Lipid-sensors, enigmatic-orphan and orphan nuclear receptors as therapeutic targets in breast-cancer

Breast-cancer is heterogeneous and consists of various groups with different biological characteristics. Innovative pharmacological approaches accounting for this heterogeneity are needed. The forty eight human Nuclear-Hormone-Receptors are ligand-dependent transcription-factors and are classified into Endocrine-Receptors, Adopted-Orphan-Receptors (Lipid-sensors and Enigmatic-Orphans) and Orphan-receptors. Nuclear-Receptors represent ideal targets for the design/synthesis of pharmacological ligands. We provide an overview of the literature available on the expression and potential role played by Lipid-sensors, Enigmatic-Orphans and Orphan-Receptors in breast-cancer. The data are complemented by an analysis of the expression levels of each selected Nuclear-Receptor in the PAM50 breast-cancer groups, following re-elaboration of the data publicly available. The major aim is to support the idea that some of the Nuclear-Receptors represent largely unexploited therapeutic-targets in breast-cancer treatment/chemo-prevention. On the basis of our analysis, we conclude that the Lipid-Sensors, NR1C3, NR1H2 and NR1H3 are likely to be onco-suppressors in breast-cancer. The Enigmatic-Orphans, NR1F1 NR2A1 and NR3B3 as well as the Orphan-Receptors, NR0B1, NR0B2, NR1D1, NR2F1, NR2F2 and NR4A3 exert a similar action. These Nuclear-Receptors represent candidates for the development of therapeutic strategies aimed at increasing their expression or activating them in tumor cells. The group of Nuclear-Receptors endowed with potential oncogenic properties consists of the Lipid-Sensors, NR1C2 and NR1I2, the Enigmatic-Orphans, NR1F3, NR3B1 and NR5A2, as well as the Orphan-Receptors, NR2E1, NR2E3 and NR6A1. These oncogenic Nuclear-Receptors should be targeted with selective antagonists, reverse-agonists or agents/strategies capable of reducing their expression in breast-cancer cells.


INTRODUCTION
Breast-cancer is heterogeneous and traditionally classified according to the expression of Estrogen-  accommodates only organic compounds and the design of new agonists/antagonists is facilitated by crystallographic data and functional screening assays. This article provides an overview of the literature available on the expression and role played by Lipid-sensors, Enigmatic-Orphans and Orphan-Receptors in breast-cancer induction/ progression. Steroid-Receptors, Heterodimeric-Receptors and the Lipid-sensors, Retinoid-X-Receptors (NR2B1/ RXRα, NR2B2/RXRβ, NR2B3/RXRγ) are excluded from our analysis given the wealth of data on their therapeutic significance in breast-cancer [4,5]. The data are complemented by an analysis of the expression levels of each NR in the PAM50 breast-cancer groups, following re-elaboration of the data in the TCGA site (The-Cancer-Genome-Atlas: http://cancergenome.nih.gov). The review supports the idea that some of the NRs considered represent unexploited therapeutic-targets in breast-cancer treatment/chemo-prevention.

LIPID-SENSORS
The active transcriptional forms of Lipid-sensors are RXR heterodimers. NR1C1-C3, NR1H2-4 and NR1I2 are permissive heterodimeric partners, as the simultaneous presence of NR1 and RXR ligands results in cooperative responses [6].

