You are here

Article Text

Article menu

Download PDFPDF

Pancreas
Formation of vitamin A lipid droplets in pancreatic stellate cells requires albumin
  1. N Kim1,
  2. W Yoo1,
  3. J Lee1,
  4. H Kim1,
  5. H Lee2,
  6. Y-S Kim3,
  7. D-U Kim4,
  8. J Oh1
  1. 1
    Laboratory of Cellular Oncology, Korea University Graduate School of Medicine, Ansan, Gyeonggi do, Korea
  2. 2
    Department of Internal Medicine, Korea University Ansan Hospital, Ansan, Gyeonggi do, Korea
  3. 3
    Department of Pathology, Korea University Ansan Hospital, Ansan, Gyeonggi do, Korea
  4. 4
    Functional Genomics Research Cencer, Korea Research Institute of Bioscience and Biotechnology, Daejeon, Korea
  1. Correspondence to Professor J Oh, Laboratory of Cellular Oncology, Institute of Biomedical Science, Korea University Graduate School of Medicine, Gojan 1-dong, Danwon gu, Ansan, Gyeonggi do 425-707, South Korea; ohjs{at}korea.ac.kr

Abstract

Objective: Quiescent pancreatic stellate cells (PSCs) store vitamin A as cytoplasmic lipid droplets, and, when activated by profibrogenic stimuli, they transform into myofibroblast-like cells characterised by the loss of vitamin A droplets. Activation of stellate cells is central to fibrogenesis, but the mechanism for the formation of vitamin A droplets and its relationship to stellate cell activation remain unclear.

Methods: With use of cultured PSCs, an attempt was made to characterise the function of albumin endogenously expressed in stellate cells.

Results: Albumin is endogenously expressed in quiescent PSCs, localised in cytoplasmic lipid droplets, and its levels are markedly reduced after stellate cell activation. Continuous albumin expression in stellate cells is sufficient to maintain their fat-storing phenotype even after cell passages and renders cells resistant to the activating effects of transforming growth factor β (TGFβ). Forced expression of albumin in PSCs after passage 2 (activated PSCs) induced the re-appearance of lipid droplets and phenotypic changes, which were previously reported with retinol treatment. Retinol increases albumin synthesis in activated PSCs and the suppression of albumin expression using small interfering RNA (siRNA) abolishes retinol-induced effects.

Conclusions: The data demonstrate a novel role for albumin in the formation of cytoplasmic vitamin A lipid droplets in stellate cells, and suggest that albumin may have a direct influence on stellate cell activation.

https://doi.org/10.1136/gut.2008.170233

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Pancreatic fibrosis, seen in chronic pancreatitis and pancreatic cancer, is characterised by excessive production and deposition of extracellular matrix (ECM) components.1 Numerous studies suggest that activated pancreatic stellate cells (PSCs) play an important role in pancreatic fibrogenesis.2 3 PSCs, in their quiescent state, can be identified by their angular appearance and the presence of vitamin A-containing lipid droplets in their cytoplasm.4 When activated by profibrogenic mediators, they transform into myofibroblast-like cells characterised by positive staining for α-smooth muscle actin (α-SMA), loss of vitamin A lipid droplets and greatly increased synthesis of ECM proteins, including type I collagen.5

    Retinoids (vitamin A and its metabolites) regulate multiple physiological activities, such as vision, reproduction, morphogenesis, cell proliferation and differentiation.6 Vitamin A (retinol) cannot be synthesised de novo by animals and is acquired from the diet. Absorbed vitamin A circulates bound to either retinol-binding protein (RBP) or serum albumin in the bloodstream and is transferred to stellate cells.7 8 9 Hepatic stellate cells (HSCs) store 60–80% of retinoids in the whole body as retinyl esters in cytoplasmic lipid droplets.7 10 Stellate cells that store retinoids in their cytoplasm have been also found in extrahepatic organs such as pancreas, lung, kidney and intestine.11 However, the mechanism of formation of vitamin A droplets is poorly understood.

    Studies showed that PSCs display similar cellular behaviour to HSCs5 12 and many of the molecular and cell phenotype changes associated with the activation of PSCs in vivo are also seen when these cells are cultured in vitro on plastic.13 Follow-up observation regarding changes in gene expression after PSC activation showed that albumin is endogenously expressed in quiescent PSCs, and our further study suggests that albumin plays an important role in the formation of cytoplasmic lipid droplets of quiescent stellate cells.

    Materials and methods

    Materials

    Male Sprague–Dawley rats of 8–10 weeks of age were purchased from Orient (Charles River Korea, Seoul, Korea) and Nagase analbuminaemic rats (NARs) of 8–10 weeks of age were purchased from Japan SLC (Hamamatsu, Japan). All rats were maintained under temperature-, humidity- and light-controlled conditions. The animals received humane care according to the institutional guidelines. Transforming growth factor β1 (TGFβ1) was generously provided by Dr H Lee (Gachon University).

