The retinol esterifying enzyme LRAT supports cell signaling by retinol-binding protein and its receptor STRA6

Abstract

Vitamin A, retinol, circulates in blood bound to retinol-binding protein (RBP). At some tissues, holo-RBP is recognized by a plasma membrane receptor termed STRA6, which serves a dual role: it mediates transport of retinol from RBP into cells, and it functions as a cytokine receptor that, on binding holo-RBP, activates JAK2/STAT5 signaling. As STAT target genes include SOCS3, an inhibitor of insulin receptor, holo-RBP suppresses insulin responses in STRA6-expressing cells. We have shown previously that the two functions of STRA6 are interdependent. These observations suggest factors that regulate STRA6-mediated retinol transport may also control STRA6-mediated cell signaling. One such factor is retinol metabolism, which enables cellular uptake of retinol by maintaining an inward-directed concentration gradient. We show here that lecithin:retinol acyl transferase (LRAT), which catalyzes esterification of retinol to its storage species retinyl esters, is necessary for activation of the STRA6/JAK2/STAT5 cascade by holo-RBP. In accordance, LRAT-null mice are protected from holo-RBP-induced suppression of insulin responses. Hence, STRA6 signaling, which requires STRA6-mediated retinol transport, is supported by LRAT-catalyzed retinol metabolism. The observations demonstrate that STRA6 regulates key cellular processes by coupling circulating holo-RBP levels and intracellular retinol metabolism to cell signaling.—Marwarha, G., Berry, D.C., Croniger, C.M., Noy, N. The retinol esterifying enzyme LRAT supports cell signaling by retinol-binding protein and its receptor STRA6.

Vitamin A, retinol (ROH), regulates multiple biological functions both during embryonic development and in the adult where it is critical for immune function, reproduction, and vision. Many of these activities are exerted by vitamin A metabolites: 11-cis-retinal, which serves as the visual chromophore, and all-trans-retinoic acid (RA), which activates the nuclear receptors RA receptors (RARs) and peroxisome proliferator-activated receptor β/δ (PPARβ/δ), thereby regulating the transcription of numerous target genes (1–3). ROH is stored in various tissues, including adipose tissue, lung, and retinal pigment epithelium (RPE) in the eye, but its major storage pool is in the liver. The vitamin is secreted from these organs bound to serum ROH-binding protein (RBP), which, in blood, is associated with another plasma protein, transthyretin (TTR), to form a ternary RBP-ROH-TTR complex that displays a 1:1:1 molar ratio under normal physiological conditions (4). How and where holo-RBP associates with TTR is incompletely understood, but it has been suggested that the interaction occurs intracellularly at the endoplasmic reticulum and that holo-RBP is secreted from the liver and enters the circulation in complex with TTR (5). Notably, it was reported that RBP expression is up-regulated in adipose tissue of obese mice and humans and that, consequently, plasma RBP level is elevated under these conditions. It was further shown that ectopically elevating RBP levels in lean mice suppressed insulin responses, indicating that RBP plays a direct role in inhibition of insulin signaling (6).

ROH dissociates from RBP prior to entering target cells and, due to its hydrophobic nature, can cross the plasma membrane and enter cells by free diffusion (reviewed in ref. 7). In some tissues (e.g., RPE, adipose tissue, and muscle), uptake of ROH from blood can also be mediated by a plasma membrane protein termed stimulated by RA 6 (STRA6). This protein binds extracellular holo-RBP, facilitates the dissociation of ROH from its carrier, and transports it into cells (8). Uptake of vitamin A by cells is also supported by cellular ROH-binding protein I (CRBP-I), a cytosolic protein that binds ROH with high affinity and delivers it to enzymes that catalyze its metabolism. It was thus shown that CRBP-I can deliver ROH to lecithin:ROH acyltransferase (LRAT), which catalyzes its conversion into the storage species retinyl esters, and to ROH dehydrogenase 10 (RDH 10), which catalyzes the first step in the generation of RA (9–15). In cells that express STRA6, apo-CRBP-I associates with an intracellular region of the receptor and serves as an acceptor for ROH entering the cell through the receptor (16). Apo-CRBP-I is thus obligatory for STRA6-mediated ROH transport. Ligation of CRBP-I disrupts its interactions with STRA6 and releases the protein to shuttle the ligand to enzymes that metabolize it. Enzymes that recognize holo-CRBP can unload ROH, regenerating apo-CRBP-I and allowing it to reassociate with STRA6 perpetuating vitamin A influx (16). Indeed, it was previously shown that expression of LRAT markedly enhances STRA6-dependent vitamin A uptake by cells (16, 17).