NR1C1
(PPARα:peroxysome-proliferator -activated-receptor-α), NRC2 (PPARβ/ δ:peroxysome-proliferator-activated-receptorβ/δ) and NRC3 (PPARγ:peroxysomeproliferator-activated-receptor-) NR1C1, NRC2 and NR1C3 control lipid homeostasis. NR1C3 is the most studied member of the NR1C subfamily, given its relevance in obesity/ diabetes. The two NR1C2 shortest protein-variants Underneath each box-plot, the tables show significant differences in the mRNA expression levels of each NR between the indicated groups. The results were obtained from the data available in the TCGA (The Cancer Genome Atlas; http://cancergenome.nih.gov). Normalization, quantification and statistical analysis on raw sequencing counts was performed using the Limma/Voom (http://bioconductor.org) package in R statistical environment. * (adjusted p < 0.01), ** (adjusted p < 0.001), *** (adjusted p < 0.0001). www.impactjournals.com/oncotarget present with incomplete DNA-and Ligand-binding domains, respectively ( Table 1). The only identified endogenous NR1C1 activator is 1-palmitoyl-2-oleoylsn-glycerol-3-phosphocholine (POGP). Prostaglandin-PGI2, 15-hydroxyeicosatetraenoic-acid (15-HETE) and polyunsaturated fatty-acids are recognized endogenous NRIC2 ligands. Certain prostaglandins and fatty-acids act as endogenous NR1C3 ligands. Some of the numerous synthetic agonists/antagonists available are listed in Table  1. NR1C1-mRNA (UNIGENE-Hs.103110), NR1C2-mRNA (UNIGENE-Hs.696032) and NR1C3-mRNA (UNIGENE-Hs.162646) expression is ubiquitous and the transcripts are detectable in mammary-glands, albeit at very low levels in the case of NR1C1. Relative to the normal counterpart, NR1C1 and NR1C3 mRNAs are down-regulated in all PAM50-classified breast-cancers ( Figure 3). Basal and Normal-like tumors show the highest NR1C1 and NR1C3 levels, respectively. In contrast, mammary-tumors express higher NR1C2 mRNA levels than the normal counterpart, due to up-regulation in Her2, The table contains basic information on the characteristics of the Lipid-Sensors group of nuclear receptors (NRs). The first column lists the human NRs considered in the review article. The official symbol of each NR is indicated in italics, while the original alias of each protein product is indicated in parenthesis. The full name of each NR is indicated underneath in italics. The second column from the left lists the human chromosome (Chr) each NR maps to. The number of exons encoding the transcripts giving rise to the corresponding NR protein-variant is indicated in the third column. The fourth column lists the accession number of each NR protein-variant. The amino acid (aa) length of each NR protein variant, the position of the DNA-binding domain (DBD) and the ligand-binding domain (LBD) are indicated in columns five, six and seven, respectively. Column eight contains a list of representative endogenous (end.) and synthetic (synth.) agonists, antagonists and reverse agonists for each NR along with an appropriate reference. The chemical structures of the listed molecules can be found in the PUBCHEM database with the use of the PUBCHEM-CID accession numbers provided. The PUBCHEM chemical structure is not available in the case of the NR2E1 agonists Ccrp-1, -2 and -3. When possible, the predicted onco-suppressive (bold) or oncogenic (black-boxed) action of the corresponding NR is indicated in the last column on the right. Synthetic agonists and antagonists of potential therapeutic interest targeting onco-suppressive and oncogenic NRs, respectively, are marked in bold and boxed in black. Finally, in the few cases where supportive data are available, the type of breast-cancer which is predicted to represent a preferential target of the NR is listed in the last column. The table contains basic information on the Enigmatic-Orphans group of nuclear receptors (NRs). The first column lists the human NRs considered in the review article. The official symbol of each NR is indicated in italics, while the original alias of each protein product is indicated in parenthesis. The full name of each NR is indicated underneath in italics. The second column from the left lists the human chromosome (Chr) each NR maps to. The number of exons encoding the transcripts giving rise to the corresponding NR protein-variant is indicated in the third column. The fourth column lists the accession number of each NR protein-variant. The amino acid (aa) length of each NR protein variant, the position of the DNA-binding domain (DBD) and the ligand-binding domain (LBD) are indicated in columns five, six and seven, respectively. Column eight contains a list of representative endogenous (end.) and synthetic (synth.) agonists, antagonists and reverse agonists for each NR along with an appropriate reference. The chemical structures of the listed molecules can be found in the PUBCHEM database with the use of the PUBCHEM-CID accession numbers provided. The structure of the NR1I3 antagonist, BDBM50422490, is available in the CHEMBL database as indicated. When possible, the predicted onco-suppressive (bold) or oncogenic (black-boxed) action of the corresponding NR is indicated in the last column on the right. Synthetic agonists and antagonists of potential therapeutic interest targeting onco-suppressive and oncogenic NRs, respectively, are marked in bold and boxed in black. Finally, in the few cases where supportive data are available, the type of breast-cancer which is predicted to represent a preferential target of the NR is listed in the last column. www.impactjournals.com/oncotarget Basal and Normal-like cancers.