    PSC isolation and culture

    Rat PSCs were isolated as described previously.5 Briefly, the pancreas was finely minced, placed in a solution of Hank’s buffered salt solution with 0.05% collagenase, 0.02% pronase and 0.1% DNase, and shaken for 20 min at 37°C. After filtrations through 150 μm mesh, cells were centrifuged on a 13.2% Nycodenz gradient at 1400 g for 20 min. PSCs were collected from the band just above the interface of the Nycodenz solution and the aqueous layer, suspended in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, and plated on non-coated plastic dishes. After reaching confluence in the primary culture, serial passages were obtained always applying a 1:3 split.

    Western blot analysis

    Cells were rinsed in ice-cold phosphate-buffered saline (PBS) twice and harvested by scraping in the lysis buffer.14 Equivalent amounts of protein were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), followed by immunoblot detection using the primary antibody. Primary antibodies were as follows: albumin (Affinity Bioreagents, Rockford, Illinois, USA), α-SMA (Sigma, St Louis, Missouri, USA), β-actin (Santa Cruz, Santa Cruz, California, USA), α-tubulin (Cell Signaling, Beverly, Massachusetts, USA), type I collagen (Calbiochem, San Diego, California, USA), Smad-2 (Cell Signaling) and Smad-3 (Zymed, San Francisco, California, USA).

    Analysis of gene expression using real-time PCR

    Total RNA was isolated using an RNeasy kit (Qiagen, Valencia, California, USA). cDNA was synthesised from 1.0 μg of total RNA with GeneAmp RNA PCR (Applied Biosystems, Foster City, California) using random hexamers. Real-time PCR was performed using LightCycler-FastStart DNA Master SYBR Green 1 (Roche, Mannheim, Germany) according to the manufacturer’s instructions. The reaction mixture (20 μl) contained LightCycler-FastStart DNA Master SYBR Green 1, 4 mM MgCl2, 0.5 μM of the upstream and downstream PCR primers and 2 μl of the first-strand cDNA as a template. To control for variations in the reactions, all PCRs were normalised against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression. The primers used were as follows: 5′-GGT GGT CTC CTC TGA CTT CAA CA-3′ (forward primer) and 5′-GTT GCT GTA GCC AAA TTC GTT GT-3′ (reverse primer) for GAPDH; 5′-CTG ATA TCT GCA CAC TCC CA-3′ (forward primer) and 5′-TCA GTG GCG AAG CAG TTA TC-3′ (reverse primer) for rat albumin.

    Protein identification by nanoLC–MS/MS

    Nano liquid chromatography–tandem mass spectrometry (nanoLC–MS/MS) analysis was performed on an Agilent 1100 Series nano-LC and LTQ mass spectrometer (Thermo Electron, San Jose, Califonia, USA). The capillary column used for LC–MS/MS analysis (150 mm×0.075 mm) was obtained from Proxeon (Odense M, Denmark) and slurry packed in-house with 5 μm, 100 Å pore size Magic C18 stationary phase (Michrom Bioresources, Auburn, California, USA). The mobile phase A for the LC separation was 0.1% formic acid in deionised water and the mobile phase B was 0.1% formic acid in acetonitrile. The chromatography gradient was set up to give a linear increase from 5% B to 35% B in 50 min, from 40% B to 60% B in 20 min and from 60% B to 80% B in 5 min. The flow rate was maintained at 300 nl/min after splitting. Mass spectra were acquired using data-dependent acquisition with full mass scan (400–1800 m/z) followed by MS/MS scans. Each MS/MS scan acquired was an average of one microscan on the LTQ. The sequest software was used to identify the peptide sequence.

    Immunofluorescence

    PSCs were plated onto glass coverslips in 12-well plates coated with gelatin. Samples were fixed in paraformaldehyde and incubated with antialbumin antibody (Affinity Bioreagents #PA1-25673) overnight at 4°C in a moist chamber, followed by fluorescein isothiocyanate-conjugated goat antichicken IgY (Santa Cruz #sc-2431). Cells were washed with PBS and mounted onto slides. Stained cells were visualised on a Zeiss AXIO Imager M1 microscope.

    Oil red O staining and autofluorescence

    PSCs were stained with oil red O to visualise lipid droplets, essentially as described by Koopman et al.15 Oil red O was diluted in triethyl phosphate instead of isopropanol. The fast-fading vitamin A-specific autofluorescence was observed with light of 330–360 nm (UV) laser.

