International Journal of Molecular Sciences Article GRP94 Is Involved in the Lipid Phenotype of Brain Metastatic Cells Naiara Santana-Codina 1,2, *,† , Anna Marcé-Grau 1 , Laia Muixí 1 , Claudia Nieva 1,3 , Mónica Marro 3 , David Sebastián 4,5,6 , Juan Pablo Muñoz 4,5,6 , Antonio Zorzano 4,5,6 and Angels Sierra 7, * 1 Biological Clues of the Invasive and Metastatic Phenotype Group, Bellvitge Biomedical Research Institute (IDIBELL), L’Hospitalet de Llobregat, E-08908 Barcelona, Spain 2 Universitat Autònoma de Barcelona (UAB), Campus Bellaterra, Cerdanyola del Vallés, E-08193 Barcelona, Spain 3 ICFO-Institut de Ciències Fotòniques, The Barcelona Institute of Science and Technology, Carl Friedrich Gauss 3, 08036 Barcelona, Spain 4 Institute for Research in Biomedicine (IRB Barcelona), The Barcelona Institute of Science and Technology, 08036 Barcelona, Spain 5 Departament de Bioquímica i Biomedicina Molecular, Facultat de Biologia, Universitat de Barcelona, 08028 Barcelona, Spain 6 CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), Instituto de Salud Carlos III, 08028 Barcelona, Spain 7 Laboratory of Molecular and Translational Oncology, Centre de Recerca Biomèdica CELLEX-CRBC-Institut d’Investigacions Biomèdiques August Pi i Sunyer-IDIBAPS, E-08036 Barcelona, Spain * Correspondence: naiarasantanacodina@gmail.com (N.S.-C.); asierrajim@gmail.com (A.S.); Tel.: +617-582-8741 (N.S.-C.); +34-93-227-55-00 (ext. 4828) (A.S.) † Current address: Department of Radiation Oncology, Dana-Farber Cancer Institute, Boston, MA 02115, USA.  Received: 23 July 2019; Accepted: 6 August 2019; Published: 9 August 2019  Abstract: Metabolic adaptation may happen in response to the pressure exerted by the microenvironment and is a key step in survival of metastatic cells. Brain metastasis occurs as a consequence of the systemic dissemination of tumor cells, a fact that correlates with poor prognosis and high morbidity due to the difficulty in identifying biomarkers that allow a more targeted therapy. Previously, we performed transcriptomic analysis of human breast cancer patient samples and evaluated the differential expression of genes in brain metastasis (BrM) compared to lung, bone and liver metastasis. Our network approach identified upregulation of glucose-regulated protein 94 (GRP94) as well as proteins related to synthesis of fatty acids (FA) in BrM. Here we report that BrM cells show an increase in FA content and decreased saturation with regard to parental cells measured by Raman spectroscopy that differentiate BrM from other metastases. Moreover, BrM cells exerted a high ability to oxidize FA and compensate hypoglycemic stress due to an overexpression of proteins involved in FA synthesis and degradation (SREBP-1, LXRα, ACOT7). GRP94 ablation restored glucose dependence, down-regulated ACOT7 and SREBP-1 and decreased tumorigenicity in vivo. In conclusion, GRP94 is required for the metabolic stress survival of BrM cells, and it might act as a modulator of lipid metabolism to favor BrM progression. Keywords: GRP94; brain metastasis; endoplasmic reticulum stress; fatty acids 1. Introduction Glycoprotein 96 (gp96, GRP94, Erp99, endoplasmin) is an endoplasmic reticulum (ER) lumen resident protein from the HSP90 family [1,2] that belongs to a set of ER chaperones originally Int. J. Mol. Sci. 2019, 20, 3883; doi:10.3390/ijms20163883 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2019, 20, 3883 2 of 17 discovered as proteins induced by glucose deprivation [3]. GRP94 often functions as a dimer providing a platform for the assembly, folding or oligomerization of loaded protein cargo [4], and its function is ATP-dependent [5]. GRP94 can be modulated by interferons [6] and by a multitude of stress situations that perturb ER function. One of its most relevant functions is stress resistance, promoting survival and protecting cells against ER stress-induced death [7] in conditions of glucose starvation and/or oxidative stress. GRP94-mediated endoplasmic reticulum stress-resistance phenotype (ERSRP) integrates a set of adaptation pathways that regulates transcription and translation of genes in an attempt to restore functions and re-establish homeostasis [8]. Chaperone overexpression is a general phenomenon that happens in tumor cells as a consequence of the increased proliferation rate and the harmful environment [9]. GRP94 has been associated with cellular transformation in a variety of cancer cell lines, rodent tumor models and human cancer biopsies correlating with increased tumorigenicity [10]. In fact, GRP94 is upregulated in CD44hi /CD24lo breast cancer stem cells, which are usually considered tumor-initiating cells with enhanced invasive properties and radiation resistance [11]. Thus, GRP94 might respond to environmental stress mediating the induction of specific ER-proteins, such as heat shock proteins and other glucose-regulated proteins, and allowing tumor-microenvironment cross-talk and metastatic growth [12,13]. Moreover, GRP94 is coupled with secretory pathways and transport of membrane proteins to the cell surface, such as integrins and toll-like receptors, contributing to regulation of immunity and metastasis progression [14,15]. The ER is also the main site for sterol and phospholipid synthesis, since many enzymes of the pathway reside in this organelle [16]. The ER stress response and specifically the unfolded protein response (UPR) pathways are closely associated with lipid metabolism through inositol-requiring enzyme 1α (IRE1α)/XBP1 or activating transcription factor 6, which can enhance phospholipid biosynthesis under ER stress conditions [17]. Therefore, ER stress has been considered an important regulator of lipid metabolism in organs like the liver [13], modulating the expression of key enzymes in the pathway [16]. Furthermore, cancer cells require a constant supply of lipids for membrane synthesis and energy production [18], suggesting ER stress and lipid metabolism might be two processes tightly connected to promote tumor growth. For example, stress derived of poor perfusion and reduced access to exogenous nutrients within the tumor microenvironment can modulate de novo fatty acid synthesis enzymes in cancer cells. Recently, the ER chaperone GRP78 has been described to regulate lipid metabolic enzymes in pancreatic cancer, leading to lipid accumulation after GRP78 depletion [19]. This link between ER stress and metabolic rewiring might be even more important in the pathogenesis of brain metastasis (BrM) due to the different energetic substrates and nutrients in the central nervous system (CNS) environment. Brain metastasis occurs as a consequence of the systemic dissemination of tumor cells, a fact that correlates with poor prognosis and high morbidity in lung (36–64%), breast (15–25%), and melanoma (5–20%) [20]. We described that GRP94 overexpression might be a hinge orchestrating ERSRP responses in BrM through activation of PERK, ATF6 and IRE1 leading to stress resistance in the hypoglycemic environment of the CNS [21,22]. Here, we investigated the role of GRP94 as a promotor of cell survival and adaptive mechanisms, ameliorating ER stress of brain metastatic cells. 2. Results 2.1. Brain Metastatic Cells Increase Total Fatty Acid (TFA) and Decrease Total Unsaturated Fatty Acid (TUFA) Content We previously defined the ERSRP of BrM cells based on a set of brain metastasis biopsies, which are freely available from the Gene Expression Omnibus (GEO) repository (GSE11078) [23]. This organ-specific metastasis signature contained 1193 genes, which clearly distinguished brain from other different metastases [24]. Protein folding enzymes and chaperones appeared as a hub connecting different functions that mediated ERSRP, GRP94 being one of the central proteins [22]. Among the organ-specific deregulated functions in brain metastases, we found three upregulated proteins (Table 1) with a role in fatty-acid (FA) synthesis (PCYT2, SLC25A1) and hydrolysis (ACOT7). Int. J. Mol. Sci. 2019, 20, 3883 3 of 17 These results suggested that BrM presented an organ-specific enrichment in genes modulating FA Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW metabolism. Previous work from our lab identified a role for GRP94 and the ERSRP in 3mediating of 17 brain metastasis (ACOT7). [22]. These In fact, suggested results GRP94 was thatoverexpressed BrM presented in an aorgan-specific brain metastatic variant enrichment in from genesbreast cancer (435-Br1) compared to the parental cell line (435-P), while there were no modulating FA metabolism. Previous work from our lab identified a role for GRP94 and the ERSRPdifferences in GRP78 expression, another chaperone in mediating involved brain metastasis in fact, [22]. In the UPR GRP94response (Supplementary was overexpressed Figure in a brain S1) and metastatic previously variant linked tofrom lipidbreast cancer (435-Br1) metabolism [19,25]. compared to the parental cell line (435-P), while there were no differences in GRP78 expression, another chaperone involved in the UPR response (Supplementary Table 1. Genes Figure S1) anddifferentially expressed previously linked to lipidinmetabolism brain metastasis [19,25]. of breast cancer patients compared to other metastases. Table 1. Genes differentially expressed in brain metastasis of breast cancer patients compared to Gene other metastases. Protein Location Function ACOT7 Acyl coenzyme A thioester hydrolase Cytoplasmic Acyl-CoA hydrolysis Gene Protein Location Function PCYT2 Ethanolamine-phosphate cytidylyltransferase Cytoplasmic Phospholipid biosynthesis ACOT7 Acyl coenzyme A thioester hydrolase Cytoplasmic Acyl-CoA hydrolysis SLC25A1 Tricarboxylate transport protein Mitochondrial Citrate/malate exchange, FA synthesis Ethanolamine-phosphate PCYT2 Cytoplasmic Phospholipid biosynthesis cytidylyltransferase Given the role of ER stress in lipid metabolism, we hypothesized that FAexchange, Citrate/malate metabolism FA might be SLC25A1 Tricarboxylate