Insufficient support for retinoic acid receptor control of synaptic plasticity through a non-genomic mechanism

Abstract

It is well established that retinoic acid receptors (RARs) function as nuclear receptors that control gene expression in response to binding of the ligand retinoic acid (RA). However, some studies have proposed that RAR-alpha (RARa) controls synaptic plasticity via non-genomic effects outside the nucleus, i.e. effects on mRNA translation of GluA1, a sub-unit of the alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor. In order to support this non-genomic mechanism, studies have reported RARa knockout mice or treatment with pharmacological levels of RA and RAR antagonists to propose that RARa is required to control normal synaptic plasticity. A major shortcoming of the non-genomic hypothesis is that there have been no mutational studies showing that RARa can bind the GluA1 mRNA to control GLUA1 protein levels in a non-genomic manner. Also, without a genetic study that removes endogenous ligand RA, it is impossible to conclude that RARa and its ligand RA control synaptic plasticity through a non-genomic signaling mechanism.

Numerous studies have established that RA functions as a ligand for nuclear RARs that bind enhancers or silencers to directly regulate genomic transcription during development and in adult organs (1–3). Some studies suggest RA can also function in a non-genomic manner to regulate synaptic plasticity in the hippocampus through a mechanism involving cytoplasmic RARa that inhibits translation of mRNA for the GluA1 sub-unit of the alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor, with such inhibition being reversed in the presence of RA (4–7). Support for this non-genomic role of RA has come from analysis of the RARa knockout as well as treatment with pharmacologic levels of RA or RAR antagonists (8–10). These studies imply a role for endogenous RA in controlling normal synaptic plasticity, however there have been no attempts to remove endogenous RA to determine if it is required for synaptic plasticity. A common method used for RA treatment is to administer RA at 10 mg/kg, a teratogenic dose resulting in micromolar levels of RA (11). In contrast, RA is normally present in the nanomolar range in various tissues including brain (12). Thus, the normal function of endogenous RA cannot be ascertained through RA treatment, but must be approached by removing endogenous RA; this can be accomplished with standard or conditional knockout studies on the RA-generating enzymes encoded by Aldh1a1, Aldh1a2, or Aldh1a3 (2). Some studies suggest that specific neurological diseases may respond to treatment with pharmacological levels of RA that can alter synaptic plasticity via a non-genomic mechanism (13). Thus, it is possible that endogenous RA does not normally exist in a concentration high enough to affect synaptic plasticity but that pharmacological levels of RA can alter synaptic plasticity through a non-genomic side-effect. It would be necessary to monitor RA levels in the hippocampus (possibly by HPLC-mass spectrometry) to determine whether endogenous RA is present at a level sufficient to perform non-genomic actions.

Recent studies provided evidence that loss of RARa in RARa knockout mice prevents neuropathic pain via altered synaptic plasticity (10). Although the authors for this study propose that this is a non-genomic effect of RARa signaling, there are no studies provided that make it clear the mechanism is non-genomic as opposed to genomic. The original study from this group suggested that RARa may act in a non-genomic fashion due to the observation that RARa can be detected cytoplasmically in dendrites and that pharmacological RA treatment (1 μM) can alter dendritic synaptic plasticity in the presence of the transcriptional inhibitor actinomycin while this action can be blocked by the protein synthesis inhibitor cycloheximide (4). Another previous study from this group reported CLIP studies (cross-linking followed by immunoprecipitation) suggesting that the F domain of RARa interacts with the 5’-untranslated region of GluA1 mRNA to potentially inhibit translation (14). However, if RARa indeed controls synaptic plasticity by regulating translation of GluA1 mRNA, it should be possible to perform mutational studies to identify specific amino acid residues in RARa required for binding to GluA1 mRNA, similar to how mutational studies were performed to clearly demonstrate that the RARa zinc finger DNA-binding domain binds retinoic acid response elements (RAREs) in genomic DNA needed for genomic control mechanisms (1). Also, mutational studies could be used to identify a specific nucleotide sequence in GluA1 mRNA required for binding to RARa. As the non-genomic mechanism for control of synaptic plasticity is proposed to be a signaling event, the endogenous ligand RA would be required as a ligand for RARa. Since endogenous RA has not been removed using a genetic loss-of-function experiment, it remains unclear whether synaptic plasticity is controlled by RARa signaling via its normal endogenous ligand RA. In summary, a general principle that ensures success in the biological sciences is the use of genetic loss-of-function studies to determine the function of any molecule (such as RA), protein (such as RARa), or gene (such as the Rara gene). Such studies will need to be performed to determine whether or not RAR signaling controls synaptic plasticity through a non-genomic mechanism.

Highlights.

Retinoic acid (RA) receptors regulate genes at the genomic level when bound to RA.

RA receptor-alpha (RARa) has been shown to control synaptic plasticity.

RARa control of synaptic plasticity has been proposed to use a non-genomic mechanism.

RARa has been suggested to control translation of the mRNA for GluA1 (AMPA receptor).

Genetic loss-of-function studies are needed to see if RAR functions non-genomically.