    Immunohistochemical analyses

    Sections (5 μm thick) of formalin-fixed, paraffin-embedded liver tissues were prepared and stained by immunohistochemical staining with the following antibodies: desmin (#M0760, Dako, Carpinteria, California, USA) and albumin (Santa Cruz #sc-58698). Antibody–antigen complexes were detected by an avidin biotin–peroxidase complex (ABP) method and then developed in diaminobenzidine–hydrogen peroxide. After counterstaining with Harris’ haematoxylin, the slides were dehydrated through graded alcohol and mounted.

    Construction of the expression vector for rat albumin

    Total RNA was isolated from rat liver tissue using an RNeasy kit (Qiagen, Valencia, California, USA) and reverse-transcribed into cDNA using GeneAmp RNA PCR (Applied Biosystems). The entire open reading frame of albumin was amplified by PCR with the designed primers and inserted into mammalian expression vector pcDNA3.1+. A point mutant (R410A/Y411A/K525A) of albumin was prepared by a PCR-based method using a Muta-direct Site-Directed Mutagenesis Kit (iNtRON, Teju, Korea). All constructs were confirmed by DNA sequencing.

    Transfection of PSCs and siRNA

    PSCs after passage 2 were transiently transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, California, USA) and subject to analysis after 36 h. For selection of stably transfected subclones, G418 (400 μg/ml) was supplemented for 5 days after cell transfection.

    Sequences for small interfering RNA (siRNA) are as follows: 5′-GAG GAU GAA GUG CUC CAG U(dTdT)-3′ (sense) and 5′-ACU GGA GCA CUU CAU CCU C(dTdT)-3′ (antisense) for albumin; 5′-CAC AGU ACU CGU GUG UGA U(dTdT)-3′ (sense) and 5′-AUC ACA CAC GAG UAC UGU G(dTdT)-3′ (antisense) for Smad2; and 5′-GUG AUG UGU AGU UCC UAG U(dTdT)-3′ (sense) and 5′-ACU AGG AAC UAC ACA UCA C(dTdT)-3′ (antisense) for Smad3. PSCs were transfected with siRNA at 30% confluence using Lipofectamine 2000 according to the manufacturer’s instructions and subject to analysis after 72 h.

    Statistical analysis

    Results are expressed as mean (SD). Paired statistical analysis was done using t tests. Comparisons were considered significant at p<0.05 and p values were two tailed.

    Results

    Albumin is expressed in quiescent PSCs

    As activation of PSCs is the central event in pancreatic fibrosis, we sought to identify proteins differentially expressed in the process of stellate cell activation. In our previous study demonstrating the post-translational processing of RECK (110 kDa),16 we reproducibly observed an immunoreactive band at the position of ∼65 kDa when whole-cell lysates of quiescent PSCs were subjected to western blotting using monoclonal antibody against RECK. To confirm the identity of the immunoreactive band, we partially purified the immunoreactive fraction by Q-Sepharose fast protein liquid chromatography (FPLC) and analysed it by nanoLC–MS/MS. Surprisingly the results obtained identified the protein in question as albumin (fig 1A). To examine the expression pattern of albumin, PSCs were isolated from rat pancreas, and whole-cell lysates were prepared from PSCs at different stages of activation in culture and analysed by western blotting. PSC activation is directly correlated with the duration of in vitro culture and the expression of α-SMA, a marker for the activated PSC phenotype13 (fig 1B). Albumin protein was found to be expressed in freshly isolated PSCs and in PSCs at day 3 after plating (PSCs-3d; preactivated PSCs), but its levels were markedly reduced in PSCs after day 5, at which point in time vitamin A-containing lipid droplets are abundantly present in cytoplasm but begin to disappear gradually. Real-time PCR analysis revealed similar changes in albumin mRNA levels during this time period of culture (fig 1C). On the other hand, RBP, a transport protein for vitamin A, was not detected in PSCs at any stage of culture activation (data not shown).

    Figure 1
    Figure 1

    Expression of albumin during the culture activation of rat pancreatic stellate cells (PSCs). (A) Albumin partially purified by Q-Sepharose fast protein liquid chromatography was separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, stained by Coomasie blue (denoted by an asterisk) and analysed by nano liquid chromatography–tandem mass spectrometry. (B) Cell lysates were prepared from freshly isolated PSCs (0d), PSCs at 3 (3d) and 5 (5d) days after plating, and from PSCs after passage 1 (P1), 2 (P2) and 3 (P3), and analysed by western blotting. α-Smooth muscle actin (α-SMA) was used as a marker for PSC activation and β-actin was used as a loading control. (C) Total RNA was isolated from PSCs at different stages of activation and analysed for albumin by real-time PCR. The data are expressed as the percentage of PSCs at 3 days after plating (PCSs-d3) and represent the means (SD) for three independent experiments. **p<0.01 compared with PSCs-d3.

