Supplementary Fig. S4).

Resistant cell lines 231DXR and HCCDXR showed similar FASN expression levels and FASN inhibitor sensitivity when compared with parental cells (Fig. 1C and Supplementary Fig. S4). Therefore, we tested the effects of EGCG and C75 in combination with doxorubicin in parental and doxorubicin-resistant models. The results of CPI ratios induced by the mono treatments versus the dual treatments are shown in Table 1 (for MTT results, see Supplementary Fig. S5). Doxorubicin (50–100–150 nmol/L) combined with C75 (40 μmol/L) exhibited synergic interaction index (Ix) in 231 and 231DXR. Furthermore, resistant cells 231DXR showed increased synergic interaction compared with 231. Similar results were observed in HCCDXR, because the addition of C75 (20 μmol/L) to the doxorubicin (50–100–150 nmol/L) treatment improved the effect of the combination compared with the parental cell line. The combination of doxorubicin (50–100–150 nmol/L) and EGCG (100 μmol/L) in 231 and 231DXR showed an additive effect in both cell lines. Interestingly, EGCG (120 μmol/L) combined with doxorubicin (50–100–150 nmol/L) also showed a major benefit in HCCDXR when compared with the parental.

Several studies reported that EGFR is a common receptor expressed in TNBC (23, 24). In vitro, EGFR activation is observed in almost all TNBC cell lines (Fig. 1A), and 70% of the TNBC tumors evaluated in this study were positively stained for EGFR protein (Fig. 1B). Furthermore, we have shown that the expression of EGFR is either maintained or increased in our doxorubicin-resistant models (Fig. 1C). Therefore, the effect of combining cetuximab and the FASN inhibitors C75 and EGCG was studied.

The combination of cetuximab (0.5–1–2 μmol/L) with EGCG or C75 showed a strong synergistic effect in all sensitive and resistant cells (Table 2). 231DXR showed lower Ix with both C75 (30 μmol/L) and EGCG (50 μmol/L) compared with parental 231 cells. However, HCC and HCCDXR showed similar ratio values for both C75 (20 μmol/L) or EGCG (25 μmol/L) when combined with cetuximab.

Interestingly, parental and resistant cells showed no differences in cell proliferation inhibition when treated with anti-FASN compounds alone. Cetuximab otherwise significantly increased CPI in vitro in the HCCDXR cell line (Supplementary Fig. S6).

To validate our in vitro findings, we then tested the antitumor activity of EGCG, cetuximab, and the combination treatment in vivo. We developed orthotopic xenograft models for 231, its chemoresistant derivative 231DXR, HCC and its resistant derivative HCCDXR. EGCG (30 mg/kg for 3 d/week) and cetuximab (0.5 mg/mice 1 d/week), as single agents reduced tumor growth in 231 xenografts over 12 days of treatment (Fig. 2A). Control animals achieved a relative volume growth of 3.3 ± 0.33, whereas EGCG reduced it to 3.0 ± 0.27 and cetuximab to 2.5 ± 0.29. Interestingly, dual FASN and EGFR blockade showed significantly higher tumor growth inhibition (2.0 ± 0.07) compared with the control (P < 0.01) or EGCG alone (P < 0.05, Fig. 2A). Under the same schedule, 231DXR xenograft showed a similar relative volume growth as the sensitive model (3.41 ± 0.4, Fig. 2B). EGCG showed tumor growth inhibition by day 12 with a relative volume growth of 2.66 ± 0.33, but cetuximab treatment did not inhibit growth. By 12 days of treatment, the relative volume growth for cetuximab was 3.52 ± 0.48, similar to the control group. Again, the combination of EGCG and cetuximab induced significant tumor growth inhibition (1.98 ± 0.17) compared with control (P < 0.01) and to cetuximab alone (P < 0.05). Despite the absence of tumor shrinkage, the combination treatment (EGCG plus cetuximab) significantly reduced tumor growth in both 231 and 231DXR orthotopic xenografts.