In ER + /MCF-7 and ER + /T47D breast-cancer cells, NR1C2-activation by GW501516 stimulates proliferation and angiogenic responses [10]. In addition, the NR1C2antagonist, SR13904, inhibits cell-growth and -survival [12]. Two NR1C2-inverse-agonists (ST247 = methyl 3-(N-(4-(hexylamino)-2-methoxyphenyl)sulfamoyl) thiophene-2-carboxylate and DG172 = (Z)-2-(2bromophenyl)-3-(4-(4-methylpiperazin-1-yl)phenyl) acrylonitrile dihydrochloride) inhibit serum-and TGFβinduced invasion of ER -/MDA-MB231 cells into threedimensional matrixes, suggesting that NR1C2 favors tumor-cell dissemination [13]. A role in ST247/DG172 action is played by Angiopoietin-Like-4 down-regulation. NR1C2 oncogenic action is supported by data obtained in MMTV-PPARβ/δ transgenic-models of mammary carcinogenesis where the NR1C2-agonist, GW501516, The table contains basic information on the characteristics of the Orphan-Receptors group of nuclear receptors (NRs). The first column lists the human NRs considered in the review article. The official symbol of each NR is indicated in italics, while the original alias of each protein product is indicated in parenthesis. The full name of each NR is indicated underneath in italics. The second column from the left lists the human chromosome (Chr) each NR maps to. The number of exons encoding the transcripts giving rise to the corresponding NR protein-variant is indicated in the third column. The fourth column lists the accession number of each NR protein-variant. The amino acid (aa) length of each NR protein variant, the position of the DNA-binding domain (DBD) and the ligand-binding domain (LBD) are indicated in columns five, six and seven, respectively. Column eight contains a list of representative endogenous (end.) and synthetic (synth.) agonists, antagonists and reverse agonists for each NR along with an appropriate reference. The chemical structures of the listed molecules can be found in the PUBCHEM database with the use of the PUBCHEM-CID accession numbers provided. The PBCHEM chemical structure is not available in the case of the NR2E1 agonists Ccrp-1, -2 and -3. When possible, the predicted onco-suppressive (bold) or oncogenic (black-boxed) action of the corresponding NR is indicated in the last column on the right. Synthetic agonists and antagonists of potential therapeutic interest targeting onco-suppressive and oncogenic NRs, respectively, are marked in bold and boxed in black. Finally, in the few cases where supportive data are available, the type of breast-cancer which is predicted to represent a preferential target of the NR is listed in the last column. reduces tumor latency [14]. Further support comes from transgenic-mice over-expressing 3-Phosphoinositide-Dependent-Kinase-1, a protein regulated by NR1C2, where GW501516 accelerates tumorigenesis [15]. NR1C2 oncogenic action may involve the EGFR/ERBB2 pathway via FABP5 (Fatty-Acid-Binding-Protein-5), which delivers ligands to NR1C2. In ER + /MCF-7 cells, EGFR/ERBB2-dependent proliferative responses are accompanied by FABP5 induction [16]. In MMTV-ErbB2/ HER2 mice, spontaneously developing breast-cancer, FABP5 ablation relieves activation of EGFR downstream signals, down-regulates NR1C2 target-genes involved in cell-proliferation and suppresses tumor-growth [17]. The retinoid pathway may also be involved in NR1C2 oncogenic activity, as NR1C2 is bound/activated by ATRA [18]. Indeed, ATRA delivery to NR1C2 and RARs is dependent on FABP5 and CRABP-II, respectively. In ATRA-resistant MMTV-neu transgenic-model of breastcancer, decreasing FABP5/CRABP-II ratio diverts ATRA from NR1C2 to RAR and suppresses tumor-growth [19]. Although all this is consistent with an oncogenic action of NR1C2 in breast-cancer, there is also evidence to the contrary. The NR1C2-agonist, GW501516, inhibits MCF-7 cells growth [20] and NR1C2 over-expression/activation reduces clonogenicity and in-vivo growth of ER + /MCF-7 and ER -/MDA-MB231 cells [21]. In conclusion, the majority of the available data indicates that NR1C2 is prooncogenic in breast-carcinoma [22].
