Transcytosis of albumin in astrocytes activates the sterol regulatory element-binding protein-1, which promotes the synthesis of the neurotrophic factor oleic acid.

We have recently reported that albumin, a serum protein present in the developing brain, stimulates the synthesis of oleic acid by astrocytes, which promotes neuronal differentiation. In this work, we gain insight into the mechanism by which albumin induces the synthesis of this neurotrophic factor. Our results show that astrocytes internalize albumin in vesicle-like structures by receptor-mediated endocytosis. Albumin uptake was followed by transcytosis, including passage through the endoplasmic reticulum, which was required to induce the synthesis of oleic acid. Oleic acid synthesis is feedback-regulated by the sterol regulatory element-binding protein-1, which induces the transcription of stearoyl-CoA 9-desaturase, the key rate-limiting enzyme for oleic acid synthesis. In our research, the presence of albumin activated the sterol regulatory element-binding protein-1 and increased stearoyl-CoA 9-desaturase mRNA. Moreover, when the activity of sterol regulatory element-binding protein-1 was inhibited by overexpression of a truncated form of this protein, albumin did not affect stearoyl-CoA 9-desaturase mRNA, indicating that the effect of albumin is mediated by this transcription factor. The effect of albumin was abolished when traffic to the endoplasmic reticulum was prevented or when albumin was accompanied with oleic acid. In conclusion, our results suggest that the transcytosis of albumin includes passage through the endoplasmic reticulum, where oleic acid is sequestrated, initiating the signal cascade leading to an increase in its own synthesis.

Astrocytes, the main glial cell population in the central nervous system, play a major role in supporting the development of neurons. In fact, astrocytes synthesize and release extracellular matrix proteins and adhesion molecules, which participate not only in the migration of neurons but also in the formation of neuronal aggregates. In addition, astrocytes produce a broad spectrum of growth factors and cytokines, which can regulate the morphology, proliferation, differentiation, and survival of neurons (for a review, see Ref. 1). We (2) have recently shown that astrocytes synthesize oleic acid from the main metabolic substrates, and oleic acid is released to the extracellular medium. Oleic acid is then incorporated in neurons, specifically into growth cones, and its presence promotes neuronal differentiation. Thus, the presence of oleic acid induces axonal growth, neuronal clustering, and expression of the axonal growth-associated protein-43 (GAP-43) by a protein kinase Cdependent mechanism. All these observations indicate that oleic acid behaves as a neurotrophic factor. The signal that triggers oleic acid synthesis in astrocytes is the serum protein albumin (2). It is well documented that albumin reaches high concentrations in the brain and the cerebrospinal fluid of newborn mammals, unlike the situation in adults (3)(4)(5). The presence of albumin in the brain may be due to the existence of a mechanism through which albumin is transferred from the blood to the brain and cerebrospinal fluid that is active only in the immature brain (6,7). In addition, the in situ synthesis of albumin can also contribute to the increased albumin levels found in the newborn brain (8).
It is also well known that oleic acid synthesis is regulated by the sterol regulatory element-binding protein-1 (SREBP-1), 1 a basic-helix-loop-helix-leucine zipper transcription factor that induces enzymes involved in oleic acid synthesis such as acetyl-CoA carboxylase, ATP citrate (pro-S)-lyase, fatty acid synthase, NADPH-producing enzymes, and stearoyl-CoA 9-desaturase (SCD) (9 -12). Like other members of the SREBP family, SREBP-1 is located in the endoplasmic reticulum (ER) and is activated by proteolysis, releasing an active mature form that binds to the promoters of genes containing sterol-responsive elements (SRE; for a review, see Ref. 13). A critical committed step in the synthesis of oleic acid is the introduction of the cis-double bond in the ⌬ 9 position, catalyzed by SCD, an enzyme also located in the ER that is transcriptionally activated by SREBP-1 because its gene contains the SRE (14). Oleic acid synthesis is feedback-regulated because oleic acid suppresses SREBP-1 activation (15)(16)(17) and SCD induction (18), both of which accompany the decrease in oleic acid synthesis (19).
