Gene expression is a fundamental part of cell biology as it initially produces an mRNA transcript that ultimately results in a functional protein. Within this process, gene transcription is a tightly controlled mechanism in cell nuclei through the action of specific DNA-binding transcription factors and an array of ancillary proteins known as co-activators or corepressors, depending on whether they act to stimulate or repress transcription. In general, most co-activators have histone acetyltransferase (HAT) activity, which acts in part to loosen the association of positively charged histones with DNA, facilitating chromatin remodeling and recruitment of the transcriptional machinery. On the other hand, corepressors often recruit histone deacetylases (HDACs) or methyltransferases, which reinforce the structural integrity of the histone–DNA association, effectively denying DNA access to the transcriptional machinery. However, evidence has accumulated showing that the precise transcriptional roles of cofactors can be promoter-specific, as many examples exist in the literature of proteins that can function as a co-activator or corepressor in different contexts. As such, it is perhaps better to globally refer to these proteins as transcriptional co-regulators.
Ligand-dependent nuclear receptor corepressor (LCoR) was originally identified by two independent laboratories through the use of cDNA library screens. Both laboratories identified LCoR as a transcriptional co-regulator. However, Fernandes et al. (2003) extensively characterized it as a ligand-dependent transcriptional corepressor of nuclear receptor–mediated transactivation, whereas Kunieda et al. (2003) showed preliminary evidence that suggested that it is a transcriptional activator through sequence similarity to mushroom body large-type Kenyon cell-specific protein 1 (Mblk-1). Future studies should provide more definitive evidence confirming that LCoR also regulates gene expression in a context-specific manner.
Domain Structure and Function
Currently, there exists no direct experimental evidence for different LCoR splice variants, or if these variants display any differences in terms of tissue/cellular localization or function. However, the National Center for Biotechnology Information (NCBI)-curated genomic Reference Sequences lists two potential isoforms of LCoR. Isoform 2 differs from isoform 1 in its 5′ UTR, 3′ UTR, and coding region, while maintaining the open reading frame. This results in similar proteins, with isoform 2 lacking 27 C-terminal amino acids. The region truncated in isoform 2 does not appear to contain any functional domain, yet there is a putative sumoylation motif. Accumulating evidence suggests that sumoylation is important for general repressor function (Geiss-Friedlander and Melchior 2007), but no experimental data exists that suggests differential LCoR isoform function. For the remainder of this entry, when referencing LCoR, it is the longer 433 (isoform 1) amino acid variant that is referenced.
LCoR also contains tandem extended N-terminal PXDLS motifs that recruit C-terminal Binding Proteins 1 (CtBP1) and 2 (Fig. 1) (Fernandes et al. 2003; Palijan et al. 2009b). CtBP1 and CtBP2 function predominanlty as transcriptional corepressors whose activity is modulated by the nuclear ratio of NADH/NAD+ (Chinnadurai 2003). Mutation of both CtBP-binding motifs is required in order to abolish CtBP binding, and attenuation of LCoR corepressor function occurs upon individual mutations at the binding sites in transcription regulated by PR and TRα, with the greatest effect observed when both sites are mutated (Palijan et al. 2009b; Wang et al. 2007). A negligible effect occurred in the transcription regulated by ERα or GR, when individual or combined mutations of the binding sites was carried out, which suggests a minimal role for CtBPs in LCoR-mediated repression of these receptors (Fernandes et al. 2003).
The helix-turn-helix motif located near the C-terminal tail (Fig. 1) also appears to be critical for LCoR function, as deletion of this motif attenuates corepression of estrogen-regulated gene expression and abolishes corepression of progesterone-regulated transcription in T47D human breast cancer cells (Palijan et al. 2009b). How this motif functions in repression, however, remains a topic for future research.
