Novel insights in the molecular mechanisms of action of retinoids and their potential repercussions on breast cancer cell proliferation

In chapter 1 of this thesis, background information on the molecular actions and effects elicited by retinoids in the context of nuclear receptor (NR) signaling and breast cancer treatment was given. In addition, the state of the art and the knowledge gap still remaining in this area was highlighted. The principal objective of the thesis was defined which was to investigate different molecular events possibly elicited after retinoid treatment to better understand the observed effects on breast cancer cell proliferation  Finally, the general outline of the present thesis was provided.  

Chapter 2 assessed whether the ligand binding domain (LBD) of the Retinoic Acid Receptor (RAR) can be used to study RAR-coregulator binding events as an alternative to the use of the full length (FL) RAR. Thus, this work aimed to characterize and compare ligand-driven coregulator interactions of both the RARα-LBD and the RARα-FL. We studied the coregulator motif interactions of RARα-FL and RARα-LBD upon incubation with the RAR agonist all-trans retinoic acid (AtRA) and the RAR antagonist Ro415253 using the Microarray Assay for Real-time Coregulator-Nuclear Receptor Interaction (MARCoNI). The results obtained showed that the coregulator binding profiles are highly similar for the RARα-FL and the RARα-LBD regarding both qualitative aspects like the type of coregulators that bind or dissociate upon (ant)agonist action and quantitative parameters such as EC50s and the maximum responses (modulation indexes at saturating ligand concentration). Moreover, based on the coregulator binding signatures, the RAR agonist AtRA could be clearly discriminated from the RAR antagonist Ro415253 using both the RARα-FL and the RARα-LBD.

Once established that coregulator binding patterns could be adequately characterised using LBDs, Chapter 3 assessed the coregulator binding profiles of the LBDs belonging to the RAR alpha (RARα), beta (RARβ) and gamma (RARγ) focusing on possible differential outcomes between the different RARs. To this end, this study used the MARCoNI coregulator binding assay to characterize coregulator motif binding patterns of the distinct RAR subtypes upon incubation with the general agonist all-trans-Retinoic Acid (AtRA); the subtype-selective agonists Am80 (RARα), CD2314 (RARβ) and BMS961 (RARγ); and the antagonist Ro415253. Chapter 3 showed multiple ligand-dependent RAR-coregulator interactions for all RARs, including binding events not described before. It also confirmed a greater ligand-independent functioning of RARβ based on both higher basal and lower ligand-induced coregulator binding. In addition, despite the generally high similarity, various coregulator motifs showed subtype-selective binding, providing a basis for a potential role of coregulator binding in the receptor subtype-specific responses. Finally, Chapter 3 also proved that subtype-selective agonists and agonistic/antagonistic actions of ligands for all the RAR variants can be easily discriminated by the MARCoNI coregulator binding assay.

After characterizing a possible role of coregulator binding in RAR subtype-specific responses, in chapter 4 and chapter 5 of this thesis, the crosstalk between retinoid and estrogen signaling in the estrogen receptor positive (ER+) breast cancer MCF7/BUS cell line was investigated. To this end, in chapter 4, the effects of AtRA on the basal and E2-induced estrogen receptor alpha (ERα) signaling were characterised using the MCF7/BUS and the U2OS-ERα-Luc cell lines. The effects of AtRA on different ERα-related events such as ERα-mediated cell proliferation and gene expression, ERα-coregulator binding and ERα subcellular localization were evaluated. While AtRA-driven inhibition of E2-induced signalling was found in the proliferation and gene expression studies, no significant effects on the E2-directed coregulator binding and subcellular distribution of ERα were observed. Strikingly, chapter 4 showed that in the absence of E2, basal ERα-mediated gene expression, ERα-coregulator binding (in lysate samples) and ERα nuclear localization were increased when exposing the cells to micromolar concentrations of AtRA. However, further coregulator binding assays making use of isolated ERα-LBDs suggested that direct binding of AtRA to ERα does not take place so that other molecular mechanisms must be involved in this phenomenon, which must be studied in the future.