Because the presence of albumin in the brain is developmentally regulated, it has been suggested that this protein could play an important role in neural cell differentiation (4). In agreement with this, albumin induces the synthesis and re-lease of oleic acid by astrocytes, which behaves as a neurotrophic factor (2). However, the mechanism by which albumin induces oleic acid synthesis and release by astrocytes is far from clear. Therefore, in this work we investigated the mechanism by which albumin promotes oleic acid synthesis in astrocytes.

EXPERIMENTAL PROCEDURES
Animals-Albino Wistar rats, fed ad libitum on a stock laboratory diet (49.8% carbohydrates, 23.5% protein, 3.7% fat, 5.5% (w/v) minerals, and added vitamins and amino acids), were used for the experiments. Rats were maintained on a 12-h light-dark cycle. Postnatal day 1 newborn rats were used to prepare astrocytes.
Astrocyte Culture-Astrocyte cultures were prepared and cultured in DMEM ϩ 10% fetal calf serum as reported previously (20). Briefly, animals were decapitated, and their brains were immediately excised. After removing the meninges and blood vessels, cells were dissociated from the forebrains. Cells were plated onto poly-L-lysine (10 g/ml)coated Petri dishes at a density of 1.0 ϫ 10 5 cells/cm 2 .
Albumin Loading and Unloading from Astrocytes-14 DIV astrocytes were loaded at 37°C in Elliot buffer (21) (11 mM sodium phosphate, 122 mM NaCl, 4.8 mM KCl, 0.4 mM KH 2 PO 4 , 1.2 mM MgSO 4 , and 1.3 mM CaCl 2 , pH 7.4) containing 5 mM glucose and 2% (w/v) bovine serum albumin conjugated to fluorescein isothiocyanate (FITC-BSA; Sigma) as described (22). After 10 min, cells were washed three times at 4°C with Elliot buffer and washed an additional three times with Elliot buffer plus 2% (w/v) BSA. For the unloading experiments, loaded astrocytes were kept in Elliot buffer plus 2% (w/v) BSA at 37°C or 4°C. At different times, the cells were visualized by confocal microscopy (excitation at 488 nm; Zeiss LSM 510), and the medium was recovered to measure fluorescence ( ex at 495 nm; em at 525 nm). To measure the FITC-BSA remaining inside the cells, fluorescence photomicrographs were captured with a digital video camera (Leica DC100), and the fluorescence was quantified using image analyzer software (NIH Image, kindly supplied by Wayne Rasband, National Institutes of Health, Bethesda, MD). For the experiments with the inhibitors, 14 DIV astrocytes were preincubated at 37°C in DMEM containing 1 M phenylarsine oxide (PAO), 30 M monensin, 50 M nocodazole, 100 M colchicine, or 40 M brefeldin A (BFA) for 20 min. For treatment with lectins or protamine, 14 DIV astrocytes were preincubated at 37°C in PBS containing 0.01% protamine, 50 g/ml Limax flavus agglutinin, 50 g/ml Glycine max agglutinin, or 50 g/ml G. max agglutinin plus 750 mM N-acetylgalactosamine (GalNAc) for 5 min. For treatment with proteases, 14 DIV astrocytes were preincubated at 37°C in PBS containing 2 mM EDTA and 0.25% trypsin or 0.1% Pronase from Streptomyces griseus type XIV (Sigma) for 5 min. Controls were carried out by preincubating cells in PBS with 2 mM EDTA. Because protease digestion disrupts the cell monolayer, astrocytes were collected, washed, and allowed to attach to the Petri dish for 1 h before loading with FITC-BSA. Apart from protamine, trypsin, or Pronase, the inhibitors were maintained at the same concentration for the rest of the experiments.
Cell Incubation-Cell incubations were carried out as described previously (20). 14 DIV astrocytes grown on Petri dishes (25 cm 2 ) were incubated at 37°C for 1 h with 1.5 ml of Elliot buffer containing cold and radiolabeled substrates in the absence or presence of BSA as indicated. For the experiments with the inhibitors, 14 DIV astrocytes were preincubated at 37°C with 1 M PAO, 30 M monensin, 50 M nocodazole, or 40 M BFA in DMEM for 20 min. The inhibitors were maintained for the rest of the time of the experiments. After the incubation, the medium was recovered for analysis.
