Retinoid Actions: A New Horizon

The findings of Lin et al. (1) that are published in this issue of Endocrinology, in an article titled “Cellular Retinoic Acid–Binding Protein 1 Modulates Stem Cell Proliferation to Affect Learning and Memory in Male Mice,” expand considerably our understanding of retinoid actions in the brain. The authors convincingly establish a role for cellular retinoic acid–binding protein 1 (CRABP1) in modulating neural stem cell (NSC) proliferation and the hippocampus-dependent functions of learning and memory. These actions of CRABP1 were found to be independent of those of the retinoic acid receptors (RARs) in mediating retinoid-dependent transcription. This led the authors to conclude that CRABP1 plays an important role in facilitating brain development and functions in a manner that is independent of the canonical retinoic acid–RAR signaling pathway. This unanticipated finding raises many intriguing questions as to whether other retinoid-responsive processes within the body may similarly require CRABP1 actions.

With the identification and characterization of the RARs in the late 1980s (2) and the identification of the related retinoid X receptors (RXRs) shortly afterward, it became obvious that retinoid actions within cells could be accounted for largely by the binding of retinoic acid to one of its nuclear receptors, thus affecting retinoid-responsive gene transcription. Some even suggested that all of the actions of retinoids within the body, aside from those in vision, could be explained by this transcriptional regulatory paradigm. However, this view began to change about 10 years ago as investigators reported nonnuclear actions of retinoids that could not be explained simply by invoking canonically that retinoic acid binding to its nuclear receptor had affected gene expression patterns [see (3) for review]. Rather, data began to accumulate that retinoic acid can directly affect intracellular signal transductions pathways that do not directly or solely affect RAR- and/or RXR-mediated transcription.

The authors' laboratory has contributed significantly to this research, and the current study grows out of the laboratory’s earlier work. Specifically, the laboratory first showed that the addition of retinoic acid rapidly (within 1 hour) activated ERK 1/2 in P19 embryonal carcinoma cells, eliciting a cascade of signaling leading to rapid posttranscriptional modification of specific proteins (4). Subsequently, they extended this by demonstrating rapid noncanonical signaling of retinoic acid in Cos-1 and CJ7 embryonic stem cells (ESCs) that did not depend on RAR signaling, because treatment of the cells with RAR antagonists failed to block the activity (5). On the basis of their earlier data, Lin et al. (1) suggested that CRABP1 mediated noncanonical signaling of retinoic acid to negatively regulate stem cell proliferation through actions on the ESC cell cycle at the G1 phase.

Making use of a newly generated mouse knockout model for Crabp1, Lin et al. (1) now report that disruption of Crabp1 augments stem cell cycle and expands ESC and NSC proliferation. The current study firmly establishes a functional role for CRABP1 in mediating the activity of retinoic acid in activating ERK 1/2. The authors further show, using immunohistochemical staining, that CRABP1 protein can be readily detected in both ESCs and NSCs in wild-type mice but not in ESCs or NSCs of Crabp1-null mice. The consequences of Crabp1 absence to the brain are striking. In Crabp1-null mice, neurogenesis and NSC proliferation are increased in the hippocampus, where CRABP1 is detected immunohistochemically in the subgranular zone, and, correspondingly, hippocampus-dependent brain functions are stronger. Cognitive function testing of Crabp1-null mice, using the novel object recognition test to assess recognition memory and the Morris water maze test to assess spatial learning and memory, demonstrated that the Crabp1-null mice performed better in these tests than matched wild-type mice. This led the authors to conclude that the deletion of Crabp1 expands the NSC pool in the hippocampus and enhances the hippocampal functions of learning and memory.

By establishing the importance of CRABP1-mediated signaling in brain function, this points toward the possibility of a direct role for CRABP1 in regulating other critical processes within the body. After CRABP1 was first identified in the 1970s, it had been speculated that CRABP1 facilitates directly retinoid-dependent actions in the body (6). However, this notion quickly waned when the RARs and RXRs were identified. The current prevailing view regarding CRABP1 actions within cells has long been that CRABP1 mediates both the metabolism of retinoic acid and the transport of the hydrophobic retinoic acid through the aqueous environment of the cell to proteins or factors where it is functionally needed (7).

The current study convincingly establishes that CRABP1 can act directly to mediate certain critical functions in specific cells and tissues. This conclusion potentially has broad implications for understanding the roles of other members of the fatty acid–binding protein (FABP) family of proteins. The FABP protein family, of which CRABPI is a member, comprises many intracellular fatty acid–binding proteins (which bind unesterified fatty acids, monoglycerides, and some other lipids) as well as other intracellular retinoid–binding proteins (which specifically bind retinol or retinoic acid) (7, 8). As has been the case for CRABP1, it is generally thought that many of these proteins facilitate the metabolism of their ligands and/or the transport of the ligand to proteins/factors that use the ligand (7, 8). However, several recently published studies, especially studies of knockout animals, strongly suggest that many of these FABP family members may have other important, albeit unknown, actions within the body. The present findings establishing the involvement of CRABP1 in maintaining brain growth and actions may help open a new round of research aimed at identifying novel physiologic actions of other FABP family members.

The demonstration by Lin et al. (1) that CRABPI plays a direct role in facilitating brain development, affecting the brain’s capacity for memory and learning, raises many additional questions that will need to be answered. The present work does not establish whether retinoic acid bound to CRBP1 is required for mediating the effects observed in the brain. One might speculate that retinoic acid binding is probably required because earlier studies showed that the addition of retinoic acid rapidly activated ERK 1/2 signaling in P19 and ESCs. The present work has not definitively established whether both retinoic acid bound to CRBP1 and apo-CRABP1 (CRABP1 without bound retinoic acid) can mediate these effects on the brain and its actions. This needs to be determined. However, this will be challenging given the promiscuity of retinoic acid binding to a spectrum of receptors and binding proteins and the complication posed by its metabolism in cells.

Another question that needs to be asked concerns the counterintuitive finding that loss or absence of Crabp1 expression gives rise to the “better” phenotype of enhanced capacity for memory and learning. This suggests that we currently do not completely understand the direct actions of CRABP1 within the body. More studies on this point are needed. Also in need of further research is how the finding that CRABP1 directly affects physiologically important processes within the hippocampus will affect the clinical use of retinoids to treat disease. 13-cis-retinoic acid (isotretinoin), which binds CRABP1 with high affinity, has been used effectively to treat acne. But 1% to 11% of patients taking the drug, depending on the study, report mental disorders, including frank depression and even suicidal ideation (9). This side effect has limited the clinical use of isotretinoin. Several published studies have proposed that the inhibition of hippocampal neurogenesis and cell death results from exposure to high concentrations of isotretinoin (9). Consequently, it is necessary to ask whether CRABPI binding to isotretinoin contributes, either positively or negatively, to these adverse effects and, if so, how?

Finally, to understand the bigger picture surrounding the present findings, investigations will be needed to identify whether the independent actions of CRABP1 may be important in other physiologic or pathophysiologic contexts. Lin et al. (1) note that CRABP1 expression dampens both ESC and NSC proliferation. This raises questions regarding CRABP1 actions in disorders of cell proliferation. Does CRABP1 have a role in dampening skin disease or cancer development? This will need to be established.

Thus, like all good research, the findings being reported in this issue of Endocrinology raise many important new questions regarding CRABP1 and retinoid actions in the body, ones that will require further research to resolve.