Retinoids and breast cancer: new clues to increase their activity and selectivity

All-trans retinoic acid and derivatives (retinoids) are promising agents in the management of certain hematologic malignancies and solid tumors, including breast cancer. Retinoids are endowed with anti-proliferative, cyto-differentiating and apoptotic effects that are largely mediated by activation of the nuclear hormone retinoic acid receptors RARα, RARβ and RARγ. These are ligand-dependent transcriptional factors controlling the expression of numerous genes. The relative importance of each receptor subtype for the anti-tumor activity of retinoids is largely unknown. Clarification of this point is of fundamental importance for the rational design of retinoid-based therapeutic approaches aimed at controlling a heterogeneous type of tumors, like breast cancer.

Th e results presented by Bosch and colleagues [1] in the previous issue of Breast Cancer Research are a major step towards the identifi cation of the critical determinants controlling the responses of breast tumor cells to alltrans-retinoic acid (ATRA), the prototype of retinoids, a promising class of anti-cancer agents. Th e authors demonstrate that the two retinoid nuclear receptors RARα and RARγ play opposite roles in controlling the growth of breast cancer cells [1]. Retinoid-dependent activation of RARα arrests the growth, while activation of the other receptor favors the proliferation of the neoplastic cell. Suppression of RARγ activity enhances the antiproliferative eff ects of ATRA in cultures of breast cancer cell lines and in an appropriate animal model of the tumor. Finally, the RARα agonist AM580 [2] is more eff ective than the pan-RAR ligand, ATRA [3,4], in controlling the growth of breast cancer cells both in vitro and in vivo. Th is is consistent with the intrinsic ability of ATRA to activate RARα and RARγ simultaneously.
ATRA is the only example of a clinically useful cytodiff erentiating agent in oncology [3,4], having changed the natural history of acute promyelocytic leukemia, a rare form of acute myelogenous leukemia [5]. Th is along with promising results obtained in preclinical models has spurred interest in the use of ATRA and natural (9-cis and 13-cis retinoic acid) or synthetic derivatives (retinoids) also for the management of solid tumors, with particular reference to breast cancer. However, the clinical experience accrued with ATRA and derivatives in breast cancer has not kept up with expectations [6]. Indeed, so far, promising results have been obtained only with the atypical synthetic retinoid fenretinide, which is eff ective in the secondary prevention of pre-menopausal breast tumors [7].
One of the reasons for the poor clinical outcome of ATRA in breast cancer may be linked to the heterogeneity of this tumor, which is a collection of distinct entities with diff erent biological and prognostic characteristics [8]. Th us, the identifi cation of specifi c markers defi ning breast cancer subtypes with particular sensitivity to ATRA and derivatives is likely to represent a priority in studies aimed at developing rational therapeutic approaches based on the use of these compounds. Defi nition of RARα and RARγ as a positive and a negative determinant of retinoid sensitivity in breast carcinoma [1] is an important step in this direction. In fact, it suggests that breast tumors with high RARα/RARγ ratios should represent optimal targets for retinoids. Two large subgroups of breast cancer with these characteristics have been identifi ed. Th e fi rst one is represented by estrogen receptor-positive tumors, where estrogens are known to increase the activity of RARA, the gene en coding RARα, while leaving unaff ected the RARγ-encoding counterpart [9,10]. Th e idea is supported by the observation that estrogen receptor-positive breast carcinoma cell lines are sensitive to ATRA, while estrogen receptornegative ones are not [11]. A second subgroup characterized by high RARα/RARγ ratios has been identifi ed among Her2/neu-positive breast cancers [12]. In fact,

Abstract
All-trans retinoic acid and derivatives (retinoids) are promising agents in the management of certain hematologic malignancies and solid tumors, including breast cancer. Retinoids are endowed with antiproliferative, cyto-diff erentiating and apoptotic eff ects that are largely mediated by activation of the nuclear hormone retinoic acid receptors RARα, RARβ and RARγ. These are ligand-dependent transcriptional factors controlling the expression of numerous genes. The relative importance of each receptor subtype for the anti-tumor activity of retinoids is largely unknown. Clarifi cation of this point is of fundamental importance for the rational design of retinoid-based therapeutic approaches aimed at controlling a heterogeneous type of tumors, like breast cancer.
approximately 20 to 25% of Her2/neu-positive tumors have been shown to present with co-amplifi cation of the RARA gene, with consequent high constitutive levels of the encoded receptor. Cell lines re capi tulating this genotype are characterized by sensitivity to ATRAdependent anti-neoplastic activity, regardless of the estrogen-receptor status [12] and sensi tivity is mediated by RARα. Despite what has been discussed above, the search for tumor subtypes with high RARα/RARγ ratios may be an oversimplifi cation of the problem. In fact, RARs are not the only nuclear receptors activated by ATRA, which has been demonstrated to stimulate peroxi some proliferator-activated receptor (PPAR)β/δ as well [13]. Interestingly, ATRA-dependent activation of PPARβ/δ is also associated with increased breast cancer cell growth, which is similar to what has been observed for RARγ [14]. Th is adds a further layer of complexity to the search for breast tumor subtypes that may benefi t from retinoid treatment.
Th e work of Bosch and colleagues supports the concept that RARα is the principal mediator of ATRA anti-tumor activity in breast cancer. Th e authors propose that targeting of RARα with selective agonists, like AM580, may be a more eff ective clinical strategy to hit breast cancer cells than a pan-RAR agonist like ATRA. Indeed, the proposed therapeutic strategy may have additional value besides the observed lack of RARγ activation. Th e use of RARα agonists should theoretically avoid activation of RARβ, whose expression in tumor-associated stromal cells is a potential determinant of breast cancer growth [15]. In addition, selective targeting of RARα may reduce certain aspects of retinoid systemic and doselimiting toxicity, such as skin rashes and intra-cranial hypertension [3,4], two eff ects that are believed to be mediated by RARγ. However, simple activation of RARα may not be an optimal strategy to counter breast cancer growth and progression, as such an approach does not take into consideration the circulating levels of endogenous ATRA (1 to 10 nM) that are suffi cient to activate tumor cells' RARγ. Hence, we propose that approaches based on combinations between RARα agonists and RARγ antagonists may represent a more rational strategy in the clinical management of breast cancer.
In conclusion, the work of Bosch and colleagues revives interest in retinoids for the treatment and chemoprevention of breast cancer. A rational use of these agents calls for more in-depth knowledge of the molecular mechanisms and determinants underlying the sensitivity/ resistance of the neoplastic cell to ATRA and derivatives. Th is is likely to result in the development of eff ective therapeutic protocols based on combinations between retinoids and other anti-neoplastic agents, which is the ultimate goal of this line of research.