Introduction
Amplification of the
ERBB2 oncogene is one of the most clinically relevant genetic changes in breast cancer and occurs in 10% to 34% of breast cancer cases.
ERBB2 overexpression is a significant predictor of both overall survival and time to relapse [
1]. The association of
ERBB2 amplification with aggressive disease and poor clinical outcome in breast cancer has made ERBB2 an attractive therapeutic target. Trastuzumab, a human monoclonal antibody targeted against the extracellular domain of ERBB2, was widely hailed as the first next generation cancer therapy when it was introduced for the treatment of estrogen receptor-negative breast cancer. Its success has been modest. When used as single-agent therapy in patients with metastatic ERBB2-positive breast cancer, response rates ranging from 11% to 26% have been observed. Because a relatively large proportion of patients do not benefit from ERBB2-targeted therapy, it is likely that factors in addition to ERBB2 itself must influence the response of these tumors to this therapy. Recently, an RNA interference screen identified regulators of fat metabolism (including the peroxisome proliferator-activated receptor [PPAR]γ-binding protein [PBP] and the nuclear receptor NR1D1 [nuclear receptor subfamily 1, group D, member 1], a PPARγ target protein) as being relevant to the survival specifically of ERBB2-positive breast cancer cells, but not that of other breast cancer cells or normal mammary epithelial cells (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data). Both genes reside on the ERBB2 amplicon and are transcriptional regulators that positively affect expression of genes such as
FASN (fatty acid synthase),
ACLY (ATP citrate lyase) and
ACACA (acetyl-coenzyme A carboxylase alpha), which are the three major enzymes of
de novo fatty acid synthesis (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data). As a result, ERBB2-positive breast cancer cells contain significantly higher amounts of cellular fats, as compared with other breast cancer cell lines or normal cells, because of concomitant overexpression of
NR1D1 and
PBP genes (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data) [
2].
The end product of
de novo fatty acid synthesis, namely palmitate, and other saturated fatty acids like it, are toxic to cells. Palmitate has been shown to generate a variety of apoptotic signals [
3,
4]. In some cases these involve the synthesis of ceramide [
5,
6], whereas in others reactive oxygen species (ROS) are produced [
3,
7]. Studies conducted in a variety of cell types, including breast cancer cell lines [
8,
9], suggest that this lipotoxicity is specific for saturated fatty acids such as palmitate. Triglyceride accumulation in nonadipose cells represents a cellular defense mechanism against lipotoxicity [
10]. Because exogenous unsaturated fatty acids have an impact on this process, it may represent a mechanism for effects of diet on cancer etiology.
Both PBP and NR1D1 are functionally related to PPARγ [
11,
12]. PPARγ expression is also higher in ERBB2-positive breast cancer cells [
13]. These cells were more sensitive to inhibition of PPARγ with antagonists such as GW9662 and T0070907, as compared with other types of breast cancer cells or normal mammary epithelial cells (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data). PPARγ inhibition in ERBB2-positive breast cancer cells resulted in cell death and apoptosis, similar to the effects of PBP and NR1D1 inhibition (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data). PPARγ is a major regulator of adipogenesis and lipid homeostasis [
14]. We sought to examine the reasons for the dependence of ERBB2-positive breast cancer cells on PPARγ for survival. In this context, we identified palmitate-induced lipotoxicity as a main effect of PPARγ inhibition in ERBB2-positive breast cancer cells.
Materials and methods
Cell culture and chemicals
Breast cancer cell lines BT474, MCF-7 and MDA-MB-361 were obtained from the American Type Culture Collection (Manassas, VA, USA). Human mammary epithelial cells (HMECs) were obtained from Cambrex (East Rutherford, NJ, USA). BT474 and MCF-7 cells were cultured in Dulbecco's modified Eagle's medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum (Hyclone) and 100 U/μl penicillin-streptomycin (Cellgro, Herndon, VA, USA); BT474 medium was also supplemented with ITS (insulin, transferring and selenium; Cellgro). MDA-MB-361 were cultured in RPMI-1640 (Hyclone) supplemented with 20% fetal bovine serum and 100 U/μl penicillin-streptomycin. HMECs were cultured in mammary epithelial growth medium (Cambrex). The PPARγ antagonist GW9662, the fatty acid palmitate, and the ceramide synthesis inhibitor fumonisin B1 were obtained from Sigma-Aldrich (St. Louis, MO, USA).
