Old Players with a Newly Defined Function: Fra-1 and c-Fos Support Growth of Human Malignant Breast Tumors by Activating Membrane Biogenesis at the Cytoplasm

A shared characteristic of tumor cells is their exacerbated growth. Consequently, tumor cells demand high rates of phospholipid synthesis required for membrane biogenesis to support their growth. c-Fos, in addition to its AP-1 transcription factor activity, is the only protein known up to date that is capable of activating lipid synthesis in normal and brain tumor tissue. For this latter activity, c-Fos associates to the endoplasmic reticulum (ER) through its N-terminal domain and activates phospholipid synthesis, an event that requires it Basic Domain (BD) (aa 139–159). Fra-1, another member of the FOS family of proteins, is over-expressed in human breast cancer cells and its BD is highly homologous to that of c-Fos with two conservative substitutions in its basic amino acids. Consequently, herein we examined if Fra-1 and/or c-Fos participate in growth of breast cancer cells by activating phospholipid synthesis as found previously for c-Fos in brain tumors. We found both Fra-1 and c-Fos over-expressed in >95% of human ductal breast carcinoma biopsies examined contrasting with the very low or undetectable levels in normal tissue. Furthermore, both proteins associate to the ER and activate phospholipid synthesis in cultured MCF7 and MDA-MB231 breast cancer cells and in human breast cancer samples. Stripping tumor membranes of Fra-1 and c-Fos prior to assaying their lipid synthesis capacity in vitro results in non-activated lipid synthesis levels that are restored to their initial activated state by addition of Fra-1 and/or c-Fos to the assays. In MDA-MB231 cells primed to proliferate, blocking Fra-1 and c-Fos with neutralizing antibodies blocks lipid-synthesis activation and cells do not proliferate. Taken together, these results disclose the cytoplasmic activity of Fra-1 and c-Fos as potential targets for controlling growth of breast carcinomas by decreasing the rate of membrane biogenesis required for growth.


Introduction
The fos and jun oncogenes are members of the family of Immediate Early Genes (IEGs) AP-1 transcription factors that are rapidly and transiently expressed in different cell types in response to a myriad of stimuli, such as growth factors, neurotransmitters, etc. [1][2][3]. The Fos proteins (c-Fos, Fra-1, Fra-2 and Fos-B) and the Jun proteins (c-Jun, JunB and JunD) share homologous regions containing a basic DNA-binding domain (BD) and a leucine zipper dimerization motif. Jun proteins form homo-and heterodimers whereas Fos proteins only form heterodimers with other IEGs, mostly Jun proteins, thus originating the variety of AP-1 transcription factors that regulate target genes expression in response to growth factors [1,4].
Although c-Fos was described as an AP-1 transcription factor more than 20 years ago, the complex consequences of its induction on cell physiology have still not been fully elucidated. It has been proposed that, upon mitogenic stimuli, c-Fos triggers and controls cell growth, differentiation and apoptosis by regulating key genes [5]. However, we have shown that in addition to its nuclear AP-1 activity, c-Fos associates to the endoplasmic reticulum (ER) and activates phospholipid synthesis as an additional response to mitogenic stimuli [6]. This cytoplasmic activity of c-Fos has been observed in vivo in light-stimulated retina ganglion and photoreceptor cells [6][7][8][9], in culture in NIH3T3 fibroblasts induced to reenter growth [10], in PC12 cells induced to differentiate [11,12], in actively growing and proliferating T98G glioblastoma multiforme-derived cells [13,14], and in human and mouse tumors from the Peripheral and Central Nervous Systems [15,16]. Although the mechanism by which c-Fos associates to the ER and activates phospholipid biosynthesis is currently not fully elucidated, it is known that c-Fos physically associates with specific, key enzymes of the pathway of phospholipid synthesis in the ER [17]. c-Fos/ER association is regulated by the phosphorylation state of c-Fostyrosine residues #10 and #30 whereas its activation capacity depends on its BD (20 amino acids spanning from 139-159) [13,14,17].
The expression of Fra-1 (the Fos-related antigen-1), another member of the Fos family of proteins, is encoded by the fos-like-1 gene (fosl1) is also induced by mitogenic stimuli [18,19]; its role in cell growth is also not well understood. Fra-1-AP-1 transcriptional activity is regulated both transcriptionally [20][21][22] and posttranslationally [22]. In exponentially growing cells, Fra-1 expression is elevated and it is hyper-phosphorylated on C-terminus Ser and Thr residues [23]. These phosphorylation sites seem to play a role both in stabilizing Fra-1 and on its AP-1 activity [22][23][24][25][26]. In spite that the leucine zipper domain of Fra-1 is homologous to that of other Fos family members, transcriptional activation studies suggest that Fra-1 is a negative regulator of AP-1 activity [25,26].
Tumor cells demand high rates of phospholipid synthesis to support membrane biogenesis for their exacerbated growth. In mice knock-out for c-Fos and knock-in for Fra-1 under the c-Fos promoter, Fra-1 rescues and substitutes for growth-dependent functions of c-Fos but not for its AP-1 activity [39], suggesting that Fra-1 and c-Fos may substitute each other in some of their biological functions. Herein we examined Fra-1 and c-Fos expression in growing breast cancer cells and their capacity to activate phospholipid synthesis. It is shown that both Fra-1 and c-Fos are over-expressed in .95% of ductal breast carcinoma tissue samples contrasting with very low or undetectable levels in normal tissue. Both Fra-1 and c-Fos were found associated to the ER and activating phospholipid synthesis in breast cancer cells in culture and in human breast cancer tissue samples, thus disclosing this activity of these proteins as a potential target for controlling growth of breast carcinomas.