NR1C3 levels are associated with improved clinical outcome and represent a prognostic factor for overall-survival in ER + /breast-cancer patients, suggesting onco-suppressive properties [23]. The synthetic NR1C3agonists, thiazolidinediones, suppress mammary-tumor growth in-vitro and in-vivo [24][25][26]. Thiazolidinedioneactivated NR1C3 interferes with ERα, Signal-Transducerand-Activator-of-Transcription-5B and Nucler-Factor-kappaB, inhibiting the proliferation of ER + / and ER -/cells which undergo differentiation [27] and apoptosis [28]. In addition, thiazolidinediones inhibit TGFβ signaling, which suppresses breast-cancer early-development [25]. Finally, NR1C3-agonists reduce mammary-tumor angiogenesis and invasion [29]. In rats, the NR1C3ligand, GW7845, inhibits carcinogen-induced breastcancer [26]. Troglitazone prevents 7,12-dimethylbenz(a) anthracene-induced transformation of murine breast-tissue [30] and NR1C3 heterozygous-deletion causes greater susceptibility to mammary-tumor development after exposure to the same anthracene-related compound [31]. The action of NR1C3 stimulation is not limited to breastcancer prevention and extends to treatment [32]. The only exception to the in-vivo data supporting NR1C3 oncosuppressive properties is represented by a study showing that the NR acts as a tumor-promoter in a transgenicmodel of breast carcinogenesis via interference with the WNT-pathway [33]. A small-sized clinical-trial reports that patients with metastatic breast-cancer fail to show any benefit from troglitazone administration [34]. An equally small and recent trial demonstrates that administration of rosiglitazone between the time of diagnostic biopsy and definitive surgery is well-tolerated although it does not alter breast-cancer cell-proliferation [35].
As for the potential role of NR1H2 and NR1H3 in breast-cancer, little and contrasting evidence is available. NR1H2/NR1H3 ligands secreted by tumor cells inhibit CCR7 expression on maturing dendritic cells, impairing immune-surveillance and favoring tumor growth [38]. In mouse breast-cancer models, 27-hydroxycholesterol augments ER-dependent mammary-tumor growth and increases NR1H2/NR1H3-dependent metastasis [39]. The two effects require conversion of cholesterol into 27-hydroxycholesterol by CYP27A1 [40]. The data obtained in these mouse models contrast with the observations made in some breast-cancer cells, where NR1H2/NR1H3 activation reduces proliferation with down-regulation of genes involved in cell cycle progression, DNA replication and other cell-growthrelated processes [41,42]. In ER + /breast-tumors the NR1H2/NR1H3 growth-inhibitory action may result from systemic effects, as the two NRs control hepatic estrogen biosynthesis. In the liver, NR1H2/NR1H3 stimulation results in sulfotransferase (an enzyme critical for estrogen deactivation) induction which inhibits breast-cancer growth in xenografted mice [43]. Non-cell autonomous growth-inhibitory effects are also suggested by in-vitro results, as culture medium from NR1H2/NR1H3 activated macrophages causes growth-inhibition and apoptosis of breast-cancer cells [44].
In conclusion, NR1H2/NR1H3 activation may represent a strategy for the treatment/chemoprevention of breast-cancer, although caution should be exercised, since NR1H2/NR1H3-agonists may favor the metastatic spread of tumor cells [39].

NR1H4 (FXR:farnesoid-X-receptor)
NR1H4 transcriptional activity is stimulated by bileacids, such as chenodeoxycholic acid [45,46]. Adrenalglands, kidney, liver and intestine express the highest NR1H4-mRNA levels (UNIGENE-Hs.282735), although the transcript is present also in mammary-glands. With the exception of the Normal-like group, all PAM50 mammarytumor sub-types express smaller NR1H4 amounts than the normal counterpart ( Figure 3). In breast-cancer, NR1H4protein levels are associated with ER-status and luminal markers [47].