In the basal-like HCC xenograft model, EGCG (30 mg/kg for 3 d/week) and cetuximab (0.5 mg/mice 1 d/week) as single agents reduced tumor growth after 10 days of treatment (Fig. 2C). Compared with the tumor volume increase observed in the control group (17.79 ± 2.76), EGCG treatment achieved a relative volume growth of 13.03 ± 2.58, whereas cetuximab significantly reduced the ratio to 6.33 ± 1.25 (P < 0.05). The combination treatment significantly reduced the tumor growth compared with the control (P = 0.001) and EGCG alone (P < 0.01), with a relative volume growth of 4.87 ± 1.06. In the HCCDXR xenograft model, control animals reached a relative volume growth of 5.02 ± 1.38 over the same period of time, significantly lower compared with the parental model (P < 0.05, Fig. 2D). EGCG and cetuximab as single agents had higher activity in the HCCDXR model than in HCC, with a relative volume growth of 3.04 ± 0.76 and 0.14 ± 0.04, respectively (cetuximab vs. control, P < 0.01; cetuximab vs. EGCG, P < 0.01). The relative volume growth in the combination setting decreased up to 0.19 ± 0.1 compared with control (P < 0.05) and EGCG alone (P < 0.05). In the HCCDXR model, cetuximab used as a single agent quickly (day 4) displayed strong antitumor activity, leading to effective tumor regression (0.76 ± 0.12). Thus, the combination of EGCG plus cetuximab was not synergic at these treatment doses in the HCCDXR xenograft. For total volume results, refer to Supplementary Fig. S7.

Tumor samples from the in vivo TNBC models were collected to evaluate drug response signaling mechanisms. 231 and 231DXR tumor samples were collected after 12 days of treatment. For HCC and HCCDXR, samples were collected after 1 cycle (4 days) because in cetuximab and EGCG plus cetuximab groups, total tumor shrinkage was observed after 10 days of treatment. 231, HCC, and HCCDXR tumor samples showed increased apoptosis (assessed by cleavage of PARP) in the combination treatment settings compared with single treatments and control groups (Fig. 2E). 231DXR however showed increased apoptosis in both single and combination treatments compared with control. Total EGFR and p-EGFR levels decreased in 231, 231DXR, and HCCDXR after cetuximab and EGCG plus cetuximab cotreatment. HCC, however, showed a decrease in p-EGFR after cetuximab and combination treatment, but total EGFR levels decreased only after cetuximab treatment. No changes in FASN protein levels were observed among mono- and cotreatments in 231, 231DXR, HCC, and HCCDXR tumors.

NRG mice treated with EGCG (30 mg/kg for 3 d/week) plus cetuximab (0.5 mg/mice 1 d/week) were weighed daily to evaluate in vivo body weight effect. With respect to initial weight, no significant changes on food and fluid intake or body weight after treatments were identified (Fig. 2F).

We have shown that EGCG combined with cetuximab was synergic in several cellular and animal models of TNBC. Therefore, we planned to study the effect of the EGCG plus cetuximab on FASN, EGFR, and downstream protein expression and activation, alone or in combination, in sensitive (231 and HCC) and resistant (231DXR and HCCDXR) cells (Fig. 3).

Cetuximab as a single agent induced a decrease in total EGFR in 231, 231DXR, and HCCDXR cellular models. On the other hand, the same treatment increased total EGFR in HCC. While cetuximab treatment decreased p-EGFR levels in HCC and HCCDXR, 231 and 231DXR showed overactivation of p-EGFR after 72 hours of treatment. Previously described studies suggest that this activation may be likely to receptor homodimerization and autophosphorylation and does not activate downstream proteins (22). Protein levels of AKT and ERK1/2 were maintained after cetuximab treatment in all cellular models. Regarding p-AKT activation levels, lower levels were observed compared with control in 231DXR and HCCDXR after 72 hours of treatment. 231 and HCC, however, showed similar or increased p-AKT levels compared with control after 72 hours of treatment, respectively. Cetuximab treatment induced apoptosis in 231 and 231DXR after 72 hours, whereas cleaved PARP is observed already after 48 hours in HCCDXR. Total PARP increases after cetuximab treatment in HCC but cleaved PARP was not observed.