STRA6 is unique in that, in addition to serving as a transporter, it functions as a cell surface cytokine receptor. On binding of extracellular holo-RBP, STRA6 recruits Janus kinase 2 (JAK2), which catalyzes phosphorylation of a tyrosine residue in the cytosolic domain of the receptor and leads to recruitment and activation of the transcription factor STAT5. Activation of STRA6 signaling by holo-RBP thus leads to induction of STAT target genes. Importantly, association with TTR prevents binding of holo-RBP to STRA6, and STRA6 mediates vitamin A uptake and cell signaling only when holo-RBP level exceeds that of TTR (18). The observations that STAT target genes whose expression is induced by holo-RBP include the gene that encodes suppressor of cytokine signaling 3 (SOCS3), a potent inhibitor of insulin signaling (19, 20), provide a clear rationale for understanding how elevated levels of RBP in blood of obese animals result in insulin resistance.

It has been shown that the two functions of STRA6, ROH transport and activation of JAK2/STAT5 signaling, are interdependent; i.e., that direct transfer of ROH from extracellular RBP to intracellular CRBP-I is required for triggering STRA6 phosphorylation and that, in turn, phosphorylation of STRA6 is critical for ROH transport by the receptor (16). These observations suggest that factors that control ROH transport by the receptor may also control STRA6 signaling. One such factor is the rate of ROH metabolism, which is important because ROH enters cells by a passive process, following a concentration gradient maintained by metabolic conversions. STRA6-mediated ROH transport may be especially sensitive to changes in metabolic rates because of the obligatory need of this process for regeneration of apo-CRBP-I. Here, to examine the connection between STRA6 signaling and intracellular ROH metabolism, we focused on LRAT, a 25-kDa endoplasmic reticulum membrane protein that catalyzes the formation of retinyl esters (reviewed in ref. 21). Using cultured cell models and LRAT-null mice, we show that LRAT is required for activation of the STRA6/JAK2/STAT5 cascade as well as for suppression of insulin signaling by holo-RBP both in cultured cells and in vivo. The observations further indicate that ablation of LRAT partially protects mice from suppression of insulin signaling brought about by diet-induced obesity. The data thus indicate that STRA6 couples “sensing” of blood level of holo-RBP and intracellular ROH metabolism to cell signaling, thereby modulating important cellular processes, including regulation of insulin responses.

MATERIALS AND METHODS

Vectors

Human STRA6 N-terminally tagged with hexahistidine tagged protein was cloned into p-Receiver M01 vector (Genecopoeia, Rockville, MD, USA). Mouse Lrat cDNA was recloned into pcDNA 3.1/HisA mammalian expression vector. LRAT shRNA expression vector was procured from Open Biosystems (Waltham, MA USA).

Antibodies

Antibodies toward JAK2, p-Tyr^1007/1008 JAK2, STAT5, p-Tyr⁶⁹⁴ STAT5, IR, p-Tyr^1150/1151 IR, Akt, and p-Ser-473 Akt were obtained from Cell Signaling (Boston, MA, USA). Antibodies against LRAT and RBP were from Abcam (Cambridge, MA. USA) and Dako (Glostrup, Denmark), respectively. STRA6 antibodies were a gift from Norbert Ghyselinck [Institute of Genetics and Cellular and Molecular Biology (IGBMC), Strasbourg, France].

