Abstract
The three-orphan nuclear receptor 4A genes are induced by diverse stressors and stimuli, and there is increasing evidence that NR4A1 (Nur77), NR4A2 (Nurr1), and NR4A3 (Nor1) play an important role in maintaining cellular homeostasis and in pathophysiology. In blood-derived tumors (leukemias and lymphomas), NR4A expression is low and NR4A1−/−/NR4A3−/− double knockout mice rapidly develop acute myelocytic leukemia, suggesting that these receptors exhibit tumor suppressor activity. Treatment of leukemia and most lymphoma cells with drugs that induce expression of NR4A1and NR4A3 enhances apoptosis, and this represents a potential clinical application for treating this disease. In contrast, most solid tumor–derived cell lines express high levels of NR4A1 and NR4A2, and both receptors exhibit pro-oncogenic activities in solid tumors, whereas NR4A3 exhibits tumor-specific activities. Initial studies with retinoids and apoptosis-inducing agents demonstrated that their cytotoxic activity is NR4A1 dependent and involved drug-induced nuclear export of NR4A1 and formation of a mitochondrial proapoptotic NR4A1–bcl-2 complex. Drug-induced nuclear export of NR4A1 has been reported for many agents/biologics and involves interactions with multiple mitochondrial and extramitochondrial factors to induce apoptosis. Synthetic ligands for NR4A1, NR4A2, and NR4A3 have been identified, and among these compounds, bis-indole derived (CDIM) NR4A1 ligands primarily act on nuclear NR4A1 to inhibit NR4A1-regulated pro-oncogenic pathways/genes and similar results have been observed for CDIMs that bind NR4A2. Based on results of laboratory animal studies development of NR4A inducers (blood-derived cancers) and NR4A1/NR4A2 antagonists (solid tumors) may be promising for cancer therapy and also for enhancing immune surveillance.
Introduction
Background
The 48 human nuclear receptors (NR) play integral roles in maintaining cellular homeostasis and in pathophysiology and NR subfamily 4 (NR4A) consists of three-orphan NRs for which there are no known physiologic ligands (1, 2). NR4A1 (Nur77), NR4A2 (Nurr1), and NR4A3 (Nor1) exhibit domain structures similar to other NRs; this includes N- and C-terminal domains containing activation function 1 (AF-1) and AF-2 [also ligand binding domain (LBD)], respectively, and they flank a DNA-binding domain (DBD) and a hinge region (Fig. 1). The sequence homology of NR4A1, NR4A2, and NR4A3 is similar in the ligand binding AF-2, hinge, and DBD but differ significantly in the N-terminal AF-1 domain (3–5) and there is evidence that this domain dictates some of the different functions of these orphan NRs (6–8). NR4A1, NR4A2, and NR4A3 are early immediate genes that exhibit overlapping and unique functions, and they characteristically are induced by diverse physical, physiologic, and pharmacologic stimuli (reviewed in refs. 9, 10). In many solid tumor-derived cancer cells that exhibit enhanced metabolic rates, NR4As are overexpressed compared with corresponding nontransformed cells, whereas most blood-derived tumors are characterized by low expression of NR4A (11, 12) and these differences will be discussed below.
Domain structure of NRs and similarities between NR4A1, NR4A2, and NR4A3.
NR4A interactions with cis-elements
NRs are ligand-activated transcription factors that bind their endogenous ligands (e.g. hormones) or synthetic ligands and the ligand-bound receptor interacts with cis element in target gene promoters (1, 2). This interaction can lead to recruitment of nuclear cofactors and results in modulation of gene expression. In addition, some receptors can act as ligand-activated nuclear cofactors (11). Structurally diverse ligands induce different conformational changes in the bound receptor and these selective receptor modulators can induce tissue- and gene-specific receptor agonist or antagonist activity. The NR4A subfamily are orphan receptors with no known endogenous ligands, and they act through both nuclear and extranuclear pathways and can be influenced not only by synthetic receptor ligands but also by other agents that do not bind the receptor (12). X-ray crystallographic and functional studies of NR4A suggest that the ligand binding pocket contains bulky hydrophobic amino acid side chains that may preclude interactions with an endogenous ligand (5, 8, 13, and 14). There is extensive evidence that NR4A alone or in combination activates ligand-independent gene expression through direct or indirect interactions with cognate cis-elements (4, 15, 16), NR4As activate gene expression through binding as monomers, homodimers, and heterodimers (with RXR) through interactions with an octanucleotide NGF1-β response element (NBRE), a Nur-responsive element (NuRE), and a DR5 motif (with RXR), respectively (Fig. 2; refs. 17–19). NR4A1, NR4A2, and NR4A3 can also form heterodimers and only NR4A1 and NR4A2 but not NR4A3 heterodimerize with RXR (20, 21). Recent high-throughput studies showed that NR4A2-bound NuRE motifs that consist of two everted palindromic octanucleotides (ERO) with no spacer between the NBREs and two inverted repeats separated by 5 nucleotides (IR5; ref. 22). Crystal structures of NR4A2–DBD interactions with ERO and IR5 have been identified as NR4A2 binding sites in multiple