Retinoid metabolism and nuclear receptor responses: New insights
into coordinated regulation of the PPAR–RXR complex
Ouliana Ziouzenkovaa, Jorge Plutzkyb,*
aDepartment of Human Nutrition, Ohio State University, Columbus, OH 43210, USA
bCardiovascular Division, Brigham and Women?s Hospital, Harvard Medical School, Boston, MA 02115, USA
Received 28 November 2007; accepted 28 November 2007
Available online 7 December 2007
Edited by Laszlo Nagy and Peter Tontonoz
regulate metabolism by activating specific nuclear receptors,
including the retinoic acid receptor (RAR) and the retinoid X
receptor (RXR). RXR, an obligate heterodimeric partner for
other nuclear receptors, including peroxisome proliferator-acti-
vated receptors (PPARs), helps coordinate energy balance. Re-
cently, many groups have identified new connections between
retinoid metabolism and PPAR responses. We found that reti-
naldehyde (Rald), a molecule that can yield RA through the ac-
tion of retinaldehyde dehydrogenases (Raldh), is present in fat
in vivo and can inhibit PPARc-induced adipogenesis. In vitro,
Rald inhibits RXR and PPARc activation. Raldh1-deficient mice
have increased Rald levels in fat, higher metabolic rates and body
temperatures, and are protected against diet-induced obesity and
insulin resistance. Interestingly, one specific asymmetric b-caro-
tene cleavage product, apo-140-carotenal, can also inhibit
PPARc and PPARa responses. These data highlight how path-
ways of b-carotene metabolism and specific retinoid metabolites
may have direct distinct metabolic effects.
Crown Copyright ? ? 2007 Published by Elsevier B.V. on behalf of
the Federation of European Biochemical Societies. All rights
Retinoids, naturally-occurring vitamin A derivatives,
Keywords: Retinoid; Retinaldehyde; PPARs; Adipogenesis;
1. Complex action of retinoids – a family of biologically active
mediators with poorly understood actions
Retinoids are a large group of naturally-occurring vitamin A
derivatives that regulate key cellular processes . Indeed, spe-
cific retinoids have been reported to be involved in cell differen-
tiation, cell cycle control, cell growth, and cellular responses to
cell injury. At the same time, retinoids are also strongly impli-
cated in various ways in the pathogenesis of obesity, diabetes,
and cardiovascular disease. Retinoids are also in clinical use
as pharmaceutical agents employed in treating specific patho-
logic conditions including leukemia and skin disorders [2,3].
Given this broad, fairly evolved portfolio – involvement in fun-
damental biologic processes, implication in pathogenesis, and
use in therapeutic applications – it is surprising how much re-
mains unknown about specific retinoid metabolites, their gen-
eration, their involvement in distinct biologic responses, or
how discrete pathways of retinoid metabolism influence disease
states or might be harnessed for clinical benefit. Here we will re-
view basic mechanisms through which retinoids exert their ef-
fects and discuss new insights into how retinoid metabolism
and handling may direct key metabolic pathways.
2. Mechanisms of retinoid action
Retinoids have been reported to exert versatile cellular ef-
fects including functional effects biomembranes , participa-
retinoylation , and phosphorylation of specific proteins .
A fundamental mechanism for retinoid action is through tran-
scriptional regulation . Retinoids regulate gene expression
by acting in a hormone-like fashion, binding to the ligand
binding domain of specific nuclear hormone receptors, and
thus activating them . In the case of the retinoic acid recep-
tor (RAR) , the highest binding affinity was seen with reti-
noic acid (RA) in the all-trans conformation . Two RA
isomers have been identified in vivo: the most abundant is
all-trans-RA, while 13-cis-RA is detected at significantly lower
concentrations in both mice and humans [9,10]. The 9-cis RA
isomer has never been identified in vivo , despite newer
methods able to quantitate fentomolar levels of this isomer
. All RA isomers bind and activate RAR in vitro, albeit
all-trans and 9-cis RA isomers bind more efficaciously than
13-cis isomer . The in vivo activation of all RAR isoforms
(RARa, RARb, RARc) is mediated by the all-trans RA iso-
mer, as reported in numerous genetic and pharmacological
The endogenous role of all-trans RA has been demonstrated
in mice lacking retinaldehyde dehydrogenase 2 (Raldh2), a key
enzyme for RA production during embryogenesis (Fig. 1). The
early growth arrest evident in Raldh2-deficient (Raldh2?/?) em-
bryosisrescuedbyadministrationof all-transRAor asynthetic
RAR ligand . In contrast 9-cis RA isomers could not com-
pensate for deficient RA synthesis in these mice . Thus, only
the all-trans RA isomer functions as an endogenous ligand for
RAR, despite otherRAisomers inpharmacologic studiesexert-
ing distinct effects in multiple pathways. Even this one example
highlights the importance of vitamin A (retinol) metabolism in
generating specific metabolites, and thus unique transcriptional
responses via activation of specific nuclear receptors. Other
*Corresponding author. Fax: +1 617 525 4366.
E-mail address: firstname.lastname@example.org (J. Plutzky).
0014-5793/$32.00 Crown Copyright ? 2007 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. All rights reserved.
