Liver X Receptor Regulation of Thyrotropin-Releasing Hormone Transcription in Mouse Hypothalamus Is Dependent on Thyroid Status

Reversing the escalating rate of obesity requires increased knowledge of the molecular mechanisms controlling energy balance. Liver X receptors (LXRs) and thyroid hormone receptors (TRs) are key physiological regulators of energetic metabolism. Analysing interactions between these receptors in the periphery has led to a better understanding of the mechanisms involved in metabolic diseases. However, no data is available on such interactions in the brain. We tested the hypothesis that hypothalamic LXR/TR interactions could co-regulate signalling pathways involved in the central regulation of metabolism. Using in vivo gene transfer we show that LXR activation by its synthetic agonist GW3965 represses the transcriptional activity of two key metabolic genes, Thyrotropin-releasing hormone (Trh) and Melanocortin receptor type 4 (Mc4r) in the hypothalamus of euthyroid mice. Interestingly, this repression did not occur in hypothyroid mice but was restored in the case of Trh by thyroid hormone (TH) treatment, highlighting the role of the triiodothyronine (T3) and TRs in this dialogue. Using shLXR to knock-down LXRs in vivo in euthyroid newborn mice, not only abrogated Trh repression but actually increased Trh transcription, revealing a potential inhibitory effect of LXR on the Hypothalamic-Pituitary-Thyroid axis. In vivo chromatin immunoprecipitation (ChIP) revealed LXR to be present on the Trh promoter region in the presence of T3 and that Retinoid X Receptor (RXR), a heterodimerization partner for both TR and LXR, was never recruited simultaneously with LXR. Interactions between the TR and LXR pathways were confirmed by qPCR experiments. T3 treatment of newborn mice induced hypothalamic expression of certain key LXR target genes implicated in metabolism and inflammation. Taken together the results indicate that the crosstalk between LXR and TR signalling in the hypothalamus centres on metabolic and inflammatory pathways.


Introduction
Obesity contributes to the aetiology of common associated metabolic diseases [1][2][3]. The causes of obesity are multifactorial, with environmental and genetic components. Westernized lifestyle, with high-caloric diets and a lack of physical exercise, is an obvious factor of obesity [4], but obesity does not develop in all individuals exposed to an obesogenic environment. Such observations emphasize the need to increase our knowledge of the molecular pathways and mechanisms involved in controlling energy balance.
In the periphery, regulation of energetic metabolism involves several types of nuclear receptors (NR). Among NRs, LXRs, activated by cholesterol metabolites, are known to be key regulators of lipid and cholesterol metabolism. The two related LXRs, LXRa (NR1H3) and LXRb (NR1H2) are part of the emerging significant newer drug targets within the NR family (for review see [5]). A second type of NR, TRs, plays a major role in controlling energy metabolism. THs are known to regulate, at a transcriptional level, all the steps of cholesterol metabolism, and TRb1 is the main receptor isoform involved [6]. Given the role played by these NR in metabolism, dysregulations of metabolic functions controlled by LXR/TRs can alter the homeostatic control circuits, contributing to the pathogenesis of many common metabolic diseases, such as obesity, insulin resistance, type 2 diabetes, hyperlipidaemia, atherosclerosis, and gallbladder disease [7]. Interactions between LXR and TR have been described in the periphery [8,9]. These interactions are based on two major points: First, LXRs and TRs both use retinoid X receptors (RXRs) as common heterodimerization partner [10], a fact that can engender competition for RXR if quantities are limiting. Second, the consensus response elements (RE) for LXRs and TRs are very similar (basically a DR4). However, so far no data are available on how such interactions affect regulations at the central level, whereas, it is well known that hypothalamus is considered as the central integrator of metabolic regulation (for review see [11]). Consequently, the key players are the central controls of food intake, relayed by neural and gene networks in different hypothalamic nuclei. Indeed, in the context of thyroid hormone (TH)-induced gene regulation, our group has focused on hypothalamic interactions between different signalling pathways controlling metabolism [12,13]. In particular, we showed that TH, through TRs, is directly involved in transcriptional regulation of melanocortin receptor type 4 (Mc4r) [12]. Moreover, the need for a detailed study on the involvement of LXR in central metabolic pathways and control of energy homeostasis is underlined by the fact that it has recently been shown that the central melanocortin pathway, particularly hypothalamic MC4R is involved in the control of hepatic cholesterol metabolism: in addition to facilitating hepatic cholesterol synthesis, the central melanocortin system influences cholesterol transport by modulating HDL cholesterol levels [14]. These results lead to the hypothesis that LXR signalling, in the hypothalamus, may interact with this pathway.