transport protein Mitochondrial differentially enriched in 435-Br1 cells as a consequence of their ERSRP synthesisand GRP94 overexpression. Therefore, weGiven the roleintracellular quantified of ER stress in lipidcontent lipid metabolism, we hypothesized by Raman spectroscopy that(RS) FA metabolism and principal might be component differentially enriched in 435-Br1 cells as a consequence of their ERSRP analysis (PCA). RS is a technique based on the inelastic dispersion of monochromatic radiation which and GRP94 overexpression. Therefore, we quantified intracellular lipid content by Raman spectroscopy (RS) and principal provides specific vibrational signatures of chemical bonds, suitable for the analysis of saturated and component analysis (PCA). RS is a technique based on the inelastic dispersion of monochromatic unsaturated fattywhich radiation acids provides [26]. To verify specificifvibrational the lipid phenotype signatures of ofchemical BrM wasbonds, an organ-specific trait, suitable for the we carried analysis out the analysis of total of saturated andfatty acids (TFA) unsaturated fattyand total acids unsaturated [26]. To verify fatty if theacid lipid(TUFA) in 435-P phenotype andwas of BrM 435-Br1 an cells, as well asorgan-specific in a metastatic variant trait, to bone we carried out (435-B1), the analysis andof we totalestablished a TFA/unsaturated fatty acids (TFA) ratio (TFA/TUFA) and total unsaturated fatty accordingacidto (TUFA) in 435-P andquantification the spectroscopic 435-Br1 cells, as of well TFAas in a cm (2845 −1 ) andvariant metastatic TUFAtobandsbone (3015 (435-B1), −1 ). we cmand Analysis established a TFA/unsaturated ratio (TFA/TUFA) according to the spectroscopic of 435-P cells (Figure 1A) revealed a high dispersion in TUFA content and the presence of two unexpected quantification of TFA (2845 cm−1) and TUFA bands (3015 cm−1). Analysis of 435-P cells (Figure 1A) revealed a high subpopulations (P1 and P2), clearly differentiated by the lower TUFA/TFA ratio of P1 with regard to P2 dispersion in TUFA content and the presence of two unexpected subpopulations (P1 and P2), clearly (p < 0.00001). P1 TFAbyappeared differentiated the lower in the sameratio TUFA/TFA regionof P1aswith 435-Br1 regard cells to P2while P2 TFAP1 (p < 0.00001). matched 435-B1 cells, TFA appeared suggesting that TUFA/TFA ratio could differentiate metastatic organ specificity. in the same region as 435-Br1 cells while P2 TFA matched 435-B1 cells, suggesting that TUFA/TFA We concluded that 435-Br1 was the ratio cell line with could higher TFA differentiate content metastatic andspecificity. organ lower percentage We concludedof TUFA, therefore that 435-Br1 waswith predominant the cell line with higher TFA content and lower percentage saturated FA profile while 435-B1 were enriched in TUFA. of TUFA, therefore with predominant saturated FA profile while 435-B1 were enriched in TUFA. Brain1.metastatic Figure 1.Figure cells present Brain metastatic an organ-specific cells present increase an organ-specific in fatty increase acidacid in fatty andand cholesterol content. cholesterol content. (A) Raman spectroscopy and PCA analysis of total fatty acids (TFA, −1 (A) Raman spectroscopy and PCA analysis of total fatty acids (TFA, 2845 cm ) and total unsaturated2845 cm −1) and total unsaturated fatty acids (TUFA, fatty cm−1(TUFA, 3015acids 3015 435-Br1 ) of 435-P, cm−1) of 435-P, 435-Br1 cells. and 435-B1 and 435-B1 cells. content (B) Lipid (B) Lipid was content was analyzed by analyzed by fluorescent staining with Nile red (TFA) and filipin (cholesterol). Nile red’s emission fluorescent staining with Nile red (TFA) and filipin (cholesterol). Nile red’s emission shifts to red when shifts to red when it associates to phospholipids (40×). Insert shows lipid droplets. it associates to phospholipids (40×). Insert shows lipid droplets. To confirm these results, we visualized the amount of TFA by means of the fluorescent stain To confirm Nile red these results, [27] (Figure 1B,we visualized upper panel). the As amount expected,of TFA by 435-Br1 means cells of theafluorescent presented higher Nile stain red Nile red [27]staining (Figureintensity 1B, upper panel). As expected, 435-Br1 cells presented a higher Nile red with regard to 435-P and 435-B1 cells. Since increased UPR activation and presencestaining intensity with regard to 435-P and 435-B1 cells. Since increased UPR activation and presence of TFA have been related to cholesterol retention in the ER [28], as well as malignancy, migration, invasion and angiogenesis [29], we analyzed free cholesterol distribution with filipin, a fluorescent probe that Int. J. Mol. Sci. 2019, 20, 3883 4 of 17 Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 4 of 17 of TFA havetobeen binds selectively free related to cholesterol cholesterol retention incholesterol but not esterified the ER [28],[30]. as well Theashigh malignancy, migration, filipin staining intensity invasion in 435-Br1 cellsand angiogenesis confirmed [29], we analyzed an increase free cholesterol in free cholesterol distribution content comparedwithtofilipin, 435-Pa fluorescent and 435-B1 cells (Figureprobe that binds selectively to free cholesterol but not esterified cholesterol [30]. The high filipin 1B, bottom panels). staining intensity in 435-Br1 cells confirmed an increase in free cholesterol content compared to Similarly, RS and PCA analysis were performed in a more enriched organ-specific metastatic cell 435-P and 435-B1 cells (Figure 1B, bottom panels). line to brainSimilarly, (BrV5) obtained RS and PCA after five consecutive analysis in vitro/in were performed vivo in a more selection enriched rounds from organ-specific 435-Br1 cells, metastatic to understand cell line toifbrain high(BrV5) lipid obtained content after and saturation were five consecutive in determinants of brain vitro/in vivo selection metastasis. rounds BrV5 cells from 435-Br1 presented cells,the to highest TFA understand if levels compared high lipid to 435-Br1 content and saturation were±determinants (366.1 44.1) and 435-P1/P2 cells (Figure of brain metastasis. BrV52A, left). cells presented the highest TFA levels compared to 435-Br1 (366.1 ± 44.1) and 435-P1/P2 Furthermore, PCA analysis differentiated 435-P cells from BrV5 by TFA and TUFA content, with BrV5 cells (Figure 2A, left). presenting the Furthermore, highest amount PCAofanalysis TFA anddifferentiated lower TUFA, 435-Pboth cells properties from BrV5 by TFA and related to TUFA content, [31]. aggressiveness with BrV5 presenting the highest amount of TFA and lower TUFA, both properties related to Thus we reproduced the low levels of TUFA in 435-Br1 and BrV5 cells (p = 0.0003) and the differences aggressiveness [31]. Thus we reproduced the low levels of TUFA in 435-Br1 and BrV5 cells (p = in TUFA between 435-P1, 64.2 ± 3.6 and 435-P2 cells, 45 ± 1.27, p = 0.0023. 0.0003) and the differences in TUFA between 435-P1, 64.2 ± 3.6 and 435-P2 cells, 45 ± 1.27, p = 0.0023. 2. The2. lipid FigureFigure phenotype The lipid phenotypeis conserved is conservedinin other brainmetastatic other brain metastatic cellcell lines. lines. 435-Br1 435-Br1 cells were cells were obtained after five obtained after rounds of in five rounds of vivo/in in vivo/invitro vitropassages passages totoobtain obtain a highly a highly brain-specific brain-specific cell cell line line BrV5. BrV5. (A) TFA (A)were TFA were analyzed analyzed by by Ramanspectroscopy Raman spectroscopy (left) and (left) principal and component principal componentanalysisanalysis showing showing an enrichment in TFA content and a decrease in TUFA in BrV5 cells. Error bars represent an enrichment in TFA content and a decrease in TUFA in BrV5 cells. Error bars represent SD of technical SD of technical replicates replicates (one representative (one representative of two of two experiments). (B) experiments). (B)Nile Nilered red(top) andand (top) filipin (bottom) filipin staining (bottom) in staining in 435-P and brain metastasis (BrM) cells 435-Br1 and BrV5 (40×, one representative of three 435-P and brain metastasis (BrM) cells 435-Br1 and BrV5 (40×, one representative of three experiments). experiments). Integrated density was quantified with Image J and normalized to cell number (error Integrated density was quantified with Image J and normalized to cell number (error bars represent SD bars represent SD of technical replicates, >150 cells). (C) Nile red (top) and filipin (bottom) staining in of technical replicates, >150 cells). (C) Nile red (top) and filipin (bottom) staining in 435-P and BrM cells 435-Br1, MDA-MB-361 and SA52. For Nile red, error bars represent SD of n = 2 (435-P, 361, SA52) or n = 4 (435-Br1) independent experiments. For filipin, pictures at 40×, one representative of two experiments, >150 cells analyzed. Error bars represent SD of technical replicates. Quantification was not possible for SA52 due to unclear nuclear visualization. For all panels, significance determined with t-test. * p < 0.05, ** p < 0.01, *** p < 0.001. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 5 of 17 435-P and BrM cells 435-Br1, MDA-MB-361 and SA52. For Nile red, error bars represent SD of n = 2 (435-P, 361, SA52) or n = 4 (435-Br1) independent experiments. For filipin, pictures at 40×, one representative of two experiments, >150 cells analyzed. Error bars represent SD of technical Int. J. Mol. Sci. 2019, 20, 3883 5 of 17 replicates. Quantification was not possible for SA52 due to unclear nuclear visualization. For all panels, significance determined with t-test. * p < 0.05, ** p < 0.01, *** p < 0.001. Next, we validated these results by Nile red and filipin staining (Figure 2B), confirming that 435-Br1 Next,had and BrV5 we avalidated higher Nile these redresults bycompared staining Nile red and filipin to 435-P = 0.003, (Figure (p staining and p = 2B),0.002,confirming that respectively). 