FEBS Letters 582 (2008) 32–38
pathways exist that can determine RA formation. Retinyl esters
tion of retinol to retinaldehyde (Rald), followed by oxidation of
Rald to yield RA  (Fig. 1). These pathways define the impor-
tance of vitamin A as a parent compound from which various
metabolites are generated. Vitamin A is stored in a transcrip-
tionally inactive esterified form predominantly in liver and adi-
posetissue[2,8].Inaddition tothe natureof thespecificretinoid
ligand that is formed, another key determinant of biologic re-
sponses via vitamin A metabolism is the biology of the nuclear
receptors targeted by these molecules.
3. Nuclear receptors as mediators of retinoid responses
RAR belongs to a large family of steroid hormone nuclear
receptors sharing a similar pattern of regulation . Upon
activation by a cognate ligand, nuclear receptors attract small
molecules generally known as co-activators while releasing
co-repressors, ultimately forming an obligate heterodimeric
complex with another retinoid receptor, the retinoid X receptor
(RXR) . Although RAR controls many metabolic genes,
many others are not RAR targets . RXR can also regulate
transcription in a homodimeric complex as well. Through its
role as a required heterodimeric partner, RXR can control
the function of many other nuclear receptors, thus integrating
a unique transcriptional network dependent on retinoid metab-
olism and RXR responses . RXR forms heterodimers with
peroxisome proliferator-activated receptors (PPAR), the vita-
min D receptor, the liver X receptor, farnesoid X receptor
(FXR), and other key transcriptional sensors of nutrients and
metabolites that help maintain homeostasis in metabolism im-
mune responses as well as other pathways .
RXR is activated by its own ligand, the identity of which has
remained controversial. An early hypothesis suggested that
RXR could be activated through the generation of 9-cis RA,
a molecule that demonstrated high-affinity RXR binding
in vitro . In the absence of its ligand, RXR can form tetra-
mers that regulate distinct response elements [16,17]. In the
presence of 9-cis RA, the RXR tetramer dissociates into
RXR homodimers with unique transcriptional effects. Other li-
gand-activated nuclear receptors can recruit RXR for heterodi-
mer formation . Although 9-cis RA remains perhaps the
most studied specific RXR ligand , the evidence establishing
9-cis RA as the physiologic endogenous agonist for RXR is less
clear [9,11]. As noted, the evidence for this compound in vivo
has been lacking despite sensitive analytical techniques [9,11].
Other studies suggest that biologic RXR responses can occur
in the absence of 9-cis RA . Interestingly, activation of hete-
rodimeric partners of RXR, such as PPARs, can be accompa-
nied by changes in RA production. For example, treatment
of mice with the PPARa synthetic ligand fenofibrate depleted
the hepatic vitamin A content by 50% as compared to mice
on regular chow . Whether an unknown endogenous reti-
noid is produced from vitamin A in these reactions and can
activate RXR, thus supporting heterodimerization with PPAR
deserves further examination. Recent work establishes that the
PPARc agonist rosiglitazone can induce RA-generating en-
zymes retinaldehyde dehydrogenase (Raldh) 1 and 2 and subse-
quent RAR activation in dendritic cells . Since PPARs are
well-established regulators of metabolic pathways in their
own right , these data highlight the potential impact reti-
noids might have through effects on PPAR responses. The con-
nections between retinoid metabolism, PPARs and function
responses have been relatively poorly understood.
PPARs are a subfamily of nuclear receptors activated by var-
chain fatty acids have been proposed as natural PPAR ligands,
and indeed, specific endogenous pathways of lipid metabolism
nists [25–27]. All three PPAR isomers – PPARa, PPARb/d, and
PPARc – are involved in gene expression in multiple aspects of
sis, various aspects of lipoprotein metabolism and transport as
well as both peroxisomal and adaptive mitochondrial b-oxida-
tion of fatty acids . PPARb/d is a key regulator of fatty oxi-
dation in muscle and fat . PPARc plays a critical role in fat,
. PPAR mechanisms have also been linked to the regulation
ingtype2 diabetes (thiazolidinediones,
dyslipidemia (fibrates, PPARa) . Possible adverse effects of
Vitamin A (Retinol)
Other NR partners:
LXR, VDR, TR...
Fig. 1. Vitamin A and b-carotene metabolism generates Rald and RA.
The major steps in the metabolism of vitamin A (retinol), retinyl esters,
and b-carotene central cleavage leading to the generation of retinalde-
hyde (Rald). Asymmetric b-carotene cleavage produces apocarotenals
including b-apo-140carotenal (apo14, see Fig. 3). Retinaldehyde (Rald)
is generated from retinol by a family of alcohol dehydrogenases (Adh1,
3, 4). Short-chain alcohol dehydrogenases catalyze reversible reactions
involving retinol oxidation to Rald and Rald?s subsequent reduction to
retinol. Two Rald molecules can be produced as a result of symmetric
b-carotene cleavage. Retinaldehyde dehydrogenases (Raldh1–4) are a
family of enzymes that produce retinoic acid (RA) from Rald.