In the current study, we analysed, at the hypothalamic level, how LXR may interfere with transcriptional regulation induced by TRs. Using in vivo gene transfer we show that activation of LXR by its synthetic agonist GW3965 represses the transcriptional activity of two known TH target genes involved in the central control of metabolism, Trh and Mc4r promoters, and this only in euthyroid mice. This repression was restored by TH treatment in hypothyroid mice. By in vivo ChIP, we showed that LXR is recruited to the Trh promoter region in the presence of T 3 but not in its absence. In contrast, RXR is recruited to the same region in the absence of T 3 . There is no simultaneous presence of RXR and LXR on Trh promoter region, suggesting that the presence of a receptor excludes the other. The interactions between the two pathways were further confirmed by qPCR, showing that T 3 treatment of newborn mice induced hypothalamic regulation of a number of key LXR target genes implicated in metabolism and inflammation.
Thus, our data provide insight into molecular pathways involved in central metabolic regulation, showing for the first time a central crosstalk between TR and LXR pathways.

Ethics statement
Animal care and experimentation were in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and approved by the Museum National d'Histoire Naturelle Animal Care and Use Committee, Paris, France.

Experimental protocol
Animals. Swiss wild-type mice were from Janvier (Le Genest St Isle, France).
T 3 treatment. To assess T 3 (Sigma-Aldrich) effect on restoring ligand-dependent repression of the promoters of Trh and Mc4r in the hypothalamus via activation of LXR by GW3965, hypothyroid pups were injected subcutaneously (s.c.) with 2.5 mg/g body weight (bw) of T 3 in 0.9% saline 24 hours before transfection. Pups were decapitated 24 h after transfection and hypothalami were dissected for luciferase assays following the manufacturer's protocol (Promega).
In the qPCR experiments assessing response to T 3 , we used a s.c. T 3 (2.5 mg/g bw) or vehicle (0.9% saline) injection 6 h and 18 h before sacrifice and dissection.
Thyroxine (T 4 ) treatment. Dams were treated with 1.2 mg/ ml of T 4 in drinking water from 24 or 48 hours before giving birth and the treatment in the drinking water was continued through lactation.
Plasmids TRH-Firefly luciferase (TRH-f.luc) was from Plasmid 2547 to + 84 Renilla luciferase (MC4R-shRNA-coding plasmids against LXRa ( ( p p p p k j k h d j k f h j f g j f g h j (pCMV-H1a b (pCMV-H1-shLXRb) were designed a, the b, the sequences is: CTC GAG GAT

Reagents
In vivo transfection and luciferase assays DNA/PEI complexes and in vivo 3 3 For transfection of both MC4R and TRH reporter plasmids, newborn mice were cotransfected bilaterally in the hypothalamic region of the brain with 4 mL (262 ml) of a 250 ng/ml solution of MC4R-r.luc (0.8 mg/pup) and TRH-f.luc (0.2 mg/pup), complexed with PEI. For shLXR experiments, one-day-old euthyroid pups were transfected with 4 ml of a solution of a transfection mix containing PEI-complexed TRH-f.luc (0.2 mg/pup)/empty CMV-H1 (0.4 mg/pup) or a mix of CMV-H1shLXRa (0.2 mg/pup) and CMV-H1shLXRb (0.2 mg/pup) with or without GW3965 at 10 26 M.
To assess the pharmacological activation of endogenous LXR, GW3965 diluted stock solution was added to the plasmids/PEI complex solution to reach final concentrations (10 27 or 10 26 M). In control animals, diluted DMSO was added to the plasmids/PEI complex solution.
Renilla and Firefly luciferase activities were measured 24 h later on each dissected hypothalamic region.