Analysis of Oleic Acid by HPLC-Oleic acid synthesized and released to the extracellular medium was determined as described previously (2). Lipids were extracted from the medium recovered after the incubation with a mixture of chloroform/methanol (2:1, v/v) as described by Folch et al. (23). Oleic acid was analyzed as 4-bromophenacyl ester derivatives by a gradient-elution HPLC method as described (24). 0.5-ml fractions of the eluate were collected, and the radioactivity was measured by liquid scintillation counting.
Electrophoretic Analysis of the Astrocyte Incubation Medium-The medium recovered after the incubation was subjected to native horizontal electrophoresis in 10% polyacrylamide gel, pH 8.9, following the manufacturer's instructions (Clean Gel, Amersham Biosciences, Inc.). When the migration had finished, the gel was cut into three portions. One portion was incubated with 0.1% (w/v) Coomassie Blue. A second gel portion was incubated with 3% (w/v) cupric acetate in 8% (v/v) orthophosphoric acid and dried at 110°C. The third gel portion was exposed for 30 days to an autoradiographic film at Ϫ70°C.
Electrophoretic Analysis of FITC-BSA and Oleic Acid Released from Astrocytes-Astrocytes were incubated at 37°C in Elliot buffer containing 5 mM glucose, 10 Ci/ml [1(2)-14 C]acetate, and 2% (w/v) FITC-BSA. After 1 h, the incubation medium was removed, and the cells were thoroughly washed with PBS, PBS ϩ 2% BSA, and PBS at 4°C. Astrocytes were then incubated at 37°C in Elliot buffer containing 5 mM glucose. After 1 h, the medium was removed and subjected to native electrophoresis in 10% polyacrylamide gel, pH 8.9. After the electrophoresis, the gels were first examined with an ultraviolet transilluminator to visualize the FITC-BSA released and were then autoradiographed to visualize radioactivity. HPLC analyses of the medium were carried out in parallel to identify the oleic acid synthesized and released.
Cell Transfection-A construction designated pGFSREBP was made by ligating a PCR-generated fragment encoding amino acids 92-410 of the SREBP-1 into the HindIII-EcoRI sites of the pEGF-C1 vector (CLONTECH), which encodes the green fluorescent protein (GFP). The fragment encoding amino acids 92-410 was obtained by PCR using pCMV-SREBP-1a (ATCC) as a template, 5Ј primer containing HindIII site (5Ј-cccaagcttgggcagcgccctcacccctg-3Ј), and 3Ј primer containing EcoRI site (5Ј-ggaattccctatgtgttccctccactgcca-3Ј). This construct results in the expression of the truncated SREBP-1 protein (92-410) as a fusion to the C terminus of the GFP. Thus, the level of expression and the location of the truncated form of SREBP-1 can readily be followed by fluorescence microscopy. Astrocytes were transfected by the calcium phosphate method with 0.9 g of DNA (pEGF-C1 or pGFSREBP)/ml of DMEM ϩ 10% fetal calf serum. After the addition of the calcium phosphate-DNA coprecipitate, 100 M chloroquine was added for 5 h. Then, cells were washed, and 15% glycerol in HEPES-buffered saline was added for 30 s. After washing, cells were incubated in DMEM ϩ 10% fetal calf serum for at least 24 h.
Western Blotting-Proteins (90 g) were extracted from the cells and applied to a 10% SDS-polyacrylamide gel under reducing conditions and transferred to a nitrocellulose sheet (Bio-Rad) as described (25). The nitrocellulose sheet was then exposed for 12 h to monoclonal SREBP-1 antibody (1:500, PharMingen). Anti-mouse Ig biotin (1:1000, Sigma) was applied for 4 h followed by avidin-peroxidase conjugate (1:2000, Sigma) for 1 h. The AEC substrate staining kit (Sigma), prepared according to the manufacturer's instructions, was applied for 10 min.