Data on the role played by NR1H4 in mammarytumors are contrasting. As bile-acids are a risk factor for post-menopausal breast-cancer [48], their high concentrations in breast-cysts/plasma of mammarytumor patients [49] suggest a role for NR1H4 in disease induction/progression. Consistent with this hypothesis, the NR1H4 agonist, deoxycholate, promotes survival and favors migration of ER -/MDA-MB-231 cells, while the inverse-agonist, guggulsterone, exerts opposite effects [50]. Similarly, farnesol, another NR1H4-agonist, exerts mitogenic effects in ER + /MCF-7 cells, but not in ER -/MDA-MB231 cells [47]. Mitogenesis may require binding/activation of ERα by NR1H4 [47]. NR1H4dependent proliferation of ER + /breast-cancer cells is stimulated by estrogen deprivation, which recapitulates menopause and aromatase-inhibitor treatment [47]. In spite of the evidence indicating that NR1H4 is oncogenic, particularly in ER + /breast-cancer, there is also evidence supporting the idea that NR1H4-activation is antioncogenic. High concentrations of the GW4064 agonist induce apoptosis of ER + /MCF-7 and ER -/MDA-MB468 cells [51]. Chenodeoxycholic-acid and GW4064 inhibit growth of tamoxifen-resistant ER + /cells. Interestingly, chenodeoxycholic-acid reduces EGF-induced growth of these cells via inhibition of the HER2 pathway [52]. On the basis of these data, it is currently difficult to establish whether NR1H4 is endowed with oncogenic or onco-suppressive properties in breast cancer and whether targeted therapeutic strategies should be aimed at inhibiting or activating the receptor.

NR1I2 (PXR:pregnane-X-receptor)
NR1I2 plays a role in xenobiotic metabolism and is activated by various synthetic ligands (Table 1) [53]. NR1I2 regulates the expression of genes involved in xenobiotic metabolism and transport [54]. NR1I2-mRNA highest levels are observed in liver although considerable amounts are measurable also in mammaryglands (UNIGENE-Hs.7303). Relative to normal tissue, NR1I2-mRNA is down-regulated in Her2 tumors, while the opposite is true in Basal tumors ( Figure 3). The TCGA data on the expression of the NR1I2 transcript in the PAM50 breast cancer subtypes are only partially in line with the reported inverse relationship between NR1I2-mRNA expression and ER-positivity [55]. NR1I2 represents a negative prognostic marker in breast-cancer, as NR1I2-protein levels correlate with labeling-index, histologic-grade and lymph-node-status [56]. In addition, NR1I2-protein over-expression and nuclear localization is associated with infiltrative-carcinoma recurrence [56]. NR1I2 plays a role in breast-cancer cell resistance to antitumor agents. In ER + /MCF-7 cells, NR1I2 is involved in induced resistance to tamoxifene via up-regulation of Multidrug-Resistance-Associated-protein-2, a membranetransporter controlling drug efflux [57]. In ER + /MCF-7 and ER -/MDA-MB231 cells, SR12813 causes docetaxel resistance and induction of the drug-resistance genes, Multidrug-Resistance-1 and Breast-cancer-Resistance-Protein B [58]. Thus, NR1I2-inhibition should be considered in the treatment of tumors with acquired resistance to anti-hormones and chemotherapy.
In mammary-glands, NR3B1 and NR3B2 are highly expressed, while NR3B3 levels are negligible ( Figure  4). Normal tissues and all PAM50-classified breastcancers express similar amounts of NR3B2-mRNA and NR3B3-mRNA. By converse, HER2-like and Basal-like mammary-tumors contain larger NR3B1-mRNA amounts than the normal tissue, which is consistent with a positive correlation between NR3B1 and HER2 expression [87]. In line with this, HER2/MAPK/AKT activation causes NR3B1 phosphorylation/activation in HER2 + /BT474 cells [88]. In addition, NR3B1 activates transcription of the genes contained in the ERBB2 amplicon observed in the majority of HER2 + breast tumors, possibly explaining the delay in tumor development observed following NR3b1 knock-out in a mouse model of ERBB2-initiated mammary cancerogenesis [89]. NR3B1 is a negative prognostic factor for breast-tumors, being associated with increased recurrence-risk and adverse clinical-outcome [90]. Consistent with NR3B1 oncogenic action, NR3B1antagonists reduce the size of ER + / and ER -/xenografts [91], while NR3B1 knock-down diminishes in-vitro www.impactjournals.com/oncotarget migration and in-vivo growth of ER -/MDA-MB-231 cells [91]. The pathways underlying NR3B1 pro-oncogenic action are obscure and may vary in ER + / and HER2 + / tumors. WNT inhibitors suppress NR3B1 transcriptional activity via β-catenin [92], reducing breast-cancer cells migration. In contrast, NR3B1-activation stimulates Vascular-Endothelial-Growth-Factor production and angiogenesis [93]. In ER + /breast-cancer cells, NR3B1 increases estrogen synthesis via aromatase induction [94]. As local synthesis of estrogens is fundamental for the growth of post-menopausal ER + /breast-cancer, NR3B1 inhibition may represent a therapeutic strategy in these patients. In breast-cancer, single study indicates that NR3B2 is a potential tumor-suppressor [95], while the potential tumor-suppressive activity of NR3B3 is supported by more studies [82,96,97]. NR3B3suppresses breast tumor growth and reverses the process of epithelialto-mesenchymal transition [97].
