HER2 (erbB2/neu) is a member of the erbB family of receptor tyrosine kinases and is involved in regulating the growth of several types of human carcinomas. HER2 represents a successful therapeutic target of the biotechnology era as exemplified by the drug Herceptin (trastuzumab), which has clinical activity in a subset of breast cancer patients. Using DNA microarrays, we identified a cohort of genes that are differentially regulated by HER2 in breast epithelial cells. One of the HER2-regulated genes discovered was fatty acid synthase (FAS), which has been shown to be overexpressed in breast cancer as well as other cancers. FAS is implicated in tumorigenesis through its role in cell proliferation and membrane lipid incorporation of neoplastic cells. Here, we demonstrate that HER2-mediated induction of FAS is inhibitable by Herceptin and tyrosine kinase inhibitors of HER2. Through a phosphatidylinositol 3′-kinase-dependent pathway, HER2 stimulates the FAS promoter and ultimately mediates increased fatty acid synthesis. Interestingly, pharmacological inhibition of FAS preferentially induced apoptosis of HER2-overexpressing breast epithelial cells relative to matched vector control cells. These studies characterize a molecular connection between two genes individually implicated in tumorigenesis but never linked together.
ErbB kinase receptors promote cell survival and can function as oncogenes. Family members include EGFR3 (erbB1), HER2 (erbB2), erbB3, and erbB4. All are receptor tyrosine kinases involved in regulating cell growth and differentiation (1). Each has been reported to be amplified or overexpressed in some forms of breast cancer, with HER2 and EGFR being the most extensively studied. Family members are activated by dimerization mediated by their respective ligands. Once activated, intracellular tyrosine kinase activity elicits diverse prosurvival signals mediated through targets including PI3K, Akt, RAS/RAF, and Janus-activated kinase/signal transducers and activators of transcription, among others. Heterodimerization between family members may allow simultaneous activation of signaling pathways unique to each receptor subtype.
HER2, one of the most well-characterized breast cancer oncogenes, is amplified in about 30% of breast cancers (2, 3). Overexpression of HER2 in tumors is associated with a poorer prognosis in terms of both relapse-free interval and overall survival (4, 5). Various strategies have emerged to block the proliferative effects of erbB family members. Monoclonal antibodies directed against HER2 have shown growth-inhibitory effects in vitro and in breast cancer xenografts (6). A humanized monoclonal antibody to HER2 (trastuzumab/Herceptin) causes tumor regression in 20% of patients with refractory breast cancer (7). This number is somewhat higher when bona fide HER2-amplified tumors are selected for therapy by fluorescence in situ hybridization (8). More recently, Herceptin has been shown to have a synergistic antitumor response with chemotherapeutic agents in women with metastatic breast cancer, increasing median relapse-free interval by 65% compared with chemotherapy alone (9). Whereas these studies suggest that Herceptin may have clear utility in select cases of advanced breast cancer, these benefits are modest and usually do not represent a cure.
Another strategy to block erbB family members involves the use of small molecule tyrosine kinase inhibitors that function by suppressing autophosphorylation of their target receptor. First-generation compounds such as PD153035 were shown to specifically inhibit EGFR (10). Recently, Pfizer has developed an erbB kinase inhibitor designated CI-1033 (11) that has been demonstrated to have potent inhibitory activity against the entire erbB family (12, 13). CI-1033 is a selective, irreversible inhibitor with an IC50 of approximately 10 nm for erbB kinases (the IC50 for other tyrosine kinases is 50,000-fold higher). In addition to being nontoxic to mice (12), CI-1033 has been shown to slow the growth of various cancer xenografts in vivo (11). Because these small molecule tyrosine kinase inhibitors target intracellular receptor phosphorylation (and thus signaling), they likely will be more potent than extracellular blocking antibodies.