    Albumin is localised in cytoplasmic lipid droplets of PSCs

    We then investigated the cellular distribution of albumin in both PSCs 3 days after plating (PCSs-d3) and PSCs after passage 2 (PSCs-P2; activated PSCs) by immunofluorescence study. PSCs-d3 are identified by the presence of cytoplasmic vitamin A-containing lipid droplets, as assessed by vitamin A autofluorescence and oil red O staining (fig 2B,D). In PSCs-d3, an intense signal for albumin was localised at the lipid droplets (fig 2C), whereas only a weak signal could be detected in the cytoplasm of PSCs-P2 (fig 2G). Intense albumin signal could not be observed without previous permeabilisation of the plasma membrane (data not shown), suggesting that albumin staining is not membrane bound. These findings show a significant change in albumin expression during the course of PSC activation.

    Figure 2
    Figure 2

    Phase contrast images (A, E), autofluorescence images (B, F), differential interference contrast images of oil red O staining (D, H) and immunofluorescence images (C, G) using anti-albumin antibody are shown for pancreatic stellate cells 3 days after plating (PCSs-d3) and PSCs after passage 2 (PSCs-P2), respectively. Scale bar = 10 μm.

    Albumin expression in PSCs isolated from NARs

    To examine if albumin plays an irreplaceable and fundamental role in quiescent PSCs, we used NARs. NARs are an albumin gene mutant strain of Sprague–Dawley rats and express <1% albumin compared with normal rats.17 First, we analysed serum protein samples by western blotting and confirmed that serum albumin was rarely detected in NARs (fig 3A). PSCs were isolated from NAR pancreas, and cell lysates of NAR-PSCs-d3 and NAR-PSCs-P1 were analysed by western blotting. Interestingly, an albumin protein (66 kDa) level comparable with that in PSCs-d3 of normal rats was clearly demonstrated in NAR-PSCs-d3 and its level reduced after cell passage (fig 3B). As seen in PSCs-d3 of normal rats, albumin in NAR-PSCs-d3 was also located in vitamin A- containing lipid droplets (fig 3D–F), and its mRNA level was ∼30% of that of normal PSCs-d3 (fig 3G) and declined with culture activation (data not shown).

    Figure 3
    Figure 3

    Albumin expression in pancreatic stellate cells (PSCs) and hepatic stellate cells (HSCs) of normal rats or Nagase analbuminaemic rats (NARs). (A) Equivalent amounts of serum protein samples obtained from normal rats and NARs were analysed for albumin by western blotting. (B) PSCs were isolated from the pancreas of normal rats or NARs, cultured, harvested at 3 days after plating (d3) or after passage 1 (P1), and analysed by western blotting. (C–F) Phase contrast images (C), autofluorescence images (D), differential interference contrast (DIC) images of oil red O staining (F) and immunofluorescence images (E) using antialbumin antibody are shown for NAR-PSCs-d3. (G) Albumin mRNA levels in PSCs-d3 of normal rats or NARs were determined by real-time PCR. The data are expressed as the percentage of normal rats and represent means (SD) for three independent experiments. **p<0.01 compared with the normal rats. (H) HSCs were isolated from normal rat liver, cultured, harvested at 3 days after plating (d3) or after passage 2 (P2), and analysed by western blotting, along with whole-cell lysates of liver mass.

    Albumin is also expressed in quiescent HSCs

    To test whether albumin is also expressed in HSCs, we isolated HSCs from normal rat liver and analysed cell lysates of HSC cultures. Albumin levels in HSCs were also found to decline with culture activation (fig 3H), and immunofluorescence study of HSCs-d3 revealed the localisation of albumin in cytoplasmic lipid droplets (data not shown). We then carried out immunohistochemical analysis on liver sections of normal rats and NARs to determine the in vivo distribution of albumin protein. Normal liver section showed abundant albumin-positive hepatocytes, whereas in the NAR liver a few cells positive for desmin, a marker for stellate cells, were positive with the albumin antibody (fig 4A, B). Thus, our data document the spatial and temporal distribution of albumin in quiescent stellate cells.

    Figure 4
    Figure 4

    Representative liver sections of a normal rat (A) or a Nagase analbuminaemic rat (NAR) (B) were immunohistochemically stained for desmin and albumin. Stellate cells are indicated by the arrow. Scale bar = 50 μm.

