Histochemistry and Cell Biology

, Volume 138, Issue 3, pp 397–406

Fatty acid-binding protein 4 (FABP4) and FABP5 modulate cytokine production in the mouse thymic epithelial cells


    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Sumie Hiramatsu
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Nobuko Tokuda
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Kazem Sharifi
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Majid Ebrahimi
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Ariful Islam
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Yoshiteru Kagawa
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Linda Koshy Vaidyan
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Tomoo Sawada
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
  • Kimikazu Hamano
    • Department of Surgery and Clinical Science, Graduate School of MedicineYamaguchi University
    • Department of Organ Anatomy, Graduate School of MedicineYamaguchi University
Original Paper

DOI: 10.1007/s00418-012-0963-y

Cite this article as:
Adachi, Y., Hiramatsu, S., Tokuda, N. et al. Histochem Cell Biol (2012) 138: 397. doi:10.1007/s00418-012-0963-y


Thymic stromal cells, including cortical thymic epithelial cells (cTEC) produce many humoral factors, such as cytokines and eicosanoids to modulate thymocyte homeostasis, thereby regulating the peripheral immune responses. In this study, we identified fatty acid-binding protein (FABP4), an intracellular fatty acid chaperone, in the mouse thymus, and examined its role in the control of cytokine production in comparison with FABP5. By immunofluorescent staining, FABP4+ cells enclosing the thymocytes were scattered throughout the thymic cortex with a spatial difference from the FABP5+ cell that were distributed widely throughout the cTEC. The FABP4+ cells were immunopositive for MHC class II, NLDC145 and cytokeratin 8, and were identified as part of cTEC. The FABP4+ cells were identified as thymic nurse cells (TNC), a subpopulation of cTEC, by their active phagocytosis of apoptotic thymocytes. Furthermore, FABP4 expression was confirmed in the isolated TNC at the gene and protein levels. To explore the function of FABP in TNC, TSt-4/DLL1 cells stably expressing either FABP4 or FABP5 were established and the gene expressions of various cytokines were examined. The gene expression of interleukin (IL)-7 and IL-18 was increased both in FABP4 and FABP5 over-expressing cells compared with controls, and moreover, the increase in their expressions by adding of stearic acids was significantly enhanced in the FABP4 over-expressing cells. These data suggest that both FABPs are involved in the maintenance of T lymphocyte homeostasis through the modulation of cytokine production, which is possibly regulated by cellular fatty acid-mediated signaling in TEC, including TNC.


Fatty acid-binding proteinThymusThymic epithelial cellsThymic nurse cellsIL-7IL-18


There is accumulated evidence that fatty acids and metabolites regulate peripheral immune cell functions, such as lymphocyte proliferation, responses to cytokine stimuli and antigen presentation (Brix et al. 2010). The n-3 polyunsaturated fatty acids (PUFA), such as docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), suppress antigen-stimulated production of interleukin (IL)-2 and subsequent clonal expansion of naïve CD4+ T lymphocytes (Pompos and Fritsche 2002), and their activation depends on changes in membrane n-3 PUFA amounts (Brix et al. 2010). Furthermore, eicosanoids, bioactive lipid metabolites produced from a membrane lipid-derived n-6 PUFA (arachidonic acid; AA) serve as highly potent inducers of inflammation and immunity (Wall et al. 2010). The mechanisms of these effects have been explained by alteration of the fatty acid composition of immune cell membranes (Gorjao et al. 2009; Rockett et al. 2010) or direct regulation of cellular signal transduction by fatty acid metabolites (Ogawa et al. 2011). However, it remains elusive how the homeostasis of fatty acids and their metabolites is controlled in the immune cells and tissues.

The thymus is a primary immune organ composed of thymic stromal cells, including thymic epithelial cells (TEC) and bone marrow derived-T lymphocyte progenitors (thymocytes). It is well known that thymocytes are induced their survival, proliferation and differentiation to the functionally mature T lymphocytes by TEC through the production of many kinds of proteins, including cytokines, chemokines and MHC molecules on their cell surface (Anderson et al. 1996, 2007; Chmurzynska 2006). Among these, IL-7 and stem cell factor (SCF) are critical for the survival and early differentiation of immature thymocytes (Massa et al. 2006; von Freeden-Jeffry et al. 1995), while IL-18 is involved in the maintenance of the thymic microenvironment through the regulation of thymic dendritic cells (Ito et al. 2006). In addition to such protein mediators, n-3 PUFA DHA and thromboxane A2 were recently shown to be involved in the apoptosis of immature thymocytes (Oida et al. 1995; Sandal et al. 2009). Therefore, it is reasonable to expect that changes in fatty acid homeostasis in the thymus may influence the T lymphocyte differentiation and/or survival.