Whereas a number of signaling pathways emanating from HER2 have been characterized, much less is known about the transcriptionally regulated HER2 genes that contribute to its tumorigenic effects. Thus, the genes that regulate the “output” of the erbB network are not well characterized. To focus on genes specifically affected by HER2, we made use of a human mammary epithelial cell line, H16N2, which ectopically overexpresses HER2 (14, 15). In addition, we examined a panel of breast cancer cell lines (relatively) recently derived from patients (e.g., SUM149). Using DNA microarrays, we defined a select set of genes induced by HER2 overexpression in breast cell lines and concomitantly repressed by HER2 inhibitors, including Herceptin and CI-1033 (11). One of the HER2-regulated genes we identified was FAS, which has been shown to be overexpressed in breast, prostate, ovarian, and colon cancer (16, 17, 18, 19, 20). Inhibitors of FAS (e.g., C75 and cerulenin) have been shown to have significant in vivo antitumor activity against breast cancer xenografts (21). Pharmacological inhibition of FAS activity induces apoptosis or cell cycle arrest of tumor cells and promotes regression of tumor xenografts. FAS inhibitors have also been shown to induce apoptosis or cell cycle arrest in a number of cancer cell lines (21, 22). Whereas FAS overexpression has been shown for breast tumor specimens, a molecular connection to HER2-mediated tumorigenesis has not been established.
Here we present evidence of molecular cross-talk between the HER2 and FAS signaling pathways at the level of transcription, translation, and biosynthetic activity. This “oncogenic connection” is mediated through PI3K via a direct effect on the cis-acting elements in the FAS promoter. Our results also demonstrate that pharmacological inhibitors of FAS preferentially kill HER2-overexpressing breast epithelial cells relative to matched vector control cells, suggesting that up-regulation of FAS may play a role in HER2-mediated tumorigenesis.
MATERIALS AND METHODS
The base medium for H16N2-PTP (vector-matched control cells), H16N2-HER2, and SUM-190PT cells was Ham’s F-12 media supplemented with 0.1% BSA, 5 mg/ml Pen-Strep, 5 mm ethanolamine, 10 mm HEPES, 5 mg/ml transferrin, 10 mm T3, 50 mm sodium selenite, and 1 mg/ml hydrocortisone. For maintenance of culture stocks, SUM-190PT cell medium was further supplemented with 5 mg/ml insulin, and H16N2-PTP cell medium was further supplemented with 10 ng/ml epidermal growth factor. The cells were propagated in serum-free media. The media for SUM-52PE, SUM-149PT, and SUM-225CWN was Ham’s F-12 supplemented with 5% fetal bovine serum, 5 mg/ml Pen-Strep, 5 mg/ml insulin, and 1 mg/ml hydrocortisone. SKBr3 cells were grown in DMEM supplemented with 10% fetal bovine serum and 5 mg/ml Pen-Strep. Stable cell lines were maintained in 100 mg/ml Geneticin (G418). All cell culture reagents were obtained from Sigma (St. Louis, MO). Detailed descriptions of the SUM cell lines can be found elsewhere.4 Treatment with various reagents was carried out on subconfluent plates (approximately 4–6 million cells/100-mm tissue culture plate) incubated at 37°C in a humidified atmosphere of 10% CO2, at the following concentrations: CI-1033 at 1 μm (Pfizer, Ann Arbor, MI); Herceptin at 200 μg/ml (Genentech, South San Francisco, CA); C75 at 5 μm (Alexis, San Diego, CA); and LY294002 at 25 μm (Calbiochem, San Diego, CA).
DNA microarray analysis of gene expression was carried out as described previously (23). Briefly, purified PCR products generated using the clone inserts as template were spotted onto poly-l-lysine-coated microscope slides using an Omnigrid robotic arrayer (GeneMachines, San Carlos, CA) equipped with quill-type pins (Majer Precision Engineering, Tempe, AZ). Based on the latest Unigene Build-136, our 10K human cDNA microarray covers approximately 5996 unique, known, named genes and 2674 expressed sequence tags. Of these, 495 genes are present in duplicates, which serve as internal controls. Protocols for printing and postprocessing of arrays are available in the public domain.5
Isolation of RNA, Labeling of cDNA Probes, and Hybridization to cDNA Arrays.
Approximately 4–6 million cells/100-mm tissue culture plate were harvested and washed with PBS. Cells were homogenized in Trizol (Invitrogen, Carlsbad, CA), and total RNA was isolated according to the manufacturer’s protocol. An extra phenol-chloroform extraction was performed to improve the quality of RNA for microarray analyses. RNA integrity was judged by denaturing formaldehyde-agarose gel electrophoresis. Total RNA was quantified and stored in aliquots at −80°C until use. Total RNA (15 μg) was routinely used as template for cDNA generation using reverse transcriptase (Invitrogen). Inclusion of amino allyl-dUTP (Sigma) in the RT reaction allowed for subsequent fluorescent labeling of cDNA using monofunctional NHS ester dyes (Amersham, Arlington Heights, IL) as described elsewhere.5 In each experiment, fluorescent cDNA probes were prepared from an experimental RNA sample (coupled to Cy5 NHS-ester) and an appropriate reference RNA sample (coupled to Cy3 NHS-ester). The labeled probes were then hybridized to 10K human cDNA microarrays at 65°C overnight as described elsewhere.5 Fluorescent images of hybridized microarrays were obtained using a GenePix 4000A microarray scanner6 (Axon Instruments, Union City, CA).