Cell cultures

NIH-3T3 fibroblast and HepG2 cells [American Type Culture Collection (ATCC), Manassas, VA, USA] were cultured in DMEM containing 5% calf serum and 10% fetal bovine serum respectively. Cells were transfected in suspension using PolyFect (Qiagen Inc., Valencia, CA, USA). Briefly, 5 μg of the plasmid was added to a tube containing 150 μl of serum-free DMEM and mixed. Transfection reagent (15 μl) was added, and the solution was incubated for 10 min at room temperature. Suspended cells were added to the mixture, and the volume was made up to 4 ml with DMEM. Cells were plated in a 100-mm dish to achieve 50% confluence. After 12 h, medium was aspirated and replaced with 10 ml DMEM for an additional 24 h.

Recombinant RBP

Bacterial expression vector for histidine-tagged RBP was expressed in Escherichia coli BL21 (DE3) and purified essentially as described previously (22). Briefly, bacteria were grown at 37°C to an optical density at 600 nm of 0.7 to 0.8, and protein expression was induced with 1 mM isopropyl-d-thiogalactopyranoside (IPTG; 4 h, 37°C). Bacteria were harvested by centrifuging the cultures at 7000 g for 15 min and lysed with B-PER lysis reagent (Pierce, Rockford, IL, USA) containing 6 M guanidine-HCL and 10 mM dithiothreitol (DTT) supplemented with a cocktail of protease/phosphatase inhibitors (1:100). Lysates were centrifuged at 14,000 g for 35 min, and the supernatant comprising the crude protein extract was affinity purified using Ni²⁺Sepharose beads (GE Healthcare, Waukesha, WI, USA), followed by elution with 250 mM imidazole. RBP was refolded with 10-fold molar excess ROH in a redox buffer (0.3 mM cystine, 3 mM cysteine, 1 mM EDTA, and 25 mM Tris-Hcl, pH 9.0) for 5 h to obtain holo-RBP. Protein was dialyzed against a buffer containing 300 mM NaCl, 100 mM Tris (pH 7.4), and 5% glycerol and was concentrated. The method typically generates an RBP-ROH complex at a 0.8–1 ROH/RBP molar ratio. Endotoxin levels in protein preparations were measured using Pierce LAL Chromogenic Endotoxin Quantitation Kit (cat. no. 88282; Pierce), following the manufacturer's protocol.

Immunblot analyses

White adipose tissue (WAT), skeletal muscle, or liver was homogenized in RIPA tissue lysis buffer supplemented with protease and phosphatase inhibitors (1:100). Protein concentrations were determined by the Bradford protein assay method. Proteins were resolved by SDS-PAGE gels, transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA), and incubated with appropriate antibodies. Blots were developed with enhanced chemiluminescence (Thermo Scientific, Rockford, IL, USA) and analyzed by Image J software.

Real-time quantitative PCR (qPCR)

Total RNA was isolated and extracted from tissues and cells using the RNeasy tissue mini kit (Qiagen Inc., Valencia, CA, USA) and TRIzol, respectively. cDNA was obtained using High Capacity RNA-to-cDNA synthesis kit (ABI, Foster City, CA, USA). qPCR was performed using TacMan chemistry using Assays-on-Demand probes (ABI) for Socs3 (Mm01249143_g1). Expression of specific transcripts amplified was normalized to the expression of 18s rRNA (4352930E).

Mouse studies

Lrat^−/− mice on a C57/Bl6 background (23) were housed in the Animal Resource Center (ARC) facility at Case Western Reserve University (CWRU), and all experiments were performed in accordance to the protocol approved by the Institutional Animal Care and Use Committee at CWRU. Male Lrat^−/− mice (12 wk old) and wild-type (WT) littermates were used. For feeding studies, mice were fed a high-fat/high-sucrose diet (HFHSD; D12331, Research Diets, New Brunswick, NJ, USA) or normal chow (LabDiet 5010 chow, PMI Nutrition International, St. Louis, MO, USA).

