review retinoic acid signaling in the nervous … · jÖrg mey and peter mccaffery the majority of...

Volume 10, Number 5, 2004 THE NEUROSCIENTIST 409Copyright © 2004 Sage PublicationsISSN 1073-8584

Retinoic acid (RA), a member of the retinoid family oflipids and mediator of vitamin A activity, is an essentialmorphogen in vertebrate development. The formation ofthe body axes and the development of a number of organsystems including retina, brain, heart, urogenital system,and lungs are dependent on RA. The developmentaldefects resulting from vitamin A deficiency as well asthe teratogenic effects of retinoid overdose weredescribed more than 50 years ago (Warkany andSchraffenberger 1946; Wilson and others 1953). Due toits effect on cell differentiation and proliferation, RA isnow being used as a therapeutic tool in dermatology andoncology. However, apart from vitamin A–related visualdeficits, the functions of retinoids in the adult nervoussystem are far less conspicuous than their role in de-velopment, and only recently has the importance of thisattracted scientific attention.

During the 1970s, it became clear that the biological-ly active principle of vitamin A is in most cases RA, withone exception being the process of phototransduction,which depends on retinaldehyde. By far, the most im-portant mechanism of RA activity is the regulation of

gene expression. This is accomplished by its binding tonuclear retinoid receptors that are ligand-activated tran-scription factors. Thus, RA acts as a transcriptionalactivator for a large number of other, downstream-regulatory molecules, including enzymes, transcriptionfactors, cytokines, and cytokine receptors. These mech-anisms account for most of the known effects of retinoidteratogenicity, hypervitaminosis A, and experimentalresults with RA. Because a number of excellent reviewson the molecular biology of RA signaling are available(Gudas and others 1994; Mangelsdorf and others 1995;Chambon 1996; Gottesman and others 2001), we willlimit ourselves to a very condensed overview of the sys-tem. Although the overwhelming majority of experimen-tal studies have focused on RA receptor–dependent me-chanisms, there have always been indications of other,nonclassical modes of action. These involve interactionswith enzymes such as protein kinase C, direct activationof electrical synapses, and the search for putative cellsurface receptors. These nonclassical mechanisms,which are presently discussed in the context of the phys-iology of vision, will not be covered in this review.

The continuing expression of retinoid receptors inmany brain regions of adult mice as well as regulatoryeffects on neurotransmitter metabolism suggests func-tions of this signaling pathway in the adult brain.Analyses of vitamin A–deprived animals and retinoidreceptor knockout mice have focused on neural plastici-ty, and these models reveal defects in spatial learningand long-term potentiation (LTP) in the hippocampus.Genetic analyses of the components of RA signalinghave led to the suggestion that RA may be involved inschizophrenia, whereas clinical observation after

Retinoic Acid Signaling in the Nervous System of Adult VertebratesJÖRG MEY and PETER MCCAFFERY

The majority of the functions of vitamin A are carried out by its metabolite, retinoic acid (RA), a potent tran-scriptional activator acting through members of the nuclear receptor family of transcription factors. In theCNS, RA was first recognized to be essential for the control of patterning and differentiation in the devel-oping embryo. It has recently come to light, however, that many of the same functions that RA directs inthe embryo are involved in the regulation of plasticity and regeneration in the adult brain. The same intri-cate metabolic control system of synthetic and catabolic enzymes, combined with cytoplasmic binding pro-teins, is used in both embryo and adult to create regions of high and low RA to modulate gene transcrip-tion. This review summarizes some of the discoveries in the new field of retinoid neurobiology including itsfunctions in neural plasticity and LTP in the hippocampus; its possible role in motor disorders such asParkinson’s disease, motoneuron disease, and Huntington’s disease; its role in regeneration after sciaticnerve and spinal cord injury; and its possible involvement in psychiatric diseases such as depression.NEUROSCIENTIST 10(5):409–421, 2004. DOI: 10.1177/1073858404263520

KEY WORDS Retinoic acid, Plasticity, Learning, Parkinson’s disease, Depression, Regeneration

From the Institut für Biologie II, Aachen, Germany (JM) and theUMMS/EK Shriver Center, Waltham, Massachusetts, and theDepartment of Cell Biology, University of Massachusetts MedicalSchool, Worcester, Massachusetts (PM).

This work was supported by a Deutsche Forschungsgemeinschaft SFB542 grant to J.M. and National Institutes of Health grants HD05515and MH66037 to P.M. Thanks to Jacob Brodsky for help with Figure 3.

Address correspondence to: Jörg Mey, Institut für Biologie II,RWTH Aachen, Kopernikusstrasse 16, 52074, Aachen, Germany(e-mail: [email protected]).

REVIEW �

410 THE NEUROSCIENTIST Retinoic Acid in the CNS

Accutane (13-cis-RA) therapy has indicated a linkbetween RA and depression. Since then, RA has alsobeen implicated in Parkinson’s disease, motoneuron dis-ease, and Huntington’s disease. The fact that retinoidsdrive the differentiation of all sorts of glial cells, neu-rons, and neoplastic cell lines in culture prompted inves-tigations of axonal regeneration. Many molecular sig-nals that convey information for neuroglial interactionsare regulated by RA. These ideas have recently been sup-ported by the discovery that retinoid receptors and bind-ing proteins participate in traumatic processes afterinjury in the sciatic nerve and spinal cord.

RA Signaling

RA exerts its effects on transcription by binding to spe-cific nuclear receptors of the steroid/thyroid hormonesuperfamily of transcriptional activators. The principalreceptors for RA are the RARs (α, β, or γ) activated byall-trans RA. The RXRs are nominally RA receptors butare general partners for several nuclear receptors includ-ing thyroid hormone receptor, vitamin D receptor, per-oxisome proliferator-activated receptor, and RAR. Al-though the 9-cis isomer of RA can bind to RXR, it hasbeen very difficult to detect this isomer in the body, andincreasing evidence suggests that RXR functions as anunliganded partner for nuclear receptors. A group ofretinoid-binding proteins (RBPs) modulates the actionof these lipids. RBP in the plasma assists in the transportof retinol. Cellular retinoic acid binding proteins I and II(CRABP-I and -II) buffer RA in the cytoplasm. CRABP-I has been suggested to promote RA catabolism (Boylanand Gudas 1992) whereas CRABP-II may assist with thetransport of RA into the nucleus to associate with theRA receptors (Delva and others 1999). Cellular retinol-binding protein (CRBP) binds retinol in the cytoplasmand may promote its metabolism to retinaldehyde(Napoli 1993) and esterification to the retinyl ester forstorage (Ghyselinck and others 1999).

The synthesis of RA from retinol was originally con-sidered to occur ubiquitously in tissues, and the distribu-tion of receptors and binding proteins determined theregionalization of RA’s actions. Studies on the develop-ing retina, however, showed that intricate patterns ofsynthetic enzymes could create regionalized zones ofsynthesis (McCaffery and others 1992; Mey and others1997). Synthesis of RA occurs via a two-step oxidativeprocess from retinol to a retinaldehyde intermediate andthen to RA (McCaffery and Dräger 2000). The first stepis catalyzed by a group of retinol dehydrogenases with awidespread and overlapping distribution (Napoli 1999;Duester 2003). Redundancy between these enzymes pre-vents major defects from occurring in null mutants ofsingle enzymes. In contrast, the second step is per-formed by retinaldehyde dehydrogenases (RALDHs)with very localized patterns of expression. RALDH-2and RALDH-3, which were both first identified in thedeveloping eye (McCaffery and Dräger 2000), deter-mine regions of RA synthesis in the developing embryo.

Null mutation of the RALDH-2 enzyme is lethal to theembryo and results in multiple brain abnormalities(Niederreither and others 1999). More recently, it hasbeen shown that the RA catabolic enzymes also have anessential role to play in the patterning of RA distribution(Abu-Abed and others 2002). CYP26A1, B1, and C1 ofthe cytochrome P450 family inactivate RA via oxidationof the fourth carbon. They occur often in a complemen-tary distribution to the synthetic enzymes (McCafferyand others 1999; Swindell and others 1999). Null muta-tion of CYP26A1 results in abnormal CNS developmentand results in death soon after birth (Abu-Abed and oth-ers 2001; Sakai and others 2001). A summary of thesemetabolic pathways is shown in Figure 1.