435-Br1 and BrV5 had a higher Nile red staining compared to 435-P Furthermore, free cholesterol levels were also higher in 435-Br1 and BrV5 compared to 435-P cells (p = 0.003, and p = 0.002, (p = 0.019 and Furthermore, respectively). p = 0.005, respectively). free cholesterol levels were To discard also higher any possible in 435-Br1 deviation andcellular of the BrV5 compared model, we to 435-P cells TFA performed (p = and 0.019cholesterol and p = 0.005, respectively). staining in two more To discard any possible brain metastatic deviation variants, of the cellular MDA-MB-361 and model, SA52 we performed (Figure 2C). TFA were TFA and cholesterol generally upregulated staining in lines in all cell two compared more brain metastatic to 435-P variants, (MDA-MB-361 vs 435-P, p = 0.037; MDA-MB-361 and SA52 SA52 vs. (Figure 435-P p =TFA 2C). were 0.004). generally Free cholesterolupregulated was alsoin all cell lines increased compared to in MDA-MB-361 compared to 435-P cells (p = 0.01). 435-P (MDA-MB-361 vs 435-P, p = 0.037; SA52 vs. 435-P p = 0.004). Free cholesterol was also increased in MDA-MB-361 compared Finally, to assess to 435-Pin if the increase cells FA (p = 0.01). levels was related to increased synthesis, we analyzed the Finally, expression oftotwoassess if the genes: increase in FA levels Sterol-Regulatory was related toProtein Element-Binding increased synthesis, awe 1 (SREBP-1), analyzed nuclear the factor expression of two genes: Sterol-Regulatory Element-Binding Protein 1 upstream of ACC or FASN [32,33] and Liver X Receptor (LXRα), an oxysterol-induced nuclear factor,(SREBP-1), a nuclear factor upstream which of ACC regulates or FASN SREBP-1 [32,33]driving expression and Liver lipidXmetabolism Receptor (LXRα),downstreaman oxysterol-induced of insulin/AKT signalingnuclear factor, in which the liver regulates [34,35]. LXRα SREBP-1 expression driving was overexpressed lipidcompared in 435-Br1 metabolism to downstream 435-P and other of insulin/AKT metastases signaling (Figure inSREBP-1 3A). the liverexpression [34,35]. LXRα was inwas overexpressed accordance to LXRα in levels, 435-Br1 as compared to 435-P brain metastatic cellsand other (435-Br1 metastases and (Figure an BrV5) showed 3A). SREBP-1 increase expression in this proteinwas in accordance compared to otherto LXRα levels, metastases as brain (Figure metastatic 3B). Similar to cells (435-Br1 and BrV5) showed an increase in this protein compared what we observed with FA levels, bone metastasis behaved like 435-P cells, with the lowest LXRα to other metastases (Figure 3B). SREBP-1 and Similar toexpression. what we observed with FA cells Lung metastatic levels, bone metastasis showed behaved an intermediate like 435-Pbetween phenotype cells, with the 435-P lowest and LXRα 435-Br1 andAll cells. SREBP-1 together expression. these data Lung metastatic suggested cellsmetastatic that brain showed an intermediate cells displayed aphenotype lipogenic between 435-P phenotype and 435-Br1 differential cells. All together among metastatic these cells to other data suggested that brain metastatic cells organs. displayed a lipogenic phenotype differential among metastatic cells to other organs. Figure3.3.BrM Figure BrMcells cellsupregulate upregulategenes genesinvolved involvedin FA synthesis. in FA (A,B) synthesis. Immunoblotting (A,B) of LXRα Immunoblotting (A) and of LXRα (A) SREBP-1 (B) in metastatic variants. (C) SREBP-1 expression was quantified vs. tubulin and SREBP-1 (B) in metastatic variants. (C) SREBP-1 expression was quantified vs. tubulin and and normalized to 435-P cells.toError normalized 435-Pbars represent cells. SEM Error bars of threeSEM represent independent experiments. experiments. of three independent Significance determined Significance with determined p < 0.05. t-test. * with t-test. * p < 0.05. 2.2. Inhibition of FA Import to the Mitochondria Impairs BrM Cell Survival after Metabolic Stress 2.2. Inhibition of FA Import to the Mitochondria Impairs BrM Cell Survival after Metabolic Stress Metastatic cells may undergo a variety of bioenergetic alterations including shift to more glycolytic Metastatic cells may undergo a variety of bioenergetic alterations including shift to more or more oxidative metabolism. Our results suggested that brain metastatic cells accumulate FAs; glycolytic or more oxidative metabolism. Our results suggested that brain metastatic cells therefore, we hypothesized that FAs, essential for the biosynthesis of membranes, might also be required accumulate FAs; therefore, we hypothesized that FAs, essential for the biosynthesis of membranes, as bioenergetic substrates in BrM cells. To analyze if FA could be used as an alternative energy source might also be required as bioenergetic substrates in BrM cells. To analyze if FA could be used as an in conditions of glucose restriction, we deprived cells of glucose and pyruvate for 72 h maintaining alternative energy source in conditions of glucose restriction, we deprived cells of glucose and glutamine (2 mM) and serum (10%) concentration (Figure 4A). Glucose deprivation impaired survival pyruvate for 72 h maintaining glutamine (2 mM) and serum (10%) concentration (Figure 4A). (15%) of 435-P cells differentially with regard to metastatic cells in brain 435-Br1 (30%), lungs 435-L3 Glucose deprivation impaired survival (15%) of 435-P cells differentially with regard to metastatic (30%) and bone 435-B1 (25%). To assess the importance of FA oxidation (FAO) in survival of BrM cells, cells in brain 435-Br1 (30%), lungs 435-L3 (30%) and bone 435-B1 (25%). To assess the importance of we used etomoxir to inhibit the mitochondrial carnitine palmitoyltransferase I (CPT1), a transporter FA oxidation (FAO) in survival of BrM cells, we used etomoxir to inhibit the mitochondrial carnitine that mediates FA import to the mitochondria and is the rate-limiting step of FAO [36]. Etomoxir palmitoyltransferase I (CPT1), a transporter that mediates FA import to the mitochondria and is the treatment in glucose starved cultures decreased survival of 435-Br1 (p = 0.001) and to a lesser extent rate-limiting step of FAO [36]. Etomoxir treatment in glucose starved cultures decreased survival of in 435-L3 (p = 0.031) with regard to 435-P cells (Figure 4A). The same experiment performed in the 435-Br1 (p = 0.001) and to a lesser extent in 435-L3 (p = 0.031) with regard to 435-P cells (Figure 4A). set of brain metastatic cells (BrV5, MDA-MB-361, SA52) proved the increased resistance to glucose The same experiment performed in the set of brain metastatic cells (BrV5, MDA-MB-361, SA52) deprivation in all these cell lines compared to the parental cells, as well as impaired survival in glucose deprivation after etomoxir (Figure 4B). These results suggested that brain metastatic cells might use FA as an alternative energy source in glucose restriction conditions, whereas other metastatic cells may depend on FA import to a lesser extent. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 6 of 17 proved the increased resistance to glucose deprivation in all these cell lines compared to the parental cells, as well as impaired survival in glucose deprivation after etomoxir (Figure 4B). These results Int.suggested that20, J. Mol. Sci. 2019, brain 3883 metastatic cells might use FA as an alternative energy source in glucose 6 of 17 restriction conditions, whereas other metastatic cells may depend on FA import to a lesser extent. Figure Figure4. 4.BrM BrMcellscellsare aremore moresensitive sensitivetotoetomoxir etomoxirininglucose glucosedeprivation deprivationconditions. conditions.(A,B) (A,B)Survival Survivalof 435-P and and of 435-P a seta ofsetmetastatic of metastatic cells to to cells different differentorgans organs(A) (A)ororto tobrain brain (B) cultured cultured without glucose± ± withoutglucose etomoxir etomoxir(200 µM).Bars (200µM). Barsrepresent representrelative relative survival survival of each cell cell line line vs. vs.its itscontrol control(non-treated) (non-treated) with with glucose, error bars represent SD of independent experiments (n = 3). glucose, error bars represent SD of independent experiments (n = 3). (C) ACOT-7 expression was(C) ACOT-7 expression was quantified quantified vs.vs.tubulin tubulin(error (errorbars barsrepresent representSEM SEMofofnn==44(435-P, (435-P,435-Br1, 435-B1),nn== 55 (435-L3) 435-Br1,435-B1), (435-L3) or n == 3 3 (BrV5) (BrV5) independent independent experiments) experiments) (D) Palmitate (D) Palmitate oxidationoxidation was after was calculated calculated 14 C-palmitate addingafter adding and measuring released 14 CO2released 14C-palmitate and measuring production14CO2 production at 5 or 0 mM glucose (error bars represent at 5 or 0 mM glucose (error bars represent SEM of three SEM of three independent independent (E) experiments). experiments). (E) Palmitate Palmitate oxidation oxidation ratio at 0 mMratio at 0 vs. glucose mM5 glucose mM glucosevs. 