Fatty acid-binding proteins (FABPs) constitute a multi-gene family of intracellular fatty acid carrier molecules with low molecular masses 14–15 kDa (Kurtz et al. 1994). Multiple FABP isoforms (FABP1–12) have been isolated from distinct tissues (Gordon et al. 1983; Hunt et al. 1986; Kurtz et al. 1994; Liu et al. 2008; Siegenthaler et al. 1993; Watanabe et al. 1991). FABPs are generally supposed to serve as molecules that promote cellular fatty acid uptake and transport toward specific metabolic pathways, and regulate gene transcription by delivering fatty acids as the ligands for specific nuclear receptors (Haunerland and Spener 2004; Schroeder et al. 2008). Thus far, we have reported that FABP5 (epidermal-type FABP) is localized in various tissues and cells including astrocytes, type II alveolar cells, splenic dendritic cells, mast cells, keratinocytes and TECs (Guthmann et al. 2004; Kitanaka et al. 2003; Owada et al. 1996, 2002a, b; Yamamoto et al. 2008). However, its roles in the thymus are mostly unknown.

In this study, we found that FABP4 (adipocyte-type FABP; aP2) was expressed by the endothelial cells and a subpopulation of TEC, thymic nurse cells (TNC), in the mouse thymic cortex with the spatial difference from expression pattern of FABP5. Furthermore, the functional significance of FABPs in cTEC was examined by analyzing the cytokine expression in the cultured thymic stromal cells stably over-expressing FABP4 or FABP5.

Materials and methods

Animals and ethical considerations

We used 8–10-week-old wild-type C57BL/6 mice that were housed in the specific pathogen-free animal care facility at the Institute of Laboratory Animals, Yamaguchi University School of Medicine. All experiments were reviewed by the Committee for Ethics in Animal Experiments of Yamaguchi University School of Medicine and were carried out under the Guidelines for Animal Experiments of the Yamaguchi University School of Medicine, and the Law and Notification requirements of the Japanese Government.

Cell lines

The mouse TNC line IT-79MTNC (Itoh et al. 1988) (a gift from Dr Tsunetoshi Itoh, Tohoku University) was maintained in DMEM containing 10 % FBS, 10 mM HEPES, 1 mM sodium pyruvate and 50 μM 2-mercaptethanol. The TSt-4/DLL-1 (a mouse thymic stromal cell line in which delta-like-1 gene was retrovirally introduced; a gift from Dr. Hiroshi Kawamoto, RIKEN) (Miyazaki et al. 2005) was maintained in complete medium (RPMI1640 containing 5 % FBS, 1 mM sodium pyruvate, 0.1 mM non-essential amino acids, 50 μM 2-mercaptethanol, 100 μg/mL streptomycin and 100 U/mL penicillin) at 37 °C, 5 % CO2 condition.

Antibodies and other reagents

For immunohistochemical analysis, we used rabbit polyclonal antibodies against mouse FABP4 (Abdelwahab et al. 2007) and rat FABP5 (Owada et al. 2002b). The other antibodies used in the experiments were as follows (see a supplementary Table 1): rat anti-mouse CD11b mAb (M1 70.15.1; Chemicon, Temecula, CA); rat anti-mouse reticular fibroblasts mAb (ER-TR7; BMA Biomedicals, Augst, Switzerland); rat anti-mouse type-1-TECs mAb (ER-TR4) (Defresne et al. 1994; van Vliet et al. 1984a, b); rat anti-mouse I-A/I-E-PE (phycoerythrin)-conjugated mAb (M5/114.15.2; BD Pharmingen, San Diego, CA); rat anti-mouse DEC205 mAb (NLDC145; Acris Antibodies Gmbh, Hiddenhousen, Germany); rat anti-mouse CD4-PE-conjugated (GK1.5) mAb; rat anti-mouse CD8-Spectral Red-conjugated mAb (53-6.7; Beckman coulter, Fullerton, CA); and rat anti-mouse cytokeratin 8 mAb (Troma I; DSHB, University of Iowa, Iowa City, IA). As secondary antibodies, Alexa Flour 488- or 568-conjugated F(ab′) fragments of goat anti-rabbit IgG, goat anti-rat Ig and goat anti-rat IgM (μ chain) (Invitrogen Corp., Carlsbad, CA) were used. For detection of the vasculature in the thymus, FITC-conjugated lectin from Bandeiraea simplicifolia (BS-1; Sigma, St. Louis, MO) which binds to vascular endothelial cells (Augustin et al. 1995), was injected intravenously, and mice were allowed to move freely for 5 min prior to euthanasia.