The expression of a large number of genes is knownto be regulated by the RA receptors (Gudas and others1994; Jonk and others 1994), and at least 40 have iden-tifiable RA response elements (McCaffery and others2001). RA provides an early signal to commit proliferat-ing stem cells to become neurons. RA acts as a differen-tiation agent for almost all neural stem cells or progeni-tors; this is so for progenitors as diverse as P19(McBurney 1993) or NTERA-2 (Andrews and others1990) stem cell lines, ES cells (Bain and others 1995),neuronal precursors such as PC12 and LAN-5 cells (Hilland Robertson 1998; Ingraham and Maness 1990;Matsuoka and others 1989), or primary stem cell cul-tures derived from rat hippocampus (Takahashi and oth-ers 1999). Because RA initiates a neural differentiationprogram, it lies upstream of a large number of neural-specific genes; many are directly induced, but a largernumber are induced secondarily or at more progressivedownstream steps. Many genes that are constituents ofthis program have been identified by subtractive tech-niques from RA-induced stem cells (Bouillet and others1995; Leypoldt and others 2001).

Less attention has been paid, however, to RA recep-tors as gene repressors. When the ligand is absent fromthe RA receptors, like many of the nuclear receptor fam-ily, they are still bound to the response element withinthe gene promoter where they can act as stronginhibitors of gene transcription. As such, the absence ofRA may act as a signal in itself by repressing geneexpression. This may be a reason why the catabolicCYP26 enzymes are so important for the control ofdevelopment. The resulting negative signal due to theabsence of the RA ligand may be just as profound a con-trol system as the activation of genes by RA.

Retinoic Acid in Learning

and Neural Plasticity

Spatial Learning Paradigms

Cognitive functions such as long-term memory requirelasting changes in synaptic plasticity, and such enduringchanges in synaptic plasticity require new protein syn-thesis. In addition to the modulation of synaptic trans-

Volume 10, Number 5, 2004 THE NEUROSCIENTIST 411

mission by local signaling pathways, systems that regu-late nuclear transcription are also involved in these proc-esses. The relevance of RA in this context has beenaddressed by examining the effects of retinoid receptormutations on performance in several learning para-digms. The first experiments were conducted in the1990s (Chiang and others 1998; Misner and others2001). Transgenic mice lacking either RARβ or RXRγor both genes were found to be deficient in the formationof spatial memory as tested with the Morris Water Maze.In this task, the animals are trained to swim to a platformthat is submerged in opaque water 1 cm beneath the sur-face. The time needed to find the hidden platform duringtraining sessions and the mean distance from the targetarea in test trials without the platform are recorded toobtain learning curves and assess memory retention.RARβ and RARβ/RXRγ double knockout mice weresignificantly impaired in accomplishing the task.Although these data may be qualified by visual deficitsof the animals (RARβ mutants show severe retinal ab-normalities; Grondona and others 1996), additionalstudies corroborated the specific function of RA for spa-tial learning: Following 12 weeks on a vitamin A–freediet, rats were trained in a radial-arm maze task. Resultsshowed that RA deficiency induced a severe deficit inspatial learning and memory and that the cognitiveimpairment was fully restored when vitamin A wasreplaced in the food (Cocco and others 2002) (Fig. 2).

Chronic ethanol consumption that produces cognitivedeficits also induces disorders in the biosynthesis of RA.In one study, an inhibition of RARβ activity was shownto reverse an alcohol-induced working-memory deficit(Alfos and others 2001).

Long-term Potentiation and Depression in the Hippocampus

Physiological and neuropsychological studies ascribespatial learning and the consolidation of long-termmemory to the hippocampal formation in the temporalforebrain. Performance in spatial learning paradigms isconsidered to be hippocampus dependent. A highexpression of retinoid receptors RARα, RARγ, andRXRβ in the hippocampus of adult mice had been de-scribed earlier (Krezel and others 1999). Potentialphysiological correlates of associative learning are long-term potentiation (LTP) and long-term depression(LTD). In the hippocampus, where these phenomenahave been studied to a great extent, CA1 pyramidal neu-rons that receive excitatory input from so-calledSchaffer collaterals of CA3 neurons show a long-lastingincrease in their excitatory postsynaptic potential whentheir synaptic excitation is paired with a high-frequencyelectrophysiological stimulation. In contrast, simultane-ous low-frequency stimulation decreases their responseto activation by Schaffer collaterals. RARβ andRARβ/RXRγ mutant mice lacked LTP almost complete-ly, whereas LTP was normal in RXRγ deficient animals.Interestingly, a severely reduced LTD was observed inall three knockout mice including the RXRγ mutants(Chiang and others 1998; Fig. 2). In aging mice, thedeclining memory performance in a radial maze task aswell as a diminished amplitude of LTP in the CA1 regionconcurred with a decrease in the levels of RARβ andRXRβ/γ expression in the brain. RA administration ledto a reversal, and the application of an RAR antagonistexacerbated the age-related impairments (Etchamendyand others 2002). In addition to affecting behavioral per-formance and LTP amplitude, the repeated injections ofRA for 4 days caused an increase in expression of RARβand RXRβ/γ in whole brain and of RXRβ/γ in the hip-pocampus. By excluding the possibility of generaleffects on affect, motivation, perception, or motor con-trol, the behavioral studies point to a specific role of RAin cognitive performance (Etchamendy and others2002).

How Could Retinoic Acid Contribute to Synaptic Plasticity and Learning?

To understand the mechanism by which RA influenceslearning and memory, it is important to differentiate theeffects of RA signaling defects on hippocampal devel-opment versus a direct effect on the physiology of learn-ing in the adult brain. In their initial report, Chiang andothers (1998) noted that the ultrastructure of hippocam-pal synapses and their physiological characteristicsappeared normal in retinoid receptor mutant mice.

Fig. 1. The retinoic acid signaling system. All-trans retinol (t-ROH) in the plasma is released from its carrier protein, retinol-binding protein (RBP), and enters the cell where it is bound bycellular retinol binding protein (CRBP). The first step of synthe-sis to retinoic acid is catalyzed by a retinol dehydrogenase(RoDH) or alcohol dehydrogenase (ADH) to all-trans retinalde-hyde (t-RAL), which remains bound to CRBP and is then oxi-dized irreversibly to all-trans retinoic acid (t-RA), which binds tocellular retinoic acid binding protein (CRABP). All-trans retinoicacid can either 1) enter the nucleus to bind to the RAR retinoicacid receptor to activate transcription; 2) diffuse out of the cellto influence neighboring cells; 3) possibly isomerize to 9-cisretinoic acid (9c-RA), which binds to the RXR retinoic acidreceptor; or 4) oxidize to 4-oxo t-RA, catalyzed by the enzymeCYP26, on the pathway to inactivation.


Therefore, the authors concluded that RAR/RXR-dependent deficiencies in LTP and LTD reflect acutedefects in the mechanisms responsible for synaptic plas-

ticity and are not due to embryonic effects of the geneknockouts that might have altered brain development.This interpretation is well corroborated by subsequent

Fig. 2. Relevance of retinoic acid (RA) for spatial learning and synaptic plasticity in the hippocampus. Long-term potentiation (LTP) (a)and long-term depression (LTD) (b) in transgenic mice that lacked RARα, RXRγ, or both receptors, collecting the field potential record-ings from hippocampal slices. The initial slope of field excitatory postsynaptic potentials is normalized to the baseline value preced-ing the induction of LTP or LTD. Each point represents the mean ± SEM (Chiang and others 1998). c, Hippocampal LTP and LTD areimpaired in vitamin A–deprived mice but can be rescued by dietary vitamin A replenishment. Summary of field potentials at differentages when the animals are raised with normal (control) or vitamin A–deficient diet (VAD). One group received vitamin A supplementafter 17 to 18 weeks VAD. Data are expressed as the mean percentage potentiation or depression 25 to 35 h after induction of LTPor LTD ± SEM (Misner and others 2001). d, Performance of VAD rats in a radial arm maze spatial learning task. Rats received normalchow (control), a VAD diet for 12 weeks (VAD) or a supplement of vitamin A after the 12-week VAD period (VAD replenished). Eachpoint represents the mean ± SEM of 10 animals (Cocco and others 2002).


studies with vitamin A–deprived animals because inthese experiments, the observed loss of synaptic plastic-ity was reversible when the animals had again access toa normal diet (Misner and others 2001; Cocco and oth-ers 2002; Fig. 2). Vitamin A–deprived mice developed acomplete loss of LTD and a severe reduction of LTP inCA1 neurons of the hippocampus. Direct application ofRA to hippocampal slices, which were used to measureLTP/LTD in vitro, also rescued the function (Misner andothers 2001).