    Albumin is involved in the formation of vitamin A lipid droplets

    The spatial and temporal distribution of albumin suggests its irreplaceable role in quiescent stellate cells. To confirm this, we carried out different experiments. First, we transfected PSCs-P2 (activated PSCs) with empty vector or expression plasmids for albumin and examined phenotypic changes. When grown in standard culture conditions, control PSCs-P2 had an elongate fibroblastoid morphology, and small, tiny dots could be observed in the perinuclear area (fig 5A–D). On the other hand, albumin expression led to the formation of autofluorescent lipid droplets occupying the major part of the cell volume and induced phenotypic changes from the myofibroblast to the fat-storing cells with a more regular polygonal shape (fig 5E–H). Such a phenotypic change is accompanied by a decrease in α-SMA levels and type I collagen (fig 5M). Previous studies have shown that retinol taken up by stellate cells is esterified to retinyl ester, mainly as retinyl palmitate, for storage and that albumin binds fatty acids via its domain III.18 19 To examine the possibility that albumin may promote the formation of vitamin A droplets via interaction with retinol-associated fatty acids, we substituted three amino acid residues, Arg410, Tyr411 and Lys525, which were reported to be in contact with the bound fatty acids,18 for the alanine residue and studied the effects of mutant expression. Interestingly, cells expressing the mutant albumin (R410A/Y411A/K525A) exhibited a flattened morphology and enlarged cell size with a slight increase in lipid droplet formation, as compared with the control PSCs (fig 5I–L). Immunofluorescence study revealed that mutant albumin staining is distributed diffusedly in the cytoplasm (fig 5K), whereas wild-type albumin is highly concentrated in lipid droplets (fig 5G). It is interesting to note that the p53 protein level is considerably increased but α-SMA is still at a low level in mutant albumin-expressing PSCs (fig 5M). Secondly, PSCs at 2.5 days after plating (PSCs-d2.5) were transfected with albumin expression vector, cultured in the presence of G418, and we examined phenotypic changes after passage 2. Unlike the case of activated PSCs-P2 (fig 5A–D), cytoplasmic lipid droplets could be persistently detected in albumin stable transfectants although their quantity varies slightly among cells (fig 6A–D), and protein levels of α-SMA were decreased (fig 6E). Interestingly, treatment of these albumin stable transfectants with TGFβ1 cause no significant morphological changes (data not shown) and protein levels of albumin and α-SMA were not altered (fig 6F), suggesting that albumin stable transfectants are resistant to the activating effects of TGFβ. Thirdly, a previous study showed that retinol treatment inhibits PSC activation20 and even induces the fat-storing phenotype.21 22 To see if albumin is involved in these retinol effects, PSCs-P2 were treated with retinol (10 μM) for 24 h and analysed for albumin by immunofluorescence, western blotting and real-time PCR. In addition to the reported phenotypic changes and reduction in α-SMA levels, retinol treatment indeed led to a >2-fold increase in both the protein and mRNA levels of albumin (fig 6H–J). Increased albumin is also located in lipid droplets as assessed by oil red O staining (data not shown). Lastly, we transfected PSCs-P2 with siRNA specific to albumin and measured the effects in retinol-induced phenotypic changes. Suppression of albumin expression resulted in loss of responsiveness to retinol (fig 6K). Therefore, these findings strongly suggest that albumin plays a direct role in the formation of cytoplasmic lipid droplets, thereby maintaining the quiescent vitamin A-storing phenotype of stellate cells.

    Figure 5
    Figure 5

    Involvement of albumin in the formation of vitamin A lipid droplets. (A–H) Pancreatic stellate cells after passage 2 were transiently transfected with either empty vector (A–D) or expression plasmids for wild-type (E–H) or mutant albumin (I–L) and examined for phenotypic changes. Phase contrast images (A, E and I), autofluorescence images (B, F and J), differential interference contrast (DIC) images of oil red O staining (D, H and L) and immunofluorescence images (C, G, and K) using antialbumin antibody are shown for the transient transfectants. Scale bar = 10 μm. (M) Levels of albumin, α-smooth muscle actin (α-SMA), type I collagen and p53 in transient transfectants were analysed by western blotting.

    Figure 6
    Figure 6

    (A–F) Pancreatic stellate cells (PSCs) 2.5 days after plating were transfected with albumin expression vector, cultured in the presence of G418 and, after passage 2, albumin stable transfectant PSCs were analysed. Phase contrast images (A), autofluorescence images (B), differential interference contrast images of oil red O staining (D) and immunofluorescence images (C) using anti-albumin antibody are shown. Scale bar = 10 μm. (E) Albumin stable transfectant PSCs were analysed by western blotting. Stable transfectants contain reduced α-smooth muscle actin (α-SMA) levels, as compared with activated pancreatic stellate cells after passage 2 (PSCs-P2). (F) Albumin stable transfectant PSCs were treated with transforming growth factor β1 (TGFβ1) for 24 h and analysed by western blotting. (G–J) PSCs-P2 were incubated in the presence or absence of retinol (10 μM) for 24 h and analysed by immunofluorescence (G, H), western blotting (I) and real-time PCR (J). (K) PSCs-P2 were transfected with small interfering RNA (siRNA) specific to albumin, then treated with retinol for 24 h, and analysed by western blotting. These figures are representatives of at least two independent experiments.