Glucose tolerance tests (GTTs)

Lrat^−/− mice and WT littermates (n=10/group) were fed HFHSD for 16 wk, deprived of food overnight, and then injected intraperitoneally with glucose (2 g/Kg). Blood was sampled from the tail vein, and glucose levels were measured at 0, 15, 30, 60, and 120 min using an UltraTouch meter. GTT was performed at the Mouse Metabolic Phenotyping Center (MMPC) of CWRU.

Food intake

To measure 24-h food intake, mice (n=10)/group were housed solitarily with ad libitum access to food and water. At the time of measurement, mice were placed in a clean cage and the food was weighed. The food and mice were weighed after 24 h, and the average food intake per mouse per cage was calculated.

Statistical analyses

The significance of differences among the samples was assessed by 1-way analysis of variance (ANOVA) followed by Tukey's post hoc test unless otherwise specified. Statistical analysis was performed with GraphPad Prism 4.01 software (GraphPad, San Diego, CA, USA).

RESULTS

LRAT is required for activation of JAK/STAT signaling and inhibition of insulin responses by holo-RBP and STRA6

3T3 fibroblasts were used to begin to examine whether intracellular ROH metabolism catalyzed by LRAT is linked to holo-RBP-induced, STRA6-mediated activation of the JAK2/STAT5 pathway. 3T3 cells express CRBP-I, which is required for STRA6 signaling, but they do not express either STRA6 or LRAT (16). Utilization of these cells thus allows for examining the individual roles of these proteins in the holo-RBP-induced activation of JAK2 and STAT5. Human RBP lacking its N-terminal secretion signal, which corresponds to circulating RBP (24), was expressed in E. coli and purified in the presence of ROH to obtain holo-RBP (see Materials and Methods). The protocol typically yields holo-RBP at a ROH/RBP ratio of 0.8-1 (Fig. 1A). The endotoxin content of protein preparations was found to be ∼0.001 EU/μl, lower than that of sterile 18 MΩ MilliQ water.

Figure 1.

LRAT and STRA6 are required for activation of JAK2 and STAT5 and suppression of insulin signaling by holo-RBP. A) Absorption spectrum of bacterially expressed holo-RBP. B) NIH-3T3 cells were transfected with vectors encoding STRA6 or LRAT or both. At 36 h post-transfection, cells were treated with holo-RBP (1 μM) or buffer for 15 min. Immunoblots of JAK2, STAT5, phosphorylated JAK2 (pJAK2), and phosphorylated STAT5 (pSTAT5) are shown. Also shown are immunoblots verifying overexpression of STRA6 and LRAT. C) HepG2 cells transfected with an empty vector (e.v.) or a vector encoding LRATshRNA. At 36 h post-transfection, cells were treated with holo-RBP (1 μM) or buffer for 15 min. Immunoblots of JAK2, STAT5, pJAK2, and pSTAT5 are shown. Also shown is an immunoblot verifying decreased expression of LRAT. D) HepG2 cells transfected with e.v. or vector encoding LRATshRNA were treated with holo-RBP (1 μM) or buffer for 8 h. Levels of Socs3 mRNA were measured by qPCR. Data are means ± sd. ***P < 0.001 vs. e.v. control; ^†††P < 0.001 vs. e.v.-transfected HepG2 cells treated with holo-RBP. E) HepG2 cells transfected with e.v. or vector encoding LRATshRNA were treated with holo-RBP (1 μM) or buffer for 8 h. Immunoblots of IR, pIR, Akt, and pAkt are shown. F) HepG2 cells transfected with e.v. or vector encoding LRATshRNA were treated with holo-RBP (1 μM) for 8 h. Cells were then treated with insulin (10 μg/ml) for 15 min. Immunoblots of IR, pIR, Akt, and pAkt are shown.