Thus, strong physiological arguments favor a directparticipation of RA in hippocampal synaptic plasticity.However, some anatomical data are difficult to reconcilewith this interpretation because several studies failed todetect RARβ expression in the hippocampus of adultmice (Krezel and others 1999; Zetterström and others1999). Immunohistochemical analysis detected RARα,RARβ, and RARγ proteins in the cortex but not thehippocampus, whereas the receptors were present in thehippocampus at embryonic days 14.5 and 16.5, a periodof extensive neurogenesis (Yamagata and others 1994).Recent results from young adult mice kept on a vitaminA–free diet also point to indirect effects that influencehippocampal functions via long-term structural changesthat are not quickly reversible (Etchamendy and others2003): In a dual-choice discrimination task in an eight-arm radial maze, 39 weeks of vitamin A deprivation wasnecessary for significant memory impairment, whereas31 weeks of deprivation was insufficient. It was unlike-ly that this was due to the time required to deplete theanimal of vitamin A because this effect was notreversible with systemic injections of RA (150 µg/kg).Because other cortical areas are also involved in spatiallearning, it is possible that RA may act on connectedsystems, even without the presence of RARβ in the hip-pocampus. Given the conclusive effects on LTP and LTDas measured in hippocampal brain slices in vitro, thelocal expression of RARα and RXR receptors, and RAactivity in the hippocampus of transgenic reporter mice(Fig. 3), a participation of RA in the physiology of learn-ing is at least a promising hypothesis.

Physiological Functions of RA in Synaptic Plasticity

Although this issue is yet to be fully resolved, one canspeculate about specific functions that RA may fulfill inthe process of memory formation. Molecular mecha-nisms that have been discovered in synaptic changesinclude Ca2+-dependent second-messenger systemsand regulated gene expression of neurotransmitter- andneurotrophin-signaling pathways. Many of these genesare targets of retinoids. Behavioral deficits after vitaminA deficiency correlated with a decrease in expression ofgenes related to retinoid signaling itself, such as RARβand RXRβ/γ, thus supporting earlier results with theknockout mice. Are RA receptors transcriptional regula-tors of genes that induce LTP or LTD? Some specificcandidates for RA targets have already been identified:In the hippocampus, expression of the Ca2+-sensitive

calmodulin-binding protein neurogranin (RC3) correlat-ed with RXRβ/γ status and could be raised by RA injec-tions, and cognitive deficits in vitamin A–deprived ani-mals were exclusively observed when RC3 was also low(Etchamendy and others 2002, 2003). RC3 releasescalmodulin kinase II in response to activation of proteinkinase C and increase of the intracellular Ca2+ concen-tration. Because RC3 is believed to be enriched in den-dritic spines, this molecule may be in a key position toconnect NMDA receptor–dependent Ca2+ increase todownstream changes in synaptic excitability and long-term structural changes. In addition to RC3, a number ofother genes involved in synaptic plasticity includingneurotrophin receptors, NMDA receptors, and thecalmodulin kinase II are also RA dependent(Etchamendy and others 2003). In conclusion, a partici-pation of retinoids in spatial learning seems well sup-ported empirically.

Song Learning in Birds

The discovery of mechanisms of how male songbirdsacquire and produce their species-specific song is aprime example of the successful integration of behav-ioral analysis and neurophysiological and molecularresearch. Juvenile male songbirds of several species,including zebra finches, canaries, and song sparrows,learn their song during a sensitive period from a conspe-cific individual, the so-called tutor. As a necessary stim-ulus in song development, the tutor’s song serves as atemplate for evaluating auditory feedback from the juve-nile’s own vocalization. During the time of song acquisi-tion, this feedback guides development of the neural cir-cuits for song production (Alvarez-Buylla and Kirn1997; Denisenko-Nehrbass and others 2000). A centralneuronal regulator of the song system is the high vocalcenter (HVC) in the forebrain, which integrates auditoryand motor activity. The HVC projects to the nucleusrobustus archistriatalis (also abbreviated RA in the liter-ature), which in turn projects to hypoglossal motor neuronsinnervating the vocal organ, the syrinx. An additionalpathway from the HVC to the RA nucleus is necessaryfor acquisition but not production of the learned song.This pathway in the anterior forebrain sequentially con-nects HVC to area X in the paleostriatum, then to thethalamic nucleus DLM, on to the lateral magnocellularnucleus of the anterior neostriatum (IMAN), and fromthere to nucleus RA. The seminal discovery of continuedproduction of neurons in the HVC in adult male song-birds triggered the recent interest concerning adult neu-rogenesis. In the HVC, neurons that project to nucleusRA are generated throughout adult life. The recruitmentof these neurons is controlled by steroids (Alvarez-Buylla and Kirn 1997).

RA Is Required for Song Learning in Zebra Finches

In an expression screen using differential display PCR,an RA-producing enzyme, homologous to mouse


RALDH-2, was found to be enriched in HVC neuronsthat project to area X. These cells show a high affinityfor retinaldehyde and generate RA (Denisenko-Nehrbassand others 2000; Denisenko-Nehrbass and Mello 2001).In addition, large neurons in another song nucleus,IMAN, were also RALDH-positive (Fig. 4). This wasobserved in adult zebra finches, canaries, song sparrows,and black-capped chickadees. During the song ac-quisition period, neurons in the RA nucleus, too,expressed transcripts of the enzyme. To test the func-tional relevance of the RA system for song learning,Denisenko-Nehrbass and colleagues (2000) implantedcrystals of the RALDH inhibitor disulfiram bilaterallyabove the ventricular surface of the HVC. The operationwas performed in juvenile zebra finches during the sen-sitive period for song acquisition and in adults with acrystallized song. A careful analysis of the bird’s vocal-izations revealed that the birds that had received disulfi-ram implants as juveniles showed disruptions and a high

variability in song production. No effects were observedafter implantation of the enzyme inhibitor at differentbrain locations or in adult birds, and no morphologicalabnormalities were apparent. Although adult zebrafinches have a high expression of RALDH in the HVC,the absence of any effect of disulfiram on song produc-tion indicates that RA was required for synaptic plastic-ity only during acquisition, not production, of motor out-put for the crystallized song. A similar observation wasmade after ablation of area X–projecting HVC neurons,which affected song learning in juvenile but not in adultbirds (Denisenko-Nehrbass and others 2000). As in thecase of spatial learning paradigms associated with thehippocampus, RA might have a direct effect on physio-logical processes of synaptic plasticity. Whether this isaccomplished via an RAR effect on gene regulation re-mains to be investigated.