Primary analysis was done using the Genepix software package (Axon Instruments). Cy5:Cy3 ratios were determined for the individual genes along with various other quality control parameters (e.g., intensity over local background). Bad spots and defective areas on the array were manually flagged. Spots with small diameters (<50 μm) and spots with low signal strengths (<350 fluorescence intensity units over local background in the more intense channel) were discarded. These data files were then imported into a Microsoft Access database and filtered for presence of data points across arrays and selected for expression levels as stated in the figure legends.
Northern Blot Analysis.
Total RNA (30 μg) was resolved by denaturing formaldehyde-agarose gel and transferred onto Hybond membrane (Amersham) by a capillary transfer set up. Hybridizations were performed as described previously (24). Briefly, prehybridization was carried out for 1 h at 65°C in Church buffer [1% BSA (fraction V), 8% SDS, 0.5 m phosphate buffer (pH 7.0), and 1 mm EDTA (pH8.0)]. Hybridization was performed in Church buffer for 12–16 h at 65°C after adding the denatured probe at 2–3 × 106 cpm/ml. Blots were washed with 2× SSC, 0.1% SDS at room temperature three times for a total period of 30 min, followed by two washes in 0.2× SSC, 0.1% SDS at 65°C for a total period of 60 min. GAPDH signal intensities were used as a loading control.
One μg of total RNA from experimental samples was reverse-transcribed with Superscript reverse transcriptase (Invitrogen) using oligo(dT) and random hexanucleotide primer for first-strand cDNA synthesis. PCRs were performed directly on 1 μl of first-strand cDNA using 500 nmol each of gene-specific primers [HER2, 5′-cgggagatccctgacctgctggaa-3′ (forward primer) and 5′-ctgctggggtaccagatactcctc-3′ (reverse primer); FAS, 5′-caccccgcaggacagccccatct-3′ (forward primer) and 5′-cgccgcccgagcccg-3′ (reverse primer); and GAPDH, 5′cggagtcaacggatttggtcgtat-3′ (forward primer) and 5′-agccttctccatggtggtgaagac-3′ (reverse primer)] in a 100-μl reaction volume comprised of 50 mm KCl, 10 mm Tris-HCl (pH 8.3), 1.5 mm MgCl2, and 200 μm each deoxynucleotide triphosphate. PCR was carried out for initial denaturation at 94°C for 5 min, followed by 25 cycles of denaturation (94°C, 1 min), annealing (55°C for GAPDH, 60°C for HER2, and 65°C for FAS, 1 min), and extension (72°C, 1 min) with 5 units of Taq polymerase (Invitrogen). This was followed by a final extension step of 72°C for 10 min. The products were analyzed on 2% agarose gels stained with ethidium bromide and visualized with UV light.
Whole-cell lysates prepared from H16N2 and SKBr3 cells treated with various combinations of CI-1033 (1 μm, 24 h), Herceptin (200 μg/ml, 24 h), LY294002 (25 μm, 24 h), and C75 (5 μm, 24 h) were separated on SDS-PAGE. The proteins were blotted onto polyvinylidene difluoride membranes (Amersham) and blocked overnight in Tris-buffered saline containing 5% nonfat milk and 0.1% Tween 20 at 4°C. The membranes were then incubated for 2 h with antibodies directed against FAS (PharMingen, San Diego, CA), PI3K 85kDa (Santa Cruz Biotechnology, Santa Cruz, CA), PARP (Cell Signaling Technology, Beverly, MA), β-tubulin (NeoMarkers, Freemont, CA), or actin (Sigma). Membranes were washed with Tris-buffered saline containing 0.1% Tween 20 and incubated with horseradish peroxidase-conjugated secondary antibody (Amersham) for 2 h at room temperature. After washing the membrane with Tris-buffered saline containing 0.1% Tween 20, signals were visualized by the enhanced chemiluminescence system (ECL; Amersham).