Retinoids function as ligands for various nuclear receptors (NRs) that
can promote (+) or repress (?) adipogenesis. All RA isomers are
natural ligands for RAR, while the 9-cis isomer can activate RXR
in vitro. RXR, as an obligate heterodimeric nuclear receptor (NR)
partner, can affect numerous other NRs involved in adipogenesis,
including the liver X receptor (LXR), the vitamin D receptor (VDR),
and the thyroidhormonereceptor(TR) tonamea few. Recentworkhas
shed new light on how retinoids are involved in PPAR responses,
including data shown here for how Rald and apo14 can inhibit RXR
and PPAR responses in vitro and in vivo in distinct ways.
O. Ziouzenkova, J. Plutzky / FEBS Letters 582 (2008) 32–38
some PPAR agonists have also been raised [33,34]. It is worth
nothing that the retinoid or RXR role in mediating the clinical
effects of PPAR agonists remains poorly understood. Recent
data continue to reveal specific and somewhat unexpected con-
nections between PPARs and RXR responses.
4. Retinoids in the integration of RXR and PPAR responses
As the data for RXR and PPARs as central regulators
jointly controlling multiple key responses emerged, proposed
models kept the biology of retinoids and fatty acids largely
separate. More recent studies have forced this separation to
be reappraised, with new insight into crosstalk between RXR
and PPAR signals. One example of this trend was already
mentioned – the evidence that PPARc could influence re-
sponses in dendritic cells via Raldh2 . Recently, Noy and
colleagues reported all-trans RA could also function as a spe-
cific ligand for PPARb/d . Moreover, RA partition between
RXR and PPARb/d depended on cellular RA binding protein
II or keratinocyte fatty acid binding protein, with these intra-
cellular transport proteins delivering RA selectively to either
RAR or PPARb/d, respectively . This channeling of RA
to RAR versus PPARb/d in different types of tissues may have
various implications, for example inhibiting or activating pro-
liferation of carcinogenic cells that affects growth of malignant
Emerging evidence has also suggested that the major vitamin
A metabolite, retinaldehyde (Rald), which had been previously
understood as mainly a unique precursor for RA, could also
function as a specific transcriptional regulator in its own right
. Rald can exist in various forms, a consequence, at least in
part, of its susceptibility to isomerization and oxidation .
These specific properties of Rald are utilized in eye, where
11-cis-Rald plays an essential role in night vision by complex-
ing with rhodopsin [1,38]. Rald?s role in vision depends on
Rald?s specific physical–chemical properties and is independent
in its conversion to RA. Outside the eye, Rald?s function has
remained essentially unknown . The fact that Rald produc-
tion from retinol is catalyzed by specific set of regulated en-
zymes [1,40] suggests a more important role for Rald in
dictating cellular responses. Rald can be generated by alcohol
dehydrogenases (Adh), short-chain dehydrogenases (Sdh or
Rdh), or central cleavage of b-carotene by b-carotene monoox-
ygenase (Bcm, Fig. 1) [1,40,41]. Sdh also catalyzes a reverse
reaction: Rald reduction to retinol . To a large part, Rald
levels depend on its subsequent catabolism by Raldh to RA
[40,42]. The differences in phenotypes among mice lacking dif-
ferent Adh and Raldh enzymes, and their respective isoforms,
reflect specific functions of these proteins . Moreover, the
expression patterns of Rald-generating and Rald catabolizing
enzymes do not overlap in different tissues, further suggesting
that Rald could play a specific metabolic role in its own right
outside of the eye .
Recently, we demonstrated that the enzymes that control
both Rald formation and catabolism are present in fat and dif-
ferentially regulated during adipogenesis . Both the induc-
tion of adipogenesis in vitro and the increased obesity evident
in ob/ob mice in vivo were accompanied by an increased ratio
of Raldh to Adh enzyme expression, changes expected to de-
crease Rald levels. Consistent with this, we found diet-induced
obesity in mice was associated with decreased Rald concentra-
tions in adipose tissue . Further examination of the effects
of Rald on transcriptional activity revealed Rald?s weak bind-
ing to both PPARc and RXR ligand binding domains and
inhibition of the activation of each of these nuclear receptors
in response to their cognate ligands. Correspondingly, in sev-
eral well established model systems, PPARc target gene
expression was reduced by Rald at low concentration ranges
(30–300 nM), corresponding to Rald levels present in Rald
fat in vivo . In keeping with Rald?s effects on both RXR
and PPARc activation in vitro, similar Rald concentrations
as those present in vivo inhibited adipogenesis, even in the
presence of a PPARc agonist. Interestingly, the effects of Rald
were both RXR-dependent and independent. For example,
Rald-mediated repression of adiponectin did not occur after
RXR silencing while triglyceride accumulation during adipo-
genesis persisted even during RXR silencing .
, although these mice had no obvious metabolic phenotype
. Given the in vitro data, we hypothesized that endogenous
Rald would regulate transcriptional responses in adipose tissue,
modulating adipogenesis and fat formation. Given the role of
the RXR–PPAR complexin energybalance,it seemedplausible
to test this idea under caloric stimulation, as seen with high fat
Consistent with the inhibition of adipogenesis seen after Rald
stimulation in vitro, endogenous increased Rald concentrations
in Raldh1?/?mice protected these mice from diet-induced obes-
ity and its associated insulin resistance . (Fig. 2). Of note,
Raldh1?/?mice remained insulin sensitive in spite of reduced
levels of adiponectin, a major adipokine implicated in the insu-
lin-sensitizing effects of thiazolidinediones .