5 mM (error glucose bars represent(error SEMbarsof 3represent independent SEMexperiments). of 3 independentFor allexperiments). For all determined panels, significance panels, significance with t-test. < 0.05, ** p with * pdetermined < 0.01, *** p* p<<0.001. t-test. 0.05, ** p < 0.01, *** p < 0.001. Ourtranscriptional Our transcriptionalstudies studiesshowed showed an increased increased expression expressionof ofACOT7 ACOT7ininbrain brainmetastatic metastaticcells, cells, anan enzymethat enzyme thathydrolyzes hydrolyzespalmitoyl-CoA palmitoyl-CoA under the the control controlof ofaasterol sterolresponse responseelement element (SRE) (SRE)[37]. [37]. Since Since palmitoyl-CoA palmitoyl-CoA is the is the substrate substrate of CPT-I, of CPT-I, it is plausible it is plausible to hypothesize to hypothesize that ACOT7 that ACOT7 could could modulate modulate lipid importlipid to theimport to the mitochondria mitochondria and FAO. ACOT7 and FAO. wasACOT7 was differentially differentially overexpressed overexpressed in 435-Br1 inand 435-Br1 and BrV5 cells compared to bone metastasis or 435-P (Figure 4C), BrV5 cells compared to bone metastasis or 435-P (Figure 4C), suggesting a possible induction suggesting a possibleof induction of mitochondrial mitochondrial FAO in brain FAO in brain metastasis andmetastasis reinforcing andthereinforcing thebetween similarities similarities thebetween parentaltheand parental and bone metastatic subtype. To further evaluate the importance of the bone metastatic subtype. To further evaluate the importance of the mitochondria in the metabolic mitochondria in the metabolic features featuressoobserved observed so far, wemitochondrial far, we assessed assessed mitochondrial morphology morphology in the setinofthe set of metastatic metastatic variants. variants. Mitochondrial dynamics is characterized by fusion and fission events that can regulate Mitochondrial dynamics is characterized by fusion and fission events that can regulate mitochondrial mitochondrial function [38,39]. Therefore, we performed fluorescent staining with Mitotracker function [38,39]. Therefore, we performed fluorescent staining with Mitotracker Green (Supplementary Green (Supplementary Figure S2A) to evaluate mitochondrial morphology. 435-P cells showed low Figure S2A) to evaluate mitochondrial morphology. 435-P cells showed low staining intensity and staining intensity and fragmented mitochondria in the perinuclear region while the BrM variants fragmented mitochondria in the perinuclear region while the BrM variants presented higher staining presented higher staining and elongated mitochondria (Supplementary Figure S2A). The rest of and elongated mitochondria (Supplementary Figure S2A). The rest of metastatic variants to lung and metastatic variants to lung and bone presented an intermediate phenotype. To quantify the bone presented an intermediate phenotype. To quantify the differential phenotype between 435-P differential phenotype between 435-P and BrV5, we calculated mitochondrial elongation (see and BrV5, we Methods) calculated [40] mitochondrial and we classified cells inelongation (see Methods) three subtypes according[40] and we classified to elongation cells in three levels: fragmented, subtypes according to elongation levels: fragmented, intermediate and tubular intermediate and tubular (Supplementary Figure S2B). BrV5 cells had a higher content of tubular(Supplementary Figure S2B). BrV5 cells had a higher content of tubular mitochondria compared to mitochondria compared to 435-P (48% vs. 28%), while 435-P cells presented mainly fragmented and 435-P (48% vs. 28%), while 435-P cells presented mainly fragmented and intermediate mitochondria. Mitochondrial fusion correlates with increased oxidative phosphorylation and mitofusin expression (Mfn1, Mfn2) [41]. BrM cell lines overexpressed Mfn2 (Supplementary Figure S2C), a protein that has been described to promote ER-mitochondria contact modulating Ca2+ homeostasis, glucose oxidation and Krebs Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 7 of 17 intermediate mitochondria. Mitochondrial fusion correlates with increased oxidative phosphorylation and mitofusin expression (Mfn1, Mfn2) [41]. BrM cell lines overexpressed Mfn2 (Supplementary Int. J. Figure S2C), a protein that has been described to promote ER-mitochondria contact Mol. Sci. 2019, 20, 3883 7 of 17 modulating Ca homeostasis, glucose oxidation and Krebs cycle [41]. Mfn1 and Mfn2 were down 2+ regulated in 435-P correlating with a fragmented phenotype, together with decreased VDAC cycle [41]. Mfn1 and Mfn2 were down regulated in 435-P correlating with a fragmented phenotype, expression (Supplementary Figure S2C), which suggested a lower mitochondrial content and together function within decreased this cell line.VDAC expression (Supplementary Figure S2C), which suggested a lower mitochondrialFinally,content and function we analyzed in thisacid FAO (fatty cell line. oxidation) after glucose withdrawal by means of a Finally, we analyzed FAO (fatty acid oxidation) radioactive assay that consists in addition of 14C-palmitate after glucose(Figure withdrawal by means 4D) and of a radioactive measurement of CO2 assay that consists in addition of 14 C-palmitate (Figure 4D) and measurement of CO production. production. In the presence of glucose (5 mM), 435-Br1 cells showed an increased release 2 of CO2 In the presence compared toof435-P glucose (5 mM), (Figure 4D, 435-Br1 left). In cells showed glucose an increased deprivation releaseallof cell conditions, CO2lines compared induced to 435-P (Figure 4D, left). In glucose deprivation conditions, all cell lines induced ß-oxidation, ß-oxidation, but these differences were more significant in 435-Br metastatic cells compared to 435-P but these differences wereMoreover, (Figure 4D). more significant in 435-Br we calculated the metastatic net increasecells compared in oxidation to 435-P (Figure subtracting the values4D).ofMoreover, released we calculated CO2 at 0 mM theglucose net increase in oxidation vs. 5 mM (Figure 4E), subtracting the values which proved of released that induction COoxidation of FA 2 at 0 mMinglucose response vs. 5 mM (Figure deprivation to glucose 4E), which proved that induction differentially occurs in of FA oxidation metastatic in response cells vs. 435-P. Alltotogether glucosethesedeprivation results suggested occurs differentially that metastatic cell lines in metastatic cellsmight respond vs. 435-P. All to metabolic together thesestress inducing results fatty acid suggested that oxidation, metastatic celleven lines though other factors might respond like ability to metabolic to face stress ER stress inducing fattyor FAoxidation, acid availabilityeven for membrane though other synthesis factors likemight abilitylimit survival. to face ER stress or FA availability for membrane synthesis might limit survival. 2.3.2.3. GRP94 GRP94 Is Required Is Required forfor Tumor Tumor Growth Growth In In Vivo Vivo and and It itModulates Modulatesthe theLipogenic LipogenicPhenotype Phenotype of of Brain Metastatic Cells Metastatic Cells Since the UPR Since the and UPRERand stressER have been previously stress have beenrelated to lipidrelated previously metabolism [13,16,19,28,33,42], to lipid metabolism we [13,16,19,28,33,42], hypothesized thatwe hypothesized of overexpression that overexpression GRP94 of GRP94 might might be associated with thebe increased associated fatty with acid the increased content and FAOfattyobserved acid content andmetastatic in brain FAO observed in evaluate cells. To brain metastatic the role cells. To evaluate of GRP94 the roleand in FA synthesis of GRP94 in we degradation, FA generated synthesis and degradation, stable knock-downs we generated of GRP94 stable in the knock-downs brain metastatic of GRP94 in the brain BRV5CA1-GFP cell linemetastatic (see MaterialBRV5CA1-GFP and Methods). cell Interestingly, line (see Material GRP94and Methods). ablation inducedInterestingly, GRP94activation a compensatory ablation induced a compensatory activation of GRP78, as previously described [43] of GRP78, as previously described [43] (Figure 5A). We pooled several control clones (CTR-pool) (Figure 5A). We pooled andseveral control selected clones (CTR-pool) two GRP94 and selected clones shGRP94-2 twoand (424-2) GRP94 clones shGRP94-2 shGRP94-8 (424-2) (424-8) for and shGRP94-8 orthotopic injection. (424-8) for orthotopic injection. The growth of shGRP94 cells was similar to parental The growth of shGRP94 cells was similar to parental and BRV5CA1 cells at 48, 72 and 96 h in standard and BRV5CA1 cells conditions culture at 48, 72 and(data 96 h not in standard shown).culture conditions 424-2 and (datashowed 424-8 cells not shown). 424-2 andtumorigenicity a decreased 424-8 cells showed and growth in vivo with regards to CTRL-pool (Figure 5B). Furthermore, tumor volume5B). a decreased tumorigenicity and growth in vivo with regards to CTRL-pool (Figure wasFurthermore, dramatically tumor volume was dramatically decreased after GRP94 ablation compared to the control, p = 0.0002 decreased after GRP94 ablation compared to the control, p = 0.0002 (Figure 5C) demonstrating that (Figure 5C) demonstrating that GRP94 is required for in vivo tumor growth. GRP94 is required for in vivo tumor growth. Figure Figure 5. Glucose-regulated 5. Glucose-regulated protein protein 94 (GRP94) 94 (GRP94) deletiondeletion reduces reduces tumorigenic tumorigenic efficiencyefficiency in an in an orthotopic orthotopic model. (A) Immunoblotting of GRP94 and actin shows efficient GRP94 model. (A) Immunoblotting of GRP94 and actin shows efficient GRP94 knockdown in BRV5CA1 cells. knockdown in BRV5CA1 cells. (B) Nude mice were injected in the mammary fat pad with control (B) Nude mice were injected in the mammary fat pad with control (BRV5CA1 CTRL-pool) or shGRP94 (BRV5CA1 (424-2 and 424-8) cells. Representative mice are shown presenting a drastic decrease in tumor burden. (C) Tumor volume was measured at the indicated times based on caliper measurements of tumor length and width as volume (mm3 ) = (Length2 × Width2 /2). Tumor volume was normalized to the volume of the first observation (day 14) for each group. Significance was determined with t-test. *** p = 0.0002. Int. J. Mol. Sci. 2019, 20, 3883 8 of 17 To analyze if the metabolic changes observed in GRP94 overexpressing cells might be reversed in GRP94 ablated cells, we measured TFA and cholesterol content by fluorescent staining (Figure 6A). Nile red revealed increased phospholipid content in shGRP94 V5CA1 cells compared to the control (clone 2, p = 0.015, clone 8, p = 0.007), together with a trend to increased cholesterol content. Next, we analyzed protein changes secondary to down-regulated GRP94 (Figure 6B). GRP94 depletion induced a significant down-regulation of ACOT7 and SREBP-1 in ablated cells with no differences in LXRα expression between depleted GRP94 clones and controls, suggesting lipid accumulation might be independent of these specific synthesis pathways. Moreover, we analyzed if GRP94 ablation decreased cell ability to survive in hypoglycemia (Figure 6C). Interestingly, shGRP94-2 (424-2) and shGRP94-8 (424-8) cells were more sensitive to glucose deprivation with a lower survival ratio than control cells (values normalized to each cell line in glucose conditions), reinforcing the role of GRP94 as a mediator of survival in glucose starvation. Etomoxir treatment did not impair survival, suggesting that GRP94 ablated cells were no longer dependent on mitochondrial FA import (Figure 6C). To test this hypothesis, we also measured FAO and we observed no differences in palmitate oxidation secondary to GRP94 ablation (Figure 6D). Although further studies are required to determine the mechanism, these data suggest that GRP94 modulates FA accumulation while it doesn’t impact fatty acid oxidation. Overall, our work identifies a lipid phenotype consisting of increased FA content and saturation common in brain metastatic cells. Furthermore, we identified GRP94 as a key mediator of BrM metabolic stress survival, likely through regulation of the ER stress response, and we describe a previously unknown Int. J.of role Mol. Sci. 2019, GRP94 in 20, x FOR lipid PEER REVIEW metabolism. 