Northern Blotting-Total RNA was extracted with Trizol reagent (Invitrogen) followed by RNA precipitation with isopropyl alcohol and purification with 75% ethanol according the manufacturer's instructions. The RNA (20 g) was then run on an agarose-formaldehyde gel and transferred to a positively charged nylon membrane (Pall). The membrane was hybridized in the ExpressHyb hybridization buffer (CLONTECH) with 32 P[dCTP]-labeled cDNA at 65°C for 1 h. The cDNA-encoding SCD (donated by Dr. J. M. Ntambi (26)) in pC3 vector was obtained by digestion with PstI. The SCD cDNA was then labeled with 32 P[dCTP] using the Random priming kit (Roche Molecular Biochemicals) according to the manufacturer's instructions. Unincorporated deoxyribonucleoside triphosphates were removed by chromatography through a Sepharose column. The nylon membrane was rinsed and then washed twice for 20 min in 2ϫ SSC with 0.05% sodium dodecyl sulfate at room temperature followed by two 20-min washes in 0.1ϫ SSC with 0.1% sodium dodecyl sulfate at 50°C before exposing to Biomax MS film (Eastman Kodak Co.).
In Situ Hybridization-The cDNA-encoding SCD, obtained as described above, was labeled with biotin-14-dCTP using the BioPrime DNA labeling kit (Invitrogen) and was used to detect SCD mRNA in astrocytes, as described (27). Cells were hybridized with 0.3 g/ml biotin cDNA at 65°C for 18 h and washed several times in decreasing concentrations of standard saline citrate buffer. Biotin-labeled SCD cDNA probe was detected with the ExtrAvidin-Cy3 conjugate (Sigma).
Statistical Analyses-Results are means Ϯ S.E. n represents the number of individual experiments carried out in duplicate or triplicate. Statistical analyses were carried out using ANOVA followed by the Scheffe F-test. (22) have previously shown that astrocytes internalize BSA by a temperature-dependent process. To gain insight into the mechanism by which albumin is taken up by astrocytes, cells were exposed to FITC-BSA in the absence or in the presence of PAO, an inhibitor of receptor-mediated endocytosis (28). After 10 min of loading, intracellular fluorescence was visualized by confocal microscopy (Fig. 1, A and C) and quantified from digitized fluorescence photomicrographs (Fig. 1E). As can be seen in Fig.  1A, astrocytes were conspicuously loaded with FITC-BSA in discrete vesicle-like structures. Under these circumstances, PAO inhibited albumin internalization because only a few fluorescence spots were observed in the phase-contrast fluorescence superimposed photomicrographs (Fig. 1C). In fact, PAO inhibited FITC-BSA uptake by about 75% (Fig. 1E). To confirm the participation of an albumin-binding protein in albumin endocytosis, cells were treated with trypsin or Pronase, proteases that inhibit endocytosis by receptor proteolysis, or with L. flavus agglutinin or G. max agglutinin, lectins that block glycoprotein-mediated endocytosis by binding to glycosyl residues (29). It should be mentioned that L. flavus agglutinin binds to N-acetylneuraminyl or N-glycolylneuraminyl residues, whereas G. max agglutinin binds to N-acetylgalactosaminyl or galactosaminyl residues. Our results showed that after 5 min of treatment with trypsin or Pronase, astrocytes lost their ability to take up albumin (Fig. 1E). In addition, G. max agglutinin (but not L. flavus agglutinin) inhibited albumin internalization by about 70%. Furthermore, the presence of the specific hapten of G. max agglutinin, GalNAc, prevented the G. max agglutinin effect. Finally, protamine, a basic protein that inhibits absorptive-mediated endocytosis (30), was also tested, but its presence did not modify albumin uptake (Fig. 1E).

Astrocytes Translocate Albumin by Transcytosis-We
When FITC-BSA-loaded astrocytes were allowed to unload, very little fluorescence remained inside the cell after 1 h (Fig.  1B). In order to know the kinetics of BSA release, loaded astrocytes were exposed to 2% (w/v) BSA and medium, and cell fluorescence was followed over 90 min at 37°C (Fig. 1G). The decrease in cell fluorescence was seen to be concurrent with an increase in the fluorescence of the medium, suggesting that FITC-BSA was gradually released from astrocytes. The fluorescence of the medium and of the cells showed a linear and inverse correlation (R 2 ϭ 0.866; p Ͻ 0.001). FITC-BSA release was abolished at 4°C (Fig. 1G), suggesting that the observed FITC-BSA release was an active process (see also Ref. 31). Furthermore, our results revealed that inhibitors of transcytosis, such as BFA (Fig. 1D), monensin, nocodazole, or colchicine (32), inhibited FITC-BSA release. Unloading was reduced to about 80% by the presence of monensin and to about 50 -60% by the presence of BFA, nocodazole, or colchicine, as judged by the amount of fluorescence remaining inside the cells and the fluorescence recovered from the incubation medium (Fig. 1F). It should be noted that albumin uptake was not modified by the presence of brefeldin A, monensin, nocodazole, or colchicine (data not shown).