human genes and are structurally different from the “classical” NuRE identified in the pro-opiomelanocortin (POMC) gene (23). Analysis of NR4A3 interactions and chromatin immunoprecipitation (ChIP)-seq and identification of other NuREs that bind NR4A2 and NR4A3 have not been reported (23) and future studies using this approach will provide some basis for understanding functional and mechanistic differences between NR4A1, NR4A2, and NR4A3. These results are consistent with direct interactions of most NRs with their cognate consensus and non-consensus response elements. There is also evidence that NR4A1 interacts with DNA-bound specificity protein (Sp) transcription factors and act as a cofactor of Sp1 or Sp4. ChIP analysis showed that NR4A1 interacted with Sp1 or Sp4 bound to GC-rich promoter sequences in the survivin, β1-, β3-, and β4-integrins, PAX-FOX01 and α5- and α6-integrin genes (24–27) and acts as a cofactor for Sp-dependent gene expression. NR4A1/Sp-regulated gene expression is decreased by knockdown of Sp1/4 and also NR4A1, and there is evidence from RNA-seq studies in rhabdomyosarcoma (RMS) cells that many key pro-oncogenic factors are regulated by an NR4A1/Sp complex. This pathway for gene regulation is not unique to NR4A1 and has previously been observed for steroid hormone receptors and several RXR-binding receptors (28). Thus, activation of gene expression by NR4A involves interactions with different partner proteins and different cis-elements, and this has been observed for other NRs. A unique feature of NR4A (particularly NR4A1) in cancer is that many apoptosis-inducing agents induce nuclear export of NR4A1, and this unusual pathway will be discussed below.
Interactions of NR4A1 with cis-elements, DNA-bound RXR, and DNA-bound-Sp.
NR4A knockout mice and cancer
Individual knockout of individual NR4As in mice did not initially indicate a role for these receptors in carcinogenesis. Both NR4A1 and NR4A3 play comparable roles in induced apoptosis of β cells and negative selection of T-lymphocytes; however, NR4A1−/− mice were viable with no obvious phenotype (29–31). Studies with NR4A2−/− mice show the importance of the receptor for induction of the dopaminergic phenotype, and these mice exhibit significant neuronal dysfunction and early mortality (32, 33). NR4A3−/− mice were generated in two laboratories, and their phenotypes were different. One study observed embryo lethality due to a failure to complete gastrulation (34) and the other report indicate that loss of NR4A3 resulted in inner ear defects (35). The role of NR4A1 in tumorigenesis was investigated in mouse models by comparing development of mouse cancer cell implants or xenografts in the presence or absence of NR4A1. For example, in a syngeneic mouse model using B16 melanoma cells, the loss of NR4A1 enhanced tumor invasion and metastasis due to increased secretion of TNFα but decreased CSF-1R expression and tumor-infiltrating migratory activity (36). In contrast, NR4A1 enhanced tumor growth in mice bearing B16F1 melanoma cells and in these cells NR4A1 enhanced angiogenesis through regulation of VEGF expression (37). Expression of NR4A1 in mouse MV3 melanoma cells enhanced circulating tumor cell survival and metastasis (38). Similar results were observed in in vivo and in vitro studies using LLC and CMT93 colon cancer cells and the loss of NR4A1 in mice resulted in decreased tumor growth and metastasis (39). Thus, with the exception of APCmin/+ mice where NR4A1 loss resulted in enhanced intestinal tumors (40), most studies suggest that NR4A1 exhibits pro-oncogenic activity in solid tumors.
Role of NR4A in Blood-Derived Tumors
Decreased NR4A and development of leukemias
Mullican and coworkers first reported the remarkable phenotypic effects observed in combined NR4A1−/−/NR4A3−/− double knockout mice; the mice were smaller in size than their wild-type counterparts and all died within 3 to 4 weeks after birth from symptoms consistent with acute myeloid leukemia (AML; ref. 41). This was accompanied by expansion of myeloid progenitors and hemopoietic stem cells, decreased expression of c-Jun, Jun-B, and proapoptotic proteins such as Fas-L and TRAIL. Leukemic blasts from AML patients also exhibited low to nondetectable levels of NR4A1 and NR4A3, and similar results were observed in various leukemia-derived cell lines (41), which complemented the in vivo results and are consistent with tumor suppressor-like activity of NR4A1 and NR4A3 in combination. These observations were supported by an elegant study on the cancer outcomes of mice with “reduced NR4A gene dosage” (42), and these included NR4A1+/+/NR4A3+/+ (wild-type), single knockouts (NR4A1+/+/NR4A3−/−, NR4A1−/−/NR4A3+/+), and knockout/heterozygotes (NR4A1+/−/NR4A3−/−, NR4A1−/−/NR4A3+/−). Analysis of peripheral blood and other histologic markers showed that wild-type and single knockout mice were normal whereas the knockout/heterozygote mice exhibited features consistent with mixed myelodysplastic/myeloproliferative neoplasms (MDS/MPN). Like the double knockout mice, NR4A1−/−/NR4A3+/− mice also exhibited AML. The MDS/MPN mice exhibited some changes in gene expression observed for NR4A1−/−/NR4A3−/− mice, including decreased expression of Jun-B, egr1, and polo-like kinase 2 (Pik2).