LCoR also recruits HDAC3 and 6 through central domains. HDAC6 recruitment occurs through a region that is delineated by amino acid 203–319 in LCoR (Fernandes et al. 2003; Palijan et al. 2009a). The interaction between HDAC6 and LCoR is somewhat surprising, given that LCoR is nuclear, but HDAC6, unlike HDAC3, is cytoplasmic in many malignant cells (Yang and Seto 2003). However, studies in normal breast epithelial cells show that HDAC6 is nuclear, and that responsiveness to endocrine therapy correlates with HDAC6 localization in breast cancer (Saji et al. 2005). Palijan et al. (2009a) also found that HDAC6 was partially nuclear in ERα-positive MCF-7 breast cancer cells. This suggests that the LCoR–HDAC6 interaction is lost in malignant cells in which HDAC6 is cytoplasmic resulting in unknown effects in tumorigenesis.
Cyclical recruitment of LCoR to the promoters of several ERα and PR target genes was demonstrated with a peak occurring 30–45 min after hormone treatment (Palijan et al. 2009a, b). Although no direct interaction was found between LCoR and CtBP-interacting protein (CtIP), they were found to interact indirectly, and colocalize on the promoters of several ERα and PR target genes (Palijan et al. 2009b). Hence, it is possible that they function as a complex on specific nuclear receptor target genes. Additionally, LCoR was shown to bind specific target gene promoters together with the PR, ERα, CtBP1, or HDAC6, substantiating previous experiments that indicate that these proteins are cofactors (Palijan et al. 2009a, b).
Furthermore, LCoR was found in a multi-subunit CtBP corepressor complex, which represses the tumor invasion suppressor E-cadherin through specific transcription factor promoter targeting and coordinated histone modifications (Shi et al. 2003). LCoR was also found to interact with lysine (K)-specific demethylase 1 (LSD1), a pivotal member of the Krüppel-like zinc finger E-box binding homeobox 1 (ZEB1)-LSD1-repressor element 1 silencing transcription factor corepressor (CoREST)–CtBP repressive complex (Wang et al. 2007). This large complex binds to the promoters of several genes such as growth hormone 1 (GH1) in developing pituitary lactotropes, and is thought to direct a precise transcriptional program that includes both activation and repression (Wang et al. 2007).
Small interfering RNA (siRNA)-mediated knockdown of LCoR revealed an increase in endogenous expression of progesterone-regulated target genes, which is consistent with reporter gene assays that show a role as a corepressor in T47D breast cancer cells (Palijan et al. 2009b). Surprisingly, results from the same type of experiment on estrogen-regulated transcription were conflicting. An increase in estrogen induced expression of an ectopic reporter gene after LCoR knockdown confirmed its role as a repressor of estrogen-mediated transcription (Palijan et al. 2009a). However, analysis of endogenous estrogen target gene expression revealed that abrogation of LCoR expression had either no effect or was required for optimal expression of specific target genes such as insulin-like growth factor binding protein 4 (IGFBP4) and cytochrome P450, family 26, subfamily B, polypeptide 1 (CYP26B1) (Kaipparettu et al. 2008; Palijan et al. 2009b). Future studies might reveal the mechanism through which LCoR functions as a co-activator/-repressor with precise tissue and gene specificity such as another member of the ligand-dependent corepressor family, RIP140 (Cavailles et al. 1995; Lee et al. 1998). The specificity of the effect on transcriptional regulation might depend on the constituents of the transcriptional regulating complex, developmental stage, and tissue. The most compelling evidence for this was presented in a study which found opposing, activator or repressor, functions for a multi-protein complex containing LCoR during mouse development (Wang et al. 2007). Which signals direct the assembly of this function-specific multifactor complex, and how this regulates transcription remain topics of future research.
Regulation of LCoR
Currently, there is very little experimental data that explains how LCoR mRNA or protein expression is regulated. It is not known if different signals direct differential isoform expression, or if this translates into altered functionality. Nevertheless, Wang and colleagues demonstrated that LCoR mRNA expression is increased by estrogen-treatment of human osteosarcoma U2OS cells (Wang et al. 2007). However, widespread and early LCoR expression suggests that there must be other signals that direct its expression, and clarifying these mechanisms might shed new light on its role in development and potentially disease.