    TGFβ downregulates albumin expression at multiple levels

    As TGFβ is a potent profibrogenic mediator,23 we examined its effects on albumin expression in PSCs. TGFβ1 treatment of PSCs-d2.5 led to a reduction in both the protein and mRNA levels of albumin by ∼50% (fig 7A,B). To characterise further the TGFβ signalling, we used siRNA to suppress the expression of Smad2 or Smad3, intracellular effectors for TGFβ, in PSCs-P2 and measured the effects by real-time PCR and western blotting. Lack of Smad2 expression caused an ∼3-fold increase in albumin mRNA levels but increased its protein levels only ∼1.5-fold (fig 7C,D). When the cysteine protease inhibitor E64 was added to the siRNA-Smad2-PSCs-P2, albumin protein levels markedly increased, indicating the presence of post-translational regulation (fig 7E). Thus our findings suggest that multiple levels of regulation are present for albumin gene expression in PSCs. As the profibrogenic effects of TGFβ are mediated in part via Ca2+ influx and oxidative stress,24 we treated PSCs-d3 with N-acetyl-l-cysteine (NAC), an antioxidant, or ruthenium red, a calcium transport inhibitor, but found no significant changes in albumin mRNA levels (data no shown).

    Figure 7
    Figure 7

    Regulation of albumin expression by transforming growth factor β (TGFβ) signal. (A, B) Pancreatic stellate cells (PSCs) at 2.5 days after plating were incubated with TGFβ1 for 24 h and analysed for albumin by western blotting (A) and real-time PCR (B). Each bar represents the standard deviation of triplicate assays. **p<0.01 compared with the untreated control. (C, D) PSCs after passage 2 (PSCs-P2) were transfected with small interfering RNA (siRNA) specific to Smad2 or Smad3 and analysed for albumin by real-time PCR (C) and western blotting (D). (E) After transfection with siRNA-Smad2, PSCs-P2 were incubated with no inhibitor, phenylmethylsulfonyl fluoride (PMSF) (1 mM), E64 (10 μM), BB94 (100 nM) and pepstatin A (Pepst.; 100 μM) for the last 24 h before harvest and analysed by western blotting.

    Discussion

    Stellate cells play pivotal roles in both the regulation of retinoid homeostasis in the whole body and the development of fibrosis.7 10 Using HSC and PSC cultures, we demonstrated that albumin is endogenously expressed in quiescent stellate cells, localised in cytoplasmic, vitamin A-containing lipid droplets. This temporal expression pattern indicates that albumin can be used as a good marker of quiescent stellate cells, and it may explain why previous investigators have not observed albumin expression in activated stellate cells.25 Also our immunohistochemical analysis on liver sections documents the unique spatial distribution of albumin in stellate cells (fig 4). Albumin is primarily synthesised in the liver, but low levels of extrahepatic expression have been documented in several other tissues, including fetal rat kidney, pancreas, lung and heart.26 Thus, this study is the first to demonstrate albumin expression in stellate cells.

    Based upon our findings and previous reports, we hypothesise that profibrogenic stimuli reduce albumin levels in quiescent stellate cells, which probably leads to the loss of vitamin A droplets, making retinoids diffuse out and become metabolised to retinoic acid (RA) isomers, such as all-trans-RA, 9-cis-RA and 9,13-di-cis-RA. This is consistent with the previous findings that the RA isomer content is increased in activated stellate cells as compared with preactivated stellate cells, whereas the contents of retinyl ester and retinol decrease.27 28 Increased RA isomers may further stimulate fibrosis.27 29

    NARs have a 7 bp deletion at the 5′ splice site of the H–I intron of the albumin gene, which affects albumin mRNA processing.30 In NARs most albumin mRNAs lack the exon H sequence and the level of intact albumin mRNA is only 1% of that of total cytoplasmic albumin mRNAs.31 We confirmed that albumin protein is rarely detected in serum protein samples (fig 3A) and liver cells (fig 4B) of NARs. Despite the genetic defect, we found that NARs-PSCs-d3 contain albumin protein (66 kDa) at levels comparable with those found in normal PSCs-d3 (fig 3B) and their albumin level declines during the activation process, as seen in normal PSCs. Thus, it appears that albumin protein persistently expressed in quiescent PSCs of NARs is equally capable of promoting the formation of lipid droplets. The molecular mechanism underlying the regulation of albumin protein expression in quiescent NAR PSCs is not clear at present, but it certainly supports the functional significance of albumin in stellate cells.

    We also found that PSCs-P2 expressing mutant albumin (R410A/Y411A/K525A) show only a slight increase in lipid droplet formation and yet express low levels of α-SMA (fig 5I,M). As mutant albumin-expressing cells display a flattened morphology and enlarged cell size (fig 5I), we checked the protein level of p53, a senescence-associated marker.32 The p53 level was low in proliferative, activated PSCs but its level increased in wild-type albumin-expressing cells and was even further enhanced in mutant albumin-expressing cells (fig 5M). Similar changes were found for p16 levels (data not shown). A previous report demonstrated that senescence of HSCs is accompanied by decreased mRNA levels for ECM proteins and cytoskeletal proteins, including α-SMA.33 Thus, it may be possible that mutant albumin synthesis produces a senescence phenotype in PSCs, leading to a reduction in α-SMA and collagen levels. Further study is required to address this issue.