Role of NR4A in lymphomas
The expression and role of NR4A has also been investigated in lymphomas, and it was reported that decreased expression of nuclear NR4A1 and NR4A3 was observed in patients with follicular lymphoma (FL) and diffuse large β-cell lymphoma (DLBCL) compared with their cells of origin. Decreased NR4A1 levels were associated with aggressive forms of FL and DLBCL and poor overall patient survival, whereas NR4A2 was detected only in some samples and levels were similar in tumor and nontumor tissue (43). NR4A3 is overexpressed in diffuse large B-cell lymphoma patients who responded favorably to chemotherapy, whereas overexpression was not observed in patients who did not respond to therapy (44). Decreased NR4A1/NR4A3 expression was associated with decreased expression of apoptotic genes such as TRAIL, Puma, and Bim, and in SuDHL4 lymphoma cells transfected with an NR4A1 expression plasmid there was a dramatic increase in apoptosis, and this was accompanied by induction of proapoptotic genes TRAIL, Bim, and Puma. An extensive analysis of DLBCL patients and subtypes showed that in germinal center B-cell-like subtype, increased patient survival was associated with high cytoplasmic NR4A1, and genomic analysis indicates that this was associated with the ERK1/2 pathway (45). In another study (46), it was reported that in aggressive lymphomas, decreased expression of NR4A3 was also associated with poor patient survival. Moreover, overexpression of NR4A3 in lymphoma cell lines induced apoptosis, suggesting that in both leukemias and lymphomas, the tumor suppressor-like activities of NR4A1 and NR4A3 are associated with their regulation of proapoptotic genes. Expression of NR4A3 is correlated with increased overall and event-frees survival of pediatric Pre-B-ALL patients but not the overall survival of AML patients (47). In contrast, analysis of NR4A1 expression in mantle cell lymphoma (MCL) showed that receptor was primarily located in the nucleus, and levels were higher than in normal B cells from lymph nodes or tonsils (48). Moreover, there was a strong correlation between NR4A1 and Bruton tyrosine kinase expression, which is a key pro-oncogenic contributor to MCL. Knockdown of NR4A1 in MCL cells did not affect cell viability but enhanced drug (ibrutinib)-induced cell killing and genomic analysis after NR4A1 knockdown was consistent with the pro-oncogenic activity of NR4A1 in MCL cells and thus differed significantly from results obtained in leukemia and other types of lymphoma cells.
Drug/ligand-induced responses
With the exception of MCL, there is evidence that the loss of NR4A1 and NR4A3 contributes to the development and expansion of blood-derived leukemias and lymphomas, and the mechanism of silencing and drug-induced activation of the receptors has been investigated (49, 50). Some studies report that histone deacetylase (HDAC) inhibitors induce expression of NR4A1 and NR4A3 in leukemia cells resulting in activation of proapoptotic pathways/genes, and this is accompanied by enhanced histone acetylation (51). This effect of gene dosage was also observed in in vivo studies (42). Genome-wide mapping of NR4A1 binding sites in human AML cells showed that NR4A1 targets 685 genes and regulates transcription in cooperation with distal ETS enhancers such as ERG and FLI-1, which promote recruitment of p300 (acetyltransferase), resulting in enhanced H3K27 acetylation (52). Analysis of the epigenetic status of NR4A1 and NR4A3 promoters showed that both promoters contained high levels of H3K4me3; however, comparative studies suggested that this was not an inhibitory mark, and this was consistent with association of pol II occupancy on both promoters (53). A chemical screening assay identified dihydroergotamine (DHE), a drug that enhances NR4A expression and inhibits AML cell growth. Mechanistic studies show that DHE acts through a mechanism that reverses the promoter-paused pol II, resulting in recruitment of the super elongation complex and increased elongation and enhanced gene expression (Fig. 3; ref. 53). A recent study showed that DHE also targeted super enhancers of pro-oncogenic factors in leukemia, and this includes MYC and results in decreased H3K27 acetylation, thus providing a novel pathway for modulation of MYC expression in leukemia (54). Fenritinimide also induced expression of NR4A1 in AML cells, and this was accompanied by nuclear export of NR4A1, interaction with bcl-2 resulting in induction of apoptosis (55). This pathway is similar to that observed in solid tumor–derived cancer cells and will be further discussed in a subsequent section of the review. Thapsigargin induces NR4A3 levels in lymphoma cells, and this drug-induced response mirrors the effect of NR4A3 overexpression resulting in inhibition of cell growth and induction of apoptosis (46). Cytosporone B (Csn-B) has been characterized as an NR4A1 ligand (56) and Csn-B induced apoptosis in lymphoma and immortalized B cells, thus demonstrating potential drug efficacy of a compound that binds NR4A1 (41).