Implication of RA in Neurological Disorders

RA’s Functions in the Corpus Striatum: A Potential Role in Huntington’s and Parkinson’s Disease

Persuasive evidence points to a role for RA in striatalfunction, and many of the protein components of the RAsignaling system are present in the corpus striatum. Ofthe binding proteins, CRABP-I is expressed in medium-sized neurons scattered throughout the striatum(Zetterström and others 1994), CRABP-II is present inthe caudate/putamen and accumbens, and CRBPI ispresent in both these structures as well as the olfactorytubercle (Zetterström and others 1999). RARβ is the pre-dominant RA receptor in the striatum, although all threeRXRs are present (Zetterström and others 1999). Thesource of RA for the striatum may be somewhat unusu-al in the form of the terminal input to the striatum fromthe substantia nigra. The dopaminergic neurons of thesubstantia nigra express very high levels of the synthet-ic enzyme RALDH-1, which is transported along thenigrostriatal fibers to the striatum where it can generateRA (McCaffery and Dräger 1994). The sink for synthe-sized RA exists in the form of the catabolic enzymeCYP26B1 (Abu-Abed and others 2002). The importanceof RA for striatal function, and its possible link toHuntington’s disease (HD), is evident from studies ofmice that are transgenic for the 5′ end of the human HDgene carrying the CAG repeat expansion and mice thatdisplay a progressive neurological phenotype (Luthi-Carter and others 2000). Striatal mRNA profiling sho-wed a decrease in RXRγ, which is of interest becausenull mutants of RXRγ lead to a decline in choline acetyl-transferase in striatal cholinergic interneurons (Saga andothers 1999). It was further shown that 22 of the 104genes that decrease in this mutant mouse were regulatedby RA, including the dopamine D2 receptor, the promo-ter of which has an RA response element (Samad andothers 1997) and which, at least during development, is

Fig. 3. A mouse transgenic for a lacZ reporter gene driven by aretinoic acid (RA) response element shows regions of RA sig-naling in the hippocampus after fluorescence immunostainingfor the lacZ protein product. a, The strongest RA signal is evi-dent in a subpopulation of granule cells in the dentate gyruswith highest levels in the ventral, infrapyramidal blade. b, Dailyinjections with 13-cis RA at 1 mg/kg result in increased reporterexpression equalizing the levels in the infrapyramidal andsuprapyramidal blades.


driven by RA (Krezel and others 1998). Similarly, thedopamine D1 receptor is also reliant on RA for expres-sion (Krezel and others 1998). Null mutation of the D2receptor in mice results in a parkinsonian-like pheno-type, and it has also been suggested that RA is associ-ated with Parkinson’s disease (Eichele 1997; Krezel andothers 1998). A further association with Parkinson’s dis-ease is suggested from the lesions of the corpus striatumthat can result from use of Antabuse (disulfiram)(Laplane and others 1992; Fisher 1998), an aldehydedehydrogenase inhibitor used as an aversion therapy for

alcoholism (Brewer 1993). Because this drug is a potentinhibitor of RALDH-1 present in the nigrostriatal fibers(McCaffery and Dräger 1994), it implicates a supportivefunction for RA in the survival of the nigrostriatal dopa-minergic neurons.

RA’s Functions in the Spinal Cord: A Potential Role in Motoneuron Disease

Intriguing results by Corcoran, So, and Maden (2002)point to a role for RA in spinal cord motoneuron disor-

Fig. 4. RALDH expression in the song system of the zebra finch. a, Diagram of a parasagittal brain section showing position of com-ponents of the song system. b, In situ analysis shows zRALDH expression restricted to the high vocal center (HVC) (arrow), IMAN, andadjacent neostriatum; inset, expression in adult female HVC. c,d, Dark and bright field views of male HVC: Expression is confined tothe cytoarchitectonic boundaries of the HVC (arrowheads in d); arrow indicates the ventricle. e,f, Expression in HVC (e) and magno-cellular nucleus of the anterior neostriatum (f) occurs in large neurons (black arrows), smaller cells are not labeled (white arrows)(Denisenko-Nehrbass and others 2000).


ders with the finding that vitamin A depletion of ratsresults in astrocytosis and a significant loss of motoneu-rons in the lumbar spinal cord. This loss resulted fromneural degeneration with an accumulation of neurofila-ments combined with a decline in RARα (but not thesynthetic enzyme RALDH-2). This implies that just asRA is necessary for survival and differentiation of neu-rons in the embryonic CNS (Quinn and De Boni 1991;Wuarin and Sidell 1991; Maden 2003), it is also impor-tant for survival of adult neurons. Of particular interestwas an analysis of spinal cord tissue from a small po-pulation (n = 10) of spontaneous amyotrophic lateralsclerosis (ALS) patients. These samples showed a sig-nificant decrease in both RARα as well as RALDH-2,and it was suggested that the loss of RALDH-2 and cor-responding decline in RA synthesis may contribute tothe loss of neurons in ALS.

RA’s Functions in the Hippocampus: A Potential Role in Neurogenesis and Depression

Depression has been suggested to be a side effect of thechronic use of Accutane (isotretinoin, 13-cis RA) whenused as an oral treatment for severe acne (Hazen and oth-ers 1983; Lamberg 1998; Hull and D’Arcy 2003),although the link is not considered conclusive by all(Jick and others 2003). Recent evidence has pointed tothe hippocampus as a focal point for 13-cis RA’s dele-terious effects on the brain. A new hypothesis for theetiology of depression proposes that this disease is theoutcome of a decline in the birth of neurons in the hip-pocampus (Vogel 2000). This concept is based on theshrinkage of this structure in depression as well as theobservations that steroid hormones, which signal stress-induced depression, repress hippocampal neurogenesiswhereas antidepressant drugs promote neuronal birth(Sheline and others 1996; Cameron and others 1998;McEwen 1999; Malberg and others 2000; Sheline 2003).The sensitivity of the hippocampus to shrinkage likelyderives from its high level of plasticity; this is reflectedin both its modulation of dendritic complexity and itscapacity to maintain a level of new neuronal birth froma stem cell population that resides in the subgranularzone of the dentate gyrus (McEwen 1999). Given the po-tent effects of RA on neurogenesis (McCaffery andDräger 2000), it may be expected that chronic exposureto 13-cis RA is detrimental to these processes. It hasbeen shown in a mouse model that chronic exposure to aclinical dose (1 mg/kg/d) of 13-cis RA suppresses hip-pocampal neurogenesis and also disrupts hippocampal-dependant memory (Crandall and others 2004). Endoge-nously synthesized RA is also likely to influence thegranule cells of the dentate gyrus within the hippocam-pal formation. RA signaling within these cells is indicat-ed by the intense induction of an RA reporter gene in asubpopulation of these cells (Fig. 3a), predominantly inthe infrapyramidal blade. The influence of chronic 13-cisRA exposure on the balance of RA signaling is shown inFigure 3b, in which 3 weeks’ exposure to 1 mg/kg/d

results in an increase in the number of granule cells ex-pressing the reporter gene, particularly evident in thesuprapyramidal blade. This suggests that suprapyramidalblade is the region in which 13-cis RA may have itsgreatest effect on hippocampal neurogenesis.

RA’s Function in the Forebrain: A Potential Role in Schizophrenia

Goodman (1998) was the first to suggest that retinoidsmay be involved in the genesis of schizophrenia. Theassociation was tentatively based on the mapping ofgenes encoding components of RA signaling, with thechromosomal location of loci suggested to be linked toschizophrenia (several only weakly linked to this dis-ease). It was also noted that some of the congenitalabnormalities resulting from embryonic RA exposure,such as enlarged ventricles, microcephaly, and cra-niofacial and digital malformations, are associated withschizophrenia. The increasing evidence that schizophre-nia is a neurodevelopmental disease implies that anabnormality in RA signaling during development mightlead to damage resulting in greater susceptibility to thedisease later in life. For instance, abnormalities in neuralcrest cell differentiation or migration resulting from RAsignaling defects could account for brain and craniofa-cial abnormalities evident in schizophrenia (LaMantia1999). A possible association between RA and schizo-phrenia derives from recent findings of protein changesin the schizophrenic brain. Reelin, a secreted proteinthat, in the adult, may promote synaptogenesis inGABAergic cells, has been convincingly shown to bedownregulated in schizophrenia (Impagnatiello and oth-ers 1998). Expression of reelin can be controlled epige-netically via methylation of its promoter and RA maycontrol reelin expression via this mechanism (Chen andothers 2002). Despite such associations, however, directevidence for a link between RA and schizophreniaremains elusive.