EGCG reduced the expression and activation of EGFR and slightly decreased FASN levels after 24 hours of treatment in all cellular models studied. Downstream proteins AKT and ERK1/2 showed similar levels at 24 hours of treatment with respect to the control. While p-AKT decreased after 24 hours of treatment, p-ERK1/2 increased considerably after both 12 and 24 hours in all cellular models. PARP cleavage is observed in all cell lines after EGCG treatment.

The combination of both drugs led to the down regulation after 24 hours of total EGFR and p-EGFR in 231, 231DXR, and HCCDXR. Otherwise, p-EGFR levels decreased in HCC, whereas total EGFR did not change during the treatment. Regarding to FASN levels, a slight decrease is observed in HCCDXR, whereas the expression level is maintained in the other cell lines. Total ERK1/2 and AKT levels were maintained during the treatment in all cell lines, except for HCCDXR, which showed decreased levels of total AKT after 48 hours of treatment. p-ERK1/2 activation status increased in 231 and 231DXR after 48 hours whereas decreased in HCC and HCCDXR. Truncated PARP was observed in all cell lines at 24 and 48 hours for the combination treatment.

In summary, the combination of cetuximab and EGCG enhances the inhibitory effect observed in the EGFR pathway by single treatments, with a general decreasing of EGFR, p-EGFR, and p-AKT levels in all cellular models. The activation of ERK1/2 observed after EGCG treatment was decreased or inhibited after combination treatment.

To further validate FASN as a potential target in TNBC, FASN expression was evaluated by immunohistochemistry (IHC) in paraffin-embedded core biopsies from 29 TNBC tumor samples. The clinicopathologic characteristics of the patients included in the study are shown in Supplementary Table S1. FASN staining was positive in all 29 TNBC samples with tumor-specific staining (Fig. 4). FASN expression positivity was graded from 0 to 2+ and scored into three categories, as described in Materials and Methods. Staining was observed in all 29 tumor samples and 31% of them with high positivity (2+). Cellular positivity within the tumor was between 66% and 100% in 26 (89.7%) of the samples and between 33% and 66% in 3 of them (10.3%). The EGFR was also evaluated in the same 29 TNBC core biopsies with a positivity of 75.9%.

TNBC lacks an approved targeted therapy, and it is mainly treated with a combination of anthracyclines and taxanes (25). The high rate of relapse after treatment urges for the development of new targeted therapies for these patients. TNBC is well-known for its heterogeneity (8). Here we show that FASN, the key enzyme for de novo lipogenesis, is not only expressed in all TNBC cell lines included in the study but also in TNBC patient samples. In vitro, basal-like cell lines express more FASN than mesenchymal-like cell lines (both at gene and protein levels) and are more sensitive to the FASN inhibitor C75. Immunostaining of FASN in 29 core biopsies from patients with TNBC shows that FASN is overexpressed in all tumor samples included in the study compared with their nontumoral surrounding breast tissues. Recent studies performed in breast and ovarian cancers have demonstrated the overexpression and specificity for FASN staining in neoplastic tissues and also its relationship with cancer aggressiveness (26, 27).