Basal and induced NR4A1 in blood-derived tumor—low expression of NR4A is observed in most blood-derived tumors due to promoter-paused RNA Pol II, and DHE induces NR4A tumor expression through recruitment of the super elongation complex (49). There is also evidence for cytoplasmic NR4A in some lymphomas (42) and the retinoid Fenretinimide also induced nuclear export of NR4A1, which formed a complex with bcl-2 (51).
In contrast, treatment of MCL cells with 1,1-bis(3′-indolyl)-1-(p-hydroxyphenyl)methane (DIM-C-pPhOH, CDIM8), a bis-indole derived NR4A1 ligand that acts as an antagonist in solid tumor–derived cells (57) also inhibited growth and synergistically enhanced ibrutinib-induced cytotoxicity. Moreover, in MCL cells, Csn did not affect cell viability, further confirming that in MCL cells NR4A1 exhibits pro-oncogenic activity (48). In summary, there is evidence that NR4A1 and NR4A3 exhibit tumor suppressive-like activities in leukemias and some lymphomas, and drug-induced expression of these receptors is a potential treatment strategy. This is an area of research with future potential for clinical applications using NR4A-inducing agents. MCL is the exception to these observations because NR4A1 exhibits tumor promoter–like activity and the underlying mechanisms that dictate these differences are not well defined. The tumor suppressor–like activity in most blood-derived cancers precludes the use of receptor ligands due to low NR4A expression, and this contrasts with most studies in solid tumors as outlined below.
Role of NR4A in Solid Tumors
The expression, prognostic value, function, compound/ligand effects, and mechanisms of action of NR4A receptors have been extensively investigated in solid tumors. However, there has been significantly more research on NR4A1 compared with NR4A2 or NR4A3. With one exception, most in vivo studies indicate that NR4A1 (58) is pro-oncogenic in solid tumors, indicating significant differences between the tumor suppressor-like activity of NR4A1 in blood-derived cancers, and the oncogenic-like activity of NR4A1 and NR4A2 in solid tumors.
NR4A expression, prognostic value, and functions
In contrast to blood-derived cancers, there is extensive evidence that NR4A1 is overexpressed in patients with multiple tumor types including breast, lung, pancreatic, ovarian, colon, endometrial, cervical and gastric cancers, rhabdomyosarcomas, and melanomas (24, 26, 59–68). Moreover, high expression of NR4A1 in lung, breast, ovarian, and colon cancers predicts poor patient survival or prognosis (60, 64–66), although a few prognostic studies are conflicting (69–74). The most convincing and consistent evidence demonstrating the pro-oncogenic activities of NR4A1 are results of gene silencing studies in solid tumor–derived cell lines which show that NR4A1 regulates one or more of cell proliferation, survival, migration/invasion, and in some cells epithelial–mesenchymal transition (Fig. 4). These effects have been observed in breast, colon, pancreatic, kidney, lung, rhabdomyosarcoma, melanoma, endometrial cancer cells (57, 59–61, 63, 65, 67, 69, 75–79). Although the pathways/genes associated with NR4A1 are complex and cell context specific, Fig. 4 illustrates the role of NR4A1 in most solid tumor–derived cancer cell lines. This includes an important function for NR4A1 in TGFβ-induced invasion of breast and lung cancer cells (65, 80, 81). Studies in our laboratory showed that two NR4A1-regulated pro-reductant genes, namely, thioredoxin domain-containing 5 (TXNDC5) and isocitrate dehydrogenase 1 (IDH1), were important for maintaining relatively high mTOR signaling and for decreasing intracellular reactive oxygen species (ROS) and ER stress (24, 25, 57). Knockdown of NR4A1 by RNA interference decreases TXNDC5 and IDH1 expression, resulting in the induction of ROS, and this is accompanied by activation of ER stress. Induction of ROS also inhibits mTOR signaling through induction of sestrin 2 (ROS-dependent), which activates AMPK, resulting in mTOR inhibition (Fig. 4). This is observed in pancreatic, breast, lung, kidney, and RMS cancer cells (25, 57, 60, 77–79) and also in endometriotic cells where mTOR signaling is also inhibited by NR4A1 knockdown (82). A tumor-specific effect of NR4A1 is observed in alveolar RMS (ARMS) where the unique PAX3–FOX01 fusion oncogene important for ARMS cell growth is also an NR4A1-regulated gene (26). As indicated in Fig. 4, NR4A1 acts as a nuclear transcription factor or nuclear coactivator to modulate target gene expression. Most studies show that NR4A2 is also pro-oncogenic in solid tumor–derived cell lines, and like NR4A1 plays a role in cancer cell proliferation, survival, and migration/invasion (76, 83–94). For example, overexpression of NR4A2 in colon cancer cells enhanced chemoresistance, and expression of NR4A2 was enhanced in colon cancer patients and was a negative prognostic factor (91). A recent study also reported that NR4A2 expression was a negative prognostic factor for glioblastoma patients, and knockdown of the receptor by RNAi resulted in decreased growth, survival, and invasion (94). Acinic cell carcinoma is a salivary gland tumor that exhibits specific rearrangements [+(4;9)(q13;q31)], which result in enhanced expression of NR4A3 (95, 96). Immunostaining of these carcinomas (97–101) shows that NR4A3 is highly expressed in most of these tumors (63/64), whereas NR4A2 (1/64) but not NR4A1 were also overexpressed compared with other salivary gland tumors (100, 101). NR4A3 exhibits pro-oncogenic activity in acinic cell carcinomas, and this is associated with increased cell proliferation and activation of NR4A-regulated genes and cooperative effects with MYB oncogene (95, 96). NR4A3 knockdown did not affect the phenotype in glioblastoma cells (94), whereas other studies show that NR4A3 exhibits tumor suppressor-like activity (102–105). Thus, in solid tumors, NR4A1 and NR4A2 exhibit pro-oncogenic, and NR4A3 exhibits tumor suppressor-like activities, and this contrasts with most blood-derived cancers where NR4A1/NR4A3 (combination) and possible NR4A2 are tumor suppressors.