9 of 17 Figure 6. Cont. Figure 6. GRP94 ablation impairs survival in glucose restriction conditions and increases lipid content. (A) Lipid content was analyzed by fluorescent staining with Nile red (TFA) and filipin (cholesterol) for V5CA1-CTR, 424-2 and 424-8 cells (40x, one representative experiment of three, error bars represent SD of >180 cells analyzed). (B) SREBP-1 and LXRα expression was quantified vs. tubulin. (C) Survival of BrM metastatic cells after GRP94 ablation cultured without glucose ± etomoxir (200 µM). Bars represent relative survival of each cell line vs. its control (non-treated) with glucose, error bars represent SD of three independent experiments. (D) Palmitate oxidation was calculated after adding 14C-palmitate and measuring released 14CO2 production at 5 or 0 mM glucose Int. J. Mol. Sci. 2019, 20, 3883 9 of 17 Figure6.6.GRP94 Figure GRP94 ablation ablationimpairs survival impairs in glucose survival restriction in glucose conditions restriction and increases conditions lipid content. and increases lipid (A) Lipid (A) content. content was Lipid analyzed content was byanalyzed fluorescent bystaining with staining fluorescent Nile red (TFA) and filipin with Nile (cholesterol) red (TFA) for and filipin V5CA1-CTR, (cholesterol) 424-2 and 424-8 cells for V5CA1-CTR, (40x, 424-2 andone representative 424-8 experiment cells (40x, one of three, representative error barsof experiment represent SD three, error of >180 bars cells analyzed). represent (B) SREBP-1 SD of >180 and LXRα cells analyzed). (B)expression SREBP-1 was and quantified vs. tubulin. LXRα expression was(C) Survival vs. quantified of BrM metastatic cells after GRP94 ablation cultured without glucose ± etomoxir tubulin. (C) Survival of BrM metastatic cells after GRP94 ablation cultured without glucose ± (200 µM). Bars represent relative etomoxir survival of each (200 µM). Barscell line vs.relative represent its control (non-treated) survival of each cellwithline glucose, vs. its error bars control represent SD (non-treated) of with three independent glucose, error barsexperiments. represent SD(D) ofPalmitate oxidation was three independent calculated(D) experiments. after adding 14 Palmitate C-palmitate oxidation was and measuring 14 CO production at 5 or 0 mM glucose calculated afterreleased adding 14C-palmitate 2 and measuring released 14CO (error bars represent 2 production at 5 or 0SEM mMofglucose three independent experiments). (right) Palmitate oxidation ratio at 0 mM glucose vs. (error bars represent SEM of three independent experiments). (right) Palmitate oxidation ratio at 5 mM glucose. For all0 panels, significance determined with t-test. * p < 0.05, ** p < 0.01, *** p < 0.001. mM glucose vs. 5 mM glucose. For all panels, significance determined with t-test. * p < 0.05, ** p < 0.01, *** p < 0.001. 3. Discussion The increase in the biosynthesis of new molecules as well as the capacity to fulfill the energetic 3. Discussion requirements of the cell are crucial for metastatic progression [44], but at the same time these processes The increase in the biosynthesis of new molecules as well as the capacity to fulfill the energetic generate a permanent state of energy demand and metabolic stress [45]. The activation of the cancer requirements of the cell are crucial for metastatic progression [44], but at the same time these stress phenotype [46] allows the correct adaptation to this stress-activating pathways to promote processes generate a permanent state of energy demand and metabolic stress [45]. The activation of cell survival in a process defined as “non-oncogene addiction” or non-oncogene co-dependency [47]. the cancer stress phenotype [46] allows the correct adaptation to this stress-activating pathways to Here we show that expression of GRP94 promotes a rewiring of cellular metabolism in BrM cells promote cell survival in a process defined as “non-oncogene addiction” or non-oncogene conferring a survival advantage in glucose deprivation conditions by activating lipid metabolism co-dependency [47]. Here we show that expression of GRP94 promotes a rewiring of cellular and glucose-independent growth. These might be crucial to brain metastasis progression given the metabolism in BrM cells conferring a survival advantage in glucose deprivation conditions by unique metabolic features of the brain interstitial space, which include lower glucose levels [48] and activating lipid metabolism and glucose-independent growth. These might be crucial to brain cholesterol enrichment by the novo synthesis that compensates for the inability of cholesterol to cross metastasis progression given the unique metabolic features of the brain interstitial space, which the blood–brain barrier [49]. Our results also suggest that BrM cells increase cholesterol metabolism, include lower glucose levels [48] and cholesterol enrichment by the novo synthesis that compensates which may act as an endogenous ligand for LXRs [50], mimicking the tissue-specific homeostasis. for the inability of cholesterol to cross the blood–brain barrier [49]. Our results also suggest that BrM Moreover, the use of RS and PCA allowed an accurate analysis of cellular lipid content, cells increase cholesterol metabolism, which may act as an endogenous ligand for LXRs [50], differentiating cells by their metabolic profile [51]. In particular, we describe a lipogenic phenotype mimicking the tissue-specific homeostasis. (increased TFA/TUFA ratio) associated to brain metastasis vs. bone metastasis. This increase in fatty acid content likely correlates with FA synthesis given the overexpression of proteins like SREBP-1 and LXRα, which induce ACC and ACLY transcription [34,35] and activate other pro-synthetic genes like FASN [52]. Increased fatty acid synthesis is a bad prognostic factor in breast cancer [53,54] as it confers resistance to therapy [55,56] and promotes metastatic and invasive capacity [51,57]. Our results are in accordance to previous studies that showed a correlation between ACOX1 and FASN expression and organ specificity to brain in a set of patients’ metastasis [58]. RS may be considered a promising strategy to stratify organ-specific metastasis according to lipid content and may be used to select patients that could benefit from therapies targeting lipid metabolism. GRP94 deletion induced further FA accumulation but it didn´t impair fatty acid oxidation. One possible explanation is that GRP94 ablation led to activation of FA import pathways that collaborate in increasing intracellular FA content. On the other hand, we cannot exclude the possibility that the lipid phenotype of BrM cells was not exclusively dependent of GRP94 levels. In fact, it may involve the action of other stress chaperones like GRP78, which was overexpressed in GRP94 ablated cells, and other UPR proteins that might in turn promote FA synthesis [7]. GRP78 has been associated with lipid accumulation in breast cancer cells due to an inhibition of mitochondrial fatty Int. J. Mol. Sci. 2019, 20, 3883 10 of 17 acid transport, resulting in a reduction of FAO [25]. Since we didn’t observe defects in FA oxidation after GRP94 deletion, this is unlikely to be the main mechanism leading to FA accumulation. In addition, GRP780 s role as a lipid regulator is not only restricted to FAO as it has been described to regulate SREBP-1 function among others [59]. Since both chaperones have been previously reported as co-regulated under a variety of stress conditions, further studies are needed to understand the role of these chaperones in lipids metabolism and metastasis progression. Interestingly, despite FA accumulation after GRP94 KD, we observed a decrease in SREBP-1 levels, the master regulator of fatty acid and triacylglycerol synthesis and a mediator of BRAF-targeted therapy resistance [60]. Since metabolism includes multiple pathways that are tuned to meet the increasing metabolic requirements during the metastatic process [61], it is possible that FA synthesis could be decoupled from SREBP-1 and engage downstream effectors like ACC-1 or FASN. GRP94 expression is critical in chaperoning lipogenesis regulator genes. Since SREBP-1 is transported from the ER to the Golgi, where it is consecutively cleaved by two proteases to be transcriptionally active [62], further experiments would be required to understand if SREBP-1 may be a client protein affected by the chaperone function of GRP94. One of the most dramatic differences induced by GRP94 depletion in BrM cells resides in the down-regulation of ACOT7, a major isoform of the ACOT family that catalyzes hydrolysis of fatty acyl-CoAs to free fatty acids that may condition the recruitment of FA to be metabolized in the mitochondria. High expression of acyl-CoA thioesterases has been implicated