Nerve Regeneration

To some extent, regeneration recapitulates the processesof embryonic development. The fact that RA guidesmany developmental events in the nervous system raisesthe possibility that it may also act as a signal for nerveregeneration. Whereas the transection of a peripheralnerve can be followed by successful restoration of func-tion, lesions in the CNS cause permanent damage.Among other causes, this difference is due to the factthat the nonneuronal constituents of the CNS (astro-cytes, oligodendrocytes, microglia) and PNS (Schwanncells, fibroblasts, macrophages) activate different path-ways of intercellular signal transduction and cause dif-ferential transcription of a large number of structural andregulatory genes (Gillen and others 1997; Raivich andothers 1999). The idea that RA plays a role in transcrip-tional control after nerve injury is supported by indirectevidence regarding the regulation of trauma-related


genes, by observations of retinoid receptors and bindingproteins after nerve injury, and by explant cultures ofdorsal root ganglia, retina, and spinal cord.

Regulation of Cytokine Expression by RA

Among the targets of transcriptional control are a multi-tude of structural genes, transcription factors, cell sur-face molecules, and also many cytokines that participatein traumatic reactions of the nervous system (Gudas andothers 1994; Mey 2001). Although these results derivealmost entirely from cell culture experiments, often withimmortalized cell lines not related to the nervous sys-tem, they do point at possible functions of RA afternerve injury. Among the pertinent proinflammatorycytokines interleukin (IL)-1α, IL-1β, IL-6, and TNFa,some of their downstream targets are suppressed orenhanced by RA depending on the cell culture system,and this is the same with various receptors for neuro-poietic cytokines (ciliary neurotrophic factor [CNTF],leukemia inhibitory factor, IL-6). The expression ofTGF−β cytokines, which often have a beneficial influ-ence on neuronal survival and regeneration, is generallyincreased with RA treatment (for review, see Mey 2001).Because the primary function of neurotrophins is seen inthe regulation of nerve cell survival and neuronal con-nectivity, their regulation by RA implies the closest con-nection between retinoids and nerve regeneration. Mostinvestigators who demonstrate a supportive effect by RAon axon growth or neuronal survival suggest the tran-scriptional activation of the neurotrophin receptor genesas the functional mechanism. This has most thoroughlybeen studied in developing sympathetic ganglia ofchicken, rat, and mouse (Rodríguez-Tébar and Rohrer1991; Holst and others 1995; Holst and others 1997;Hwa and others 1997; Kobayashi and others 1998; Wyattand others 1999). Although there seem to be major dif-ferences in the effect of RA on expression of receptortyrosine kinases in sympathetic neuroblasts of birds andmammals, experiments with specific agonists and antag-onists suggest that endogenous RA acts on Trk genes viaactivation of RARα in both classes of vertebrates (Holstand others 1995; Wyatt and others 1999). The receptor ofglial cell line–derived neurotrophic factor, GF-Rα1, isalso upregulated by way of the nuclear receptor RARα(Thang and others 2000). A different principle of signal-ing is based on direct cell-cell contacts, mediated via celladhesion molecules of the immunoglobulin superfamily,Ca2+-dependent cadherins, and integrins, the receptorsfor extracellular matrix molecules. These, too, are tran-scriptional targets of RA, and biological effects such asneuritic growth on laminin have been observed (Rossand others 1994; DiProspero and others 1997;Whitesides and others 1998; Medhora 2000).

RA-Induced Differentiation of Neurons and Glial Cells

During embryonic development of the nervous system,RA can regulate the differentiation of oligodendrocytes

(Noll and Miller 1994; Staines and others 1996), astro-cytes (Wuarin and others 1990), and nerve cells (Henionand Weston 1994; Thang and others 2000). The effect ofRA on neuronal differentiation has been studied exten-sively with embryonal carcinoma cells, in which a largenumber of RA-induced and repressed genes were identi-fied (Bouillet and others 1995; Oulad-Abdelghani andothers 1998; Cheung and Ip 2000; Cheung and others2000). Neuronal development, as well as regeneration,requires the outgrowth of axons. In cell culture experi-ments with embryonic tissues, RA increased axonal out-growth from the spinal cord (Wuarin and others 1990;Hunter and others 1991; Wuarin and Sidell 1991; Madenand others 1998), dorsal root ganglia (DRG) (Haskelland others 1987; Corcoran and others 2000), cerebellum(Yamamoto and others 1996), and sympathetic ganglia(Rodríguez-Tébar and Rohrer 1991; Holst and others1997). In one study, RA exerted a specific effect onBMP-induced dendrite development but not axonalelongation or survival of rat sympathetic neurons. It hastherefore been suggested that RA functions as an en-dogenous morphogen for neuronal cell shape and polar-ity (Chandrasekaran and others 2000).

RA Promotes Axonal Regeneration In Vitro

In addition to inducing neurite outgrowth from develop-ing neurons, cell culture studies have now shown thatRA also promotes regeneration of axons from differenti-ated nerve cells. In these experiments, attention was paidto possible synergistic interactions between RA andnerve growth factor (NGF), BDNF, and NT-3. Followinginjection of all-trans RA onto the chorio-allantoic mem-brane of stage E16 chick embryos, the retinas wereexplanted 24 h later, and axonal regeneration was moni-tored in organ cultures. Although RA alone did notincrease neurite outgrowth of retinal ganglion cells, itproduced a synergistic effect with application of BDNFthat was subsequently applied in vitro (Mey andRombach 1999). With neurons from DRG, RA seems toact downstream from a neurotrophin. In this case, NGFinduced neurite outgrowth from the DRG by activatingthe synthesis of all-trans RA: DRG cultured in the pres-ence of NGF-blocking antibodies and RA showed neu-rite outgrowth equivalent to that obtained with neurotro-phin alone, whereas the action of NGF was abolished bya blocker of RA-synthesizing aldehyde dehydrogenases(Corcoran and Maden 1999). On stimulation with RA,the NGF- and NT-3–responding cells upregulated theexpression of RARα1 and RARβ2. Because the inductionof neurite outgrowth was also accomplished with anRARβ-specific agonist but not with agonists for RARαor RARγ, the crucial transducer of the RA signal wasprobably RARβ2 (Corcoran and others 2000). This con-clusion was then tested with spinal cord explant cultures.In these experiments, RA-dependent neurite outgrowthfrom embryonic day 13.5 and 10-month-old mice or 3-month-old rats correlated positively with the expressionof RARβ. When adult spinal cords were then transfected


with RARβ2, axonal regeneration occurred even fromthese explants that otherwise produced no regeneratingneurites. Controls including the transduction withRARβ4 were negative, and in the absence of RARβ,NGF did not cause neurite extension (Corcoran, So,Barber, and others 2002). Therefore, the presence ofRARβ2 seems to be a prerequisite for axonal growth, atleast for some neuronal populations.

RA Signaling after Nerve Injury In Vivo

In the mammalian PNS, where axonal regenerationremains possible throughout life, the entire RA signal

transduction cascade is expressed: Transcripts of allretinoid receptors, of CRBP-I, CRABP-I, and CRABP–II;RA-synthesizing enzymes; and CYP26A1 have beenreported in rat sciatic nerves. Using a zymographybioassay, the activity of one aldehyde dehydrogenase,RALDH-2, was also detected in the PNS (Zhelyaznikand others 2003). Experiments with a transgenic reportermouse indicate that transcriptional activity of RA,although absent in normal nerves, is induced by a nervecrush (Fig. 5). Reporter gene activity distal from the le-sion site was highest 7 days after injury, that is, duringthe time of axonal regeneration. This reaction correlatedwith an increase in expression of CRBP-I and CRABP-

Fig. 5. Upregulation of CRABP-II after peripheral nerve injury. a, Relative mRNA expression of CRABP-II in crushed or transected sci-atic nerves of adult rats, distal from the site of the lesion. After different periods of time, expression was measured by quantitativereverse transcriptase–PCR, normalized to GAPDH and plotted in comparison to expression levels of the nonlesioned contralateralnerve. b, Relative CRABP-II immunoreactivity in crushed or transected sciatic nerves of adult rats, distal from the site of the lesion.Protein levels were quantified from signal strength in Western blots and compared to the IR of contralateral control nerves. Bars repre-sent means ± 95% levels of confidence. c, Transcriptional activity of retinoic acid in an RAREhspLacZ reporter mouse is observed 7days after sciatic nerve crush (blue staining) but not in nonlesioned nerves.