Ethics Statement
Freshly excised human breast tumor and matched benign specimens were obtained from female patients with written consent following the procedures indicated by the Research Ethics Board of the Hospital Nacional de Clínicas, Universidad Nacional de Córdoba, Argentina, and with the Helsinki Declaration of 1975, as revised in 1983. Samples were processed anonymously. Patient ages ranged from 38-82 years old. All the procedures used for this study were approved by the Ethics Committee of CIQUIBIC-CONICET.
Cells were transfected 100 mg/ml Fra-1 or c-Fos antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), using the BioPORTER QuickEasy Protein Delivery Kit (Sigma-Aldrich, Saint Louis, MO, USA). Transfection efficiency .75% was controlled using FITC-IgG antibody. AP-1 nuclear localization signal peptide (NLSP) (American Peptide Company, Sunnyvale, CA, USA) that blocks AP-1 nuclear import [11] was added to cultures at 75 mM final concentration in 5 ml of medium. Control cells received the same volume of delivery medium.

Sub-cellular Fractionation
This was as previously described [13]. Briefly, total homogenates (TH) (rinsed cultured cells or tissue samples) prepared in PBS plus protease inhibitor cocktail (Sigma-Aldrich), were processed with an Ultra-turrax homogenizer and centrifuged for 1 h at 100,0006g to separate the microsomal (MF) and supernatant (SF) fractions. For stripping of TH, prior to centrifugation, TH's were made 1 M with KCl, left to stand for 10 min and centrifuged at 100,0006g for 1 h to separate into MF and SF. MF was re-suspended in the initial volume of PBS plus protease inhibitor cocktail.

Cell Proliferation Assays
CyQUANTH Cell Proliferation Assay Kit (Invitrogen) was used.

Phospholipid Synthesis Determination
In vitro phospholipid labeling determinations in tumors, cells and sub-cellular fractions was as described [11] using 100ug of tumor/ cell homogenate protein. When stated, recombinant His-tagged Fra-1 or c-Fos (1.5 ng/mg or 1.0 ng/mg of initial TH protein, respectively) were added to assays re-suspended in 300 mM imidazole/8 M urea; control incubates received the same volume of vehicle.