After four decades, doxorubicin is still the most used and effective chemotherapy agent in the clinic for the treatment of breast cancer (28). For this reason, two doxorubicin-resistant models representative of the two major molecular subtypes in TNBC, 231DXR (mesenchymal-like) and HCCDXR (basal-like), were developed in our laboratory. Both models showed significant resistance when exposed to high doses of doxorubicin compared with the parental cell lines. Several mechanisms have been described for doxorubicin resistance in the last few years (28, 29), some of them regarding changes in the EGFR pathway. HCCDXR showed overexpression of EGFR and overactivation of AKT and ERK1/2 when compared with its parental cell line. Overexpression of EGFR in patients with breast cancer after chemotherapy has been already described to be related with higher local recurrence, metastasis, and so poor outcome (30, 31). On the other hand, AKT and ERK1/2 have been shown to be overactivated and also linked to chemoresistance in several cancers (32, 33). Interestingly, similar levels of FASN were observed in parental cell lines compared with their resistant derivate (DXR). Recent studies pointed out that FASN activity plays an important role in drug resistance because it allows fast synthesis of new phospholipids for membrane remodeling and plasticity (34, 35). Here, we performed combinatorial experiments with doxorubicin and the FASN inhibitors C75 or EGCG in our sensitive and resistant cell lines to determine whether FASN inhibition may be useful to overcome doxorubicin resistance in TNBC. Although the EGCG antitumor effects may be due to binding probably with other cellular targets than FASN (36), we and others have previously identified EGCG as an inhibitor of FASN activity, able to induce apoptosis in several tumor cell lines, and also to reduce the size of mammary tumors in animal models (17–19, 37, 38). Our results showed that the interaction index for both combinations was improved in both resistant models (231DXR and HCCDXR) compared with the parental lines. Interestingly, C75 and EGCG showed same CPI between parental or DXR derivate when treated alone, suggesting that FASN activity could play an important role in doxorubicin resistance in TNBC. In addition, FASN inhibition has been described to reverse chemoresistance in ovarian and hepatocellular cancer cells (39, 40).

Several studies have shown that EGFR is overexpressed in TNBC (22, 23). Our results showed a high percentage of immunostaining positivity for EGFR in the core biopsies from patients with TNBC, in concordance with previous findings. The anti-EGFR compound cetuximab showed low levels of CPI in vitro in all cell lines except for HCCDXR, which showed significantly increased CPI and apoptosis compared with the parental cell line. In vivo, while none, low, or moderate antitumor effects were observed in 231DXR, 231, and HCC xenografts, respectively, cetuximab treatment obtained almost complete tumor shrinkage after two cycles of treatment in HCCDXR. These results are consistent with the EGFR overexpression in HCCDXR cell line, probably as an adaptation to overcome doxorubicin-induced apoptosis. Paradoxically, this mechanism of chemoresistance could sensitize the cell line to anti-EGFR treatments. This is in agreement with the results of a phase II study combining cetuximab with cisplatin in metastatic TNBC, where the overall response rate was doubled and the progression-free survival increased from 1.5 to 3.7 months when compared with cisplatin monotherapy (41).

The common overexpression of EGFR in TNBC and its relationship with cancer progression after therapy, together with FASN expression in all TNBC and in DXR models, provided the rationale to test combined FASN inhibitors and EGFR blockade in these settings. We found that simultaneous treatment of cetuximab and EGCG or C75 resulted in strong synergism in both parental and resistant models. Strong cooperative growth inhibition between FASN and EGFR has been also observed in vitro for ovarian cancer cell lines (42). The synergism between FASN and EGFR may be due to the mechanisms involving FASN blockade that results in blockage of lipid synthesis, lipid raft destabilization, and further EGFR degradation (42). Furthermore, EGCG has been described to be closely associated with destabilization and inhibition of EGFR receptor and downstream signaling pathways (43, 44). In addition, Bollu and colleagues recently found that FASN-dependent palmitoylation of EGFR is required for EGFR dimerization and kinase activation (45). The same authors, in a previous study, described that FASN-dependent palmitoylation also activates mitochondrial EGFR and promotes cell survival whereas its inhibition leads to apoptosis (46). Furthermore, phosphorylation and activation of FASN by direct interaction with members of the EGFR family have already been described (46).