3T3 cells were transfected with vectors encoding STRA6 or LRAT or both and treated with holo-RBP (1 μM, 15 min.). The effect of the treatment on phosphorylation levels of JAK2 and STAT5 was examined. The data (Fig. 1B) showed that treatment with holo-RBP enhanced the phosphorylation of both JAK2 and STAT5 in cells transfected with both STRA6 and LRAT, but not in cells transfected with either STRA6 or LRAT alone. The observations thus indicate that LRAT is necessary for STRA6-dependent signaling. HepG2 hepatocarcinoma cells, which endogenously express CRBP-I as well as both STRA6 and LRAT, were used to further examine the role of LRAT in holo-RBP-induced activation of JAK2 and STAT5. Treatment of these cells with holo-RBP effectively enhanced the phosphorylation levels of both JAK2 and STAT5 and the activity was abolished on reducing the expression of LRAT using corresponding shRNA (Fig. 1C). In accordance with activation of STAT5, holo-RBP markedly increased the expression level of the STAT target gene SOCS3 but had no effect on expression of this gene in HepG2 cells with reduced expression of LRAT (Fig. 1D). It was previously shown that SOCS3, which inhibits insulin receptor (IR) signaling (19, 20), mediates the ability of holo-RBP to suppress insulin responses (25). In accordance with up-regulation of SOCS3, treatment of HepG2 cells with holo-RBP markedly decreased the phosphorylation status of insulin receptor (IR) and its downstream effector Akt (Fig. 1E). Strikingly, reducing the expression level of LRAT diminished the ability of holo-RBP to decrease the phosphorylation of IR and Akt (Fig. 1E). The effect of the decreased expression of LRAT on inhibition of insulin responses by holo-RBP was further examined in HepG2 cells treated with insulin. Pretreatment with holo-RBP attenuated insulin-induced phosphorylation of IR and Akt, and the effect was diminished in cells with reduced LRAT expression (Fig. 1F). Taken together, the observations indicate that LRAT is required for enabling holo-RBP to induce STRA6-mediated activation of JAK2/STAT5 pathway as well as subsequent inhibition of insulin signaling.

LRAT is required for holo-RBP-induced activation of JAK2 and STAT5 and inhibition of insulin responses in vivo

Mice lacking LRAT were used to examine the involvement of LRAT in activation of JAK2/STAT5 signaling by holo-RBP in vivo. Lrat^−/− mice and WT littermates were treated with holo-RBP by intraperitoneal injections. Buffer or holo-RBP (0.1 mg) was injected 3 times at 2-h intervals, and mice were euthanized 1 h after the last injection. Immunoblot analyses showed that RBP in serum of treated mice displayed two bands, corresponding to endogenous and administered recombinant histidine-tagged RBP, respectively (Fig. 2A). In accordance with previous reports (23, 26), the basal level of RBP in plasma of Lrat^−/− mice was lower than that observed in WT mice (Fig. 2A, B), likely reflecting the deficiency in vitamin A liver stores in these mice. Injection of holo-RBP into WT mice increased the total serum level of RBP by ∼2 fold, approximating the increase in RBP levels observed in blood of obese mice (ref. 6 and Fig. 4B). Ectopic administration of RBP resulted in a similar total RBP levels in WT and Lrat^−/− mice (Fig. 2B). Treatment of WT mice with holo-RBP increased the phosphorylation levels of both JAK2 (Fig. 2C, D) and STAT5 (Fig. 2E, F) in WAT and skeletal muscle. In contrast, the treatment did not affect JAK2 and STAT5 phosphorylation in the liver, an organ that, unlike WAT and muscle, does not express STRA6. The basal levels of phosphorylation of JAK2 and STAT5 in WAT and muscle were lower in Lrat^−/− mice, and the effect of RBP was markedly reduced as compared to effects observed in WT animals (Fig. 2C–F). In accordance, expression of the STAT target gene Socs3 was efficiently induced on treatment with holo-RBP in WAT and skeletal muscle of WT mice but not in mice lacking LRAT (Fig. 2G).

Figure 2.