Many potential inputs could contribute to the observed lean,
insulin sensitive phenotype of these mice. One possibility would
be a change in circulating adipokines, for example RBP4, a pro-
tein reported to be altered in insulin resistance states in both
mice and humans [44,45]. Interestingly, Rald binds directly to
RBP4 , raising the possibility that Rald may be involved in
the phenotypes seen in association with altered RBP4 levels
are the systemic shifts in retinoid metabolism deriving from
pense of RAproduction .Imbalance inthese retinoids could
white fat of high fat fed Raldh1?/?mice was characterized by
markedly smaller adipocytes than wild-type (WT) mice, a
change reported in other mouse models to correlate with in-
creased insulin sensitivity . The changes in adipocyte size
seen in Raldh1?/?mice corresponded directly with Rald levels
seen in adipocytes. It remains to be defined if these Rald-associ-
ated smaller adipocytes in vivo resulted from impaired PPARc
signaling and decreased adipocyte differentiation or some other
effect. In vitro, preadipocytes isolated from Raldh1?/?mice
demonstrated impaired differentiation and decreased PPARc
target gene expression as compared to similarly differentiated
cells from WT mice, thus suggesting possible adipocyte autono-
mous effects . Of note, these changes in adipogenesis in
trations of the PPARc agonist rosiglitazone , further sup-
porting Rald-mediated PPARc inhibition. Interestingly, prior
O. Ziouzenkova, J. Plutzky / FEBS Letters 582 (2008) 32–38
reports on the effects of a synthetic PPARc antagonist in mice
had similar systemic metabolic effects as Rald, including sup-
pression of fat formation and improved insulin sensitivity as
seen with increased Rald levels . Ultimately, the protection
against diet-induced obesity evident in Raldh1-deficient mice
must be understood in terms of energy balance. Raldh1?/?mice
ate a similar amount of food and water as WT mice, suggesting
some other mechanism for their protection from obesity during
high fat feeding . Metabolic cage studies revealed that
Raldh1?/?mice had increased metabolic rates, with significant
increasesinbasalbodytemperature (Fig.2). Thesedataidentify
increased dissipation of energy, and a fundamental shift in en-
ergy balance in the presence of increased Rald concentrations.
Are these effects specific to Raldh1 deficiency or are they evi-
dent in other models in which Rald levels are manipulated? In
ob/ob mice, administration of Rald, but not retinol or RA,
reproduced the phenotype seen with Raldh1 deficiency .
This divergence between administration of Rald and RA ar-
gues strongly that Rald exerts its own effects that are distinct
from simply serving as a precursor for RA formation. Of note,
administration of the general Raldh inhibitor citral to ob/ob
mice had similar effects as seen with Raldh1 deficiency, includ-
ing inhibition of fat formation and improved insulin sensitivity
. Further research will be required to more fully understand
the role of Rald in vivo and its specific mechanisms of action.
Such data will also help address whether Raldh1 inhibition
might be a therapeutic target for treating obesity and insulin
resistance. The prospect that Rald might serve as a potential
target for intervention would be bolstered by additional evi-
dence that Rald levels govern energy balance. Other studies
have begun to provide such support.
ifest obesity during feeding of a vitamin A-deficient diet .
Rdh1 is a proximal enzyme in RA generation that catalyzes a
reversible oxidation of retinol to Rald. This reaction depends
on the substrate concentrations as well as the redox status of
ity and Rald metabolism. A deficiency in Bcm (or CMO1), an-
Weight increase (g)
RALDH1– / –
High Fat Diet
(cal/min/g body weight)
0 80 160
Raldh1– / –
Fig. 2. Raldh1 deficiency protects against diet-induced obesity and insulin resistance by increasing metabolic rates. (A) Age- and sex-matched
Raldh1?/?mice (n = 9) and wild-type (WT, n = 10) mice received a high-fat diet (45 kcal%) for 180 days before undergoing metabolic studies. Two
representative X-ray imagesof each genotypeobtainedof randomly selected femalesfrom WT and Raldh1?/?groupsare shownand demonstrate body
mass differences evident after 120 days of high-fat feeding. There was no significant difference in the total calories consumed and water intake between
Raldh1?/?and WT mice. (B) Raldh1?/?mice (filled circles) weighed significantly less than WT (open circles) mice at 180 days (*P < 0.001, Wilcoxon
rank test), differences that were also evident throughout the study (data not shown). Significant differences in adiposity between WT and Raldh1?/?
mice were evident in fat pads (subcutaneous and omental) as well as liver (data not shown, P < 0.001, Mann–Whitney test). (C) Metabolic cage studies
were performed on WT (n = 5) and Raldh1?/?females (n = 4, black bars, d) on similar high fat diet (45 kcal%) for 180 days. Metabolic rates
(means ± S.D.) were measured over 24 h in WT mice (open circles, dashed line) and Raldh1?/?mice (black circles, solid line;*indicate significance
levels of P < 0.05, Wilcoxon rank test). Rectal temperature was significantly higher in Raldh1?/?mice versus WT mice (Wilcoxon rank test; not
shown).(D)InsulintolerancetestswereperformedonWT andRaldh1?/?mice asin (C)usingstandardapproaches.Thesestudiesrevealeda significant
improvement in insulin sensitivityin Raldh1?/?mice (black circles,solid line) versus WT mice (opencircles,dashed line;both tests performed after 16 h
fast). Data are presented as means ± S.E.,*P < 0.05, Mann–Whitney test. Glucose tolerance tests revealed similar results (not shown).