Cell viability: proliferation assays
Cell viability after small hairpin RNA transfection or chemical treatments was assessed live cell counts after trypsinization and trypan blue staining using a hemocytometer. For high-throughput experiments, cells grown on 96-well plates were washed once with 1× phosphate-buffered saline (PBS), fixed with 2.5% formaldehyde, stained with Hoechst 33342 (Molecular Probes-Invitrogen, Carlsbad, CA, USA) and analyzed with an In Cell Analyzer 1000 (GE Healthcare, Piscataway, NJ, USA) high-content imaging system; cell counts and statistics were performed using In Cell Investgator 3.4 software (GE Healthcare).
Reverse transcription polymerase chain reaction
Total RNA was extracted from cells using TRizol (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized by reverse transcription of 2 μg of RNA in a 20 μl reaction using Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI, USA) at 42°C for 1 hour. PCR reactions were performed by using standard Taq polymerase (Fisher BioReagents, Fairlawn, NJ, USA) with the following primer pairs (forward and reverse, respectively): PPARγ, 5'-AGCCTCATGAAGAGCCTTCCA-3' and 5'-ACCCTTGCATCCTTCACAAGC-3'; fatty acid binding protein 4 (FABP4; aP2), 5'-GCATGGCCAAACCTAACATGAT-3' and 5'-CCTGGCCCAGTATGAAGGAAA-3'; hormone sensitive lipase (HSL), 5'-TACAAACGCAACGAGACAGGC-3' and 5'-TGTGATCCGCTCAAACTCAGC-3'; adipose tryglyceride lipase (ATGL), 5'-AGCTCATCCAGGCCAATGTCT-3' and 5'-GGTTGTCTGAAATGCCACCAT-3'; carnitine palmitoyltransferase 1 (CPT-1), 5'-TCACATTCAGGCAGCAAGAGC-3' and 5'-AATCGTGGATCCCAAAAGACG-3'; and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 5'-GCAAATTCCATGGCACCGT-3' and 5'-TCGCCCCACTTGATTTTGG-3'.
After the initial denaturation step (95°C for 3 minutes), PCR reactions consisted of 30 to 35 cycles of a 95°C step (15 seconds), a 52 to 55°C step (15 seconds), and a 72°C step (20 seconds), followed by a final elongation step at 72°C (5 minutes). PCR products were separated on 2% agarose-ethidium bromide gels. For quantitative determination of PCR product, a real-time reverse transcription PCR (RT-PCR) was performed on an ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA), using SYBR Green PCR Master Mix (Applied Biosystems). Primer pairs were the same as those used in regular RT-PCR. PCR reactions consisted of an initial incubation at 95°C (2.5 minutes) and 40 cycles of a 95°C step (15 seconds) and a 60°C step (60 seconds). Product levels were calculated after normalization with GAPDH or β-actin controls.
Immunofluoresence
Immunofluoresence was performed on cells grown and treated either in 96-well plates or on cover slips in 24-well plates. In all cases, cells were fixed after treatment with 2.5% formaldehyde, washed with 1× PBS, permeabilized with 0.1% Triton-X 100 (Fisher Chemicals, Fairlawn, NJ, USA), blocked with 3% normal goat serum (Sigma-Aldrich), incubated with a 1:50 to 1:200 dilution of the primary antibody, washed with 1× PBS, incubated with a 1:800 dilution of the secondary antibody, washed again with 1× PBS, and finally stained with Hoechst 33342 (Molecular Probes-Invitrogen). Cells stained on 96-well plates were imaged using the In Cell Analyzer 1000 (GE Healthcare) and signal measurements and statistics were performed by the In Cell Investigator 3.4 software (GE Healthcare). Cells immunostained on cover slips were imaged by using a Leica TCS SP5 confocal microscope system (Leica Microsystems Inc., Bannockburn, IL, USA). Antibodies used were anti-activated Bax (6A7; BD Pharmingen, San Jose, CA, USA) and Alexa Fluor 568 goat anti-rabbit IgG (#A-11011; Invitrogen).