LCoR mRNA is expressed as early as the two-cell stage during development, and is ubiquitously expressed in a variety of fetal and adult human tissues. Highest expression is observed in the placenta, the cerebellum and corpus callosum of the brain, the adult kidney, and a number of fetal tissues (Fernandes et al. 2003). Analysis of placental tissue revealed that LCoR predominantly localizes to the syncytiotrophoblast layer of terminally differentiated cells, which is critical for controlling maternal hormonal signals that regulate fetal metabolism and development (Pepe and Albrecht 1995).
Curiously, generation of a mouse LCoR knockout (Lcor−/−) model revealed that even though LCoR is expressed early during development it is not essential for it as most Lcor−/− animals survived to birth, but failed to suckle and died within hours. Additionally, a growth difference was noticeable starting at the 18th day of gestation (E18.5) as Lcor−/− mouse embryos were slightly (∼15%), but significantly, smaller than wild-type littermates (Dr. Yaacov Barak, unpublished results). Furthermore, immunohistochemical analysis of neonatal mouse pituitary tissue showed that LCoR expression increased postpartum potentially through signaling by estrogen, which is consistent with the mouse knockout model, which suggests a critical role for LCoR postpartum (Wang et al. 2007). Similarly, CtBP1 knockout mice were viable, but were small and died early (Chinnadurai 2003). This suggests that LCoR and CtBP1 might be co-regulating the same genes during development. These results coupled with the observation that LCoR localizes to and represses the GH1 promoter in a mouse pituitary cell line (Wang et al. 2007) indicate that although its role in early development is not critical, signal-dependent postpartum induction of LCoR and its subsequent regulation of genes is essential in later development.
Immunohistochemical analysis has shown that LCoR localizes to discrete nuclear bodies in the cell. The same type of analysis has also shown that these bodies contain CtBP1, CtBP2, CtIP, and polycomb ring finger oncogene (BMI1) (Fernandes et al. 2003; Palijan et al. 2009a, b). CtIP and BMI1, integral components of polycomb group repressor (PcG) complexes (Chinnadurai 2006; Fasano et al. 2007; Sewalt et al. 1999), associate indirectly with LCoR (Palijan et al. 2009b), and as such these observations suggest a possible role for LCoR in PcG complexes, which play an important role in cellular identity during development (Schuettengruber and Cavalli 2009).
Little is known concerning LCoR and its expression in disease states. In bladder cancer, a study found that LCoR mRNA expression was significantly (more than 0.5 fold) lower in more aggressive human cell lines compared with a less aggressive cell line (Abedin et al. 2009). A similar analysis of LCoR expression in a panel of breast cancer cell lines did not establish a correlation between LCoR and expression of ERα, or of cofactors CtBP1 or histone deacetylase 6 (HDAC6), nor did there appear to be a correlation with cancer aggresiveness (Dr. Sylvie Mader, unpublished results). Nevertheless, caution must be exercised when drawing conclusions from mRNA data as it may not correlate well with protein expression. Indeed, early and ubiquitous expression of LCoR, its critical role postpartum, and the possibility of various pleiotropic effects through interactions with different nuclear receptors, suggest that LCoR might be an important player in disease, particularly cancer development. However, determination of whether it acts in a tumor suppressor or oncogene fashion requires more experimental data.
A survey of the literature on LCoR reveals that it has been characterized extensively as a ligand-dependent nuclear receptor corepressor, which is expressed early in development and ubiquitously. Additionally, its developmental role becomes crucial postpartum. However, no information exists on the specific mechanisms controlling LCoR expression and turnover neither in the embryo nor the adult.
LCoR functions by recruiting the NAD(H)-dependent repressors, CtBPs, and histone deacetylases. The fact that CtBP activity is dependent on the NADH/NAD+ ratio suggests that LCoR repression could be affected by metabolic activity. Furthermore, it is not known how LCoR recruits cofactors, and if it is dependent on other signals. Accumulating evidence shows that LCoR may also have co-activator roles. Differential co-activator/corepressor roles enacted by transcriptional regulating complexes must depend at least in part in endogenous signals. These signals could result in posttranslational modifications that direct the assembly of complexes with transcriptional roles specific to different tissues. How LCoR is modified by these signals and participates in these multiprotein complexes remains poorly understood and highlights the fact that much remains to be accomplished in order to elucidate the mechanisms surrounding LCoR function.