O. Ziouzenkova, J. Plutzky / FEBS Letters 582 (2008) 32–38
other Rald-generating enzyme, similarly results in obesity .
molecules . In these animal studies [47,48], Rald concentra-
tion in fat and expression of Rald metabolizing enzymes in adi-
pose tissue did not directly examine the connection between
Rald and obesity but these findings are consistent with Rald
another example of crosstalk between PPAR and vitamin A
metabolism, since Bcm is a PPARc-regulated target gene .
5. Apocarotenals: novel transcriptional modulators?
The evidence that certain retinoids may inhibit PPAR and
RXR responses and thus help regulate metabolism is not re-
stricted to Rald as a unique metabolite in the vitamin A path-
way. We identified that a separate pathway involved in
metabolizing b-carotene produces another aldehyde that can
also repress certain nuclear receptor responses, including those
of PPARc and PPARa . It has long been recognized that
which can be subsequently reduced to retinol or undergo oxida-
tion to form two RA molecules  (Fig. 3), thus explaining b-
carotene also being called provitamin A. More recent work
has established that b-carotene can also undergo asymmetric
cleavage, producing a series of molecules with varying chain
length known as apocarotenals (Fig. 3) . Although apoca-
for these molecules had previously been identified . We pos-
tulated apocarotenals might also be involved in transcriptional
responses. We found that one specific apocarotenal, b-apo140-
carotenal (apo14), but not other structurally similar apocarote-
nals, could inhibit PPARc, PPARa, or RXR activation and
well-established responses of these nuclear receptors . In
contrast to Rald, apo14 does not activate RAR. Inhibition of
PPARc and RXR by apo14 in adipocytes resulted in the sup-
pression of PPARc target genes, Fabp4 and adiponectin, and
adipogenesis . In endothelial cells, apo14 inhibited PPARa
activation and augmented expression of vascular cell adhesion
molecule-1 (VCAM-1), a known PPARa target gene (Fig. 4).
As predicted from the in vitro data, apo14 promoted inflamma-
tion in vivo, increasing leukocyte recruitment fivefold in the air
Rald, whose production is known to occur under physiologic
conditions, the asymmetric cleavage of b-carotene and genera-
tion of apo14 has been associated with pathologic conditions
such as oxidative stress [53,54]. Perhaps generation of apo14
contributes to the significant increase in mortality noted among
smokers taking b-carotene supplementation . Inhibition of
nuclear receptors by this b-carotene cleavage product could
potentially shed light on the failure of b-carotene supplementa-
tion to decrease cardiovascular events, perhaps by apo14 pro-
moting inflammation and atherosclerosis . In meta-
analysis studies of vitamin A and E supplementation trials, an
average 16% increase in mortality was found . These num-
bers arealarminggiven theincreasing numberofdairyproducts
supplemented with lipophylic vitamins .
6. The RXR–PPAR network
Nuclear receptors – like RXR and PPARs – by controlling
the expression of multiple different target genes, can integrate
upstream signals into coordinate systemic and programmatic
9-cis Retinoic acid
PPARα, γ, δ
Multiple Other NRs
Fig. 3. Pathways of b-carotene metabolism. The metabolism of b-carotene can generate specific molecules with different cellular effects. Symmetric
cleavage of b-carotene ultimately produces two molecules of retinoic acid (RA), a known agonist for RAR and RXR (through 9-cis RA isomers).
RXR can homodimerize or serve as an obligate heterodimeric partner for various other nuclear receptors, including all PPAR isotypes. Alternatively,
b-carotene may also undergo asymmetric cleavage yielding apocarotenals, a specific series of molecules named according to the length of their side
chains. Apocarotenals are thought to be generated under conditions of oxidative stress and/or inflammation.
O. Ziouzenkova, J. Plutzky / FEBS Letters 582 (2008) 32–38
responses. The fact that the input for activating both RXR and
PPARs derive from the metabolism of dietary nutrients yield-
ing b-carotene metabolites, retinoids, and fatty acids identifies
these receptors as potential links between diet, energy balance,
and pathologic conditions influenced by nutrition such as dia-
betes and atherosclerosis. Understanding how RXR and
PPARs are involved in such conditions, and how these recep-
tors might be best targeted for therapeutic benefit, will require
considerably more studies. Two important but poorly under-
stood aspects of this RXR–PPAR network are now coming
under increased scrutiny, namely the coordinated interplay be-
tween RXR–PPAR signals and how natural mediators may
modulate RXR–PPAR responses. The concept of natural
PPAR and RXR inhibitors is relatively new [37,51] but adds
to the list of candidate mechanisms involved in helping termi-
nate if not regulate these important transcriptional mediators.