For detection of neutral fat stores, cells were grown on 96-well plates, fixed with 2.5% formaldehyde, washed with 1× PBS, stained with 10 μg/ml 4,4-difluoro-1,3,5,7,8-pentamethyl-4-bora-3a,4a-diaza-s-indacene (BODIPY 493/503; Molecular Probes), and counter-stained with Hoechst 33342 (Molecular Probes) for nuclei identification. Cells were imaged using the In Cell Analyzer 1000 and pictures were analyzed using the In Cell Investigator 3.4 software.
For fatty acid detection and quantification, total cellular lipids were extracted in accordance with a procedure described previously [
15]. Briefly, approximately 10
7 cells were pelleted and re-suspended in 3 ml chloroform:methanol (1:2) weight/0.05% butylated hydroxytoluene (BHT) to extract lipids. To monitor the recovery of fatty acids, 100 μg of heptadecanoic acid was added to each sample, before lipid extraction. Samples were centrifuged to remove cellular debris and chloroform and distilled water was added to form a biphasic solution. The two phases were separated by centrifugation and organic phases were transferred to a new tube and dried under nitrogen gas. Fats were then re-suspended in 250 μl toluene + 500 μl 1% sulfuric acid in methanol and incubated at 50°C overnight under nitrogen gas, followed by two steps of 1.25 ml 5% NaCl – 1.25 ml hexane extraction. Hexane extractions were combined, washed with 1.0 ml 2% NaHCO
3, run through Na
2SO
4 columns, and dried under nitrogen gas. The fatty acid methyl esters were re-suspended in 1.0 ml methyl acetate and were analyzed by gas chromatography/mass spectrometry using an Agilent 6890 series gas chromatograph (Agilent Technologies Inc., Santa Clara, CA, USA) equipped with a 5873 mass-selective detector. Part of the extraction was used for quantification of total and individual fatty acids under mass spectrometry and part for thin layer chromatography separation of triglycerides by a toluene-based system.
Reactive oxygen species assay
O2- generation was measured by using hydroethidine (Molecular Probes-Invitrogen, Carlsbad, CA, USA). Cells were incubated with a final concentration of 10 μmol/l of the dye for 30 minutes, and then fixed with 2.5% formaldehyde, stained with Hoechst 33342 and imaged with an In Cell Analyzer 1000. The hydroethidine signal was quantified using the In Cell Investigator 3.4 software.
Statistical analysis
The Student's two-tailed t-test was employed for the calculation of P values.
Discussion
ERBB2-positive cells possess significantly higher amounts of tryglyceride stores than do other cell types (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data; Figure
1b–e). It has been shown that PBP and NR1D1 promote adipogenesis, and therefore fat storage, because of their functional relationship with PPARγ (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data) [
11,
12]. PPARγ activity is necessary for the viability specifically of ERBB2-positive breast cancer cells, in a similar manner to PBP and NR1D1 (Kourtidis A, unpublished data; Carkner RD, Eifert C, Brosnan MJ, Conklin DS; Figure
1a). Our findings indicate that apoptosis due to PPARγ inhibition is consistent with endogenous palmitate toxicity. PPARγ enables ERBB2-positive breast cancer cells to convert fatty acids to triglycerides in order to avert lipotoxicity caused by the significantly high levels of fats that these cells produce. The decreased ability of BT474 and MDA-MB-361 cells to accumulate more fats after exogenous supplementation of palmitate (Figure
2b,d) is in agreement with the idea that these cells have near toxic levels of endogenously produced palmitate. Exogenous palmitate that has no effects on MCF-7 cell or HMEC viability is lethal to the ERBB2-positive cells. These results underscore the importance of triglyceride accumulation as a cellular defense against lipotoxicity [
10] in cells that have abnormally high levels of fatty acid synthesis activity.