NR4A-regulated pathway/gene expression in solid tumors. These results were derived primarily from knockdown studies (of NR4A1) and were observed in multiple cancer cell lines (55–59, 61, 63, 65, 71–75).
Although NR4As are structurally similar and by definition are trans-acting factors that modulate gene expression (e.g., Fig. 4), their paradoxical cell context-dependent activities and differential effects of drugs/ligands are due in part to their intracellular interactions with other factors. Kurakula and coworkers (106) summarized the interactome for NR4A1, NR4A2, and NR4A3, and identified 64, 25, and 13 interacting factors, respectively, with only limited studies on interactions with NR4A2 and NR4A3. The NR4A interaction surfaces include all receptor domains; the three receptors interact with many different factors but only with a few proteins in common and this contributes to their different cell/tissue context-dependent activities. In addition, another key distinguishing feature of NR4As is that after activation by agents, ligands, or stimuli, their effects are due to both nuclear and/or extranuclear NR4A. Among the first reports on the role of NR4A1 in cancer were studies showing that the effects of structurally diverse apoptosis-inducing agents were due to the nuclear export of NR4A1 (107), whereas this has not been a distinguishing feature of NR4A2 or NR4A3. The following sections of the review will focus on agent/stimuli-induced extranuclear and nuclear activation/inactivation of NR4A with most of the published studies on NR4A1.
Extranuclear NR4A action
There is extensive evidence from knockdown and overexpression studies that NR4A1 regulates cancer cell proliferation, survival, and migration/invasion (Fig. 4), and studies with retinoids and other apoptosis-inducing agents identified an important proapoptotic role for NR4A1 (107). Initial reports showed that the retinoid 6-[3-(1-adamantyl)-4-hydroxyphenyl]-2 naphthalene carboxylic acid (AHPN, CD437) induced both retinoid receptor-dependent or -independent apoptosis in different cancer cell lines (107–113). Moreover, there was also evidence that AHPN-mediated induction of apoptosis in cancer cells was dependent on NR4A1 but did not require the DBD of the receptor. This response was blocked by leptomycin B, a nuclear export inhibitor, and it was shown that AHPN-induced apoptosis was due to nuclear export of NR4A1 and mitochondrial targeting of the receptor (111). Similar effects were observed in cancer cells treated with many other structurally diverse compounds including phobol esters, etoposide, cadmium, cholic acid derivatives, etoposide, HDAC inhibitors, dibutyltin derivatives, coumarin analogues, bile acids, oxidized analogues of bis-indole derived compounds, acetylshikonin analogues, and n-butylenephthalide and related compounds (111–124). Many of these agents not only induced nuclear export and apoptosis in cancer cells but also increased overall levels of NR4A1. Mechanistic studies with AHPN revealed that extranuclear NR4A1 interacted with mitochondrial bcl-2, resulting in formation of a proapoptotic complex that induced mitochondrial disruption, cytochrome c release, and activation of the intrinsic apoptosis pathway (Fig. 5; refs. 111, 113). Subsequent studies report that NR4A1 binding to bcl-2 requires the loop region between the BH4 and BH3 domains of bcl-2 (113) and a site adjacent to the BH3 peptide binding crevice was recently shown to be involved in NR4A1–bcl-2 binding (125). Paclitaxel and a short NR4A1 peptide that interacts with bcl-2 mimic the proapoptotic effects of NR4A1 (126, 127), suggesting that design of small molecules that target the NR4A1-interacting sites of bcl-2 represents a novel class of apoptosis-inducing agents (126, 127). The extranuclear proapoptotic functions of NR4A2 have not been reported; whereas transfection of NR4A3 into breast cancer cells induced apoptosis and interactions with bcl-2 were also observed (102). Thus, identification of ligand-dependent nuclear export of NR4A3 may also have some clinical potential for killing cancer cells (Fig. 5). There is also evidence that nuclear export of NR4A1 plays an integral role in many other pathways in cancer cells. Several agents, including butyrate, sulindac, and 5-fluorouracil, also induce nuclear export of NR4A1, which is accompanied by induction of bax, cytochrome c release, and apoptosis in colon cancer cells; however, NR4A1 did not target the mitochondria in these studies (Fig. 5; ref. 128). It was assumed that other cytosolic factors were involved in the induction of apoptosis. Insulin-like growth factor binding protein 3 (IGFBP3) interacts directly with NR4A1, resulting in nuclear export and mitochondrial targeting of the receptor, and this is associated with activation of Jun N-terminal kinase (JNK) and inhibition of Akt (129, 130).