G3965 treatment
To assess the pharmacological activation of endogenous LXR, GW3965 was either added to the transfection mix (10 27 or 10 26 M), or subcutaneously injected at 12.5 mg/Kg or 25 mg/Kg of body weight 24 hours before the transfection.  [12]. For shLXR coding sequence was purchased from MWG Eurofins. The cont-cloned into CMV-H1 as already described [12]. Each shRNA-according to validated shRNA sequences published in [16] and shLXR ) and against LXR r.luc) was already published [12]. To knock-down endogenous rase-coding sequence [15]. MC4R-bp of the TRH promoter cloned upstream of the Firefly lucife-Factory (manufacturing code: pF312) and contains

In vivo chromatin immunoprecipitation
Pups were treated with T or vehicle and sacrificed 20 h later. 3 Dorsal hypothalamic regions including the PVN were dissected. Samples were fixed in 1% formaldehyde solution and sonicated. Control and T 3 treated samples were used for ChIP with anti-LXR or anti-RXR antibodies or without antibody (Ab, negative cont-control). Precipitated DNA fragments were purified. Primers spanning the most conserved nTRE identified in Trh (site 4) promoter were used in qPCR to measure enrichment of DNA samples. Negative controls comprised primers spanning irrelevant sequences in gene sequence. Detailed ChIP protocol is provided in SI of [12].
Hypothalami of pups were dissected. Total RNA was extracted. Concentration of total RNA was measured, and RNAs were stored in Tris 10 mM/EDTA 0.1 mM (pH 7.4) at 280uC. To quantify mRNAs in hypothalami, 1 mg of total RNA was reverse-transcribed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). Selected LXR target genes were: Stearoyl-CoA desaturase 1 (Scd1), Peroxisome proliferatoractivated receptor a (Ppara), ATP-binding cassette (ABC) transporter (Abcg1), Lipoprotein lipase (Lpl), Sterol regulatory element binding transcription factor 1 (Srebp1), Liver X receptors (Lxra and Lxrb), ATP-binding cassette sub-family A (Abca1), Peroxisome proliferator-activated receptor c (Pparc) and Mlxipl (Chrebp, Carbohydrate response element binding protein). We tested also the Brain-derived neurotrophic factor (Bdnf) that could have a link with LXR signalling, and genes related to inflammation: Tumor necrosis factor-alpha (Tnfa), Interleukin-1 alpha (Il-1a), Interleukin 6 (Il-6), Cyclooxygenase-2 (Cox-2) and v-Rel avian reticuloendotheliosis viral oncogene homolog A (RELA, also known as Nf-kb 3 or p65). Primers and taqman probes for the detection of LXR target genes and control assays (Mm01197142_m1 for Scd1, Mm00440939_m1 for Ppara , . Direct detection of the PCR products was monitored by measuring the increase in fluorescence generated by the TaqMan probes. Samples containing 2 mL of cDNA, 1 mL of specific probe, and 10 mL of 26TaqMan universal PCR Master Mix (Applied Biosystems) were prepared in a final volume of 20 mL. The gene-specific PCR products were measured continuously using ABI PRISM 7300 Sequence Detection System (Applied Biosystems) during 40 cycles. All experiments were run using the same thermal cycling parameters (one cycle at 95uC for 10 min and 40 cycles at 95uC for 15 s and 60uC for 1 min). Nontemplate controls were used to detect non-specific amplification. The threshold cycle (CT) of each target product was determined and kept constant for all data analysis, and DCT between target and endogenous control (Gapdh) was calculated. The difference in DCT values of two groups (DDCT) was used to calculate the fold increase (F = 2 2DDCT ) and to determine the changes in target gene expression between control and treated group.

Statistical Analysis
For in vivo gene transfer, non-parametric permutation test (Cytel Studio software) was used to assess statistical significance. Each experiment was carried out with n$10, repeated at least three times providing the same results. Differences were consid-ered significant at p,0.05 with *, p,0.05; **, p,0.01; ***, p, 0.001. QPCR results: Statistical analysis for qPCR data compared DDCT values using non-parametric ANOVA, followed by a permutation test to compare control and treated groups. Independent experiments (7#n#8) were repeated two times providing similar results and data were pooled.