II. After cellular uptake of vitamin A, CRBP-I isinvolved in the oxidation of retinol to retinal. CRABP-II,on the other hand, facilitates the transport of RA to thenucleus (Budhu and Noy 2002). Because protein levelsof RALDH-2 did not change significantly after injury, itis possible that the transcriptional effect of RA was trig-gered through upregulation of CRBP-I and CRABP-II.In degeneration experiments in which axonal growthwas prevented, mRNA and protein expression ofCRABP-II remained at elevated levels, whereas axonalregeneration was followed by downregulation of CRBP-I and CRABP-II to control levels (Fig. 5). Similar to hip-pocampal learning paradigms, the research into themolecular consequences of injury-related RA signalingis just beginning. In the PNS, Schwann cells are likelytargets because receptors and binding proteins areexpressed in nerves severed from the spinal cord andDRG, so that increased expression after injury cannot beattributed to the neurons. In primary Schwann cell cul-tures, RA caused a decrease in the expression of CNTF,a response similar to what is observed in vivo (Johannand others 2003). As mentioned above, a number ofother signals that are involved in axonal regenerationcount among the possible targets of RA. The questionabout the precise role of RA signaling after injury ofCNS or PNS must now also be addressed.

References

Abu-Abed S, Dollé P, Metzger D, Beckett B, Chambon P, Petkovich M.2001. The retinoic acid-metabolizing enzyme, CYP26A1, is essen-tial for normal hindbrain patterning, vertebral identity, and devel-opment of posterior structures. Genes Dev 15:226–40.

Abu-Abed S, MacLean G, Fraulob V, Chambon P, Petkovich M, DolléP. 2002. Differential expression of the retinoic acid-metabolizingenzymes CYP26A1 and CYP26B1 during murine organogenesis.Mech Dev 110:173–7.

Alfos S, Boucheron C, Pallet V, Higueret D, Enderlin V, Béracochéa D,and others. 2001. A retinoic acid receptor antagonist suppressesbrain retinoic acid receptor overexpression and reverses a workingmemory deficit induced by chronic ethanol consumption in mice.Alcohol Clin Exp Res 25:1506–14.

Alvarez-Buylla A, Kirn JR. 1997. Birth, migration, incorporation, anddeath of vocal control neurons in adult songbirds. J Neurobiol33:585–601.

Andrews PW, Nudelman E, Hakomori S, Fenderson BA. 1990.Different patterns of glycolipid antigens are expressed followingdifferentiation of TERA-2 human embryonal carcinoma cellsinduced by retinoic acid, hexamethylene bisacetamide (HMBA) orbromodeoxyuridine (BUdR). Differentiation 43:131–8.

Bain G, Kitchens D, Yao M, Huettner JE, Gottlieb DI. 1995.Embryonic stem cells express neuronal properties in vitro. DevBiol 168:342–57.

Bouillet P, Oulad-Abdelghani M, Vicaire S, Garnier J-M, Schuhbaur B,Dolfi P, and others. 1995. Efficient cloning of cDNAs of retinoicacid-responsive genes in P19 embryonal carcinoma cells and char-acterization of a novel mouse gene, Stra1 (mouse LERK-2/Epgl2).Dev Biol 170:420–33.

Boylan JF, Gudas L. 1992. The level of CRABP-I expression influ-ences the amounts and types of all-trans-retinoic acid metabolitesin F9 teratocarcinoma stem cells. J Biol Chem 267:21486–91.

Brewer C. 1993. Recent developments in disulfiram treatment.Alcohol 28:383–95.

Budhu AS, Noy N. 2002. Direct channeling of retinoic acid betweencellular retinoic acid-binding protein II and retinoic acid receptorsensitizes mammary carcinoma cells to retinoic acid-inducedgrowth arrest. Mol Cell Biol 22:2632–41.

Cameron HA, Tanapat P, Gould E. 1998. Adrenal steroids and N-methyl-D-aspartate receptor activation regulate neurogenesis in thedentate gyrus of adult rats through a common pathway.Neuroscience 82:349–54.

Chambon P. 1996. A decade of molecular biology of retinoic acidreceptors. FASEB J 10:940–54.

Chandrasekaran V, Zhai Y, Wagner M, Kaplan PL, Napoli JL, HigginsD. 2000. Retinoic acid regulates the morphological development ofsympathetic neurons. J Neurobiol 42:383–93.

Chen Y, Sharma RP, Costa RH, Grayson DR. 2002. On the epigeneticregulation of the human reelin promoter. Nucleic Acids Res30:2930–9.

Cheung WM, Chu AH, Ip NY. 2000. Identification of candidate genesinduced by retinoic acid in embryonal carcinoma cells. JNeurochem 68:1882–8.

Cheung WMW, Ip NY. 2000. Interaction of neurotrophins and retinoicacid in neural differentiation. Mol Cells 8:496–502.

Chiang M-Y, Misner D, Kempermann G, Schikorski T, Giguère V,Sucov HM, and others. 1998. An essential role for retinoid recep-tors RARβ and RXRγ in long-term potentiation and depression.Neuron 21:1353–61.

Cocco S, Diaz G, Stancampiano R, Diana A, Carta M, Curreli R, andothers. 2002. Vitamin A deficiency produces spatial learning andmemory impairment in rats. Neuroscience 115:475–82.

Corcoran J, Maden M. 1999. Nerve growth factor acts via retinoic acidsynthesis to stimulate neurite outgrowth. Nat Neurosci 2:307–8.

Corcoran J, Shroot B, Pizzey J, Maden M. 2000. The role of retinoicacid receptors in neurite outgrowth from different populations ofembryonic mouse dorsal root ganglia. J Cell Sci 113:2567–74.

Corcoran J, So PL, Barber RD, Vincent KJ, Mazarakis ND,Mitrophanous KA, and others. 2002. Retinoic acid receptor beta2and neurite outgrowth in the adult mouse spinal cord in vitro. J CellSci 115:3779–86.

Corcoran J, So PL, Maden M. 2002. Absence of retinoids can inducesmotoneuron disease in the adult rat and a retinoid defect is presentin motoneuron disease patients. J Cell Sci 115:4735-41.

Crandall J, Sakai Y, Zhang J, Koul O, Mineur Y, Crusio WE, and oth-ers. 2004. 13-cis-retinoic acid suppresses hippocampal cell divi-sion and hippocampal-dependant learning in mice. Proc Natl AcadSci USA 100:5111-16.

Delva L, Bastie J-N, Rochette-Egly C, Kraiba R, Balitrand G, DespouyG, and others. 1999. Physical and functional interactions betweencellular retinoic acid binding proteins II and the retinoic acid-dependent nuclear complex. Mol Cell Biol 19:7158–67.

Denisenko-Nehrbass NI, Jarvis E, Scharff C, Nottebohm F, Mello CV.2000. Site-specific retinoic acid production in the brain of adultsongbirds. Neuron 27:359–70.

Denisenko-Nehrbass NI, Mello CV. 2001. Molecular targets of disulfi-ram action on song maturation in zebra finches. Mol Brain Res87:246–50.

DiProspero NA, Meiners S, Geller HM. 1997. Inflammatory cytokinesinteract to modulate extracellular matrix and astrocytic support ofneurite outgrowth. Exp Neurol 148:628–39.

Duester G. 2003. Alcohol dehydrogenase as a critical mediator ofretinoic acid synthesis from vitamin A in the mouse embryo. J Nutr128:459S–62S.

Eichele G. 1997. Retinoids: from hindbrain patterning to Parkinsondisease. Trends Genet 13:343–5.