Drug/ligand/TGFβ-induced nuclear export of NR4A1 and mitochondrial/cytosolic interactions; various inputs are indicated (1–12) and resulting outputs are also illustrated as 1–12 (circled). NR4A1 ligands such as Celastrol and THPN also bind the receptor. The drug/agent-induced outputs illustrate the multiple interaction of extranuclear NR4A with mitochondrial targets as well as the endoplasmic reticulum and a proteasome complex.
In melanoma cells, fatty acid oxidation (FAO) is important for generating ATP and cellular oxidants, and this is observed in cells maintained under low glucose conditions. Glucose deprivation also resulted in nuclear export of NR4A1 to mitochondria and formation of a complex with TPβ a subunit of a mitochondrial functional protein (131). TPβ is important for FAO and NR4A1 protected TPβ from oxidation and thereby enhanced FAO and survival of melanoma cells maintained in low glucose medium (38). The novel NR4A1 binding compound 1-(3,4,5-trihydroxyphenyl)nonan-1-one (THPN) induces NR4A1 nuclear export to the mitochondria in melanoma cells, where it interacts with Tom40 and Tom70, resulting in disruption of the VDAC1 transition pore complex and activation of autophagic cell death pathways (132). Two recent studies identified 2-imino-6-methoxy-2H-chromene-3-carbothioamide (IMCA; ref. 131) and celastrol (133), a naturally occurring triterpenoid as NR4A1 ligands and the former compound induced nuclear export to mitochondria, possibly inducing apoptosis via interaction with bcl-2. Celastrol is a potent anticancer agent, and treatment of liver cancer cells with this compound induces nuclear-to-mitochondrial translocation of NR4A1, which interacts with TRAF2; NR4A1 is ubiquitinated and interacts with p62/SQSTM1 resulting induction of autophagy (134). Plexin D1 is a receptor that interacts with its ligand Sema 3E to promote breast cancer survival, and the Sema domain of Plexin D1 (SD1) acts as ligand binding trap and activates cell death pathways (135). Unliganded Plexin D1 interacts with cytosolic NR4A1 to induce apoptosis; however, the precise mechanism of this response and the role of bcl-2 was not defined. Phorbol esters induce ER stress in liver cancer cells, and this is associated with nuclear export of NR4A1 and formation of an NR4A1-translocation-associated protein subunit γ (TRAPγ) complex (135). TRAIL-induced apoptosis in liver cancer cells involves NR4A1 nuclear export, formation of an NR4A1–bcl-2 mitochondrial complex which was dependent on interaction of NR4A1 with interferon stimulated gene 12a (ISG12a) for both nuclear export and enhanced bcl-2 interactions (136). Two digitalis-like compounds induced NR4A1 expression in colon cancer cells, and this was accompanied by nuclear export of NR4A1 to the cytosol, where it interacts with β-catenin resulting in degradation of β-catenin (137). These examples of agent-induced nuclear export of NR4A1 (Fig. 5) illustrate the diverse pathways for nuclear export of NR4A1 and cell killing which are due to both mitochondrial disruption and nonmitochondrial effects. In contrast, TGFβ-induced invasion and metastasis of breast and lung cancer cells is an example of a pro-oncogenic pathway that is also dependent on NR4A1 and its nuclear export (65, 80, 81). This process involves TGFβ-induced nuclear export of NR4A1, which forms a cytosolic complex containing NR4A1, axin2, Arkadia, and RNF12, which are necessary for proteasome-dependent degradation of the inhibitory SMAD-7. The loss or decrease of SMAD7-dependent inhibition of TGFβ signaling enhances TGFβ-induced cancer cell invasion. Most of the pathways associated with nuclear export of NR4A1 involve activation/inactivation of kinases, which can be modulated by various kinase inhibitors; however, the role of individual kinases is both agent- and cell context dependent. For example, TGFβ-induced nuclear export of NR4A1 in breast cancers cells is p38-MAPK14–dependent and blocked by p38 inhibitors. Whereas this same response is activated by c-JNK in lung cancer cells and blocked by JNK inhibitors. TGFβ-dependent phosphorylation of NR4A1-Ser351 was observed after activation of different kinase pathways that induce nuclear export of NR4A1 in both breast and lung cancer cells (80, 81). Phosphorylation of nuclear NR4A1 is important for nuclear export of the receptor and a recent study suggests a possible role for NR4A1 sumoylation in this process (138). These results demonstrate the remarkable and highly variable effects of agent/biological-induced nuclear export of NR4A1, which interacts with multiple factors to induce apoptosis but also facilitates TGFβ-induced invasion of cancer cells (Fig. 5).