In summary, while NR3B1 suppression/inhibition is likely to be of therapeutic value in Her2, Basal and postmenopausal or tamoxifen resistant ER + /tumors, activation or induction of NR3B3 may represent a viable therapeutic strategy in ER -/tumors.
In normal mammary-glands, NR5A2 levels are much higher than NR5A1 levels. Basal cancer overexpresses NR5A1 relative to the other tumor subtypes and normal tissue. In contrast NR5A2-mRNA is downregulated in all breast-tumors, regardless of the PAM50 classification ( Figure 4). The strongest NR5A2-mRNA expression is observed in Luminal-A Luminal-B and Normal-like which may be consistent with NR5A2-gene control by ERα [99][100][101].
NR5A1 has never been the object of studies in breast-cancer while data on NR5A2 are available. NR5A2 knock-down inhibits estrogens proliferative action in ER + /MCF-7 cells and down-regulates ERα target-genes [99]. In ER + / and ER -/breast-cancer cells, NR5A2 is a mitogen and this action may involve NR5A2-dependent stimulation of Growth-Regulation-by-Estrogen-in-Breastcancer-1 (GREB-1) transcription [102]. In ER + /MCF-7 and ER -/MDA-MB231 cells, NR5A2 increases motility, a key process in metastatic spread [103]. Finally, NR5A2 may represent a negative prognostic marker, as a targetgenes signature is associated with poor outcome in highgrade mammary-tumors [104]. Thus, NR5A2 is likely to be an oncogene and reduction of NR5A2-antagonists may produce anti-tumor effects.

ORPHAN RECEPTORS
The fifteen members of the Orphan-receptors family are NRs for which endogenous-ligands are not identified. The sole exceptions are represented by NR1D1, NR1D2 and NR4A3 for which two endogenous-agonists are hypothesized.

NR2C1 (TR2:testicular-receptor-2) and NR2C2 (TR4:testicular-receptor-4)
NR2C1 and NR2C2 can form heterodimers and act as transcriptional activators/repressors of other NRs. NR2C1 and NR2C2 are involved in early embryonic development and embryonic stem cells [113]. NR2C1-mRNA (UNIGENE-Hs.108301) and NR2C2-mRNA (UNIGENE-Hs.555973) are expressed ubiquitously and are measurable in mammary-glands. NR2C1 and NR2C2 levels are lower in breast-cancer than normal tissue ( Figure 5). NR2C1 is over-expressed in Basal relative to Her2 and Normal-like tumors. In ER + /breast-cancer cells, NR2C1 suppresses ERα-mediated transcriptional activity, blocking ERα-binding to DNA via formation of an ERα-NR2C1 heterodimer. This inhibits estrogeninduced cell-growth and G(1)/S transition [114]. NR2C1 may indirectly contribute to ATRA anti-estrogenic activity in ER + /breast-cancer, as the retinoid controls the activity of this NR in other cellular contexts [115,116]. NR2C1 suppresses androgen-mediated AR transactivation, which may be of therapeutic interest in AR + /mammary-tumors [117]. The few available data do not provide clues as to the relevance of NR2C1 and NR2C2 in breast-cancer, but it can be suggested that the two receptors are anti-oncogenic.