Other nuclear receptors have already established the relevance
of negative inputs to biologic responses. For example, the reg-
ulation of FXR, another RXR heterodimeric partner, occurs
in part through the production of specific, natural bile acids
with either agonistic and antagonistic properties [58,59]. Our
studies discussed here reveal that vitamin A and b-carotene
metabolism can produce transcriptionally active metabolites
with counteracting properties on specific nuclear receptors
and hence the metabolic pathways regulated by those receptors
[37,51]. Further insight into metabolism of vitamin A and its
bioactive metabolites will likely continue to reveal new targets
and mechanisms that will help explain how these pathways
control energy balance and might ultimately be manipulated
in the treatment of metabolic disorders.
Acknowledgement: The authors recognise support from the AHA SDG
0530101N (O.Z.) and the NIH RO1 071745, PD 048743 (J.P.).
 Napoli, J.L. (1996) Retinoic acid biosynthesis and metabolism.
FASEB J. 10, 993–1001.
 Bonet, M.L., Ribot, J., Felipe, F. and Palou, A. (2003) Vitamin A
and the regulation of fat reserves. Cell. Mol. Life. Sci. 60, 1311–
 Villarroya, F., Iglesias, R. and Giralt, M. (2004) Retinoids and
retinoid receptors in the control of energy balance: novel
pharmacological strategies in obesity and diabetes. Curr. Med.
Chem. 11, 795–805.
 Stillwell, W. and Nahmias, S. (1983) Effect of retinol and retinoic
acid on P/O ratios of coupled mitochondria. Biochem. Int. 6, 385–
 Kubo, Y., Wada, M., Ohba, T. and Takahashi, N. (2005)
Formation of retinoylated proteins from retinoyl-CoA in rat
tissues. J. Biochem. (Tokyo) 138, 493–500.
 Kambhampati, S., Verma, A., Li, Y., Parmar, S., Sassano, A. and
Platanias, L.C. (2004) Signalling pathways activated by all-trans-
retinoic acid in acute promyelocytic leukemia cells. Leuk. Lym-
phoma 45, 2175–2185.
 Mark, M., Ghyselinck, N.B., Wendling, O., Dupe, V., Mascrez,
B., Kastner, P. and Chambon, P. (1999) A genetic dissection of
the retinoid signalling pathway in the mouse. Proc. Nutr. Soc. 58,
 Repa, J.J., Hanson, K.K. and Clagett-Dame, M. (1993) All-trans-
retinol is a ligand for the retinoic acid receptors. Proc. Natl. Acad.
Sci. USA 90, 7293–7297.
 Kane, M.A., Chen, N., Sparks, S. and Napoli, J.L. (2005)
Quantification of endogenous retinoic acid in limited biological
samples by LC/MS/MS. Biochem. J. 388, 363–369.
 Tang, G.W. and Russell, R.M. (1990) 13-cis-Retinoic acid is an
endogenous compound in human serum. J. Lipid Res. 31, 175–
 Mic, F.A., Molotkov, A., Benbrook, D.M. and Duester, G.
(2003) Retinoid activation of retinoic acid receptor but not
retinoid X receptor is sufficient to rescue lethal defect in retinoic
acid synthesis. Proc. Natl. Acad. Sci. USA 100, 7135–7140.
 Germain, P. et al. (2006) International Union of Pharmacology.
LX. Retinoic acid receptors. Pharmacol. Rev. 58, 712–725.
 Shulman, A.I. and Mangelsdorf, D.J. (2005) Retinoid x receptor
heterodimers in the metabolic syndrome. N. Engl. J. Med. 353,
 Mukherjee, R. et al. (1997) Sensitization of diabetic and obese
mice to insulin by retinoid X receptor agonists. Nature 386, 407–
 Heyman, R.A., Mangelsdorf, D.J., Dyck, J.A., Stein, R.B.,
Eichele, G., Evans, R.M. and Thaller, C. (1992) 9-cis Retinoic
acid is a high affinity ligand for the retinoid X receptor. Cell 68,
Apo14: Before WY After WY
Leukocyte count, x106
Fig. 4. Apo14, but not other related apocarotenals, inhibits PPARa responses in vitro and in vivo. (A) Apo14 promotes TNFa-mediated vascular
adhesion molecule-1 expression in vitro. Apo14 increased TNFa (20 ng/ml, 15 h) induced VCAM1 mRNA expression in human endothelial cells as
shown by Northern blotting while apo8 had no such effects (10 mM, 3 h for both apocarotenals). Stimulation of similarly treated endothelial cells with
apo14 reversed known PPARa-mediated repression of TNFa-induced VCAM1 mRNA expression in a manner dependent on both apo14 incubation
time and order of stimulation. Apo14 prestimulation (10 lM) for 3 h, but not 15 min, reversed the known repression of TNFa-induced VCAM-1
mRNA expression in human endothelial cells seen with the PPARa agonist WY14643 (250 lM, 3 h). Using the same concentrations/conditions as
before, the opposite order of stimulation – WY first followed by apo14 treatment – had no effect on VCAM1 expression. Such data would suggest
possible competition between the PPARa WY and apo14 for PPARa activation. (B) The air pouch model is a well-established method for measuring
inflammatory cell recruitment to a subcutaneous air pouch generated in the dorsum of mice. In this model, apo14 increased leukocyte recruitment
in vivo in a PPARa-dependent manner. Pre-treatment with apo14 (10 lM, 2 h) increased TNFa-induced (25 ng/ml for 2 h) leukocyte recruitment
in vivo in WT mice (P < 0.001). In contrast, apo14 had no effect in PPARa-deficient mice. As seen in many other studies, PPARa deficiency is
associated with increased basal inflammation, as evident in the increased recruitment of inflammatory cells to the air pouch in the absence of PPARa.