H16N2 cells were split and grown overnight in 60-mm tissue culture plates (Costar, Acton, MA) at a density of 500,000 cells/well. Preincubation with 1 μm CI-1033 (Pfizer) or 25 μm LY294002 (Calbiochem) for 24 h was followed by incubation with 2 μCi/ml [1,2-14C]acetic acid (specific activity, 100 mCi/mmol; ICN Radiochemicals, Costa Mesa, CA) for 30-min, 2-h, or 6-h pulse intervals. Fatty acid synthesis was assayed by Folch extraction (25) of the cell lysates followed by scintillation counting in Econo-Safe (Research Products International Corp., Mount Prospect, IL). All determinations were made in triplicate.
Luciferase Reporter Promoter Studies.
H16N2 cells were plated on day 0 in 6-well tissue culture plates at 250,000 cells/well. The next day, transfection was carried out using FuGENE 6 transfection reagent (Roche, Indianapolis, IN) with 1 μg/well FAS promoter-luciferase reporter construct (kindly provided by Dr. J. V. Swinnen; Ref. 26). On day 2, cultures were fed fresh media, and cells were incubated with CI-1033 (1 μm) or LY294002 (25 μm). On day 3, the cells were washed with PBS, lysed in 400 μl of 1× reporter lysis buffer, and stored at −70°C (Promega, Madison, WI). Aliquots of 100 μl of cleared lysate were assayed for luciferase activity using a luciferase assay kit (Promega). The mean of triplicate measurements was used for analysis. Identical transfection and quantitation protocol was carried out with a SV40 promoter-luciferase reporter construct (a kind gift of Dr. J. V. Swinnen) to normalize the transfection efficiency across experimental samples.
Analysis of Cell Death by Fluorescein-labeled Caspase Inhibitors.
Cells were grown in 6-well dishes at a density of 50,000 cells/well, and upon treatments, cells were harvested by trypsinization and centrifuged at 400 × g, and the cell pellet was resuspended in 2 ml of culture medium. Caspase activity was detected in situ, according to the manufacturer’s protocol (Serologicals Corp., Norcross, GA). Briefly, the cells were centrifuged at 400 × g and resuspended in 150 μl of medium. Cells were incubated with 1× fluorescein-labeled pan-caspase inhibitor FAM-VAD-FMK at 37°C for 1 h in a humidified atmosphere of 10% CO2. Cells were then washed twice with wash buffer, finally resuspended in 200 μl of wash buffer, fixed in 1% paraformaldehyde for 15 min at room temperature or overnight at 4°C, and mounted on glass microscope slides for counting and photography under fluorescence microscope using a long pass filter (excitation, 490 nm; emission, ≥520 nm). Fluorescein-labeled inhibitors of caspases (FLICA)-labeled cells emitted green fluorescence, which was recorded in the context of unstained cells visualized by simultaneous bright-field microscopy. Three independent experiments were performed for each data point, with exactly 200 cells counted per sample. The percentage of apoptosis was calculated as the mean ± SE.
Transcriptome Analysis of HER2-mediated Gene Expression.
To identify genes induced by HER2 overexpression, we used a 9984 element (10K) cDNA micorarray consisting of approximately 5996 known, named genes and 2674 expressed sequence tags [Unigene Build 136 (23, 27)]. A human mammary epithelial cell line, H16N2, stably overexpressing the HER2 gene (H16N2-HER2), was compared against the control vector-transfected cell line (H16N2-vector) to identify a set of candidate genes presumably regulated by HER2. The list of genes obtained was narrowed down to a subset in which the expression profile was attenuated by treatment with negative regulators of HER2, including Herceptin or CI-1033. In addition, expression profiles of two HER2-positive and six HER2-negative breast cancer cell lines were also interrogated on microarrays. Using the aforementioned selection criteria, a consolidated list of genes that are up- or down-regulated in H16N2-HER2 and other HER2-positive breast cancer cell lines was determined (Fig. 1). Not surprisingly, the gene with the most increased transcript expression in our analysis of H16N2-HER2 cells was HER2 itself. Other notable genes in this select list include IGFBP2, PI3K regulatory subunit 3, RhoGDIβ, SLUG, biglycan, desmoplakin, and, surprisingly, several genes involved in fatty acid metabolism including FAS. Because independent reports support a role for HER2 and FAS in tumorigenesis, we decided to explore this connection in detail.
Overexpression of HER2 Induces Expression of the FAS Transcript and Protein.