Etchamendy N, Enderlin V, Marighetto A, Pallet V, Higueret P, JaffardR. 2003. Vitamin A deficiency and relational memory deficit inadult mice: relationships with changes in brain retinoid signalling.Behav Brain Res 145:37–49.

Etchamendy N, Enderlin V, Marighetto A, Vouimba R-M, Pallet V,Jaffard R, and others. 2002. Alleviation of a selective age-relatedrelational memory deficit in mice by pharmacologically inducednormalization of brain retinoid signaling. J Neurosci 21:6423–9.

Fisher CM. 1998. ‘Catatonia’ due to disulfiram toxicity. Arch Neurol46:789–804.

Ghyselinck NB, Bavik C, Sapin V, Mark M, Bonnier D, Hindelang C,and others. 1999. Cellular retinol-binding protein I is essential forvitamin A homeostasis. EMBO J 18:4903–14.

Gillen C, Korfhage C, Müller HW. 1997. Gene expression in nerveregeneration. Neuroscientist 3:112–22.


Goodman AB. 1998. Three independent lines of evidence suggestretinoids as causal to schizophrenia. Proc Natl Acad Sci U S A95:7240–44.

Gottesman ME, Quadro L, Blaner WS. 2001. Studies of vitamin Ametabolism in mouse model systems. Bioessays 23:419.

Grondona JM, Kastner P, Gansmuller A, Décimo D, Chambon P. 1996.Retinal dysplasia and degeneration in RARbeta2/RARgamma2compound mutant mice. Development 122:2173–88.

Gudas LJ, Sporn MB, Roberts AB. 1994. Cellular biology and bio-chemistry of the retinoids. In: Sporn MB, Roberts AB, GoodmanDS, editors. The retinoids: biology, chemistry, and medicine. NewYork: Raven.

Haskell BE, Stach RW, Werrbach-Perez K, Perez-Polo JR. 1987. Effectof retinoic acid on nerve growth factor receptors. Cell Tissue Res247:67–73.

Hazen PG, Carney JF, Walker AE, Stewart JJ. 1983. Depression: a sideeffect of 13-cis-retinoic acid therapy. J Am Acad Dermatol9:278–9.

Henion PS, Weston JA. 1994. Retinoic acid selectively promotes thesurvival and proliferation of neurogenic precursors in cultured neu-ral crest cell populations. Dev Biol 161:243–50.

Hill DP, Robertson KA. 1998. Differentiation of LA-N-5 neuroblas-toma cells into cholinergic neurons: methods for differentiation,immunohistochemistry and reporter gene introduction. Brain ResBrain Res Protoc 2:183–90.

Holst AV, Lefcort F, Rohrer H. 1997. TrkA expression levels of sym-pathetic neurons correlate with NGF-dependent survival duringdevelopment and after treatment with retinoic acid. Eur J Neurosci9:2169–77.

Holst Av, Rodríguez-Tébar A, Michaille J-J, Dhouailly D, BäckströmA, Ebendal T, and others. 1995. Retinoic acid-mediated increase inTrkA expression is sufficient to elicit NGF-dependent survival ofsympathetic neurons. Mol Cell Neurosci 6:185–98.

Hull PR, D’Arcy C. 2003. Isotretinoin use and subsequent depressionand suicide: presenting the evidence. Am J Clin Dermatol4:493–505.

Hunter K, Maden M, Summerbell D, Eriksson U, Holder N. 1991.Retinoic acid stimulates neurite outgrowth in the amphibian spinalcord. Proc Natl Acad Sci U S A 88:3666–70.

Hwa V, Oh Y, Rosenfeld RG. 1997. Insulin-like growth factor bindingprotein-3 and -5 are regulated by transforming growth factor-betaand retinoid acid in the human prostate adenocarcinoma cell linePC-3. Endocrine 6:235–42.

Impagnatiello F, Guidotti AR, Pesold C, Dwivedi Y, Caruncho H, PisuMG, and others. 1998. A decrease of reelin expression as a putativevulnerability factor in schizophrenia. Proc Natl Acad Sci USA95:1571823.

Ingraham CA, Maness PF. 1990. Tyrosine phosphorylated proteinsdecrease during differentiation of neuronal and glial cells. DevNeurosci 12:273–85.

Jick SS, Kremers HM, Vasilakis-Scaramozza C. 2003. Isotretinoin useand risk of depression, psychotic symptoms, suicide, and attempt-ed suicide. Arch Dermatol 136:1231–6.

Johann V, Jeliaznik N, Schrage K, Mey J. 2003. Retinoic acid down-regulates the expression of ciliary neurotrophic factor (CNTF) inrat Schwann cells. Neurosci Lett 339:13–6.

Jonk LJ, de Jonge ME, Vervaart JM, Wissink S, Kruijer W. 1994.Isolation and developmental expression of retinoic-acid-inducedgenes. Dev Biol 161:604–14.

Kobayashi M, Fujii M, Kurihara K, Matsuoka I. 1998. Bone morpho-genetic protein-2 and retinoic acid induce neurotrophin-3 respon-siveness in developing rat sympathetic neurons. Mol Brain Res53:206–17.

Krezel W, Ghyselinck NB, Samad TA, Dupé V, Kastner P, Borelli E,and others. 1998. Impaired locomotion and dopamine signaling inretinoid receptor mutant mice. Science 279:864–7.

Krezel W, Kastner P, Chambon P. 1999. Differential expression ofretinoid receptors in the adult mouse central nervous system.Neuroscience 89:1291–300.

LaMantia AS. 1999. Forebrain induction, retinoic acid, and vulnerabil-ity to schizophrenia: insights from molecular and generic analysisin developing mice. Biol Psychiat 46:19–30.

Lamberg L. 1998. Acne drug depression warnings highlight need forexpert care. JAMA 279:1057.

Laplane D, Attal N, Sauron B, de Billy A, Dubois B. 1992. Lesions ofbasal ganglia due to disulfiram neurotoxicity. J Neurol NeurosurgPsychiatry 55:925–9.

Leypoldt F, Lewerenz J, Methner A. 2001. Identification of genes up-regulated by retinoic-acid-induced differentiation of the humanneuronal precursor cell line NTERA-2 cl.D1. J Neurochem76:806–14.

Luthi-Carter R, Strand A, Peters NL, Solano SM, Hollingsworth ZR,Menon AS, and others. 2000. Decreased expression of striatal sig-naling genes in a mouse model of Huntington’s disease. Hum MolGenet 9:1259–71.

Maden M. 2003. Role and distribution of retinoic acid during CNSdevelopment. Int Rev Cytol 209:1–77.

Maden M, Keen G, Jones G. 1998. Retinoic acid as a chemotactic mol-ecule in neuronal development. Int J Dev Neurosci 16:317–22.

Malberg JE, Eisch AJ, Nestler EJ, Duman RS. 2000. Chronic antide-pressant treatment increases neurogenesis in adult rat hippocam-pus. J Neurosci 20:9104–10.

Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schütz G,Umesono K, and others. 1995. The nuclear receptor superfamily:the second decade. Cell 83:835–9.

Matsuoka I, Mizuno N, Kurihara K. 1989. Cholinergic differentiationof clonal rat pheochromocytoma cells (PC12) induced by retinoicacid: increase of choline acetyltransferase activity and decrease oftyrosine hydroxylase activity. Brain Res 502:53–60.

McBurney MW. 1993. P19 embryonal carcinoma cells. Int J Dev Biol37:135–40.

McCaffery P, Andreola F, Giandomenico V, De Luca LM. 2001.Vitamin A and gene expression. In: Moustaid Moussa N, BerdanierC, editors. Nutrient-gene interaction in health and disease. BocaRaton (FL): CRC Press. p 283–320.

McCaffery P, Dräger UC. 1994. High levels of retinoic acid-generatingdehydrogenases in the meso-telencephalic dopamine system. ProcNatl Acad Sci U S A 91:7772–6.

McCaffery P, Dräger UC. 2000. Regulation of retinoic acid signalingin the embryonic nervous system: a master differentiation factor.Cytokine Growth Factor Rev 11:233–49.