    We previously demonstrated that RECK, a membrane-anchored matrix metalloproteinase (MMP) inhibitor, is protected from proteolytic degradation by TGFβ signalling in activated PSCs, promoting ECM accumulation.16 Albumin is the main determinant of plasma oncotic pressure and reportedly has many other biological properties, such as free radical scavenging and molecule transportation.34 In this study, we demonstrated a novel role for albumin in the formation of vitamin A lipid droplets in stellate cells. It was also shown that albumin stable transfectant PSCs are resistant to the activating effects of TGFβ and the suppression of albumin expression blocks retinol-induced phenotypic changes in PSCs (fig 6F,K), which strongly suggests that albumin has a direct influence on stellate cell activation. Thus, a better understanding of the regulation of protein expression for the genes, such as albumin and RECK, related to stellate cell activation, would have the potential to lead to the development of therapeutic strategies to prevent or retard the fibrotic process.

    Acknowledgments

    We thank Dr H Lee for providing TGFβ1. We are also grateful to Dr William G Stetler-Stevenson for encouragement and discussion.

    REFERENCES

    1. ↵
      1. Apte MV,
      2. Wilson JS
      . Mechanisms of pancreatic fibrosis. Dig Dis 2004;22:273–9.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Haber PS,
      2. Keogh GW,
      3. Apte MV,
      4. et al
      . Activation of pancreatic stellate cells in human and experimental pancreatic fibrosis. Am J Pathol 1999;155:1087–95.
      OpenUrlPubMedWeb of Science
    3. ↵
      1. Jaster R
      . Molecular regulation of pancreatic stellate cell function. Mol Cancer 2004;3:26.
      OpenUrlCrossRefPubMed
    4. ↵
      1. Bachem MG,
      2. Schneider E,
      3. Gross H,
      4. et al
      . Identification, culture, and characterization of pancreatic stellate cells in rats and humans. Gastroenterology 1998;115:421–32.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Apte MV,
      2. Haber PS,
      3. Applegate TL,
      4. et al
      . Periacinar stellate shaped cells in rat pancreas: identification, isolation, and culture. Gut 1998;43:128–33.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Sporn MB,
      2. Roberts AB,
      3. Goodman DS
      . The retinoids: biology, chemistry, and medicine. New York: Raven Press, 1994.
    7. ↵
      1. Senoo H
      . Structure and function of hepatic stellate cells. Med Electron Microsc 2004;37:3–15.
      OpenUrlCrossRefPubMed
    8. ↵
      1. Kawaguchi R,
      2. Yu J,
      3. Honda J,
      4. et al
      . A membrane receptor for retinol binding protein mediates cellular uptake of vitamin A. Science 2007;315:820–5.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Fortuna VA,
      2. Martucci RB,
      3. Trugo LC,
      4. et al
      . Hepatic stellate cells uptake of retinol associated with retinol-binding protein or with bovine serum albumin. J Cell Biochem 2003;90:792–805.
      OpenUrlCrossRefPubMed
    10. ↵
      1. Senoo H,
      2. Kojima N,
      3. Sato M
      . Vitamin A-storing cells (stellate cells). Vitam Horm 2007;75:131–59.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Nagy NE,
      2. Holven KB,
      3. Roos N,
      4. et al
      . Storage of vitamin A in extrahepatic stellate cells in normal rats. J Lipid Res 1997;38:645–58.
      OpenUrlAbstract
    12. ↵
      1. Benyon RC,
      2. Iredale JP
      . Is liver fibrosis reversible? Gut 2000;46:443–6.
      OpenUrlFREE Full Text
    13. ↵
      1. Kruse ML,
      2. Hildebrand PB,
      3. Timke C,
      4. et al
      . Isolation, long-term culture, and characterization of rat pancreatic fibroblastoid/stellate cells. Pancreas 2001;23:49–54.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Oh J,
      2. Seo DW,
      3. Diaz T,
      4. et al
      . Tissue inhibitors of metalloproteinase 2 inhibits endothelial cell migration through increased expression of RECK. Cancer Res 2004;64:9062–9.
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Koopman R,
      2. Schaart G,
      3. Hesselink MK
      . Optimisation of oil red O staining permits combination with immunofluorescence and automated quantification of lipids. Histochem Cell Biol 2001;116:63–8.
      OpenUrlPubMedWeb of Science
    16. ↵
      1. Lee H,
      2. Lim C,
      3. Lee J,
      4. et al
      . TGF-beta signaling preserves RECK expression in activated pancreatic stellate cells. J Cell Biochem 2008;104:1065–74.
      OpenUrlCrossRefPubMed
    17. ↵
      1. Nagase S,
      2. Shimamune K,