Ablation of Lrat inhibits holo-RBP-induced JAK2/STAT5 signaling in vivo. A) Immunoblots of RBP in 3 individual WT and Lrat^−/− mice injected with holo-RBP or buffer (see text for details). B) Quantitation of immunoblots shown in A. C) Immunoblots of JAK2 and pJAK2 in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice injected with holo-RBP or buffer. D) Quantitation of immunoblots shown in C. E) Immunoblots of STAT5 and pSTAT5 in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice injected with holo-RBP or buffer. F) Quantitation of immunoblots shown in E. G) Levels of Socs3 mRNA in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice injected with holo-RBP or buffer, measured by qPCR. Data are means ± sd. *P < 0.05, **P < 0.01, ***P < 0.001 vs. buffer-treated WT mice; ^†††P < 0.001 vs. holo-RBP treated WT mice.

Figure 4.

Effects of HFHSD on body weight and plasma levels of RBP and TTR in WT and Lrat^−/− mice. A) Percentage body weight gain in Lrat^−/− mice and WT littermates fed HFHSD for 16 wk. B) Immunoblots of RBP and TTR in plasma of mice fed an HFHSD or normal chow for 16 wk. C, D) Quantitation of levels of RBP (C) and TTR (D) in plasma of mice fed a HFHSD or normal chow diet for 16 wk. E) Quantitation of RBP/TTR ratio in plasma of mice fed an HFHSD or normal chow diet for 16 wk. Data are means ± sd. **P < 0.01, ***P < 0.001 vs. normal chow-fed WT mice; ^††P < 0.01 vs. HFHSD-fed WT mice; ^‡‡‡P < 0.001 vs. normal chow-fed Lrat^−/− mice.

It was previously shown that up-regulation of SOCS3, a potent inhibitor of the insulin receptor (IR) (19), underlies the ability of holo-RBP to inhibit insulin signaling (25). Hence, the observations that holo-RBP does not induce the expression of SOCS3 in the absence of LRAT suggest that the enzyme is required for enabling holo-RBP to inhibit insulin signaling. In agreement, treatment with holo-RBP markedly reduced the phosphorylation level of both IR and its downstream effector Akt in WAT and skeletal muscle of WT mice but had no effect on the phosphorylation of either IR or Akt in Lrat^−/− mice (Fig. 3). Attesting to the involvement of STRA6 in mediating suppression of insulin signaling by holo-RBP, the treatment did not affect the phosphorylation level of either IR or Akt in the liver. Taken together, the data demonstrate that LRAT is an essential component of the STRA6/JAK2/STAT5 signaling cascade and is necessary for suppression of insulin signaling by holo-RBP in vivo.

Figure 3.

Ablation of Lrat inhibits holo-RBP-induced suppression of insulin signaling in vivo. A) Immunoblots of IR and pIR in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice injected with holo-RBP or buffer. B) Quantitation of immunoblots shown in A. C) Immunoblots of Akt and pAkt in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice injected with holo-RBP or buffer. D) Quantitation of immunoblots shown in C. Data are means ± sd. *P < 0.05, ***P < 0.001 vs. buffer-treated WT mice; ^††P < 0.01, ^†††P < 0.001 vs. holo-RBP treated WT mice; ^‡P < 0.05, ^‡‡‡P < 0.01 vs. buffer-treated Lrat^−/− mice.