Thymi were removed after perfusion with 4 % paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB, pH 7.4), cryoprotected with 30 % (w/v) sucrose in 0.1 M PB and then embedded into freezing medium (OCT compound; Sakura Finetek, Torrabce, CA). Cryosections (10-μm thickness) were cut with a cryostat (CM1900; Leica, Nussloch, Germany) and permeabilized with 0.3 % Triton X-100 in phosphate-buffered saline (PBS-) for 30 min at room temperature. The sections were then incubated with primary antibodies over night at 4 °C, washed and incubated with fluorochrome-conjugated secondary antibodies. DAPI or TOTO-3 iodide (Invitrogen Corp.) was used for nuclear staining. After several washes, sections were mounted using GEL/MOUNT mounting medium (Biomeda, Foster City, CA). Images were acquired by confocal laser microscope (LSM510; Carl Zeiss, Oberkochen, Germany). For double staining of FABPs, we used Zenon Alexa Fluor 594 Rabbit IgG Labeling Kit (Invitrogen Corp.) according to the manufacturer’s instructions.

Isolation of TNCs

Isolation of TNCs was performed as described previously (Wekerle and Ketelsen 1980; Wekerle et al. 1980). Briefly, mice were euthanized by cervical dislocation, and thymi were removed and minced into small pieces in cold PBS. The thymic fragments were digested with 0.25 % trypsin and 0.04 mg/mL DNase I (Roche Diagnostics, Mannheim, Germany) in Ca2+, Mg2+-free HBSS (Sigma) for 20 min at 37 °C by gentle passages through an 18G needle. After several rounds of enzymatic digestion, the cells were collected by centrifugation and resuspeneded in DMEM (Sigma) containing 10 mM HEPES. The cell suspension was carefully overlaid onto heat-inactivated fetal calf serum (FCS) in 50-mL conical tube. The tube was stood at room temperature for 15 min to allow sedimentation of TNCs under gravity. The sediment (mainly TNCs) in the FCS fraction was carefully collected, washed with PBS- and overlaid onto FCS again. After four rounds of sedimentation, isolated TNCs were smeared onto glass slides, air dried and fixed with 4 % PFA in 0.1 M PB (pH 7.4) for 5 min. Immunohistochemical staining of the isolated TNCs was performed as described above.

Induction of apoptosis of thymocytes and detection by TUNEL method

Apoptosis of thymocytes was induced by injection of dexamethasone (Dex). The Dex solution was prepared in 16 % ethanol in PBS and injected into peritoneal cavity at a dose of 12.5 mg Dex/kg body weight. For TUNEL staining, thymus was sampled at 8 h after injection and sections were prepared as described above. The sections were preincubated with buffer containing 100 mM sodium cacodylate, 1 mM cobalt (II) chloride and 50 μg/mL gelatin for 5 min at 37 °C, and then reacted with the same buffer containing 100 U/mL terminal deoxynucleotidyl transferase (TAKARA, Tokyo, Japan) and 10 nmol/mL biotin-16-dUTP (Roche Diagnostics) for 60 min at 37 °C. After several washes, the sections were incubated with anti-FABP antibodies followed by incubation with Alexa Flour 488-conjugated streptavidin (Invitrogen Corp.). The sections were stained with a nuclear dye (DAPI) and mounted, and images were acquired by confocal laser microscope.


For RT-PCR and quantitative real-time RT-PCR (qRT-PCR), total RNA was extracted using TRIzol reagent (Invitrogen Corp.) and reverse-transcribed to cDNA by using Transcriptor High Fidelity cDNA Synthesis Kit (Roche Diagnostics) with oligo-dT primer. RT-PCR was performed under optimal conditions with gene-specific primers as described previously (Hart et al. 2007; Muller et al. 2005; Murphy et al. 2005; Zubkova et al. 2005). Gene-specific primers for β-actin were used as an internal control. The sequences of primers used in this study were in Supplementary Table 2. For qRT-PCR, cDNAs were prepared as described above and reactions were performed on the StepOnePlus Real-time PCR System (Applied Biosystems, Foster City, CA). The TaqMan probes used were as follows: Fabp4, Mm00445878_m1; Fabp5, Mm00783731_s1; IL-7, Mm00434291_m1; IL-18, Mm00434225_m1; β-actin, Mm02619580_g1 (Applied Biosystems). Fold expression was calculated by ΔΔCT method and β-actin was used as a reference gene.