GW3965 represses the transcription from Trh and Mc4r promoters only in euthyroid newborn mice
We analysed by in vivo gene transfer the effects of a synthetic LXR ligand, GW3965, on Trh and Mc4r transcriptional activity by subcutaneous or intracerebroventricular injection of newborn mice. One-day-old euthyroid or hypothyroid pups were s.c. injected by GW3965 at 12.5 mg/Kg or 25 mg/Kg of body weight and were co-transfected 24 h after in the hypothalamic region with TRH-f.luc (0.2 mg/pup)/MC4R-r.luc (0.8 mg/pup) as described in material and methods section. Hypothalami were recovered 24 h later and Firefly and Renilla luciferase activities were measured. We showed by in vivo gene transfer of TRH-and MC4R-luciferase reporter plasmids in absence or presence of GW3965 that activation of LXR by its synthetic ligand GW3965 induces repression of the transcriptional activity of Trh Thyroid hormone treatment of hypothyroid newborn mice or dams restores the GW3965-dependent repression from Trh promoter, but not that from Mc4r promoter in PVN To determine whether thyroid hormone plays a role in the LXR/TR dialogue and if its presence is able to restore the transcriptional regulation of Trh or Mc4r by LXR in hypothyroid newborn mice hypothalami, one-day-old hypothyroid pups were s.c. injected by T 3 (2.5 mg/g bw) 24 h before they were cotransfected by TRH-f.luc (0.2 mg/pup)/MC4R-r.luc (0.8 mg/pup) mix as in the previous experiment. Hypothalami were dissected 24 h later and Firefly and Renilla Luciferases measured. Subcutaneous triiodothyronine treatment of hypothyroid newborn mice restored in the hypothalamus the GW3965-dependent repression of the Trh promoter (p: 0.0025) (Fig. 2) but not that of Mc4r promoter (data not shown). For the same purpose, we treated dams with T 4 at 12 mg/ml in the drinking water for 24 h before giving birth, and pups (d1) were co-transfected in the hypothalamus by TRH-f.luc (0.2 mg/pup)/MC4R-r.luc (0.8 mg/ pup) simultaneously with GW3965 at 10 26 M. Hypothalami were recovered 24 h later and Firefly and Renilla luciferases activities 0.04401 (without CMV-H1, second column) and 0.02952 (in presence of CMV-H1, fourth column)) in the hypothalamus of euthyroid mice and this before knocking-down LXR. By adding shLXRa-b in the transfection mix, we observed an activation of the transcriptional activity of Trh promoter (Fig. 4A, p:0.00174 and p:1e-005 compared respectively to without (third column) or with (fourth column) GW3965). Interestingly, this activation of TRH-Luc by knocking-down LXRa-b was observable even in the absence of GW3965 (Fig. 4B, third column compared to second p: 0.0073).

LXR and RXR are not detected simultaneously on Trh promoter
In order to uncover the transcriptional mechanism underlying the regulations described above, we analysed the occupancy of Trh promoter by LXR and RXR by using the technique of in vivo ChIP. LXR or RXR binding was analysed on the most conserved nTRE identified in Trh (site 4) regulatory region. ChIP using anti-LXR or anti-RXR antibodies was carried out on hypothalami from hypothyroid newborn mice treated or not with T 3 for 20 h. Results represent the occupancy of LXR and RXR at the TREsite 4 in Trh promoter (Fig. 5). These experiments showed that LXR is recruited to the Trh promoter region in the presence of T 3 (Fig. 5, third column from the right) but not in its absence (Fig. 5, fourth column from the right). In contrast, RXR is recruited to the same region in the absence of T 3 (Fig. 5, second column from the right) as already described [18]. We also note that there is no significant simultaneous recruitment of LXR and RXR to Trh promoter, suggesting that the presence of one NR excludes the other.
Some key LXR target genes involved in metabolism and inflammation are regulated by thyroid status in the hypothalamic PVN of newborn mice LXR activation represses the transcriptional activity of Trh and Mc4r promoters, confirming a crosstalk between TR and LXR and thus, regulates genes undergoing feedback by T 3 in the hypothalamic PVN and involved in metabolism. TH in the periphery regulates several LXR target genes. Thus, we hypothesized that T 3 could also regulate some LXR target genes within the hypothalamus. To test this possibility, eu-vs hypothyroid newborn mice (day 1) were either treated with T 3 (2,5 mg/g bw) or vehicle 6 or 18 hours before dissection. Hypothalami (PVN) were dissected and effects of thyroid status on some LXR target gene mRNA expression were quantified by qPCR (Fig. 6). T 3 treatment of euthyroid pups significantly increased Abcg1 (6 h after T 3 treatment, p,0.001, p: 0.00035), Lpl (6 h after T 3 treatment, p, 0.05, p: 0.0339; 18 h after T 3 treatment, p,0.01, p: 0.0014), Ppara (6 h after T 3 treatment, p,0.01, p: 0.00684) and Scd1 (6 h after T 3 treatment, p,0.05, p: 0.02102) mRNA levels, versus   We looked at other genes involved in cholesterol homeostasis (Srebp, Lxra and Lxrb) or encoding the ATP-binding cassette  transporters (Abca1) as well as genes implicated in lipid metabolism (Pparc and Chrebp) and we noticed no significant effect of T 3 on mRNA expression levels (data not shown).