Duplicate elements on our microarrays suggested that the FAS transcript was approximately 4-fold up-regulated by overexpression of HER2 in H16N2 cells (Figs. 1 and 2). Importantly, both Herceptin (200 μg/ml, 24 h) and CI-1033 (1 μm, 24 h) attenuated expression of the FAS transcript in H16N2-HER2 cells. Cell lines that endogenously overexpress HER2 (i.e., SKBr3, SUM190, and SUM225) also exhibited down-regulation of the FAS transcript upon HER2 inhibition (Fig. 2). Of note, we did observe a marginal decrease in the levels of the HER2 transcript upon treatment with CI-1033 (1 μm, 24 h), which was especially prominent in the H16N2-HER2 cell line (Figs. 1 and 2). Similar results with respect to the regulation of the FAS and HER2 transcripts were obtained by RT-PCR analysis of selected samples (Fig. 2, inset).
To further validate our DNA microarray findings, we carried out Northern blot analysis for the FAS transcript in H16N2-HER2 cells and matched vector control cells. Importantly, we observed up-regulation of the FAS transcript when HER2 was ectopically expressed in H16N2 cells (Fig. 3,A). Consistent with HER2 dependence, Herceptin attenuated expression of the FAS transcript. To confirm that this observation is not a peculiarity of stable cell lines, we examined the breast carcinoma cell line SKBr3, which naturally expresses high levels of HER2. As expected, incubation of SKBr3 cells with Herceptin down-regulated the FAS transcript (Fig. 3 B). Interestingly, the SUM-149 cell line, which is HER2 negative, expressed low levels of FAS transcript.
Because mRNA transcript levels do not always correlate with protein levels, we examined FAS protein levels in the same set of cells. Importantly, as with FAS transcript, we observed a similar HER2 dependence of the FAS protein in both H16N2-HER2 cells and SKBr3 cells (Fig. 3, C and D). Both Herceptin (200 μg/ml, 24 h) and CI-1033 (1 μm, 24 h) treatment attenuated levels of the FAS protein.
Enforced Expression of HER2 Increases Fatty Acid Synthesis.
To determine the effect of HER2 expression on fatty acid synthesis, H16N2 cells were metabolically labeled with [U-14C]acetic acid in the presence or absence of HER2 signaling inhibitors [i.e., CI-1033 (1 μm, 24 h), LY294002 (25 μm, 24 h)]. As demonstrated in Fig. 4, HER2 overexpression facilitates increased fatty acid metabolism. Furthermore, upon treatment with inhibitors of HER2 signaling, fatty acid synthesis is markedly decreased, strengthening the observed association between the HER2 and the fatty acid biosynthetic pathway.
Involvement of HER2 in the Regulation of FAS Promoter-mediated Transcription.
To test whether HER2 overexpression leads to activation of the FAS promotor, we transfected a well-characterized 178-bp FAS promoter-reporter construct (28) into the H16N2-HER2 and vector cell lines. Importantly, we observed a 2–3-fold increased FAS reporter activity in the HER2-overexpressing cell line relative to the vector control (Fig. 5). HER2-mediated activation of the FAS promotor was significantly reduced in the presence of HER-signaling inhibitors. Taken together, these observations indicate that HER2-mediated stimulation of FAS expression takes place, at least in part, at the level of transcription and is mediated by cis-acting elements present in the FAS promoter.
Functional Relevance of the HER2-FAS Connection.
We next determined whether inhibition of FAS had any impact on HER2 function. FAS inhibitors, such as cerulenin and C75, have been shown to selectively induce cell death in FAS-overexpressing tumors as opposed to normal cells. We hypothesized that because HER2-overexpressing cells induce the FAS transcript and protein, they should be more sensitive to the toxic effects of FAS inhibitors relative to normal (vector control) cells. Remarkably, H16N2-HER2 cells were markedly more sensitive to apoptosis induced by the FAS inhibitor, C75, than the matched vector control cell line (Fig. 6). H16N2 cells expressing HER2 exhibited increased caspase activity in response to C75 (5 μm, 24 h), whereas vector control cells were relatively resistant (Fig. 6, A and B). To establish that the sensitivity to C75 was indeed due to HER2-mediated effects, we pretreated H16N2-HER2 cells with CI-1033 (1 μm, 24 h) before exposure to C75 (5 μm, 24 h). Consistent with the notion of a HER2-FAS tumorigenic axis, C75-mediated toxicity was blocked upon pretreatment of HER2-overexpressing cells with CI-1033 (1 μm, 24 h). Furthermore, we observed increased cleavage of the death substrate, PARP, in H16N2-HER2 cells treated with C75 (5 μm, 24 h) relative to vector control cells (Fig. 6, C and D). As predicted, the FAS inhibitor-induced cleavage of PARP was attenuated when cells were pretreated with CI-1033 (1 μm, 24 h).