In Chapter 5 the studies performed in chapter 4 using AtRA were extended to the RAR subtype-selective ligands. Thus, it was investigated whether the distinct effects of RAR subtype-selective agonists on breast cancer cell proliferation could partially be driven by differential effects on the ERα pathway. To test that hypothesis, evaluation of the effects driven by the general agonist AtRA and the subtype-selective agonists Am80 (RARα), CD2314 (RARβ) and BMS961 (RARγ) on basal ERα-related events was carried out using the MCF7/BUS and the U2OS-ERα-Luc cell lines. The results obtained indicated that proliferation, ERα-mediated gene expression, ERα-coregulator binding and ERα subcellular localization can be affected by some of the retinoids in a compound-specific manner. Nonetheless, the responses did not reveal a clear distinction between the different effective (anti-proliferative) AtRA and Am80 and the more ineffective retinoids CD2314 and BMS961. In addition, the coregulator binding studies with the isolated ERα-LBD showed that none of the compounds modulated ERα by direct binding (ligand-receptor) suggesting other pathways as underlying causal factors of the distinct effects of RAR subtype-selective ligands on ERα signaling. From all retinoids tested, the general RAR-agonist AtRA and the RARβ agonist CD2314 displayed the most remarkable effects on ERα signaling. Upon AtRA and CD2314 treatment at micromolar concentrations up-regulation of ER-mediated transcription was observed. However these retinoids differed when comparing  other ERα-related molecular events. Thus, while AtRA increased ERα nuclear localization and coregulator binding,  CD2314 exerted no significant effects on these molecular endpoints.

To obtain further insight in the crosstalk between retinoid and estrogen signaling in breast cancer, in chapter 6 of the present thesis, the effects of the general RAR agonist AtRA on the cell proliferation and the related gene expression profiles upon treatment with E2 were studied in the ER+ breast cancer cell line T47D. For this purpose, different assays determining the effects on cell proliferation and the transcriptomic pattern (RNA-seq) were used. In addition, the effects of AtRA on the specific ERα- and ERβ-mediated gene expression were assessed by means of reporter gene assays with the T47D-ERα-Luc and U2OS-ERβ-Luc cell lines. The results obtained from the proliferation and reporter gene studies pointed at AtRA as a repressor of the E2-induced breast cancer growth and as inhibitor of the E2-driven expression mediated by ERα and ERβ. To further understand the underlying modes of action of RAR-ER crosstalk, in a next step chapter 6 acquired further insights on the possible interplay between estrogen and retinoid signalling by means of a transcriptomic analysis. A high proportion of estrogen-responsive genes were found to be modulated by a single treatment with AtRA. In addition, RNA-seq data showed that AtRA can exert a variable modulation of the E2-induced transcriptomic responses. In other words, AtRA both counteracted and enhanced gene expression responses induced by E2. Further analysis showed that the estrogen-responsive genes whose E2-driven expression was counteracted upon co-exposure to AtRA were more often related to biological processes involved in breast cancer cell proliferation, differentiation or apoptosis. Altogether, the results of chapter 6 supported the existence of genomic crosstalk between estrogen and retinoid signaling eliciting distinct responses depending on the gene considered. However, co-treatment with AtRA more frequently presented inhibiting/counteracting effects on E2-induced genes involved in breast cancer growth.

In Chapter 7, the studies assessing the effects of the general agonist AtRA on the proliferation and related transcriptomic patterns in the T47D breast cancer cell line were extended to the RAR subtype-selective agonists. The effects of the different retinoids on RAR or other NR pathways of importance in breast cancer cell proliferation (e.g. ER pathway) were assessed by analyzing gene expression patterns. This study showed that the general AtRA and the RARα agonist Am80 highly hampered T47D breast cancer cell proliferation while the RARβ agonist CD2314 and the RARγ agonist BMS961 showed to be ineffective anticancer agents which even slightly increased proliferative rates. The transcriptomic profiles clearly differentiated between the effective and ineffective retinoids showing multiple differences in gene expression. Chapter 7 also suggested RARα as the main RAR subtype involved in the effects of retinoids on gene expression as shown by the higher and relatively similar effects observed on the transcriptomic signature elicited by AtRA and Am80. Differences between the anti-proliferative retinoids (AtRA and Am80) and the ineffective retinoids (CD2314 and BMS961) mainly occurred in terms of magnitude of expression. Thus, relevant biological processes such as proliferation, differentiation and apoptosis and pathways such as the RAR, the ER, the VDR and the GR pathways were differentially modulated by the effective and ineffective retinoids. Altogether, the main conclusion drawn in this chapter was that small differences in effects on multiple pathways regulating breast cancer cell proliferation may sum up leading to the distinct responses elicited by the different retinoids in T47D breast cancer cells.

Finally, in Chapter 8 we first provided an overview on the principal results generated throughout the present thesis. Next, we went in detail with a discussion of all the findings of the thesis. In this analysis, comparison with previous literature, implications of the results obtained and follow-up studies to be performed in the future, were incorporated.