Regarding expression of LXR target genes involved in inflammation, T 3 treatment of euthyroid pups significantly increased Tnfa (18 h after T 3 treatment, p,0.05, p: 0.01643) and Il1 (6 h after T 3 treatment, p,0.001, p: 0.0049) mRNA levels (Fig. 6). Hypothyroidism and T 3 treatment of hypothyroid newborn mice had no significant effect on the expression of Tnfa and Il1 mRNA levels. The other genes studied (Il6, Nfkb and Cox2) were not significantly affected by either treatment (data not shown).

Discussion
Both TRs and LXRs play important roles in metabolic regulation [19][20][21][22]. Crosstalks between pathways regulated by these two NRs have been reported, especially regarding lipid metabolism-related genes [8] but also in other physiological systems such as the central nervous system [23] as LXRs are expressed in, notably, the brain stem, the hypothalamus and the cortex [24]. 24-(S)-hydroxycholesterol-liganded LXR regulates cholesterol availability in the brain playing an important role in hypothalamic cholesterol homeostasis, which is crucial for brain physiology [25]. In addition, the study of Perez-Tilve et al. [14] shows a role for the melanocortin system in controlling hepatic cholesterol metabolism. Inhibition of the melanocortin system increases circulating HDL cholesterol levels by reducing its uptake by the liver. Thus, there could be a crosstalk between pathways governed by LXR and TR in the hypothalamus, in the context of maintenance of lipid homeostasis.
Thus, our objective was a greater understanding of the physiological links between TR and LXR signalling systems, in the context of central control of energy homeostasis. In particular, our hypothesis was that LXR might modulate TR-dependent Trh and Mc4r transcriptions. This was verified by in vivo gene transfer and shRNA strategy to knock-down LXRs. Our data show that hypothalamic LXR, when activated by its synthetic ligand, has a repressive action on Trh and Mc4r promoters. Interestingly, when both LXRs where knocked-down, the repression on Trh promoter transcriptional activity was lost (with or without the ligand GW3965), resulting in an activation of transcription from the promoter compared to controls (with no GW3965, but with LXRs, fig. 4B). This result indicates that LXR is able to repress Trh promoter activity even when it is not activated by synthetic ligand. Thus, endogenous LXR ligand could activate LXR to repress Trh promoter. When LXRs are knocked-down, this repression is released, leading to a relative activation of the Trh promoter. This result can be linked with previous results showing LXR to repress Dio2 transcription [26]. In both cases, the repressive effect of LXR would lead to a decrease in T 3 production, thus showing a global inhibitory effect of LXR on T 3 driven pathways.