Immunohistochemistry
Breast Tumor Tissue Array (BioChain Institute Inc., Hayward, CA, USA) specimens were de-waxed and re-hydrated as described [15] and incubated overnight at 4uC with primary antibodies diluted in blocking buffer as follows: rabbit anti-Fra-1, rabbit antic-Fos, mouse anti-PCNA and goat anti-calnexin (all 1/300) antibodies. Anti-goat Alexa 546, anti-rabbit Alexa 488 and antimouse Alexa 688 secondary antibodies were applied (1/500) and slides mounted with FluorSave (Calbiochem, San Diego, CA, USA). Visualization and image analysis was as described for cells.

Small Interfering RNA (siRNA) Transfection
To repress human Fra-1 or/and c-Fos, predesigned doublestranded siRNAs (ON-TARGET plus SMART pool; Dharmacon Inc., Lafayette, CO, USA) were used. A siCONTROL nontargeting siRNA pool (Dharmacon Inc.) was used as a negative control. MDA-MB231 cells were grown under standard culture

Statistical Analysis
Student's two tailed t test of Infostat software (Universidad Nacional de Córdoba, Argentina) was used.

Fra-1 and c-Fos Expression is Up-regulated in the Cytoplasm of Growing Beast Cancer Cells and Both Proteins Co-localize with ER Markers
Fra-1 and c-Fos expression was analyzed by immunofluorescence and immunoblot in MDA-MB231 and MCF7 human breast adenocarcinoma cell lines. MDA-MB231 are highly invasive, Estrogen Receptor a negative, E-cadherin negative and Vimentin positive cells whereas MCF7 are weakly invasive, Estrogen Receptor a positive, E-cadherin positive and Vimentin negative cells [28]. In quiescent MDA-MB231 ( Figure 1A Interestingly, in growing cells, c-Fos and Fra-1 were present both in the nucleus and in the cytoplasm although both were more abundant in the cytoplasm. When cells were immuno-stained for the ER marker calnexin, the cytoplasmic fraction of both proteins clearly co-localized with the ER marker ( Figure 1 A-B, middle columns and merged images to the right). This sub-cellular localization was confirmed by WB after subjecting total homogenate (TH) from quiescent (-FBS) and growing cells (+FBS) to subcellular fractionation by centrifugation at 100,0006g for 1 h to isolate the microsomal (MF) and the supernatant (SF) fractions. Both MDA-MB231 ( Figure 1C, left panel) and MCF7 cells ( Figure 1C, right panel) clearly showed higher Fra-1 and c-Fos immuno-reactivity in total homogenate (TH) from growing cells (+FBS) as compared to quiescent (-FBS) cells and both proteins were recovered in the MF. When the MF was stripped of its associated proteins by treatment with 1M KCl, Fra-1 and c-Fos were no longer recovered in MF but rather were recovered in the SF ( Figure 1C, last two lanes).  We previously showed that the BD of c-Fos (amino acids 139-159) is required to activate phospholipid synthesis [11,17]. Considering that the BD of Fra-1 and c-Fos are conserved showing only two conservative substitutions in their BDs (see schematic representation in Figure 2) and that both proteins colocalize with the ER marker calnexin, we next studied if Fra-1, as previously described for c-Fos in other cells [6][7][8][9][10][11][12][13][14][15][16][17], activates phospholipid synthesis in actively growing cells. 32 P-labeling of phospholipids was determined in vitro using TH prepared from quiescent and growing MDA-MB231 and MCF7 cells. TH and the MF of growing MDA-MB231 cells (+FBS, second column) showed a significant increase in 32 P-phospholipid labeling as compared to quiescent cells (-FBS, first column)( Figure 3A). Furthermore, a comparable activation of 32 P-phospholipid label- ing was observed when recombinant Fra-1 (third column) or c-Fos (fourth column) (1.5 or 1.0 ng/mg of TH protein, respectively) were added to TH of quiescent (-FBS) cells. Similarly, when the TH from growing cells (+FBS) was stripped with 1M KCl (thus containing negligible levels of c-Fos and Fra-1 as shown in Figure 1), 32 P-phospholipid labeling decreased markedly mirroring the levels in TH from quiescent cells. Addition of recombinant Fra-1 or c-Fos to these stripped membranes restored 32 Pphospholipid synthesis to the initial activated levels (last two columns). Similar results were observed with MCF7 cells ( Figure 3B). Altogether, these results show that Fra-1 is capable of activating phospholipid synthesis to levels comparable to those obtained with c-Fos.