      3. Shumiya S
      . Albumin-deficient rat mutant. Science 1979;205:590–1.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Simard JR,
      2. Zunszain PA,
      3. Ha CE,
      4. et al
      . Locating high-affinity fatty acid-binding sites on albumin by x-ray crystallography and NMR spectroscopy. Proc Natl Acad Sci USA 2005;102:17958–63.
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Ha JS,
      2. Theriault A,
      3. Bhagavan NV,
      4. et al
      . Fatty acids bound to human serum albumin and its structural variants modulate apolipoprotein B secretion in HepG2 cells. Biochim Biophys Acta 2006;1761:717–24.
      OpenUrlPubMed
    20. ↵
      1. McCarroll JA,
      2. Phillips PA,
      3. Santucci N,
      4. et al
      . Vitamin A inhibits pancreatic stellate cell activation: implications for treatment of pancreatic fibrosis. Gut 2006;55:79–89.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Pinzani M,
      2. Gentilini P,
      3. Abboud HE
      . Phenotypical modulation of liver fat-storing cells by retinoids. Influence on unstimulated and growth factor-induced cell proliferation. J Hepatol 1992;14:211–20.
      OpenUrlCrossRefPubMed
    22. ↵
      1. Margis R,
      2. Borojevic R
      . Retinoid-mediated induction of the fat-storing phenotype in a liver connective tissue cell line (GRX). Biochim Biophys Acta 1989;1011:1–5.
      OpenUrlPubMed
    23. ↵
      1. Leask A,
      2. Abraham DJ
      . TGF-beta signaling and the fibrotic response. FASEB J 2004;18:816–27.
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Yang KL,
      2. Chang WT,
      3. Chuang CC,
      4. et al
      . Antagonizing TGF-beta induced liver fibrosis by a retinoic acid derivative through regulation of ROS and calcium influx. Biochem Biophys Res Commun 2008;365:484–9.
      OpenUrlCrossRefPubMed
    25. ↵
      1. Koda H,
      2. Okuno M,
      3. Imai S,
      4. et al
      . Retinoic acid-stimulated liver stellate cells suppress the production of albumin from parenchymal cells via TGF-beta. Biochem Biophys Res Commun 1996;221:565–9.
      OpenUrlCrossRefPubMed
    26. ↵
      1. Nahon JL,
      2. Tratner I,
      3. Poliard A,
      4. et al
      . Albumin and alpha-fetoprotein gene expression in various nonhepatic rat tissues. J Biol Chem 1988;263:11436–42.
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Okuno M,
      2. Sato T,
      3. Kitamoto T,
      4. et al
      . Increased 9,13-di-cis-retinoic acid in rat hepatic fibrosis: implication for a potential link between retinoid loss and TGF-beta mediated fibrogenesis in vivo. J Hepatol 1999;30:1073–80.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Blomhoff R,
      2. Wake K
      . Perisinusoidal stellate cells of the liver: important roles in retinol metabolism and fibrosis. FASEB J 1991;5:271–7.
      OpenUrlAbstract
    29. ↵
      1. Okuno M,
      2. Moriwaki H,
      3. Imai S,
      4. et al
      . Retinoids exacerbate rat liver fibrosis by inducing the activation of latent TGF-beta in liver stellate cells. Hepatology 1997;26:913–21.
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Esumi H,
      2. Takahashi Y,
      3. Sato S,
      4. et al
      . A seven-base-pair deletion in an intron of the albumin gene of analbuminemic rats. Proc Natl Acad Sci USA 1983;80:95–9.
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Shalaby F,
      2. Shafritz DA
      . Exon skipping during splicing of albumin mRNA precursors in Nagase analbuminemic rats. Proc Natl Acad Sci USA 1990;87:2652–6.
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Collado M,
      2. Serrano M
      . The power and the promise of oncogene-induced senescence markers. Nat Rev Cancer 2006;6:472–6.
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Schnabl B,
      2. Purbeck CA,
      3. Choi YH,
      4. et al
      . Replicative senescence of activated human hepatic stellate cells is accompanied by a pronounced inflammatory but less fibrogenic phenotype. Hepatology 2003;37:653–64.
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Evans TW
      . Review article: albumin as a drug—biological effects of albumin unrelated to oncotic pressure. Aliment Pharmacol Ther 2002;16(Suppl 5):6–11.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Funding This work was supported by the 21C Frontier Functional Human Genome Project (grants FG-08-21-16) from the Ministry of Science and Technology of Korea.

    • Competing interests None.

    • Provenance and Peer review Not commissioned; externally peer reviewed.

    • See Commentary, p 1319

    Linked Articles