O. Ziouzenkova, J. Plutzky / FEBS Letters 582 (2008) 32–38
 Kersten, S., Reczek, P.R. and Noy, N. (1997) The tetramerization Download full-text
region of the retinoid X receptor is important for transcriptional
activation by the receptor. J. Biol. Chem. 272, 29759–29768.
 Yasmin, R., Yeung, K.T., Chung, R.H., Gaczynska, M.E.,
Osmulski, P.A. and Noy, N. (2004) DNA-looping by RXR
tetramers permits transcriptional regulation ‘‘at a distance’’. J.
Mol. Biol. 343, 327–338.
 Vivat-Hannah, V., Bourguet, W., Gottardis, M. and Gronemeyer,
H. (2003) Separation of retinoid X receptor homo- and hetero-
dimerization functions. Mol. Cell Biol. 23, 7678–7688.
 Germain, P. et al. (2006) International Union of Pharmacology.
LXIII. Retinoid X receptors. Pharmacol. Rev. 58, 760–772.
 Lei, Z., Chen, W., Zhang, M. and Napoli, J.L. (2003) Reduction
of all-trans-retinal in the mouse liver peroxisome fraction by the
short-chain dehydrogenase/reductase RRD: induction by the
PPAR alpha ligand clofibrate. Biochemistry 42, 4190–4196.
 Szatmari, I. et al. (2006) PPARgamma controls CD1d expression
by turning on retinoic acid synthesis in developing human
dendritic cells. J. Exp. Med. 203, 2351–2362.
 Michalik, L. et al. (2006) International Union of Pharmacology.
LXI. Peroxisome proliferator-activated receptors. Pharmacol.
Rev. 58, 726–741.
 Krey, G., Braissant, O., L?Horset, F., Kalkhoven, E., Perroud,
M., Parker, M.G. and Wahli, W. (1997) Fatty acids, eicosanoids,
and hypolipidemic agents identified as ligands of peroxisome
proliferator-activated receptors by coactivator dependent receptor
ligand assay. Mol. Endocrinol. 11, 779–791.
 Kliewer, S.A., Forman, B.M., Blumberg, B., Ong, E.S., Borg-
meyer, U., Mangelsdorf, D.J., Umesono, K. and Evans, R.M.
(1994) Differential expression and activation of a family of murine
peroxisome proliferator-activated receptors. Proc. Natl. Acad.
Sci. USA 91, 7355–7359.
 Ziouzenkova, O. et al. (2003) Lipolysis of triglyceride-rich
lipoproteins generates PPAR ligands: evidence for an antiinflam-
matory role for lipoprotein lipase. Proc. Natl. Acad. Sci. USA
 Chawla, A. et al. (2003) PPARdelta is a very low-density
lipoprotein sensor in macrophages. Proc. Natl. Acad. Sci. USA
 Ahmed, W., Orasanu, G., Nehra, V., Asatryan, L., Rader, D.J.,
Ziouzenkova, O. and Plutzky, J. (2006) High-density lipoprotein
hydrolysis by endothelial lipase activates PPARalpha: a candidate
mechanism for high-density lipoprotein-mediated repression of
leukocyte adhesion. Circ. Res. 98, 490–498.
 Grimaldi, P.A. (2007) Lipid sensing and lipid sensors: peroxisome
proliferator-activated receptors as sensors of fatty acids and
derivatives. Cell Mol. Life. Sci. 64, 2459–2464.
 Barish, G.D., Narkar, V.A. and Evans, R.M. (2006) PPAR delta:
a dagger in the heart of the metabolic syndrome. J. Clin. Invest.
 Rosen, E.D., Sarraf, P., Troy, A.E., Bradwin, G., Moore, K.,
Milstone, D.S., Spiegelman, B.M. and Mortensen, R.M. (1999)
PPAR gamma is required for the differentiation of adipose tissue
in vivo and in vitro. Mol. Cell. 4, 611–617.
 Ziouzenkova, O. and Plutzky, J. (2004) Lipolytic PPAR
activation: new insights into the intersection of triglycerides
and inflammation? Curr. Opin. Clin. Nutr. Metab. Care. 7,
 Brown, J.D. and Plutzky, J. (2007) Peroxisome proliferator-
activated receptors as transcriptional nodal points and therapeu-
tic targets. Circulation 115, 518–533.
 Nissen, S.E. and Wolski, K. (2007) Effect of rosiglitazone on the
risk of myocardial infarction and death from cardiovascular
causes. N. Engl. J. Med. 356, 2457–2471.
 Singh, S., Loke, Y.K. and Furberg, C.D. (2007) Long-term risk of
cardiovascular events with rosiglitazone: a meta-analysis. JAMA
 Shaw, N., Elholm, M. and Noy, N. (2003) Retinoic acid is a high
affinity selective ligand for the peroxisome proliferator-activated
receptor beta/delta. J. Biol. Chem. 278, 41589–41592.
 Schug, T.T., Berry, D.C., Shaw, N.S., Travis, S.N. and Noy, N.