Blocking Fra-1 and c-Fos Abrogates Phospholipid Synthesis Activation and Cell Proliferation
The participation of nuclear and cytoplasmic c-Fos in driving tumor cell proliferation and growth was examined in growing MDA-MB231 cells cultured in the presence or the absence of an AP-1 Nuclear Localization Sequence Peptide (NLSP) that blocks the nuclear import of c-Fos and Fra-1 as AP-1 dimers [42]. Figure 4A shows that after culturing cells +FBS for 36 h, cell number roughly doubled. However, as described previously [14], adding NLSP to the cultures at 0 or 6 h after priming cells to proliferate (+FBS) blocks cell proliferation, whereas at later times (16 h) it is no longer effective (column 5), indicating that nuclear AP-1-c-Fos/Fra-1 is required to trigger proliferation at early stages. By contrast, cytoplasmic c-Fos and Fra-1 are required at all time points because blocking their activity with specific neutralizing antibodies delivered at any time after feeding FBS,+or -NLSP, specifically blocks proliferation ( Figure 4A, columns 6,7,8). These results highlight the need of AP-1 to trigger the genomic events for proliferation and growth and that of cytoplasmic Fra-1 and c-Fos to activate lipid synthesis required for membrane biogenesis to sustain growth.
To confirm the importance of Fra-1/c-Fos-dependent phospholipid synthesis activation during proliferation and growth, this was assessed in FBS-stimulated MDA-MB231 cells cultured +NLSP or -NLSP and transfected with c-Fos and Fra-1 neutralizing antibodies at different times. Accordingly, higher rates of 32 P-phospholipid labeling were detected in growing cells (+FBS) when compared to quiescent cells (-FBS) ( Figure 4B). Feeding cells with NLSP at 0 h or 6 h after FBS priming, abrogated 32 P-phospholipid labeling activation, whereas primed cells-activated levels were observed when NLSP was added at 16 h. These results further support that nuclear AP-1 is needed to trigger the genomic events for proliferation and growth whereas cytoplasmic Fra-1 and c-Fos are required to sustain growth.
Taking into consideration that MDA-MB231 and MCF7 cells contain both c-Fos and Fra-1, it was deemed of interest to determine if both proteins are required to support cell proliferation or if one is enough and can substitute for the other. For this, quiescent MDA-MB231 cells were transfected with siRNA against c-Fos or against Fra-1 or both. At 96 h of transfection, cells were induced to proliferate by the addition of FBS to the culture medium and proliferation determined 36 h later. Figure 5A shows that proliferation of mock-transfected cells was essentially the same as that of non-transfected cells whereas transfection of siRNA against either c-Fos or Fra-1 partially blocked proliferation. However, blocking the expression of both proteins completely blocks proliferation indicating that either c-Fos or Fra-1 can support proliferation provided that protein expression levels are high enough. Fig. 5B shows that treatment of cells with siRNA effectively blocked c-Fos and Fra-1 expression as determined by WB assays.