It is interesting to note that the repressive action of LXR on Trh and Mc4r promoters is only observed in euthyroid newborn mice. Thus, LXR-dependent regulation of Trh and Mc4r transcription seems to be influenced by thyroid status in the PVN. Thyroid status is also known to specifically regulating Mc4r expression in the PVN, with hypothyroidism increasing endogenous Mc4r expression [12], which parallels the well-established rise in PVN Trh expression also induced by hypothyroidism [27,28]. We further explored the importance of the thyroid status by studying the concomitant effect of T 3 and GW3965 on TR/LXR dialogue. TH replacement in hypothyroid hypothalamus restored the GW3965-dependent repression of the Trh promoter by LXR but not that of Mc4r. These different responses could be explained by a difference in the mechanisms involved, in particular, regarding TR-induced regulations. Data from many in vitro studies on positively regulated genes suggest a model wherein TRs bind to pTREs with or without its ligand, T 3 [29]. In vivo, the nTREs studied by Decherf et al. in the Mc4r and Trh promoters showed distinct TRb recruitment patterns as a function of T 3 presence [12]. Mc4r TRE1 recruited only low levels of TRb in the absence of T 3 , whereas T 3 induced a large increase of TRb binding. In contrast, TRb was found on Trh TRE site 4 without hormone, as previously shown [30], and T 3 induced the dissociation of TRb from Trh TRE site 4 (see Fig. 2B and Supporting Information, Fig. S4 in [12]). These differences may contribute to the differences observed on LXR mediated regulations in addition to the different sensitivities to T 3 -dependent repression of these two genes. In the present study, the in vivo ChIP experiments showed that LXR is recruited to the Trh promoter region in the presence of T 3 but not in its absence. In contrast, RXR is recruited to the same region in the absence of T 3 (as found by Decherf et al [18]). We also note that there is no significant simultaneous recruitment of LXR and RXR to the site 4 of Trh promoter, suggesting that the presence of a NR excludes the other. However, in periphery, LXRs generally function as permissive heterodimer with RXR. But it is worth to note that this heterodimer is observed on LXRE, which is a DR4 (that means, two half-sites present for binding of NR) whereas we studied LXR binding on the most conserved nTRE identified in Trh, the site 4, a half-site, which could explain the absence of RXR. Furthermore, the Trh site 4 preferentially binds TR/RXR heterodimers [31].  Consequently, we could propose a model where, in hypothyroid mice, TRb would be recruited to the Trh site 4 as a heterodimer with RXR to activate the ligand-independent transcription. After T 3 treatment, TRb could dissociate from the site 4 where LXR would be then recruited and represses Trh transcription. It has been reported [12] for Mc4r, that in a hypothyroid state, TRE1 recruited low levels of TRb but T 3 treatment induced a large increase of TRb binding. The recruitment of LXR could consequently be inhibited and then, prevent the regulation of Mc4r transcription by LXR, observed in a euthyroid state. These data provide a basis for a model for LXR interference with Trh and Mc4r transcription. Thus, the data show that LXR represses the transcription of T 3 regulated genes involved in central control of metabolism in the hypothalamic PVN. To further test the hypothesis that T 3 could also regulate key LXR target genes in the hypothalamus we used qPCR to analyse potential central TR/LXR crosstalk.
In the periphery, LXRs regulate a variety of genes encoding diverse enzymes (for review see, [5,8]) involved in cholesterol metabolism such as ABCG1 and Cyp7A1 [32][33][34] as well as in hepatic lipogenesis (e.g. FAS and SCD1) [35,36] and in lipoproteins metabolism (LPL, CETP, etc) [37]. Many of these LXR target genes are also regulated by TH in the periphery [8,[38][39][40] and then could be regulated by T 3 in the hypothalamus. We selected some of these LXR target genes to investigate potential central T 3 regulation, which could in turn be linked to central control of metabolism. As for in vivo gene transfer experiments, where we only saw effects of LXR ligand in euthyroid animals, again in qPCR analysis of T 3 regulation of LXR target genes expression, we observed more regulations in the euthyroid than in the hypothyroid group, emphasising the importance of the thyroid status in the crosstalk between LXR and TR signalling. Our data show that T 3 significantly increased Scd-1 mRNA levels in PVN from euthyroid and hypothyroid pups. Moreover, Scd-1 mRNA level was lower in hypothyroid than in euthyroid pups. Thus, Scd-1 is positively regulated by TH in the PVN region of newborn mice. SCD-1 is one of the major enzymes involved in fatty acid metabolism and converts saturated fatty acids into monounsaturated fatty acids [41][42][43]. At the central level, a possible link between SCD-1 and T 3 could be related to anti-inflammatory action and control of food intake. Indeed, hypothalamic inflammation is an early signal of the onset of obesity, due to abnormal control of caloric intake (review in [44]). Moreover, Cintra and colleagues [45] show that linolenic and oleic fatty acids (unsaturated fatty acids) inhibit the AMPK/ ACC pathway (as does T 3 ) [46] while increasing SCD1 expression in the hypothalamus, again, as does T 3 in our study. Furthermore, fatty acids injections significantly reduce spontaneous food intake and enhance the anorexigenic effect of leptin [45].