It has been established that cancer cells depend on an altered cellular physiology. As an example, cancer cells favor aerobic glycolysis instead of oxidative phosphorylation for energy production, a phenomenon described by Warburg several decades ago [
20]. Fatty acid synthesis has been proposed to facilitate this mode of energy production in several cancer cells [
21,
22]. Over-expression of
PBP and
NR1D1 in ERBB2-positive breast cancer cells causes these cells to store fatty acids at 10 times the level of other breast cancer cells (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data) [
2]. These two genes act coordinately to upregulate
de novo fatty acid synthesis, enabling ERBB2-positive cells to regenerate their NAD
+ by consuming nicotinamide adenine dinucleotide phosphate (NADPH) that is necessary for fatty acid synthesis. As a result, the cells can continue to catabolize glucose and maintain their metabolic balance (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data). ERBB2 has been also shown to have a positive bidirectional relationship with FASN, in this way influencing
de novo fatty acid synthesis [
23]. The tight genetic linkage between
PBP,
NR1D1, and
ERBB2 on the 17q12-21 amplicon commonly found in breast cancers [
24] suggests that ERBB2-positive breast cancer cells are genetically preprogrammed to depend on fatty acid synthesis for energy production.
The generation of ROS and the biosynthesis of ceramide have been proposed as possible mechanisms of lipotoxicity [
4]. Increased ROS production in BT474 cells following PPARγ inhibition, along with cell rescue after treatment with antioxidants, support the notion that ROS production is the mediator of palmitate toxicity in BT474 cells. However, inhibition of ceramide synthesis under the same conditions resulted in only a 27% increase in BT474 cell viability, showing that palmitate toxicity is mainly independent of ceramide production.
Several studies have shown that supplementation of saturated rather than unsaturated fatty acids is toxic to cells [
25]. When ERBB2-negative breast cancer cells like MDA-MB-231 were involved, only palmitate, but not oleate, was shown to induce apoptosis [
9]. It has been proposed that this is because unsaturated fats are more efficiently metabolized into triglycerides [
26] and that they can rescue cells from lipotoxicity by channeling saturated fats like palmitate to triglyceride stores. Nevertheless, a number of reports have indicated that the Mediterranean diet, which is rich in the unsaturated oleate-containing olive oil, has anti-oncogenic properties against ERBB2-positive breast cancer [
27,
28]. Similar effects were seen for other unsaturated fatty acids, like the omega-3 polyunsaturated docosahexaenoic acid [
29] and the omega-6 polyunsaturated γ-linolenic acid [
30]. PPARγ inhibition in BT474 cells resulted in higher levels mainly of saturated rather than unsaturated fats (Figure
1d,e), and we confirmed that the saturated fatty acid palmitate is specifically toxic to these cells. However, the unsaturated oleate was also increased by PPARγ inhibition (Figure
1e), and we observed that oleate treatment produced similar toxic effects on BT474 cells (data not shown). This indicates that it is not the type of fatty acid that is responsible for the toxicity in these cells, but rather the overall high levels of fats caused by the significantly upregulated
de novo fatty acid synthesis. ERBB2-positive cells have maximized their ability to store fats, and therefore additional supplementation of fatty acids, or interruption of their ability to store them by PPARγ inhibition, is lethal to these cells.
Excess of palmitate in ERBB2-positive cells could also feedback to inhibit fatty acid synthesis. It has been shown that high-fat diet downregulates both FASN and malic enzyme 1 [
31]. Both enzymes are necessary for
de novo fatty acid synthesis, because FASN is the enzyme that catalyzes palmitate synthesis from acyl-coenzyme A and malonyl-coenzyme A using NADPH, whereas malic enzyme 1 supplies FASN with NADPH. ERBB2-positive breast cancer cells are sensitive to inhibition of both FASN and malic enzyme 1 genes (Kourtidis A, Carkner RD, Eifert C, Brosnan MJ, Conklin DS; unpublished data). In this case, PPARγ activity is not only important for detoxifiying the cells but also for securing palmitate into fat stores, in order for FASN and malic enzyme 1 to continue to function. Further examination is needed to confirm this hypothesis and establish this feedback loop, which could provide several missing links in the regulation of this metabolic process and, therefore, further ways to eradicate this particularly aggressive form of cancer.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
AK carried out all cell-based studies, participated in fatty acid analysis studies, and drafted the manuscript. RS participated in the ROS studies. MJB and RC carried out fatty acid analysis. DC conceived of the study, participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.