Malignant Human Breast Tumors Show Abundant Fra-1 and c-Fos Expression Associated to the ER and Activated Phospholipid Synthesis
Considering the results described above, we analyzed Fra-1 and c-Fos expression and phospholipid synthesis activation capacity in malignant human breast tumor samples. Immunohistochemistry assays revealed a marked over-expression of Fra-1 and c-Fos in 100% of 210 tissue samples from different human breast tumors [invasive ductal carcinoma (n = 200), Medullary carcinoma (n = 2), Phyliodes sarcoma (n = 2), Mucinous carcinoma (n = 2), lobular carcinoma in situ (n = 2) and squamous cell carcinoma (n = 2)], contrasting with the low or undetectable levels of Fra-1 and c-Fos detected in the non-pathological samples (n = 37). Noteworthy, both proteins were over-expressed in 95% of the tumor samples analyzed, which also showed significant PCNA staining which evidenced their active proliferative status. When assessing the subcellular localization of these proteins, Fra-1 was found mainly in the cytoplasm: 69% of tumor samples showed only cytoplasmic Fra-1 whereas the remaining tumor samples contained both nuclear and cytoplasmic-localized Fra-1 (31%). c-Fos was also preferentially present in the cytoplasm of the tumor samples: 100% showed cytoplasmic c-Fos whereas, of these, 63% also contained nuclear c-Fos. Remarkably, in all cases cytoplasmic Fra-1 and c-Fos co-localized with the ER marker calnexin ( Figure 6A, B). Figure 7A shows immunoblot experiments of 5 normal (N1 to N5) and 8 tumor samples (T1 to T8) randomly selected from the breast tissue collection. Fra-1 and c-Fos over-expression is clearly observed in actively proliferating tumor samples (as evidenced by the high levels of PCNA immuno-reactivity) as compared to their non-pathological counterparts. Tubulin staining was used as a loading control ( Figure 7A). Subjecting tumor samples 5 and 6 to subcellular fractionation and WB confirmed that Fra-1 and c-Fos expression is associated to MF and that this association is reversed by 1M KCl-treatment ( Figure 7B).

Fra-1 and c-Fos Activate Phospholipid Biosynthesis in
Human Breast Tumors 32 P-Phospholipid labeling was assayed in vitro using fresh human breast tumor and non-pathological samples. Figure 8A shows high 32 P-phospholipid labeling in TH from all the tumor samples as compared to the mean value from the non-pathological samples ( Figure 8A). This elevated phospholipid synthesis correlates with the elevated content of c-Fos/Fra-1 and the highly proliferative status of tumor cells ( Figure 7A). Subcellular fractionation of tumors 5 and 6 confirmed that the activated-32 P-phospholipid labeling capacity was preponderantly present in MF and that this activity was abrogated when membranes were stripped with 1M KCl ( Figure 7B-C, fourth columns). Both, Fra-1 and c-Fos activate phospholipid synthesis as the addition of either Fra-1 or c-Fos to these stripped membranes restores phospholipid synthesis to the initial rates observed in the untreated tumor MF ( Figure 8B-C). These results confirm data obtained with MDA-MB231 and MCF7 cells and further support the role of Fra-1 and c-Fos as activators of phospholipid synthesis to support membrane biogenesis for tumor cell growth.