Astrocytes Release Oleic Acid Along with Albumin-Our results show that inhibitors of albumin endocytosis, such as PAO, inhibit oleic acid synthesis and release by astrocytes by about 90% (Fig. 2A). Likewise, oleic acid synthesis and release was also inhibited by several inhibitors of transcytosis, such as monensin (94%), nocodazole (46%), or BFA (43%) (Fig. 2A). In addition, our observations suggest that oleic acid leaves astrocytes as a fatty acid-albumin complex. Thus, polyacrylamide gel electrophoretic analysis of the medium after incubation of astrocytes for 1 h with 10 mM lactate ϩ 100 -150 dpm/nmol [ 14 C]lactate in the presence of 2% (w/v) BSA revealed that the putative fatty acid-BSA complex had the same R F , regardless of whether the gel was developed for proteins, fatty acids, or radioactivity (Fig. 2B). These results clearly indicate that the radiolabeled oleic acid was bound to albumin in the medium recovered after the incubation. However, the possibility that the oleic acid-albumin complex could be formed in the extracellular medium after the release of free oleic acid from the cells cannot be excluded because albumin was always present in the extracellular medium throughout the experiment. To confirm that the oleic acid-BSA complex is indeed formed inside the cell and then released to the extracellular medium, astrocytes were loaded with FITC-BSA in the presence of radiolabeled substrates. After washing, FITC-BSA was allowed to unload from astrocytes, and the presence of oleic acid complexed to FITC-BSA was analyzed. Our results revealed that in electrophoresis, the FITC-BSA released from astrocytes migrates together with radiolabeled oleic acid (Fig. 2C). In addition, monensin, which inhibits albumin exocytosis, dramatically reduced the amount of FITC-BSA and radiolabeled oleic acid released from astrocytes to the extracellular medium (Fig.  2C). These results suggest that the complex is formed inside the cell before being released to the extracellular medium.
Albumin Induces Stearoyl-CoA 9-desaturase through the Activation of the Sterol Regulatory Element-binding Protein-To gain insight into the mechanism by which albumin enhances oleic acid synthesis, we investigated the possible changes in mature/precursor forms of SREBP-1, a transcription factor that stimulates oleic acid synthesis (11,12). Oleic acid protects SREBP-1 proteolysis by preventing the cleavage and release of low molecular mass SREBP-1 mature form (15)(16)(17), which induces the enzymes involved in oleic acid synthesis (11,12,18). In this context, our results show that albumin increases the amount of the low molecular mass SREBP-1 mature form, whereas it decreases that of the high molecular mass SREBP-1 inactive precursor (Fig. 3A). The effect of albumin on SREBP-1 proteolysis was avoided by the presence of 100 M oleic acid or 50 M nocodazole (the latter inhibits the transport of proteins to the ER (33)) (Fig. 3A). . After 1 h, the incubation medium was analyzed by polyacrylamide gel electrophoresis together with 2% (w/v) BSA, 2% (w/v) BSA ϩ 100 M oleic acid, or ϩ 1 Ci of [1-14 C]oleic acid. Gel portions were developed with Coomassie Blue to visualize proteins and cupric acetate to visualize fatty acids or were exposed to an x-ray film to visualize radioactivity. The only protein detected in the incubation medium was BSA. Astrocytes were preincubated in the absence (Control) or presence of 30 M monensin (Mo) for 20 min, and this was maintained at the same concentration up to the end of the experiment (C). Thereafter, astrocytes were incubated in Elliot buffer containing 5 mM glucose, 10 Ci/ml [1(2)-14 C]acetate, and 2% (w/v) FITC-BSA. After 1 h, the incubation medium was removed, and the cells were washed and incubated in Elliot buffer containing 5 mM glucose for 1 h. The compounds released to the extracellular medium were analyzed by polyacrylamide gel electrophoresis. The gels were examined with an ultraviolet transilluminator to visualize the FITC-BSA released and were autoradiographed to visualize radioactivity, which was identified as oleic acid according to HPLC analysis.