These in vitro results together with TNBC patient sample study supported the evaluation of the antitumor efficacy of cetuximab in combination with the FASN inhibitor EGCG in doxorubicin-sensitive and -resistant TNBC orthoxenografts in vivo. We and others have demonstrated that EGCG displays antitumor activity without affecting food intake or weight loss, which are common side effects of other FASN inhibitors (37, 38, 47). We have previously shown that EGCG produces apoptosis in vitro and in vivo in breast cancer (19, 37, 38). Here, EGCG showed strong CPI and high apoptosis in vitro and moderate antitumor activity in vivo. EGFR inactivation and downregulation are observed in all cell lines after EGCG treatment in vitro, consistent with previous studies (43, 44, 48). At the same time, slight decreases in FASN levels were observed in all cell lines, suggesting that the EGFR pathway could be implicated in FASN expression regulation in TNBC (38, 49). Furthermore, p-AKT decreased its activation levels, whereas p-MAPK increased considerably after EGCG treatment, maybe due to a negative crosstalk between these two proteins (33, 42, 38).

Here, we clearly validate our in vitro results by showing that the combination of EGCG with the EGFR inhibitor cetuximab results in synergistic tumor growth reduction in both doxorubicin-sensitive and -resistant xenograft models. Tumor analysis showed decreased EGFR and p-EGFR in the cetuximab and combination settings in all models. However, this event was not observed in the EGCG treatment in vivo possibly due to the unstable nature of EGCG in serum (50). Nevertheless, increased apoptosis was observed in the combination settings in all models compared with the single treatments, even after only one cycle in the HCC and HCCDXR xenograft. In vitro study interactions also reveal that p-ERK1/2 levels decreased in the combination settings in the HCC and HCCDXR cell lines and reduced the loop effect in p-ERK1/2 in the combination treatment in 231 and 231DXR.

In summary, we found that FASN is expressed in all TNBC primary tumor samples tested. We then showed therapeutic benefit of FASN inhibition in sensitive and chemoresistant TNBC preclinical models, as well as strong synergistic antitumor effect of FASN inhibitors in combination with cetuximab. Our findings provide a rationale for further investigation of FASN inhibitors in combination with anti-EGFR signaling agents in patients with TNBC and who progressed after chemotherapy.

No potential conflicts of interest were disclosed.

Conception and design: G. Viñas, T. Puig

Development of methodology: A. Giró-Perafita, S. Palomeras, D.H. Lum, A. Blancafort, G. Viñas, G. Oliveras, F. Pérez-Bueno

Acquisition of data (provided animals, provided facilities, etc.): A. Giró-Perafita, D.H. Lum, G. Viñas, A.L. Welm

Analysis and interpretation of data: A. Giró-Perafita, F. Pérez-Bueno, A. Sarrats, A.L. Welm, T. Puig

Writing, review and/or revision of the manuscript: A. Giró-Perafita, D.H. Lum, A. Sarrats, T. Puig

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Giró-Perafita

Study supervision: G. Viñas, T. Puig

The authors thank Dr. F. Soler (Division of Pharmacy of Catalan Institute of Oncology, Girona, Spain) for kindly supplying doxorubicin. They also thank Oklahoma Medical Research Foundation (OMRF) PDX-PCT Core for their support in the animal experiments. They thank Drs. E. López (Pathology Department Director, Hospital Josep Trueta, Girona, Spain) and J. Brunet (Medical Oncology Director, Catalan Institute of Oncology, Girona, Spain) for their support in the patient studies.

This work was supported in part by the Spanish Instituto de Salud Carlos III (ISCIII; FIS PI11/00692 and PI14/00329; to' T. Puig), Fundación Ramón Areces (to T. Puig), the support of Catalonian Government (2014SGR00868), U.S. Department of Defense Breast Cancer Research Program Award (#W81XWH1210077; to A.L. Welm), Xarxa de Bancs de Tumors de Catalunya sponsored by Pla Director d'Oncologia de Catalunya (XBTC), and Metavivor Research and Suport Inc (#50502690, to D.H. Lum). The University of Girona awarded A. Giró-Perafita, S. Palomeras, and A. Blancafort with a predoctoral grant.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.