PI3K Activity Potentially Links HER2-mediated Signaling to FAS.
One of the primary arms of the HER2 signaling pathway is presumed to be mediated by activation of PI3K. Interestingly, we noted up-regulation of PI3K regulatory subunit 3 in our microarray analyses (Fig. 1). HER2-associated FAS expression and function were found to be dependent on PI3K activity, as seen in Figs. 3,C, 4, and 5. Furthermore, we observed increased protein levels of PI3K (p85 subunit) induced by HER2, which were inhibitable by Herceptin (200 μg/ml, 24 h) and CI-1033 (1 μm, 24 h; Fig. 7).
Breast cancer is the most common malignancy of American females, with an estimated 203,500 new cases diagnosed and 40,000 expected to die in 2002 (29). Selected genetic and biochemical features of breast cancer have been individually characterized and include BRCA1 and BRCA2 (30, 31, 32), HER2 (3), and FAS (16, 17, 20), among others. Large-scale gene expression profiling studies of breast cancer have also been undertaken (33, 34). Classifications correlating with erbB2 status, estrogen receptor status, and clinical prognosis were reported (33, 35, 36). Hereditary forms of breast cancer have also been studied using microarrays (37). Gene expression profiles that correlated with BRCA1 or BRCA2 mutations in hereditary breast cancer were distinguishable from each other and from sporadic breast cancer. Transcriptome analysis of breast cancer cell lines (38) has been used to study the signaling pathways emanating from IFNs, death receptor ligands, and chemotherapeutics. Whereas large-scale transcriptome analyses of breast cancer have been undertaken, only a handful of the differentially expressed genes have been followed up with functional or biochemical studies. The present study attempts to test a focused hypothesis generated from global transcriptome analysis of HER2 overexpression in breast cancer cells.
Based on our gene expression survey, HER2 mediated the up-regulation of several genes that are known tumor markers or have recently been found to be associated with tumorigenic phenotypes. Examples include IGFBP2, PI3K, RhoGDI, kininogen, and chaperonin. IGFBP2 has been reported by several groups to be up-regulated in prostate cancer (39, 40), high-grade glioblastoma multiforme tumors (41), and diffuse astrocytomas (42). PI3K regulatory subunit is constitutively elevated in H16N2-HER2 cells (43, 44) and is required for the proliferation of H16N2 cells induced by Heregulin-β (45). Small GTPases of the Rho family such as RhoGDIβ, RhoA, Rac, and Cdc42 proteins are involved in the organization of the microfilamental network, cell-cell contact, and malignant transformation, especially in breast tissue. Progression of breast tumors from WHO grade I to grade III was shown to be accompanied by an overall increase in the amount of Rho GTPases, supporting the notion that Rho GTPases are involved in carcinogenesis (46). Kininogen, an endogenous inhibitor of cathepsins, displayed altered expression, processing, and localization in malignant human tumor tissue compared with benign tissue counterparts. An imbalance between cathepsins and kininogen, found to be associated with the metastatic phenotype, presumably facilitates tumor cell invasion and metastasis (47). HSP1 (chaperonin; Mr 60,000; HSP60; GROEL), HSP47, and HSP60 are known markers and pharmacological targets in human tumors (48).
Interestingly, some of the HER2-associated down-regulated genes are also known to be repressed in the context of a tumor phenotype, supporting the view that HER2 mediates tumorigenic activities by both up- and down-regulating key genes. On top of our list of genes in this category is SLUG, a zinc finger protein, which is recognized as an antiapoptotic transcription repressor that promotes cell survival and eventual malignant transformation of mammalian pro-B cells otherwise slated for apoptotic death (49). SLUG transcript was reportedly down-regulated in human breast cancer cell lines compared with normal mammary epithelial cell lines (50). Another molecule in this category, biglycan, is a small leucine-rich proteoglycan, which interacts with transforming growth factor-β and other extracellular matrix molecules to influence the attachment and migration of cells (51). The expression of biglycan was found to be strongly down-regulated in basal cell carcinomas examined by immunohistochemistry (52).