To investigate whether the effect of albumin on SREBP-1 might be associated with changes in the mRNA synthesis of SCD, an enzyme located in the ER that catalyzes the formation of the ⌬ 9 -cis-double bond of oleic acid, Northern blot analysis was performed with RNA extracts from astrocytes under the same experimental conditions as those used to investigate SREBP-1 proteolytic changes. Albumin increased SCD mRNA, whereas the presence of 100 M oleic acid or 50 M nocodazole clearly counteracted its effect (Fig. 3B).
To confirm whether SREBP-1a is directly involved in the stimulation of oleic acid synthesis caused by albumin, we used a truncated form of the SREBP protein that binds to the SRE-1 site of the SCD gene, preventing the stimulation of SCD transcription caused by SREBP-1a. The truncated form of SREBP-1a containing amino acids 92-410 has been reported to be an efficient inhibitor of the transcriptional activity of SREBP-1a. In fact, this truncated protein retains the ability to translocate to the nucleus and bind to SRE-1, preventing the binding of SREBP-1a (34). Accordingly, we designed experiments to overexpress this truncated form of SREBP-1a in astrocytes and to test the effect of albumin under these circumstances. A construction designated pGFSREBP was made that results in the expression of the truncated SREBP-1 protein (92-410) as a fusion to the C terminus of the GFP. Astrocytes were then transfected with pEGFP or with pGFSREBP. After 24 h, the cells were incubated with albumin for 8 h, and the level of expression of the SCD mRNA was analyzed by fluorescence in situ hybridization (Fig. 3, C-F). Astrocytes transfected with pGFSREBP (Fig. 3, C and E) showed a clear translocation of the truncated protein to the nucleus (Fig.  3C, arrows). In addition, SCD mRNA was not increased in pGF-SREBP-transfected astrocytes ( Fig. 3E; note the absence of SCD mRNA in the transfected astrocyte, indicated by arrows), whereas it was strongly enhanced in non-transfected astrocytes (other cells in Fig. 3E) by the presence of albumin. The overexpression of GFP (Fig. 3D) did not modify the effect of albumin on SCD mRNA expression (Fig. 3F). DISCUSSION We (2) have recently shown that astrocytes synthesize and release oleic acid, which induces neuronal differentiation. The signal that triggers oleic acid synthesis and release by astrocytes is albumin (2), a serum protein that is specifically taken up by the brain during development (3,4,6,7). In this work, we show that astrocytes take up albumin by receptor-mediated endocytosis and that this is followed by translocation (see also Ref. 31) and exocytosis (Fig. 1). Thus, the existence of absorptive-mediated endocytosis can be precluded because albumin uptake was not affected by the presence of protamine (Fig. 1E), an inhibitor of this process (30). On the other hand, albumin internalization was inhibited by PAO (Fig. 1E), an inhibitor of receptor-mediated endocytosis (28). Our results are consistent with the idea that an albumin-binding protein located on the astrocytic plasma membrane would be involved in albumin uptake. In consonance with this, a brief treatment with proteases prevented albumin internalization in astrocytes (Fig. 1E), suggesting that albumin internalization is mediated by a protein that acts as a putative albumin receptor. This would probably be a glycoprotein containing N-acetylgalactosaminyl or galactosaminyl residues because albumin endocytosis was inhibited by G. max agglutinin (Fig. 1E), a lectin that binds to these glycoprotein residues (29). Our results also suggest that, once inside the astrocyte, albumin moves by vesicle-mediated transcytosis (Fig. 1A). In fact, FITC-BSA release was a temperature-dependent process (Fig. 1G) and was inhibited by microtubule-disrupting agents such as nocodazole or colchicine (Fig. 1F), suggesting that albumin is transported by microtubule-guided vesicles from the Golgi to the ER (33). Moreover, albumin release by astrocytes was inhibited by BFA and monensin (Fig. 1, D and F), both of which prevent vesicular trafficking from the ER to the Golgi (35) or from this to the plasma membrane (36), respectively.