in regulatory mechanisms involved in ER stress in cancer [63]. Moreover, low levels of ACOT7 prevented human breast cancer by activating p53-p21 signaling pathway and cell cycle-arrest [64]. Further studies are required to dissect the relationship between the observed down-regulation of ACOT7 in GRP94 ablated cells and low tumorigenicity. Different tumoral subtypes acquire different bioenergetic alterations, either towards a more glycolytic or more oxidative metabolism. This tissue-specific context invalidates the idea of a unique common transition in all cancer types, as Warburg hypothesized [65]. Our results confirmed that brain metastatic cells were more resistant to glucose deprivation than the parental cell line and other metastasis. These data, in addition to etomoxir susceptibility and mitochondrial morphology, suggested that carcinoma cells growing in the brain might be more dependent on mitochondrial metabolism, mainly through FA import and oxidation. This metabolic shift is in agreement with previous studies showing gain of glucose-independent growth [48] and increased oxidative phosphorylation and beta-oxidation in breast cancer BrM [66]. The arrival and establishment in the new microenvironment might give preference to some metabolic functions in an organ-specific manner, and maintenance of oxidative metabolism may be critical for BrM establishment and progression. Further work is required to evaluate the role of GRP94 in maintaining respiration and ATP production [67] in BrM. From these studies, we conclude that GRP94 is critical in the metabolic reprogramming of BrM cells, rewiring lipid metabolism and mediating glucose-independent survival during metastasis progression. Further work using untargeted lipidomic analysis of brain metastatic cells might reveal lipid species altered favoring metastasis progression. 4. Materials and Methods 4.1. Cell Culture and Treatments MDA-MB-435 (435-P) breast cancer cells and their metastatic variants to brain (435-Br1), lung (435-L3), bone (435-B1) and liver (435-Lv1) were obtained from successive in vivo/in vitro passages [68]. MDA-MB-361 cells were obtained from Laboratoire d’Oncogénetique (Centre Rue Huguenin, Saint-Cloud, France). SA52 cells were obtained from Laboratory of Experimental Cancerology (University of Ghent, Ghent, Belgium). BrV5 and BRV5CA1 were obtained after several rounds of in vitro/in vivo injection of BR-eGFP-CMV/Luc cells, as described elsewhere [69]. All cell lines were maintained under standard conditions in DMEM/F12 medium (Invitrogen, Carlsbad, CA, USA) Int. J. Mol. Sci. 2019, 20, 3883 11 of 17 supplemented with 10% FBS, 1 mM pyruvate and 2 mM L-glutamine at 37 ◦ C in a humidified 5% CO2 incubator. For glucose deprivation experiments, cells were seeded and after 24 h, they were starved of glucose (0 or 4.5 mg/mL) using DMEM without glucose (Invitrogen) supplemented with 10% SBF and L-Gln 2 mM for 48 or 72 h. Treatment with etomoxir (Sigma, St. Louis, MO, USA) was performed at a final concentration of 200 µM for 72 h. 4.2. Cell Viability Cells were fixed in paraformaldehyde 4% and stained with crystal violet (0.2% in 2% ethanol) for 20 min. Dye was extracted with SDS 10% and relative proliferation was determined by measuring OD at 595 nm to evaluate survival. 4.3. Fluorescent Stainings Nile red and filipin staining: 6–8 × 104 cells were seeded in 24-well plates and left to adhere for 24 h on coverslips coated with gelatin 2%. Cells were fixed in paraformaldehyde 4%, washed with PBS1X and stained with Nile red (Sigma, 1 µg/mL in PBS 1X) or filipin (Sigma, 20 µg/mL in PBS 1X) for 1 or 2 h, respectively. Coverslips were mounted with Vectashield with (Nile red) or without (filipin) DAPI (Vector laboratories). Samples were visualized with an Olympus BX60 (Olympus, Tokyo, Japan) microscope at 40× or 60×. Integrated density was quantified with Image J using a constant area and was normalized to cell number (Int Dens/cell). Immunofluorescence: For immunofluorescence, 1 × 105 cells were seeded in 24-well plates containing gelatin-coated cover slips and they were fixed after 24 h using paraformaldehyde 4% for 15 min at 4 ◦ C. Blocking and permeabilization was performed using PBS1x-20%FBS containing 0.2% Triton X-100 for 30 min at room temperature. The antibodies used were: GRP94 (1:800, Santa Cruz, CA, USA), and GRP78 (1:400, Santa Cruz). Alexa Fluor® 555 anti-mouse and Alexa Fluor® 488 anti-goat (Invitrogen) were used as secondary antibodies at the adequate concentration (1:1000). Vectashield (Vector laboratories, Burlingame, CA, USA) with DAPI was used for nucleus visualization. Mitochondrial staining with Mitotracker green: Cells were stained with Mitotracker green (M-7514, Molecular probes Invitrogen) at 100 nM for 30 min, washed and analyzed in complete media without fixation in a confocal microscope (SP2 Leica, 63×, excitation at 488 nm, emission at 500–540 nm). Elongation was calculated with Image J (v 1.52a) as previously described. Briefly, images were transformed to greyscale and inverted, threshold was set to 90 and the function Analyze Particles was used to obtain circularity values (values of 1 were excluded). Elongation was defined as the inverse of circularity. Frequency analysis was used to define the abundance of each phenotype normalized to total number of mitochondria analyzed. An average of 90 cells were analyzed for each cell line. 4.4. Stable GRP94 Protein Knockdown GRP94 stable knock-down: Five constructs containing different short hairpin RNA (shRNA) against GRP94 (TRCN0000029424, TRCN0000029425, TRCN0000029426, TRCN0000029427, TRCN0000029428) and a non-target shRNA control vector, were obtained commercially from Sigma Aldrich MISSION® shRNA. BRV5CA1-GFP cells were transfected with the five shGRP94 constructs or shControl, respectively, using Lipofectamine 2000 (Invitrogen) as a transfection agent. Clonal selection was performed and Western blot and immunofluorescence were used to validate GRP94 knockdown. 4.5. Protein Expression To assess protein expression in the different experiments, cell lines were lysed in a 1% SDS (v/v) extraction buffer containing an anti-protease cocktail (Roche, Vilvoorde, Belgium). Protein concentrations were determined using the Bradford assay (MicroBCA, Pierce, Belgium). After resolution by SDS-PAGE, electrophoresed proteins were transferred to polyvinylidene fluoride (PVDF) membranes that were blocked and probed with the following antibodies: ACOT7 (1:1000, Abcam, Cambridge, UK), LXRα (1:1000, PPMX, Tokyo, Japan), SREBP-1 (1/500, Abcam), GRP94 (1/1000, Sta Cruz, CA, USA), Mfn1 Int. J. Mol. Sci. 2019, 20, 3883 12 of 17 (1:1000, Santa Cruz, CA, USA,), Mfn2 (1:2000, Abcam), VDAC (1:5000, Abcam), α-tubulin (1:10,000, Sigma), actin (1:2000, Sigma) and the corresponding peroxidase-conjugated secondary antibody at 1:2000: Peroxidase conjugated anti-rabbit secondary Ab (Amersham, Little Chalfont, UK), Peroxidase conjugated Antimouse secondary Ab (Pierce, Perbio Science Ltd., Cheshire, UK). Immunoreactive bands were quantified using a VersaDoc™ (BioRad Laboratories, Hercules, CA, USA) Imaging System using the Super Signal west-Pico (Pierce Biotechnology, Rockford, IL, USA). MWs were established with See Blue Plus2 prestained Standford (Invitrogen). 4.6. 14 C-Pamitate Oxidation Fatty acid oxidation was assessed by oxidation of 14 C-palmitate, as previously described [70]. The 1 × 10 5 cells were seeded in 12-well plates and left to adhere for 24 h. Cells were washed with 1 mL of 0.1% BSA fatty acid free in KRBH (NaCl 135 mM, KCl 3.6 mM, NaH2 PO4 0.5 mM, MgSO4 0.5 mM, CaCl2 1.5 mM, NaHCO3 2 mM and HEPES 10 mM, pH 7.4). Cells were incubated in a sealed system with 500 µL of 0.1% BSA fatty acid free in KRBH for 30 min and washed again with the first solution. The mix reaction (0.9 mM carnitine, 0.25 mM palmitate and 1 µCi/mL (1-14 C) palmitate bound to 1% BSA, in KRBH with (5 mM) or without glucose) was added. Released 14 CO2 was fixed in a Whatman filter moistened in KOH. After 4.5 h at 37 ◦ C in a CO2 -free incubator, 100 µL of perchloric acid at 40% was injected in the well and left overnight. The potassium hydroxide-saturated filters were immersed in 5 mL of scintillation fluid. Protein quantification for each cell line was performed by BCA assay. Results are expressed as nmol palmitate × mg−1 prot × h−1 = (cpm sample-cpm blank) × 125/(total cpm × mg prot × h). 4.7. Animal Models Athymic Nude-Foxn1nu female mice weighing 22–28 g were purchased from Harlan Laboratories S.A. (Barcelona, Spain) and were housed in the IDIBELL facility in SFP conditions, at 20–24 ◦ C, 60% relative humidity, and 12 h light and dark periods. Animals were allowed free access to UV-irradiated water and an adequate sterile diet. All animal-related procedures were performed in accordance with the National Institute of Health Guidelines for the Care and Use of Laboratory and with the approval of the animal care committee (Reference 9703 (protocol number)). Mammary tumors were induced by inoculation of 1 × 106 of cells in 0.05 mL of medium without serum in the inguinal mammary gland (IMFP). The xenografts were implanted in animals under isoflurane-induced anesthesia. Tumor volume was measured twice a week based on caliper measurements of tumor length and width as volume (mm3 ) = (Length2 × Width2 /2). Animals were euthanized at day 72 after injection when tumors reached ~1000 mm3 . 