McCaffery P, Lee M-O, Wagner MA, Sladek NE, Dräger UC. 1992.Asymmetrical retinoic acid synthesis in the dorsoventral axis of theretina. Development 115:371–82.

McCaffery P, Wagner E, O’Neil J, Petkovic M, Dräger UC. 1999.Dorsal and ventral retinal territories defined by retinoic acid syn-thesis, break-down and nuclear receptor expression. Mech Dev82:119–30.

McEwen BS. 1999. Stress and hippocampal plasticity. Ann RevNeurosci 22:105–22.

Medhora MM. 2000. Retinoic acid upregulates beta(1)-integrin in vas-cular smooth muscle cells and alters adhesion to fibronectin. Am JPhysiol Heart Circ Physiol 279:382–7.

Mey J. 2001. Retinoic acid as a regulator of cytokine signaling afternerve injury. Z Naturforsch C 56c:163–76.

Mey J, McCaffery P, Dräger UC. 1997. Retinoid acid synthesis in thedeveloping chick retina. J Neurosci 17:7441–9.

Mey J, Rombach N. 1999. Retinoic acid increases BDNF-dependentregeneration of chick retinal ganglion cells in vitro. Neuroreport10:3573–7.

Misner DL, Jacobs S, Shimizu Y, de Urquiza AM, Solomin L,Perlmann T, and others. 2001. Vitamin A deprivation results inreversible loss of hippocampal long-term synaptic plasticity. ProcNatl Acad Sci U S A 98:11714–9.

Napoli JL. 1993. Biosynthesis and metabolism of retinoic acid: rolesof CRBP and CRABP in retinoic acid—roles of CRBP andCRABP in retinoic acid homeostasis. J Nutr 123:362–6.

Napoli JL. 1999. Retinoic acid: its biosynthesis and metabolism. NuclAcid Res Mol Biol 63:139–88.

Niederreither K, Subbarayan V, Dollé P, Chambon P. 1999. Embryonicretinoic acid synthesis is essential for early mouse post-implantationdevelopment. Nat Genet 21:444–8.


Noll E, Miller RH. 1994. Regulation of oligodendrocyte differentia-tion: a role for retinoic acid in the spinal cord. Development120:649–60.

Oulad-Abdelghani M, Chazaud C, Bouillet P, Mattei M-G, Dollé P,Chambon P. 1998. Stra3/lefty, a retinoic acid-inducible novel mem-ber of the transforming growth factor-β superfamily. Int J Dev Biol42:23–32.

Quinn SDP, De Boni U. 1991. Enhanced neuronal regeneration byretinoic acid of murine dorsal root ganglia and of fetal murine andhuman spinal cord in vitro. In Vitro Cell Dev Biol 27A:55–62.

Raivich G, Bohatschek M, Kloss CUA, Werner A, Jones LL,Kreutzberg GW. 1999. Neuroglial activation repertoire in theinjured brain: graded response, molecular mechanisms and cues tophysiological function. Brain Res Rev 30:77–105.

Rodríguez-Tébar A, Rohrer H. 1991. Retinoic acid induces NGF-dependent survival response and high-affinity binding NGF recep-tors in immature chick sympathetic neurons. Development112:813–20.

Ross S, Ahrens RA, De Luca LM. 1994. Retinoic acid enhances adhe-sivenes, laminin and integrin gamma1 synthesis and retinoic acidreceptor expression in F9 teratocarcinoma cells. J Cell Physiol159:263–73.

Saga Y, Kobayashi M, Ohta H, Murai N, Nakai N, Oshima M, and oth-ers. 1999. Impaired extrapyramidal function caused by the targeteddisruption of retinoid X receptor RXRgamma1 isoform. GenesCells 4:219–28.

Sakai Y, Meno C, Fujii H, Nishino J, Shiratori H, Saijoh Y, and others.2001. The retinoic acid-inactivating enzyme CYP26 is essential forestablishing an uneven distribution of retinoic acid along the ante-rio-posterior axis within the mouse embryo. Genes Dev15:213–25.

Samad TA, Krezel W, Chambon P, Borelli E. 1997. Regulation ofdopaminergic pathways by retinoids: activation of the D2 receptorpromoter by members of the retinoic acid receptor-retinoid Xreceptor family. Proc Natl Acad Sci U S A 94:14349–54.

Sheline YI. 2003. Neuroimaging studies of mood disorder effects onthe brain. Biol Psychiat 54:338–52.

Sheline YI, Wang PW, Gado MH, Csernansky JG, Vamier MW. 1996.Hippocampal atrophy in recurrent major depression. Proc NatlAcad Sci U S A 93:3908–13.

Staines WA, Craig J, Reuhl K, McBurney MW. 1996. Retinoic acidtreated P19 embryonal carcinoma cells differentiate into oligoden-drocytes capable of myelination. Neuroscience 71:845–53.

Swindell EC, Thaller C, Sockanathan S, Petkovic M, Jessell TM,Eichele G. 1999. Complementary domains of retinoic acid produc-tion and degradation in the early chick embryo. Dev Biol216:282–96.

Takahashi J, Palmer TD, Gage FH. 1999. Retinoic acid and neu-rotrophins collaborate to regulate neurogenesis in adult-derivedneural stem cell cultures. J Neurobiol 38:65–81.

Thang SH, Kobayashi M, Matsuoka I. 2000. Regulation of glial cellline-derived neurotrophic factor responsiveness in developing ratsympathetic neurons by retinoic acid and bone morphogeneticprotein-2. J Neurosci 20:2917–25.

Vogel G. 2000. Neuroscience: new brain cells prompt new theory ofdepression. Science 290:258–9.

Warkany L, Schraffenberger E. 1946. Congenital malformationsinduced in rats by maternal vitamin A deficiency: I. Defects of theeye. Arch Ophthalmol 35:150–69.

Whitesides J, Hall M, Anchan R, LaMantia A-S. 1998. Retinoid sig-naling distinguishes a subpopulation of olfactory receptor neuronsin the developing and adult mouse. J Comp Neurol 394:445–61.

Wilson JG, Roth CB, Warkany J. 1953. An analysis of the syndrome ofmalformations induced by maternal vitamin A deficiency: effectsof restoration of vitamin A at various times during gestation. Am JAnat 92:189–217.

Wuarin L, Sidell N. 1991. Differential susceptibilities of spinal cordneurons to retinoic acid-induced survival and differentiation. DevBiol 144:429–35.

Wuarin L, Sidell N, De Vellis J. 1990. Retinoids increase perinatalspinal cord neuronal survival and astroglial differentiation. Int JDev Neurosci 8:317–26.

Wyatt S, Andres R, Rohrer H, Davies AM. 1999. Regulation of neu-rotrophin receptor expression by retinoic acid in mouse sympathet-ic neuroblasts. J Neurosci 19:1062–71.

Yamagata T, Momoi MY, Yanagisawa M, Kumagai H, Yamakado M,Momoi T. 1994. Changes of the expression and distribution ofretinoic acid receptors during neurogenesis in mouse embryos. DevBrain Res 77:163–76.

Yamamoto M, McCaffery P, Dräger UC. 1996. Influence of the choroidplexus on cerebellar development: analysis of retinoic acid synthe-sis. Dev Brain Res 93:182–90.

Zetterström RH, Lindquist E, Mata de Urquiza A, Tomac A, ErikssonU, Perlmann T, and others. 1999. Role of retinoids in the CNS: dif-ferential expression of retinoid binding proteins and receptors andevidence for presence of retinoic acid. Eur J Neurosci 11:407–16.

Zetterström RH, Simon A, Giacobini MM, Eriksson U, Olson L. 1994.Localization of cellular retinoid-binding proteins suggests specificroles for retinoids in the adult central nervous system.Neuroscience 62:899–918.

Zhelyaznik N, Schrage K, McCaffery P, Mey J. 2003. Activation ofretinoic acid signaling after sciatic nerve injury: upregulation ofcellular retinoid binding proteins. Eur J Neurosci 18:1033–40.

review retinoic acid signaling in the nervous … · jÖrg mey and peter mccaffery the majority of...

Documents