NR2E1 (TLX:drosophila-tailless-homolog) and NR2E3 (PNR:photo-specific-nuclear-receptor)
NR2E1 plays a role in neuronal stem-cells homeostasis [118], while NR2E3 is involved in retinal visual function [119]. NR2E1-mRNA (UNIGENE-Hs.157688) and NR2E3-mRNA (UNIGENE-Hs.187354) expression is restricted to eye and brain and eye and muscle, respectively. Low levels of NR2E1-mRNA and NR2E3-mRNA are detectable in mammaryglands and NR2E1 is further down-regulated in breastcancers ( Figure 5). Higher NR2E1 expression levels are observed in Basal relative to Luminal tumors, while the opposite is true for NR2E3. Limited functional data on NR2E1 and NR2E3 are available and only few pertain to the breast-cancer realm. NR2E1 represses target-gene expression in neuronal-stem-cells [118], is an oncogene in glioblastoma and inhibits senescence in different cell types [120]. A recent and seminal paper by Lin et al. [121] supports the idea that NR2E1 is a potential drug target for the treatment of ERbreast cancer. In fact high levels of NR2E1 expression in ERare a negative prognostic factor in this type of cancer. Consistent with this targeted knock-down of NR2E1 inhibits the growth of different ERbreast cancer cell lines. By converse, overexpression of the nuclear receptor stimulates the formation of mammospheres, the growth and the invasive behavior of ER -MDA-MB231 cells [121]. NR2E3 is an ERα transcriptional activator [122] and it may contribute to hormone-dependent growth of ER + /tumors, although high NR2E3 levels are associated with favorable responses to tamoxifen. NR2E3 oncogenic action may extend to ER -/ tumors, inducing MDA-MB-231 cell migration in-vitro and metastatic spread in-vivo [122].
Smaller NR4A1-mRNA amounts are observed in all PAM50 mammary-tumor types relative to the normal gland ( Figure 5). Luminal-A and Normal-like show higher NR4A1 mRNA levels than Luminal-B tumors. NR4A1agonists induce apoptosis in mammary-tumor cells [129], although NR4A1 apoptotic activity is not necessarily related to NR4A1 transcriptional activity. In ER + /MCF-7 cells, the apoptotic action of a natural coumarin and Plexin-D1 [130] requires JNK-dependent phosphorylation of NR4A1 and translocation from the nucleus to the cytoplasm, where the NR binds and inhibits BCL-2. NR4A1 activation reduces breast-cancer cell-migration, [131], although NR4A1-silencing inhibits TGF-β-induced EMT suggesting an opposite effect. The effect on EMT is consistent with the observation that inflammatory cytokines induce NR4A1 and enhance TGF-β-dependent breast-cancer cell-invasiveness in-vitro and in-vivo. Thus, NR4A1 induction/activation may reduce breast-cancer growth, although this beneficial effect may be counterbalanced by increased metastatic-spread.
NR4A2-mRNA is down-regulated in all PAM50 sub-types relative to the normal mammary tissue and, consistent with data obtained at the protein level [132], Basal tumors contain the smallest NR4A2 amounts ( Figure 5). In primary breast-cancer, NR4A2 expression is inversely correlated with lymph-node metastases and directly correlated with increased relapse-free survival, suggesting onco-suppressive properties. In spite of this, NR4A2 silencing in Basal cell-lines decreases xenograft growth [133]. NR4A2 inhibits aromatase expression in mammary-gland stromal adipocytes [134] and this action may have implications for breast-cancer prevention, as obesity and estrogen production are breast-cancer riskfactors. The few data available in mammary-tumors support therapeutic strategies based on NR4A2-agonists. As the NR4A2 E-region is occupied by hydrophobic molecules, which prevents synthetic ligand accessibility [135], the observation that 6-mercaptopurine activates NR4As by targeting the N-terminal portion discloses new avenues in the design of agonists [136].
Although NR4A3-mRNA is generally downregulated in breast-cancer relative to the normal gland, Basal contain larger amounts than Luminal-A or Luminal-B tumors ( Figure 5), confirming the results of a study showing NR4A3 up-regulation in TN relative to Luminal breast-tumors [137]. This suggests an oncosuppressive role of NR4A3 in mammary-tissue, consistent with the NR4A3 growth-inhibitory action in other cellular contexts [138,139]. Onco-suppression in breast-cancer is supported by NR4A3 up-regulation during apoptosis in MCF-7 cells [140]. NR4A3 induction in MCF-7 cells by the cyto-differentiating agent, ATRA, is also consistent with NR4A3 onco-suppressive potential [141].