Leptin is also well known to regulate TRH, either directly [47], or indirectly, through MC4R activation [48], and both TH and leptin signalling activation induce anorexigenic pathways. Lpl was also regulated by T 3 in the newborn mice PVN. LPL is a multifunctional enzyme that plays a major role in the metabolism and transport of lipids in peripheral tissues [49,50]. In the brain, LPL is important to energy balance and body weight regulation, influencing the same pathways as SCD, i.e unsaturated fatty acid control of food intake. In particular, neuronal LPL deficiency leads to hypothalamic AgRP up-regulation (review in [51]), thereby activating orexigenic pathways. Hypothetically, LPL would provide important lipid-derived regulatory signals such as polyunsaturated fatty acids (PUFAs), which in turn regulate AgRP expression in the hypothalamus and thus energy balance and body weight. However, AgRP signalling inhibits HPT axis (review in [11]) and T 3 up-regulates Lpl expression. Thus, these results confirm that AgRP and T 3 act in an antagonistic manner, the former stimulating orexigenic pathways, whilst the latter inhibits them.
Thus, it appears that LXR and TR have common target genes and pathways in the hypothalamus. Taken together our results suggest that LXR and TR crosstalk within the hypothalamus could be involved in the central control of food intake, and more generally, in the regulation of energetic homeostasis.
Another common link between these gene networks could be their implication in neurodegenerative disease and inflammation. In fact, LXRs have emerged as important regulators of the innate and adaptive immune system and inflammation [32]. LXR signalling impacts the development of Alzheimer's disease (AD) pathology and LXRs are promising therapeutic targets for AD treatment because of their ability to affect components of the disease such as cholesterol content, Ab clearance, APP processing, ABCA1, etc. [52]. It has been recently demonstrated that TR and LXR competitively up-regulate the human selective AD indicator-1 (Seladin-1) gene promoter at the transcriptional levels and both receptors share a positive TRE/LXRE [23]. Indeed, BDNF, like LXR, is involved in cholesterol metabolism and in neurodegenerative disease [53]. In hypothyroid pups we report that T 3 treatment increases Bdnf mRNA levels. Crupi et al. [54] have also reported that T 3 significantly enhanced the post-traumatic brain injury expression of the neuroprotective neurotrophin BDNF, showing a potential anti-inflammatory effect of T 3 . Further, we observed a significant T 3-dependent increase in Ppara mRNA levels in the hypothalamic region. Thus, Ppara is centrally regulated by T 3 and may be involved in the central regulation of lipid metabolism. Indeed, it is established that, in the periphery, Ppara is a LXR/TR target, and extensive data establish the importance of PPARa in inflammation [55]. Thus, Ppara may be involved at the central level in inflammation. In addition, synthetic LXR agonists have been shown to have anti-inflammatory features [56]. The anti-inflammatory activities of LXR were described in 2003 [57] using a cutaneous inflammatory mouse model in which activation of LXR by GW3965 or 22-hydroxycholesterol inhibited production of TNFa and IL-1a. Interestingly, it has also been demonstrated that a hyperthyroid state in the rat increases circulating levels of TNF-a by actions exerted at the Kupffer cell level, and this is related to the oxidative stress status established in the liver by T 3 -dependent calorigenesis [58]. Our results show that T 3 also increases Tnfa and Il1 in the hypothalamic region. Thus, the crosstalk between LXR and TH is also involved in central regulation of inflammation. Indeed, those two pathways could act synergistically to reduce inflammation, LXR inhibiting Trh transcription and thus T 3 secretion, leading to a subsequent reduction of TNFa and IL1 production, reinforcing the repressive effect of LXR on these two factors [57].
Thus, LXRs could be attractive drug targets for therapeutic intervention in metabolic disorders and inflammatory diseases even at the central level. RXR also plays an important role in the regulation of inflammation [59]. Finally, TRs act also as inhibitors of inflammation [60]. These data together lead us to suggest that it could exist an interaction between TR/RXR and LXR at the was normalized with Gapdh. n = 8 per group, a pool of two independent experiments is shown. Non-parametric ANOVA followed by permutation test with strata was used to assess statistical significance. *, p,0.05, **, p,0.01; ***, p,0.001. doi:10.1371/journal.pone.0106983.g006 central level to regulate inflammation in addition to the metabolism regulation.
In conclusion, this study represents the first in vivo report of a central LXR/TR signalling crosstalk, in a brain region relevant to metabolic homeostasis, the hypothalamus. The dual control of central metabolic and inflammatory pathways could lead to a fine tuning, allowing for synergetic regulations. Moreover, furthering the understanding of the molecular mechanisms of LXR/TR interaction in these regulations could be important for therapeutic intervention.