Ablation of LRAT partially protects mice from diet-induced insulin resistance

Plasma RBP levels are elevated in obese mice, leading to insulin resistance. While holo-RBP/STRA6 signaling is not the sole pathway that suppresses insulin responses in obese animals, it is an important contributor to the effect (27). Hence, the observations that LRAT is required for activating the pathway suggest that Lrat^−/− mice may be partially protected from diet-induced insulin resistance. To examine this possibility, Lrat^−/− mice and WT littermates were fed an HFHSD for 16 wk, a dietary regime that results in obesity and insulin resistance (28). Lrat^−/− mice were leaner than WT counterparts prior to initiation of HFHSD feeding, but they gained weight at a similar rate (Fig. 4B). Food intake was not significantly different between the groups (data not shown). Basal plasma RBP level was lower in lean Lrat^−/− vs. WT mice, but was markedly increased upon HFHSD feeding of both groups (Fig. 4B, C). Plasma RBP level in obese Lrat^−/− mice was ∼40% lower than that of obese WT littermates (Fig. 4C). Like RBP, the plasma level of its binding partner in blood, TTR, was lower in lean Lrat^−/− vs. WT mice (Fig. 4D). However, unlike RBP, HFHSD intake had no effect on plasma level of TTR either in WT or in Lrat^−/− mice (Fig. 4D). Consequently, the RBP/TTR ratio in plasma of both obese Lrat^−/− and WT mice was elevated to a similar degree (Fig. 4E). As RBP and TTR circulate in blood of lean mice at an approximately equimolar concentration (4), these observations suggest that obesity results in a similar elevation in free holo-RBP in WT and Lrat^−/− mice.

The phosphorylation levels of JAK2 and STAT5 and, accordingly, the expression level of SOCS3 were higher in WAT, skeletal muscle as well as in liver of WT obese mice as compared to lean counterparts (Fig. 5). The activation of the JAK2/STAT5 cascade in the liver clearly reflects the activity of factors other than STRA6, which is not expressed in this organ. For example, the cascade may be activated by leptin, another cytokine whose plasma level is elevated in obese animals and which activates JAK/STAT signaling in tissues that express the leptin receptor (LR). Notably however, obesity-induced activation of JAK2 and STAT5 as well as the increase in SOCS3 expression were significantly smaller in WAT and muscle of Lrat^−/− vs. WT animals (Fig. 5). In accordance with up-regulation of SOCS3, the phosphorylation level of IR and its downstream effector Akt were markedly reduced in WAT, muscle, and liver of obese WT mice (Fig. 6A–D). The effect was significantly smaller in WAT and muscle, but not in liver, of obese Lrat^−/− vs. WT mice (Fig. 6A–D). These observations indicate that the diminished signaling activities of holo-RBP and STRA6 in Lrat^−/− mice partially protects these animals from obesity-induced activation of JAK2 and STAT5 and suppression of insulin responses. Indeed, while the effect was modest, GTTs showed that obese Lrat^−/− mice are more glucose tolerant than obese WT littermates (Fig. 6E).

Figure 5.

Ablation of Lrat partially protects mice from obesity-induced JAK2/STAT5 activation. A) Immunoblots of JAK2 and pJAK2 in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice fed normal chow or HFHSD. B) Quantitation of immunoblots shown in A. C) Immunoblots of STAT5 and pSTAT5 in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice fed normal chow or HFHSD. D) Quantitation of immunoblots shown in C. E) Levels of Socs3 mRNA in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice fed normal chow or HFHSD. Data are means ± sd. *P < 0.05, ***P < 0.001 vs. normal chow-fed WT mice; ^††P < 0.01, ^†††P < 0.001 vs. HFHSD-fed WT mice; ^‡‡P < 0.01, ^‡‡‡P < 0.001 vs. normal chow-fed Lrat^−/− mice.

Figure 6.

Ablation of Lrat partially protects mice from obesity-induced suppression of insulin signaling. A) Immunoblots of IR and pIR in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice fed normal chow or HFHSD. B) Quantitation of immunoblots shown in A. C) Immunoblots of Akt and pAkt in WAT, skeletal muscle, and liver of 3 individual WT and Lrat^−/− mice fed normal chow or HFHSD. D) Quantitation of immunoblots shown in C. Data are means ± sd. *P < 0.05, **P < 0.01, ***P < 0.001 vs. normal chow-fed WT mice; ^†P < 0.05, ^††P < 0.01 vs. HFHSD-fed WT mice; ^‡‡P < 0.01, ^‡‡‡P < 0.001 vs. normal chow-fed Lrat^−/− mice. E) GTT carried out in WT and Lrat^−/− fed an HFHSD for 16 wk. Data are means ± sd. **P < 0.01 vs. Lrat^−/− mice; n = 10/group.