Our results show that the inhibition of endocytosis, transcytosis, or exocytosis by protein trafficking inhibitors is associated with a significant inhibition of oleic acid synthesis and/or release by astrocytes ( Fig. 2A). Therefore, it may be concluded that the passage of albumin through the astrocyte promotes oleic acid synthesis. Moreover, both BFA and nocodazole inhib- ited the intracellular trafficking of FITC-BSA (Fig. 1), indicating that albumin would pass through the ER (33). Because albumin shows a high affinity for oleic acid (37) and because this fatty acid is synthesized in the ER, it is reasonable to suggest that albumin sequestrates oleic acid in the ER (Fig. 4) to be released as an oleic acid-albumin complex by active exocytosis (Fig. 2, B and C, and Fig. 4).
Oleic acid synthesis must be feedback-regulated because oleic acid suppresses SREBP-1 activation (15)(16)(17) and SCD induction (18), both of which accompany the decrease in oleic acid synthesis (19). SREBP-1 is the transcription factor that induces the enzymes involved in oleic acid synthesis, in particular SCD (11,12), an enzyme located in the ER that catalyzes the formation of the ⌬ 9 -cis-double bond of oleic acid. SREBP-1 is also located in the ER and is activated by proteolysis, releasing an active mature form which binds to the promoters of genes containing SRE (for a review, see Ref. 13). In this context, our results show that albumin promotes SREBP-1 cleavage, resulting in an increase in the SREBP-1 mature form (Fig.  3A). SREBP-1 proteolysis promoted by albumin was associated with the induction of SCD, as judged by the increase in SCD mRNA concentrations observed in the presence of albumin (Fig. 3B). In fact, the truncated form of SREBP-1, which retains the ability to translocate to the nucleus (Fig. 3C) and bind to SRE-1, preventing the binding of SREBP-1a (34), inhibited the effect of albumin on SCD mRNA expression (Fig. 3E). Therefore, it may be concluded that SREBP-1 mediates the effect of albumin on oleic acid synthesis.
Because oleic acid protects SREBP-1 from proteolysis (15)(16)(17), it is reasonable to suggest that the effect of albumin would be brought about by sequestration of the fatty acid in the ER. In agreement with this, when the sequestration of oleic acid was prevented by the addition of albumin already containing oleic acid, SREBP-1 cleavage and the increase in SCD mRNA were abolished (Fig. 3, A and B). Furthermore, when the passage of albumin through the ER was prevented by nocodazole, a drug that inhibits the transport of proteins to the ER (33), the effect of albumin on SREBP-1 activation and SCD mRNA was also avoided (Fig. 3, A and B), indicating that albumin has to reach the ER to activate the signal cascade leading to oleic acid synthesis.
Taken together, our results indicate that the transcytosis of albumin through the astrocyte involves its passage through the ER (Fig. 4). Once in the ER, albumin would sequestrate newly synthesized oleic acid, activating the transcription factor SREBP-1. The mature form of SREBP-1 would migrate to the nucleus, activating the transcription of SCD in order to increase oleic acid synthesis. The oleic acid-albumin complex would then be transported from the ER to the plasma membrane, where it would be released by active exocytosis to access neurons. It should be mentioned that the end feet of astrocytes completely surround blood capillaries, whereas the other processes of astrocytes contact with neurons (38). Because the transfer of albumin from the blood to the brain across the blood-brain barrier is developmentally regulated (6), it is easy to speculate that during development, the end feet of astrocytes take up albumin, which is then translocated to the ER, inducing oleic acid synthesis. The oleic acid-albumin complex would then be released to the extracellular medium in contact with neurons to promote its differentiation (Fig. 4).

FIG. 4. Regulation of oleic acid synthesis by albumin in astrocytes.
Removal of oleic acid caused by transcytosis of albumin through the ER activates SREBP-1, which induces the transcription of stearoyl-CoA 9-desaturase, promoting the synthesis of oleic acid. The oleic acid-albumin complex is then released to the extracellular medium by active exocytosis. ABP, albumin-binding protein.