NR6A1 (GCNF:germ-cell-nuclear-factor)
NR6A1 plays a crucial role in embryonic-stem-cell (ESC) homeostasis [142]. In ESC, NR6A1 is a positive determinant of pluripotency being down-regulated by the differentiating-agent, ATRA [99]. In adults, NR6A1-mRNA tissue-specific expression is limited to ovary and testis (UNIGENE-Hs.586460). Mammaryglands contain very low amounts of NR6A1 which are generally up-regulated in breast-cancer ( Figure 5). The PAM50 subtypes showing the highest levels of NR6A1 up-regulation are Basal and Luminal-B tumors. The first observation is supported by a recent study indicating NR6A1 gene-expression enrichment in TN/ and ER + / tumors [143]. These data suggest that NR6A1 may be endowed with oncogenic properties in breast-cancer. www.impactjournals.com/oncotarget CONCLUSION Integration of the expression and functional data available allows a reliable prediction of the oncosuppressive or oncogenic role played by many of the NRs considered in breast-cancer (Tables 1, 2, 3). Within the Lipid-Sensors group, NR1C3, NR1H2 and NR1H3 are likely to play an onco-suppressive action. NR1F1, NR2A1 and NR3B3 (Enigmatic-Orphans) as well as NR0B1, NR0B2, NR1D1, NR2F1, NR2F2 and NR4A3 (Orphan-Receptors) seem to exert a similar activity. These NRs represent viable candidates for the development of therapeutic strategies aimed at increasing their expression or activating them in tumor cells. The availability of pharmacological agonists for NR1C3, NR1H2, NR1H3, NR1F1, NR0B2, and NR4A3 should boost pre-clinical studies in this direction. For all the remaining NRs, efforts should be oriented towards the design and synthesis of selective and high-affinity agonists. Except for NR1D1 and NR2F1, whose levels are significantly lower in Basallike relative to the other PAM50 subgroups, the expression profiles of the transcripts encoding the above mentioned NRs do not indicate any expression specificity in terms of breast-cancer subtypes. Nevertheless, as indicated in Table 2 and on the basis of the available literature, NR1F1 may be of particular significance as a therapeutic target in ERbreast cancer. The group of NRs endowed with potential oncogenic properties in breast-cancer is smaller and consists of the Lipid-Sensors, NR1C2 and NR1I2, the Enigmatic-Orphans, NR1F3, NR3B1 and NR5A2, as well as the Orphan-Receptors, NR2E1, NR2E3 and NR6A1. To obtain anti-tumor effects, oncogenic NRs should be targeted with selective antagonists, reverseagonists or agents/strategies capable of reducing their expression in breast-cancer cells. At present only synthetic antagonists targeting NR1C2, NR1I2 and NR1F3 are available. On the basis of the expression profiles in the PAM50 subgroups, we propose that NR1C2, NR1I2 and NR2E1 are pharmacological targets of particular interest in Basal-like tumors. In the case of NR2E1, this is line with the evidence present in the literature which indicates that targeting of the receptor is particularly promising in ERmammary tumors [121]. A similar relevance in ERis predicted also for NR1F3, as indicated in Table 2. Similar considerations suggest that NR1C2 and NR3B1 may be a useful targets in Her2 breast cancer as well (Tables 1,  2). Finally the interest of NR3B1 may not be limited to this last group, as it may extend to ERmammary tumors regardless of HER2-positivity (Table 2).
In the case of both onco-suppressive and oncogenic NRs there are some general points that should be considered and were touched upon in this review. Studies focusing on NR-targeting should be aimed at establishing the therapeutic potential in specific types of breast cancer given the heterogeneity of the disease. It is, indeed, highly unlikely that each of the identified NRs plays the same role in all mammary tumor-subtypes and targeting it results in similar anti-tumor effects, as recently demonstrated for the activation of RARs by ATRA and derived retinoids [5]. Thus, studies should not be limited to evaluating differential effects in the ER + / and ER -/cellular context, as traditionally done, but should take into consideration other identified breast cancer sub-groups such as the PAM50 classes used in this review. In addition, when different protein-variants of a specific NR are known, it is important to gather information as to the specific forms predominantly expressed in each breast-cancer subtype. In fact, this type of information is not available and should be gathered, as different protein-variants may have opposite effects in terms of oncogenic or onco-suppressive activity. Finally, the side effects potentially triggered by targeting a specific NR in breast-cancer should be considered before designing any targeted therapeutic strategy. With respect to this, a preliminary analysis of the data available on the tissue-distribution and physiological function of each NR is likely to be helpful in the selection of the specific NR to be targeted. www.impactjournals.com/oncotarget