Discussion
Fra-1 is a Fos family member that is over-expressed in diverse types of human cancers including brain [37,43], lung [44], oesophagus [36], thyroid [38], colorectal, skin, ovary, etc. (reviewed in [24,27]). Of the four Fos family members, Fra-1 is likely to be the most frequently expressed in different forms of human cancer [24,27]. Several studies have shown that Fra-1 over-expression in breast tumor cell lines is associated with an aggressive behavior of these cells [28,45]. Although only a few reports have addressed Fra-1 over-expression in clinical tumors, this was found in all malignant breast tissues [31,33,34]. Furthermore, in all these studies, nuclear and cytoplasmic Fra-1 was found although in many cases only nuclear-containing Fra-1 cells were considered as Fra-1-positive cells [34]. It was reasoned that only nuclear-localized Fra-1 is important due to its transcription factor function, disregarding the possibility of other functions for this protein. By contrast, Song et al. [31] found that 90.2% of all breast carcinomas analyzed (n = 445) showed nuclear and cytoplasmic Fra-1 over-expression whereas only 9.2% of the samples (n = 41) showed only nuclear Fra-1 over-expression; of the 20 benign tumors examined, only 3 showed weak cytoplasmic immuno-reactivity that co-existed with the nuclear reactivity that was present in all samples. However, a clear shift from nuclear to a simultaneous nuclear/cytoplasmic localization of Fra-1 was noticed in ,90% of breast carcinomas examined which led them to conclude that a non-transcriptional function of Fra-1 remained to be demonstrated [31].
We previously showed that over-expressed, cytoplasmic c-Fos activates phospholipid synthesis in brain tumor cell lines and in human brain tumors thus supporting the elevated rates of membrane biogenesis required for the exacerbated growth of these tumors [15]. To activate phospholipid synthesis, the basic domain of c-Fos (aa 139-160) is required [16,17]. Furthermore, point mutations of this conserved BD in basic amino acids #139 or #144 (K139N, R144N) has no effect on BD's lipid synthesis activating capacity whereas that of basic amino acid #146 (R146N) completely abolishes this activity further supporting the importance of BD for the activation of specific enzymes in the pathway of synthesis of lipids [17]. Consequently, considering the results described above regarding Fra-1 cytoplasmic over-expression and the fact that Fra-1 and c-Fos share their basic domains, we herein addressed the possibility that Fra-1 may also activate phospholipid synthesis in growing human breast tumors. Using breast tumor cell lines and clinical tumor samples, it is demonstrated that this is indeed the case. Briefly, we found that growing MDA-MB231 and MCF7 cells exhibit high rates of phospholipid biosynthesis as compared with quiescent cells. Furthermore, stripping membranes prepared from growing cells of its non-integral proteins by 1M KCl-treatment resulted in quiescent cell phospholipid synthesis rates, which was restored to its initial activated rates by the addition of recombinant Fra-1 or c-Fos to the assay medium. These results were mirrored when analyzing human breast tumor samples: phospholipid synthesis was also significantly higher in tumors as compared to normal tissue and was significantly reduced when subjecting tumor samples to 1M KCl-treatment. Addition of Fra-1 or c-Fos to stripped samples restored phospholipid synthesis activation to initial rates. Finally, the previously shown dual function (nuclear and cytoplasmic) of c-Fos (14,11) was also demonstrated for Fra-1 in experiments in which cells were fed NLSP at different times after inducing them to re-enter growth. AP-1-Fra-1 and/or AP-1c-Fos were required to trigger proliferation only at early stages of cell proliferation. In contrast, at later stages of cell proliferation (16 h after priming cells), cytoplasmic Fra-1 and/or c-Fos (that activate phospholipid biosynthesis) were necessary to sustain proliferation. Blocking Fra-1 and c-Fos with intracellular neutralizing antibodies abrogated both phospholipid synthesis activation and cell proliferation.
The cytoplasmic regulatory role of c-Fos or Fra-1 is not the only AP-1 independent activity of an IEG. c-Jun was also shown to protect cells from apoptosis independently of its AP-1 activity [46]. In addition, cytoplasmic p53 is capable of triggering apoptosis in the absence of transcription [47] whereas translocation of nuclear p21Cip1/WAF1 to the cytoplasm promotes resistance to apoptotic stimuli [48]. Consequently in light of these results and of ours shown herein, a new concept is emerging in which the biological activity of these proteins results from the combination of their nuclear and cytoplasmic activities [49]. These data may highlight the significance of therapy based on the blockade of Fra-1 and/or c-Fos functions in breast cancer. Therapies directed towards blocking Fra-1 and c-Fos expression are promising moreover if it is considered that these proteins are normally down-regulated in normal breast cell growth but become up-regulated during, and are causally related to, breast cancer progression.