4.8. Raman Spectroscopy For measurements in the 2820–3030 cm−1 range, 9 × 104 cells were seeded in Petri dishes with #0 coverglass (Mattek, Ashland, MA). Cells were maintained under standard culture conditions. Hypoglycemia (0 mg/mL) was induced after 24 h of cell seeding and maintained for 48 h until they were analyzed. They were fixed for 15 min at 4 ◦ C with paraformaldehyde at 4% in PBS and, after 3 PBS washes, maintained at 4 ◦ C in PBS. Raman spectra were acquired using an InVia Raman microscope (Renishaw, UK) with a backscattered configuration. Raman excitation was performed with a 514 nm optical beam focused through a 60× objective and with 9 mW power [51]. The spectra were background subtracted with a custom-written Labview program and the Gaussian fits for total fatty acids (TFA) and total unsaturated fatty acids (TUFA) bands (2845 cm−1 and 3015 cm−1 respectively) were performed in Matlab (Mathworks Inc., Natick, MA) allowing the quantification of the two types of fatty acids in the cytoplasm [26]. Normalization under all Raman spectra was performed to correct for the different amplification in the signal. This normalization can be based on the fact that the spectral region used (the CH stretching region) can be considered as the total biomass present in our confocal volume [71]. Int. J. Mol. Sci. 2019, 20, 3883 13 of 17 4.9. Statistical Analysis Principal component analysis (PCA) was performed over the pre-processed Raman spectra in order to evaluate the spectral differences between the cancerous cell lines studied and to develop a model allowing their discrimination and classification [72]. PCA operates in an unsupervised manner (no previous knowledge of the samples under study is provided) and finds an alternative set of coordinates, the principal components (PCs), to reduce the dimensionality and complexity of the data set. All the spectra can then be explained in a much simpler fashion through a small number of PCs that accounts for the maximum variance in the data [51]. PCA was carried out in Raman experiments in the range of 2820 to 3030 cm−1 to differentiate cell groups according to fatty acid content and type. Multivariate statistical analysis was performed using the PLS toolbox (Eigenvector Research, Wenatchee, WA) in the Matlab (Mathworks Inc., Natick, MA) programming environment. SPSS (Statistical Package for the Social Sciences) for Windows was used for the statistics of TFA and TUFA quantification. In all the analyses, differences were considered significant when student’s “t” was lower than 0.05. Student’s t-test was used to compare groups of in vitro and in vivo experiments. In all the analyses, differences were considered significant when p < 0.05. p-Values lower than 0.05 were considered significant. Microsoft Excel and GraphPad were used to plot the graphs and perform statistical analysis. Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/16/ 3883/s1. Author Contributions: Conceptualization, N.S.-C. and A.S.; methodology, N.S.-C., L.M., C.N., M.M.; formal analysis, N.S.-C., C.N., M.M., L.M.; investigation, N.S.-C., A.M.-G., L.M.,C.N., M.M., D.S., J.P.M.; resources, A.S., A.Z.; writing—original draft preparation, N.S.-C. and A.S.; writing—review and editing, N.S.-C., A.S., A.Z.; visualization, N.S.-C.; supervision, A.S., A.Z.; project administration, A.S., N.S.-C.; funding acquisition, A.S., A.Z. Funding: This research was supported by grants from the Spanish Ministry of Health and Consumer Affairs FIS-PI10/00057, FIS-PI14/00336 and FIS-PI18/00916, from the I+D+I National Plan with the financial support from ISCIII-Subdirección General de Evaluación and the Fondo Europeo de Desarrollo Regional (FEDER), and by grants from Generalitat de Catalunya 2017_SGR_1305, Fundació Privada Cellex, Fundació Privada Froi and the AECC (Spanish Association Against Cancer). Acknowledgments: The authors would like to acknowledge the technical support in in vivo models provided by Mr Ruben Foj. Conflicts of Interest: A.S. is inventor of the patent US 9,645,150 B2 and EP 2440931 named “Methods for determining the risk for developing brain metastasis, and a kit to carry out said method”, which refers to GRP94´s potential as biomarker to predict brain metastasis and to target therapies. All other authors have declared that no conflicts of interest exist. Abbreviations GRP94 Glucose-Regulated Protein 94 FA Fatty Acids ER Endoplasmic Reticulum UPR Unfolded Protein Response ERSRP ER stress-resistance phenotype IRE1 Inositol-requiring enzyme 1 CNS Central Nervous system TFA Total Fatty Acid TUFA Total Unsaturated Fatty Acid RS Raman Spectroscopy PCA Principal Component Analysis SREBP-1 Sterol Regulatory Element-Binding Protein- 1 LXR-α Liver X Receptor BrM Brain metastasis FAO Fatty Acid Oxidation Mfn Mitofusin KD Knock-down Int. J. Mol. Sci. 2019, 20, 3883 14 of 17 References 1. Maharaj, K.; Que, N.; Hong, F.; Huck, J.; Gill, S.; Wu, S.; Li, Z.; Gewirth, D. Exploring the functional complementation between Grp94 and Hsp90. PLoS ONE 2016, 11, e0166271. [CrossRef] [PubMed] 2. Hoter, A.; El-Sabban, M.; Naim, H. The HSP90 family: Structure, regulation, function, and implications in health and disease. Int. J. Mol. Sci. 2018, 19, 2560. [CrossRef] [PubMed] 3. Yang, Y.; Li, Z. Roles of heat shock protein gp96 in the ER quality control: Redundant or unique function? Mol. Cells 2005, 20, 173–182. [PubMed] 4. Huck, J.; Que, N.; Hong, F.; Gewirth, D. Structural and functional analysis of GRP94 in the closed state reveals an essential role for the pre-N domain and a potential client-binding site. Cell Rep. 2017, 20, 2800–2809. [CrossRef] [PubMed] 5. Ansa-Addo, E.; Thaxton, J.; Hong, F.; Wu, B.; Zhang, Y.; Fugle, C.; Metelli, A.; Riesenberg, B.; Williams, K.; Gewirth, D.; et al. Clients and oncogenic roles of molecular chaperone gp96/grp94. Curr. Top. Med. Chem. 2016, 16, 2765–2778. [CrossRef] [PubMed] 6. Chen, Y.; Ashok, B.; Liu, X.; Garikapaty, V.; Mittelman, A.; Tiwari, R. Induction of heat shock protein gp96 by immune cytokines. Cell Stress Chaperones 2003, 8, 242–248. [CrossRef] 7. Zhu, G.; Lee, A. Role of the unfolded protein response, GRP78 and GRP94 in organ homeostasis. J. Cell Phisiol. 2015, 230, 1413–1420. [CrossRef] 8. Tsai, Y.C.; Weissman, A.M. The unfolded protein response, degradation from endoplasmic reticulum and cancer. Genes Cancer 2010, 1, 764–778. [CrossRef] 9. Clarke, R.; Cook, K.L.; Hu, R.; Facey, C.O.; Tavassoly, I.; Schwartz, J.L.; Baumann, W.T.; Tyson, J.J.; Xuan, J.; Wang, Y.; et al. Endoplasmic reticulum stress, the unfolded protein response, autophagy, and the integrated regulation of breast cancer cell fate. Cancer Res. 2012, 72, 1321–1331. [CrossRef] 10. Lee, A.S. Glucose-regulated proteins in cancer: Molecular mechanisms and therapeutic potential. Nat. Rev. Cancer 2014, 14, 263–276. [CrossRef] 11. Ghosh, S.; Shinogle, H.; Galeva, N.; Dobrowsky, R.; Blagg, B. Endoplasmic reticulum-resident heat shock protein 90 (HSP90) isoform glucose-regulated protein 94 (GRP94) regulates cell polarity and cancer cell migration by affecting intracellular transport. J. Biol. Chem. 2016, 291, 8309–8323. [CrossRef] [PubMed] 12. Luo, B.; Lee, A.S. The critical roles of endoplasmic reticulum chaperones and unfolded protein response in tumorigenesis and anticancer therapies. Oncogene 2013, 32, 805–818. [CrossRef] 13. Zheng, Z.; Zhang, C.; Zhang, K. Role of unfolded protein response in lipogenesis. World J. Hepatol. 2010, 2, 203–207. [CrossRef] [PubMed] 14. Liu, B.; Yang, Y.; Qiu, Z.; Staron, M.; Hong, F.; Li, Y.; Wu, S.; Li, Y.; Hao, B.; Bona, R.; et al. Folding of toll-like receptors by the HSP90 paralogue gp96 requires a substrate-specific cochaperone. Nat. Commun. 2010, 1, 79. [CrossRef] [PubMed] 15. Gutierrez, T.; Simmen, T. Endoplasmic reticulum chaperones and oxidoreductases: Critical regulators of tumor cell survival and immunorecognition. Front. Oncol. 2014, 4. 16. Han, J.; Kaufman, R. The role of ER stress in lipid metabolism and lipotoxicity. J. Lipid Res. 2016, 57, 1329–1338. [CrossRef] [PubMed] 17. Wang, M.; Kaufman, R.J. The impact of the endoplasmic reticulum protein-folding environment on cancer development. Nat. Rev. Cancer 2014, 14, 581–597. [CrossRef] [PubMed] 18. Mancini, R.; Noto, A.; Pisanu, M.; De Vitis, C.; Maugeri-Saccà, M.; Ciliberto, G. Metabolic features of cancer stem cells: The emerging role of lipid metabolism. Oncogene 2018, 37, 2367–2378. [CrossRef] [PubMed] 19. Dauer, P.; Sharma, N.; Gupta, V.; Durden, B.; Hadad, R.; Banerjee, S.; Dudeja, V.; Saluja, A.; Banerjee, S. ER stress sensor, glucose regulatory protein 78 (GRP78) regulates redox status in pancreatic cancer thereby maintaining “stemness”. Cell Death Dis. 2019, 10. [CrossRef] 20. Lin, X.; DeAngelis, L. Treatment of brain metastasis. J. Clin. Oncol. 2015, 33, 3475–3484. [CrossRef] 21. Martin, B.; Aragues, R.; Sanz, R.; Oliva, B.; Boluda, S.; Martinez, A.; Sierra, A. Biological pathways contributing to organ-specific phenotype of brain metastatic cells. J. Proteome Res. 2008, 7, 908–920. [CrossRef] [PubMed] 22. Sanz-Pamplona, R.; Aragues, R.; Driouch, K.; Martin, B.; Oliva, B.; Gil, M.; Boluda, S.; Fernandez, P.L.; Martinez, A.; Moreno, V.; et al. Expression of endoplasmic reticulum stress proteins is a candidate marker of brain metastasis in both ErbB-2+ and ErbB-2− primary breast tumors. Am. J. Pathol. 2011, 179, 564–579. [CrossRef] [PubMed] Int. J. Mol. Sci. 2019, 20, 3883 15 of 17 23. Martínez-Aranda, A.; Hernández, V.; Guney, E.; Muixí, L.; Foj, R.; Baixeras, N.; Cuadras, D.; Moreno, V.; Urruticoechea, A.; Gil, M.; et al. FN14 and GRP94 expression are prognostic/predictive biomarkers of brain metastasis outcome that open up new therapeutic strategies. Oncotarget 2015, 6, 44254–44273. [CrossRef] [PubMed] 24. Landemaine, T.; Jackson, A.; Bellahcene, A.; Rucci, N.; Sin, S.; Abad, B.M.; Sierra, A.; Boudinet, A.; Guinebretiere, J.M.; Ricevuto, E.; et al. A six-gene signature predicting breast cancer lung metastasis. Cancer Res. 2008, 68, 6092–6099. [CrossRef] [PubMed] 25. Cook, K.; Soto-Pantoja, D.; Clarke, P.; Cruz, M.; Zwart, A.; Wärri, A.; Hilakivi-Clarke, L.; Roberts, D.; Clarke, R. Endoplasmic reticulum stress protein GRP78 modulates lipid metabolism to control drug sensitivity and antitumor immunity in breast cancer. Cancer Res. 2016, 76, 5657–5670. [CrossRef] 26. Freudiger, C.W.; Min, W.; Saar, B.G.; Lu, S.; Holtom, G.R.; He, C.; Tsai, J.C.; Kang, J.X.; Xie, X.S. Label-free biomedical imaging with high sensitivity by stimulated Raman scattering microscopy. Science 2008, 322, 1857–1861. [CrossRef] 27. Greenspan, P.; Fowler, S.D. Spectrofluorometric studies of the lipid probe, nile red. J. Lipid Res. 1985, 26, 781–789. 