Nuclear functions of NR4As and their ligands
RNA interference and overexpression experiments demonstrate that in the absence of ligands, NR4A1 regulates pro-oncogenic pathways and genes associated with cancer cell proliferation, survival, and migration/invasion (Fig. 4), and in limited studies similar results have been observed for NR4A2. In contrast, NR4A3 exhibits oncogenic and tumor suppressor-like activity and has not been extensively studied in cancer. The identification of NR4As with important functional activities in cancer and other diseases has spurred development of receptor ligands that target NR4A as either agonists or antagonists. Wu and coworkers first identified cytosporone β (Csn-B) as an NR4A1 ligand (Fig. 6); however, Csn-B also induced apoptosis in cancer cells via nuclear export of the receptor to the mitochondria, and this was also observed for THPN (55, 56, 132). Several Csn-B analogues also bound NR4A1 with KD values in the low micromolar range and not only induced mitochondrial localization of NR4A1 but also acted within the nucleus to repress expression of brain and reproductive organ-expressed protein (BRE) in gastric cancer cells (56). BRE is an NR4A1-regulated antiapoptotic/survival gene, indicating that in gastric cancer cells Csn-B and its analogues are acting as an antagonist for this gene. In contrast, treatment of gastric cancer cells with Csn-B and related compounds induced activity in cells transfected with an NR4A1-responsive construct. Thus, the Csn-B compounds act as selective NR4A1 modulators, and their agonist or antagonist activities are gene- and function dependent, and this is typical of selective receptor modulators for many other nuclear receptors (139).
Structures of ligands that bind NR4As. The KD values for receptor binding by cytosporone B (1.68 μmol/L; ref. 56), celastrol (292 nmol/L; ref. 133), and chloroquine (0.27 μmol/L; ref. 151) have been reported. KD values for prostaglandin A2 and CDIM12 have not been reported and the KD value for CDIM8 is 0.56 μmol/L (unpublished results).
Studies in the Safe laboratory have focused on bis-indole derived compounds (CDIM), which bind NR4A1 and 1,1-bis(3′-indolyl)-1-(p-hydroxyphenyl)methane (CDIM8; DIM-C-pPhOH), and the p-carbomethoxyphenyl derivatives have been used as representative NR4A1 ligands (77–82). These compounds both induce and repress NR4A1-regulated genes but in terms of their functional responses they antagonize the nuclear NR4A1-regulated pro-oncogenic pathways/genes illustrated in Fig. 4. Treatment of many solid tumor–derived cell lines with NR4A1 antagonists or knockdown of NR4A1 inhibits the pathways and gene illustrated in Fig. 4. However, genomic or proteomic analysis demonstrate significant cell context–dependent differences in gene expression in pancreatic, kidney and breast cancer, and rhabdomyosarcoma cells (24, 26, 79, 140). CDIM/NR4A1 antagonists also inhibit TGFβ-induced nuclear export of NR4A1 and cell invasion in lung and breast cancer cells, suggesting that binding of the CDIM ligand inhibits some elements of nuclear export process (81, 82). 1,1-Bis(3′-indolyl)-1-(4-chlorophenyl)methane (CDIM12) is an NR4A2 ligand, and molecular modeling studies indicate that CDIM12 does not directly interact with the ligand binding. Modeling studies suggest that CDIM12 binds the coactivator region of NR4A2, and this needs to be further investigated (141). Although transactivation studies in pancreatic cancer cells show that CDIM12 activates multiple genes and NR4A2-responsive constructs (142, 143), CDIM12 exhibits functional antagonist activities and inhibits cancer cell growth and survival (94, 144). Prostaglandin E2 (PGE2) is a cyclooxygenase-2 (COX-2)–derived gene product that is pro-oncogenic in colon cancer and induces NR4A2 in both in vivo and in vitro models. The resulting changes in gene expression include enhanced osteopontin expression and FAO (145–147). PGE2-mediated activation of NR4A2 enhanced NR4A2/RXR-mediated expression of prolactin in stromal cells, and the tumor–stromal prolactin signaling initiates prostate cancer, and this is blocked by COX-2 inhibitors (148). Prostaglandin A2 (PGA2) has been identified as an NR4A3 ligand (149) (Fig. 6) and induces apoptosis and inhibits growth of breast and cervical cancer cells (150). In some cells, it is possible that PGA2 enhances NR4A3-dependent tumor suppressor activity and acts as an agonist (151).