DNA microarray analysis of HER2-induced gene expression led us to investigate FAS. Fatty acid metabolism is an important part of normal physiology. Fatty acid synthesis and degradation (oxidation) are controlled by distinct pathways localized in the cytosol and mitochondria, respectively (53). Fatty acids or their derivatives comprise biological membranes, serve as hormones and intracellular messengers, and act as fuel molecules to provide energy. In well-nourished individuals, fatty acid synthesis pathways are down-regulated due to high levels of dietary fat and are thus underused by normal cells (53). Importantly, however, high levels of FAS protein have been demonstrated in a number of solid tumors (16, 17, 18, 19, 20). This difference between normal and tumor cells with respect to FAS expression has been pursued as a therapeutic target. FAS inhibitors (e.g., C75 and cerulenin) have been shown to induce apoptosis or cell cycle arrest in a number of cancer cell lines (19, 22, 54). The mechanism of how FAS inhibition leads to preferential killing of tumor cells is unclear. However, it has been hypothesized that accumulation of malonyl CoA, the primary substrate of FAS, is the mediator of cell cytotoxicity (55, 56). Increased cellular levels of malonyl CoA lead to inhibition of fatty acid oxidation, a likely contributing factor to cell death induced by FAS inhibition. Normal cells, relative to cancer cells, use primarily dietary fatty acids rather then synthesizing fatty acids de novo, possibly accounting for the differential with malignant cells. Taken together, these findings suggest that HER2-overexpressing breast cancer cells (e.g., H16N2-HER2) exhibit high levels of FAS that make them more susceptible to apoptosis by FAS inhibition.
We contend that our approach to studying HER2-mediated gene expression changes is free from the obscuring effects of the in situ tumor milieu and associated coamplified genes. This could allow for the characterization of novel pathways modulated by HER2. Interestingly, overexpression of FAS has been noted as an early event in the development of prostate (57) and breast cancer (58). This raises the possibility of a link between growth factor signaling and FAS pathways. Growth factor-associated FAS overexpression has been recently reported by Swinnen et al. (59), who observed stimulation of FAS expression by epidermal growth factor in the prostate cancer cell line LNCaP through the SREBP-mediated pathway. SREBPs, a family of transcription factors that activate genes involved in the synthesis of cholesterol and fatty acids (60, 61), are known to be regulated by extracellular signal-regulated kinase-mitogen-activated protein kinase and PI3K pathway (62, 63). Recently, a human mammary epithelial cell line, MCF10A, transformed with the oncogene h-RAS has been reported to demonstrate induction of FAS expression, mediated by activation of PI3K and mitogen-activated protein kinase pathways, which in turn increased SREBP-1 levels (64). In light of the evidence that HER2 overexpression also induces PI3K activity in H16N2-HER2 cells (15, 44, 45), it is understandable that the FAS overexpression noted in these cells might result from a PI3K-mediated mechanism. Our microarray results, FAS activity assay, FAS promoter-reporter assay, and PI3K immunoblot analysis support the hypothesis that HER2-mediated PI3K activity facilitates activation of the FAS pathway. The observation that Herceptin and CI-1033 markedly repress FAS expression suggests that this may contribute to the antitumorigenic effects of these drugs. Moreover, measurement of FAS activity post-HER2-targeted therapy may have predictive potential clinically.
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.
Supported in part by a grant from the Breast Cancer Research Foundation (to M. E. L. and A. M. C.) and a Bioinformatics Pilot Grant Foundation (to A. M. C.).
The abbreviations used are: EGFR, epidermal growth factor receptor; FAS, fatty acid synthase; PI3K, phosphatidylinositol 3′-kinase; Pen-Strep, penicillin-streptomycin; RT, reverse transcription; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PARP, poly(ADP-ribose) polymerase; IGFBP, insulin-like growth factor-binding protein; RhoGDI, Rho GDP dissociation inhibitor; HSP, heat shock protein; SREBP, sterol regulatory element-binding protein; 10K, 10,000; NHS, N-hydroxysuccinimidyl.
We thank Max Wicha for providing Herceptin antibody and helpful discussions, Johannes Swinnen (Leuven, Belgium) for the FAS promoter-reporter constructs, and Terrence Barrette for bioinformatics support. We also appreciate the help extended by Mark Demming at Pathology Imaging for photographic requirements.