28. Pineau, L.; Colas, J.; Dupont, S.; Beney, L.; Fleurat-Lessard, P.; Berjeaud, J.M.; Berges, T.; Ferreira, T. Lipid-induced ER stress: Synergistic effects of sterols and saturated fatty acids. Traffic 2009, 10, 673–690. [CrossRef] 29. Raghu, H.; Sodadasu, P.K.; Malla, R.R.; Gondi, C.S.; Estes, N.; Rao, J.S. Localization of uPAR and MMP-9 in lipid rafts is critical for migration, invasion and angiogenesis in human breast cancer cells. BMC Cancer 2010, 10, 647. [CrossRef] 30. Kruth, H. Histochemical detection of esterified cholesterol within human atherosclerotic lesions using the fluorescent probe filipin. Atherosclerosis 1984, 51, 281–292. [CrossRef] 31. Rysman, E.; Brusselmans, K.; Scheys, K.; Timmermans, L.; Derua, R.; Munck, S.; Van Veldhoven, P.P.; Waltregny, D.; Daniels, V.W.; Machiels, J.; et al. De novo lipogenesis protects cancer cells from free radicals and chemotherapeutics by promoting membrane lipid saturation. Cancer Res. 2010, 70, 8117–8126. [CrossRef] [PubMed] 32. Brown, M.S.; Goldstein, J.L. The SREBP pathway: Regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 1997, 89, 331–340. [CrossRef] 33. Basseri, S.; Austin, R.C. Endoplasmic reticulum stress and lipid metabolism: Mechanisms and therapeutic potential. Biochem. Res. Int. 2012, 2012, 841362. [CrossRef] [PubMed] 34. Maxwell, K.N.; Soccio, R.E.; Duncan, E.M.; Sehayek, E.; Breslow, J.L. Novel putative SREBP and LXR target genes identified by microarray analysis in liver of cholesterol-fed mice. J. Lipid Res. 2003, 44, 2109–2119. [CrossRef] [PubMed] 35. Okazaki, H.; Goldstein, J.L.; Brown, M.S.; Liang, G. LXR-SREBP-1c-phospholipid transfer protein axis controls very low density lipoprotein (VLDL) particle size. J. Biol. Chem. 2010, 285, 6801–6810. [CrossRef] [PubMed] 36. Samudio, I.; Harmancey, R.; Fiegl, M.; Kantarjian, H.; Konopleva, M.; Korchin, B.; Kaluarachchi, K.; Bornmann, W.; Duvvuri, S.; Taegtmeyer, H.; et al. Pharmacologic inhibition of fatty acid oxidation sensitizes human leukemia cells to apoptosis induction. J. Clin. Investig. 2010, 120, 142–156. [CrossRef] [PubMed] 37. Takagi, M.; Suto, F.; Suga, T.; Yamada, J. Sterol regulatory element-binding protein-2 modulates human brain acyl-CoA hydrolase gene transcription. Mol. Cell Biochem. 2005, 275, 199–206. [CrossRef] 38. Liesa, M.; Palacin, M.; Zorzano, A. mitochondrial dynamics in mammalian health and disease. Physiol. Rev. 2009, 89, 799–845. [CrossRef] 39. Legros, F.; Lombes, A.; Frachon, P.; Rojo, M. Mitochondrial fusion in human cells is efficient, requires the inner membrane potential, and is mediated by mitofusins. Mol. Biol. Cell. 2002, 13, 4343–4354. [CrossRef] 40. Dagda, R.K.; Cherra, S.J., 3rd; Kulich, S.M.; Tandon, A.; Park, D.; Chu, C.T. Loss of PINK1 function promotes mitophagy through effects on oxidative stress and mitochondrial fission. J. Biol. Chem. 2009, 284, 13843–13855. [CrossRef] 41. Zorzano, A.; Liesa, M.; Sebastian, D.; Segales, J.; Palacin, M. Mitochondrial fusion proteins: Dual regulators of morphology and metabolism. Semin. Cell Dev. Biol. 2010, 21, 566–574. [CrossRef] [PubMed] Int. J. Mol. Sci. 2019, 20, 3883 16 of 17 42. Wei, Y.; Wang, D.; Topczewski, F.; Pagliassotti, M.J. Saturated fatty acids induce endoplasmic reticulum stress and apoptosis independently of ceramide in liver cells. Am. J. Physiol. Endocrinol. Metab. 2006, 291, E275–E281. [CrossRef] [PubMed] 43. Ni, M.; Lee, A.S. ER chaperones in mammalian development and human diseases. FEBS Lett. 2007, 581, 3641–3651. [CrossRef] [PubMed] 44. Deberardinis, R.J.; Sayed, N.; Ditsworth, D.; Thompson, C.B. Brick by brick: Metabolism and tumor cell growth. Curr. Opin. Genet. Dev. 2008, 18, 54–61. [CrossRef] [PubMed] 45. Mathew, R.; White, E. Eat this, not that! How selective autophagy helps cancer cells survive. Mol. Cell. Oncol. 2014, 2, e975638. [CrossRef] [PubMed] 46. Luo, J.; Solimini, N.L.; Elledge, S.J. Principles of cancer therapy: Oncogene and non-oncogene addiction. Cell 2009, 136, 823–837. [CrossRef] [PubMed] 47. Galluzzi, L.; Kepp, O.; Vander Heiden, M.; Kroemer, G. Metabolic targets for cancer therapy. Nat. Rev. Drug Discov. 2013, 12, 829–846. [CrossRef] [PubMed] 48. Chen, J.; Lee, H.J.; Wu, X.; Huo, L.; Kim, S.J.; Xu, L.; Wang, Y.; He, J.; Bollu, L.R.; Gao, G.; et al. Gain of glucose-independent growth upon metastasis of breast cancer cells to the brain. Cancer Res. 2015, 75, 554–565. [CrossRef] 49. Dietschy, J. Central nervous system: Cholesterol turnover, brain development and neurodegeneration. Biol. Chem. 2009, 390, 287–293. [CrossRef] 50. Villa, G.; Hulce, J.; Zanca, C.; Bi, J.; Ikegami, S.; Cahill, G.; Gu, Y.; Lum, K.; Cravatt, B.; Mischel, P. An LXR-cholesterol axis creates a metabolic co-dependency for brain cancers. Cancer Cell 2016, 30, 683–693. [CrossRef] 51. Nieva, C.; Marro, M.; Santana-Codina, N.; Rao, S.; Petrov, D.; Sierra, A. The lipid phenotype of breast cancer cells characterized by Raman microspectroscopy: Towards a stratification of malignancy. PLoS ONE 2012, 7, e46456. [CrossRef] [PubMed] 52. Shaw, R.J.; Cantley, L.C. Decoding key nodes in the metabolism of cancer cells: Sugar & spice and all things nice. F1000 Biol. Rep. 2012, 4, 2. [PubMed] 53. Kuhajda, F.P.; Jenner, K.; Wood, F.D.; Hennigar, R.A.; Jacobs, L.B.; Dick, J.D.; Pasternack, G.R. Fatty acid synthesis: A potential selective target for antineoplastic therapy. Proc. Natl. Acad. Sci. USA 1994, 91, 6379–6383. [CrossRef] [PubMed] 54. Flavin, R.; Peluso, S.; Nguyen, P.L.; Loda, M. Fatty acid synthase as a potential therapeutic target in cancer. Futur. Oncol. 2010, 6, 551–562. [CrossRef] [PubMed] 55. Menendez, J.A.; Lupu, R.; Colomer, R. Inhibition of tumor-associated fatty acid synthase hyperactivity induces synergistic chemosensitization of HER −2/neu-overexpressing human breast cancer cells to docetaxel (taxotere). Breast Cancer Res. Treat. 2004, 84, 183–195. [CrossRef] [PubMed] 56. Warmoes, M.; Jaspers, J.E.; Xu, G.; Sampadi, B.K.; Pham, T.V.; Knol, J.C.; Piersma, S.R.; Boven, E.; Jonkers, J.; Rottenberg, S.; et al. Proteomics of genetically engineered mouse mammary tumors identifies fatty acid metabolism members as predictive markers for cisplatin resistance. Mol. Cell. Proteomics 2013, 12, 1319–1334. [CrossRef] [PubMed] 57. Scott, K.E.; Wheeler, F.B.; Davis, A.L.; Thomas, M.J.; Ntambi, J.M.; Seals, D.F.; Kridel, S.J. Metabolic regulation of invadopodia and invasion by acetyl-CoA carboxylase 1 and de novo lipogenesis. PLoS ONE 2012, 7, e29761. [CrossRef] [PubMed] 58. Jung, Y.Y.; Kim, H.M.; Koo, J.S. Expression of lipid metabolism-related proteins in metastatic breast cancer. PLoS ONE 2015, 22, 554–565. [CrossRef] [PubMed] 59. Yang, Y.; Cheung, H.; Tu, J.; Miu, K.; Chang, W. New insights into the unfolded protein response in stem cells. Oncotarget 2016, 7, 54010–54027. [CrossRef] [PubMed] 60. Talebi, A.; Dehairs, J.; Rambow, F.; Rogiers, A.; Nittner, D.; Derua, R.; Vanderhoydonc, F.; Duarte, J.A.G.; Bosisio, F.; Van den Eynde, K.; et al. Sustained SREBP-1-dependent lipogenesis as a key mediator of resistance to BRAF-targeted therapy. Nat. Commun. 2018, 9, 2500. [CrossRef] [PubMed] 61. Luo, X.; Cheng, C.; Tan, Z.; Li, N.; Tang, M.; Yang, L.; Cao, Y. Emerging roles of lipid metabolism in cancer metastasis. Mol. Cancer 2017, 16, 76. [CrossRef] [PubMed] 62. Horton, J.D.; Goldstein, J.L.; Brown, M.S. SREBPs: Activators of the complete program of cholesterol and fatty acid synthesis in the liver. J. Clin. Investig. 2002, 109, 1125–1131. [CrossRef] [PubMed] Int. J. Mol. Sci. 2019, 20, 3883 17 of 17 63. Liu, K.; Chou, S.; Yen, M.; Kuo, P. Regulatory mechanism of fatty acid-CoA metabolic enzymes under endoplasmic reticulum stress in lung cancer. Oncol. Rep. 2018, 40, 2674–2682. [CrossRef] [PubMed] 64. Jung, S.; Lee, H.; Hwang, H.; Parl, H.; Moon, Y.; Kim, B.; Lee, H.; Kim, K.; Kin, Y.; Lee, B.; et al. Acyl-CoA thioesterase 7 is involved in cell cycle progression via regulation of PKCζ–p53–p21 signaling pathway. Cell Death Dis. 2017, 8, e2793. [CrossRef] [PubMed] 65. Wallace, D.C. Mitochondria and cancer. Nat. Rev. Cancer 2012, 12, 685–698. [CrossRef] [PubMed] 66. Chen, E.I.; Hewel, J.; Krueger, J.S.; Tiraby, C.; Weber, M.R.; Kralli, A.; Becker, K.; Yates, J.R., 3rd; Felding-Habermann, B. Adaptation of energy metabolism in breast cancer brain metastases. Cancer Res. 2007, 67, 1472–1486. [CrossRef] [PubMed] 67. Huang, C.; Lee, C.; Tu, C.; Wu, C.; Huang, M.; Wei, P.; Chang, Y. Glucose-regulated protein 94 mediates progression and metastasis of esophageal squamous cell carcinoma via mitochondrial function and the NF-kB/COX-2/VEGF axis. Oncotarget 2018, 9, 9425–9441. [CrossRef] 68. Mendez, O.; Fernandez, Y.; Peinado, M.A.; Moreno, V.; Sierra, A. Anti-apoptotic proteins induce non-random genetic alterations that result in selecting breast cancer metastatic cells. Clin. Exp. Metastasis 2005, 22, 297–307. [CrossRef] 69. Martinez-Aranda, A.; Hernandez, V.; Picon, C.; Modolell, I.; Sierra, A. Development of a preclinical therapeutic model of human brain metastasis with chemoradiotherapy. Int. J. Mol. Sci. 2013, 14, 8306–8327. [CrossRef] 70. Collins, C.L.; Bode, B.P.; Souba, W.W.; Abcouwer, S.F. Multiwell 14CO2-capture assay for evaluation of substrate oxidation rates of cells in culture. Biotechniques 1998, 24, 803–808. [CrossRef] 71. Bocklitz, T.; Walter, A.; K, H.; Rösch, P.; Popp, J. How to pre-process raman spectra for reliable and stable models? Anal. Chim. Acta. 2011, 704, 47–56. [CrossRef] [PubMed] 72. Hilvo, M.; Denkert, C.; Lehtinen, L.; Muller, B.; Brockmoller, S.; Seppanen-Laakso, T.; Budczies, J.; Bucher, E.; Yetukuri, L.; Castillo, S.; et al. Novel theranostic opportunities offered by characterization of altered membrane lipid metabolism in breast cancer progression. Cancer Res. 2011, 71, 3236–3245. [CrossRef] [PubMed] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).