The high expression of NR4A receptors in solid tumors makes them potential targets for receptor ligands, and this is supported by studies on NR4A1 where Csn-B and related compounds bind and induce nuclear export and apoptosis (Figs. 5 and 6) whereas CDIMs primarily inactivate nuclear NR4A1-regulated pro-oncogenic pathway/genes (Figs. 4 and 6). Ligands for NR4A2 and NR4A3 have not been extensively investigated (Fig. 6) but may also be effective as anticancer agents. Applications of these selective NR4A modulators do not have to be confined to cancer because some of these compound (e.g., CDIMs) show promise for treating endometriosis (82), Parkinson's disease (141, 152, 153), and for enhancing learning and memory (154, 155).
NR4A in Cancer Immune Surveillance
NR4A receptors are immediate early genes induced by T-cell receptor signaling and play an important role in T-cell development and immune responses (156). NR4A is expressed in Tregs and in CD4+/CD8+ T cells, and levels of NR4A dictate, in part, responsiveness to immunotherapy targeting tumor and immune cell checkpoints, and in CAR-T cell therapy (157–159). Analysis of tumor-infiltrating lymphocytes (TIL) shows that in exhausted T cells, NR4A is overexpressed, and this is accompanied by enhanced expression of PD-L1 and Tim3, decreased expression of cytokines, and low levels of cell killing (159). Loss of NR4A receptors partially reversed exhaustion resulting in tumor regression and increased survival (160). Hibino and coworkers (161) showed comparable results in mice lacking NR4A1 and NR4A2; moreover, after treatment or wild-type mice with campothecin or the cyclooxygenase-2 (COX-2) inhibitor SC-236 to decrease NR4A1 levels, there was also a decrease in tumor volumes and an increase in CD8+/CD4+ ratios (161). Thus, loss of NR4A1 by genetic or pharmacologic means enhanced immune surveillance. This was also observed using 1,1-bis(3′-indolyl)-1-(3-chloro-4-hydroxy-5-methoxyphenyl)methane (CDIM8–3-Cl-5-OCH3), which inhibited mammary tumor growth in both xenograft and syngeneic mouse models (140, 162). Moreover, treatment with the CDIM/NR4A1 antagonist enhanced CD8+/CD4+ ratios in TILs, and this was primarily due to decreased levels of CD4+. Thus, NR4A1 antagonists represent a novel class of drugs that enhance immune surveillance in a syngeneic mouse model using mouse 4T1 breast cancer cells, and mechanistic studies also showed that PD-L1 is an NR4A1/Sp1–regulated gene that is downregulated by CDIM/NR4A1 antagonists (162).
Summary
NR4A are structurally related NRs that activate gene expression through targeting common cis-elements. NR4A1, NR4A2, and NR4A3 are activated by diverse stressors; however, their roles in cancer are diverse and paradoxical. In leukemias and most lymphomas, NR4A1 and NR4A3 are tumor suppressor genes, whereas the role of NR4A2 is not well defined. The mechanisms of NR4A-mediated genes and responses involve nuclear NR4A and require further investigation. In solid tumors, both NR4A1 and NR4A2 are tumor promoter-like genes, whereas NR4A3 exhibits tumor suppressor and promoter activities. NR4A1 has been intensively studied in solid tumor–derived human cancer cell lines and exhibits activity as a nuclear transcription factor and a nuclear cofactor. In addition, drug-induced nuclear export of NR4A1 results in the induction of apoptosis, and this nuclear export pathway has primarily been observed for NR4A1 and not for NR4A2 or NR4A3. The extranuclear proapoptotic activity of NR4A1 in cancer cells and mouse models not only involves a unique interaction with bcl-2 but also interactions with variety of other partners in the mitochondria and cytosol. In contrast, TGFβ also induces nuclear export of NR4A1, where it exhibits tumor promoter-like activity and plays a role in degradation of inhibitory SMAD-7 in breast and lung cancer cells. The development of NR4A ligands is ongoing, and it is clear that their activity as agonists or antagonists can be used to target blood-derived and solid tumors, and these compounds can also be effective as enhancers of immune surveillance.
Authors' Disclosures
S. Safe has a research agreement with Systems Oncology for CDIM compound. No disclosures were reported by the other author.
Acknowledgments
Funding from the Kleberg Foundation (S. Safe), Texas A&M AgriLife Research (S. Safe), the Sid Kyle Chair Endowment (S. Safe), Systems Oncology (S. Safe) and NIH (P30-ES029067; S. Safe) is gratefully acknowledged.
Footnotes
Mol Cancer Res 2021;19:180–91
- Received August 13, 2020.
- Revision received September 22, 2020.
- Accepted October 19, 2020.
- Published first October 26, 2020.
- ©2020 American Association for Cancer Research.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.↵
- 130.↵
- 131.↵
- 132.↵
- 133.↵
- 134.↵
- 135.↵
- 136.↵
- 137.↵
- 138.↵
- 139.↵
- 140.↵
- 141.↵
- 142.↵
- 143.↵
- 144.↵
- 145.↵
- 146.↵
- 147.↵
- 148.↵
- 149.↵
- 150.↵
- 151.↵
- 152.↵
- 153.↵
- 154.↵
- 155.↵
- 156.↵
- 157.↵
- 158.↵
- 159.↵
- 160.↵
- 161.↵
- 162.↵