(2007) Opposing effects of retinoic acid on cell growth result from
alternate activation of two different nuclear receptors. Cell 129,
 Ziouzenkova, O. et al. (2007) Retinaldehyde represses adipogen-
esis and diet-induced obesity. Nat. Med. 13, 695–702.
 Jang, G.F., McBee, J.K., Alekseev, A.M., Haeseleer, F. and
Palczewski, K. (2000) Stereoisomeric specificity of the retinoid
cycle in the vertebrate retina. J. Biol. Chem. 275, 28128–28138.
 Desvergne, B. (2007) Retinaldehyde: more than meets the eye.
Nat. Med. 13, 671–673.
 Duester, G., Mic, F.A. and Molotkov, A. (2003) Cytosolic
retinoid dehydrogenases govern ubiquitous metabolism of retinol
to retinaldehyde followed by tissue-specific metabolism to retinoic
acid. Chem. Biol. Interact., 201–210.
 Lakshman, M.R. (2004) Alpha and omega of carotenoid cleavage.
J. Nutr. 134, 241S–245S.
 Molotkov, A. and Duester, G. (2003) Genetic evidence that
retinaldehyde dehydrogenase Raldh1 (Aldh1a1) functions down-
stream of alcohol dehydrogenase Adh1 in metabolism of retinol
to retinoic acid. J. Biol. Chem. 278, 36085–36090.
 Duester, G. (2001) Genetic dissection of retinoid dehydrogenases.
Chem. Biol. Interact., 469–480.
 Yang, Q. et al. (2005) Serum retinol binding protein 4 contributes
to insulin resistance in obesity and type 2 diabetes. Nature 436,
 Graham, T.E. et al. (2006) Retinol-binding protein 4 and insulin
resistance in lean, obese, and diabetic subjects. N. Engl. J. Med.
 Rieusset, J., Touri, F., Michalik, L., Escher, P., Desvergne, B.,
Niesor, E. and Wahli, W. (2002) A new selective peroxisome
proliferator-activated receptor gamma antagonist with antiobesity
and antidiabetic activity. Mol. Endocrinol. 16, 2628–2644.
 Zhang, M., Hu, P., Krois, C.R., Kane, M.A. and Napoli, J.L.
(2007) Altered vitamin A homeostasis and increased size and
adiposity in the rdh1-null mouse. FASEB J. 21, 2886–2896.
 Hessel, S. et al. (2007) CMO1 deficiency abolishes vitamin A
production from beta-carotene and alters lipid metabolism in
mice. J. Biol. Chem. 282, 33553–33561.
 Krinsky, N. and Russell, R.M. (2001) Regarding the conversion
of beta-carotene to vitamin A. Nutr. Rev. 59, 309.
 Boulanger, A., McLemore, P., Copeland, N.G., Gilbert, D.J.,
Jenkins, N.A., Yu, S.S., Gentleman, S. and Redmond, T.M.
(2003) Identification of beta-carotene 15, 150-monooxygenase as a
peroxisome proliferator-activated receptor target gene. FASEB J.
 Ziouzenkova, O. et al. (2007) Asymmetric cleavage of beta-
carotene yields a transcriptional repressor of retinoid X receptor
and peroxisome proliferator-activated receptor responses. Mol.
Endocrinol. 21, 77–88.
 Wang, X.D., Tang, G.W., Fox, J.G., Krinsky, N.I. and Russell,
R.M. (1991) Enzymatic conversion of beta-carotene into beta-
apo-carotenals and retinoids by human, monkey, ferret, and rat
tissues. Arch. Biochem. Biophys. 285, 8–16.
 Handelman, G.J., van Kuijk, F.J., Chatterjee, A. and Krinsky,
N.I. (1991) Characterization of products formed during the
autoxidation of beta-carotene. Free Radic. Biol. Med. 10, 427–
 Wu, Z., Robinson, D.S., Hughes, R.K., Casey, R., Hardy, D. and
West, S.I. (1999) Co-oxidation of beta-carotene catalyzed by
soybean and recombinant pea lipoxygenases. J. Agric. Food
Chem. 47, 4899–4906.
 Rapola, J.M., Virtamo, J., Ripatti, S., Huttunen, J.K., Albanes,
D., Taylor, P.R. and Heinonen, O.P. (1997) Randomised trial of
alpha-tocopherol and betacarotene supplements on incidence of
major coronary events in men with previous myocardial infarc-
tion. Lancet 349, 1715–1720.
 Bjelakovic, G., Nikolova, D., Gluud, L.L., Simonetti, R.G. and
Gluud, C. (2007) Mortality in randomized trials of antioxidant
supplements for primary and secondary prevention: systematic
review and meta-analysis. JAMA 297, 842–857.
 Penniston, K.L. and Tanumihardjo, S.A. (2006) The acute and
chronic toxic effects of vitamin A. Am. J. Clin. Nutr. 83, 191–201.
 Makishima, M. et al. (1999) Identification of a nuclear receptor
for bile acids. Science 284, 1362–1365.
 Yu, J. et al. (2002) Lithocholic acid decreases expression of bile
salt export pump through farnesoid X receptor antagonist
activity. J. Biol. Chem. 277, 31441–31447.
O. Ziouzenkova, J. Plutzky / FEBS Letters 582 (2008) 32–38