DISCUSSION

STRA6 mediates cellular uptake of ROH from circulating holo-RBP (8) and also functions as a cell-surface signaling receptor, which is phosphorylated on binding of holo-RBP and, in turn, activates a JAK2/STAT5 cascade to up-regulate the expression of STAT target genes. STRA6 thus regulates various cellular responses, including inhibition of insulin signaling (16, 18, 25, 27, 29). We previously demonstrated that the two functions of STRA6 are closely coupled and critically interdependent (16). These observations suggest that factors that control STRA6-mediated ROH transport may also control STRA6-mediated signaling. As LRAT catalyzes ROH metabolism, and thus facilitates its rate of uptake, we examined whether it is also involved in STRA6 signaling and ensuing modulation of cellular responses.

The data demonstrate that LRAT plays a central role in enabling holo-RBP to activate the STRA6-induced phosphorylation cascade in cultured cells (Fig. 1) as well in vivo. Hence, while treatment with holo-RBP triggered phosphorylation of JAK2 and STAT5 and up-regulated the STAT target gene SOCS3 in WAT and muscle of WT mice, the effect was markedly diminished in Lrat^−/− counterparts (Fig. 2). The observations that the activity was not completely abolished indicate that, while LRAT is critical for enabling efficient STRA6 activation, other enzymes may sustain STRA6 signaling although to a much lesser extent. One enzyme that may function in this capacity is RDH-10 which, similarly to LRAT, can metabolize ROH delivered to it by CRBP-I (12, 30). Another enzyme that may compensate for the loss of LRAT in Lrat^−/− mice is DGAT1, an acyl CoA:ROH acyltransferase (ARAT), which catalyzes retinyl ester formation in tissues such as intestine and skin (31–33), and is believed to catalyze ROH esterification in adipose tissue of LRAT^−/− mice (34). Whether RDH10 or DGAT1 can sustain the signaling activity of STRA6 remains to be clarified. The observations that induction of SOCS3 on STRA6 activation is compromised in the absence of LRAT suggest that the enzyme is necessary for suppression of insulin responses by holo-RBP. Indeed, the ability of holo-RBP to decrease the phosphorylation levels of IR and its downstream effector Akt was markedly diminished in Lrat^−/− mice (Fig. 3).

The elevated blood level of RBP in obese mice results in insulin resistance (6). The observations that RBP suppresses insulin signaling only when its blood levels are high may be understood in view of the observations that heterodimerization of RBP with its blood partner TTR inhibits its interactions with STRA6 (18). Consequently, in lean mice, where holo-RBP and TTR circulate at near equimolar levels (4), TTR prevents excessive STRA6 signaling. In contrast, as HFHSD feeding increases plasma level of RBP but does not alter the level of TTR (Fig. 4 and ref. 35), the protection by TTR is lost, enabling RBP to activate STRA6 and suppress insulin responses. While the levels of both RBP and TTR were found to be lower in Lrat^−/− mice, the RBP/TTR ratio in blood of these animals was similar to that observed in WT mice fed either regular chow or an HFHSD (Fig. 4E). Nevertheless, the phosphorylation levels of JAK2, STAT5, IR, and Akt were significantly lower in STRA6-expressing tissues of Lrat^−/− vs. WT mice (Fig. 5, Fig. 6A–D). The data thus indicate that ablation of LRAT protects mice from obesity-induced suppression of insulin signaling. Indeed, GTT showed that Lrat^−/− mice are more insulin-tolerant than their WT littermates (Fig. 6E). Notably, ablation of LRAT only partially rescued the mice from HFHSD-induced suppression of insulin responses. These observations likely reflect that factors other than holo-RBP are involved in obesity-induced inhibition of insulin responses. For example, the adipokine leptin, whose plasma level is elevated in obese animals, functions like RBP to activate a JAK/STAT cascade and up-regulate SOCS3 thereby leading to insulin resistance (36). Nevertheless, the data surprisingly show that, by supporting STRA6 signaling, vitamin A metabolism by LRAT significantly contributes to the link between obesity and insulin resistance.