Western blotting analysis

Total proteins were extracted from freshly isolated thymi and other organs by homogenization in 2× SDS-PAGE sample buffer containing protease inhibitors (Roche Diagnostics). The concentration of extracted proteins was determined by Bradford assay using BSA as a standard. The proteins were electrophoresed in 12 % polyacrylamide gels and then transferred onto Immobilon-PSQ PVDF membrane (Millipore, Bedford, MA). After blocking with 0.1 % Tween-20 in Tris-bufferd saline containing 5 % skim milk and 1 % BSA, the membranes were reacted with primary antibodies against FABP4 or FABP5 at a final concentration of 0.5 μg/mL. After washing, the membranes were reacted with HRP-conjugated goat anti-rabbit Ig (GE Healthcare, Buckinghamshire, UK) for 1 h at room temperature, and then signals were detected by ECL-Western Blotting Detection System (GE Healthcare).

Construction of expression vector and transfection

The coding regions of mouse FABP4 and FABP5 cDNA were amplified by PCR and each of amplified cDNA was subcloned into the pcDNA3 mammalian expression vector (Invitrogen Corp.). The constructed expression vector (pcDNA3 for mock, pcDNA3/FABP4 or pcDNA3/FABP5) was transfected into TSt-4/DLL1 cells in 24-well plate format using Lipofectamine 2000 (Invitrogen Corp.) following the manufacturer’s instructions. Transformants were selected for 2 weeks in complete medium supplemented with 0.5 mg/mL G418 (Sigma). After isolation and expansion of the resistant clones, the over-expression of FABP4 or FABP5 was confirmed at gene and protein levels by qRT-PCR and western blot analysis, respectively.

DNA array analysis

To analyze the changes in cytokine gene expressions induced by over-expression of FABP4 or FABP5 gene, the total RNAs were purified by RNeasy MinElute Cleanup Kit (Qiagen, CA, USA), and the gene expressions was examined on a focused DNA array (Allergy Tip Genopal, Mitsubishi Rayon, Tokyo, Japan). Data were analyzed by Array Navigator Viewer software (DYNACOM, Chiba, Japan).

Fatty acid stimulation

Established cell lines over-expressing FABP4 or FABP5 (TSt-4/DLL1-FABP4 and TSt-4/DLL1-FABP5) were cultured in a semi-confluent state in complete medium containing 250 μg/mL G418 and starved of FBS for 24 h. The ethanol solutions of fatty acids (FAs): docosahexaenoic acids (DHA, 22C6, n-3), arachidonic acids (AA, 20C4, n-6) or stearic acid (SA, 18C0) (Cayman chemical, Ann Arbor, MI) were dispersed into the complete medium containing 5 % FBS at the final concentrations of 0, 50 and 100 μM, and cells were incubated with each medium. After 24 h of incubation, the cytokine gene expressions were examined by qRT-PCR as described above.

Statistical analysis

All numerical data are shown as mean ± SEM. Statistical comparisons of means were made by F test and Student’s t test. For multiple comparisons in gene expressions, One-way ANOVA or two-way factorial ANOVA followed by Tukey’s post hoc test were made using SPSS 20.0J (SPSS Japan Inc., Tokyo, Japan). P < 0.05 was considered significant.


Among the FABP family members examined by RT-PCR analysis, we detected gene expressions of FABP4 and FABP5 in the adult mouse thymus (Fig. 1a, left), and the protein expression of these two FABPs were further confirmed by western blot analysis (Fig. 1a, right). By immunohistochemisty, FABP5+ cells showed a fine meshwork throughout the cortex and positive cells were also observed in the medulla (Fig. 1b), consistent with our previous findings (Owada et al. 2002a). In contrast, FABP4+ cells showed a honeycomb-like structure enclosing many thymocytes with their thin projections, and they were scattered within the cortex (arrowheads in Fig. 1c). FABP4+ cells were also detected in the medullar and cortex, some of which showed a tubular structure (Fig. 1c). These FABP4+ tubular structures were identified as blood vessels by the co-localization with FITC-BS-1 lectin-positive staining (Fig. 1e), and the FABP4+ cells surrounded the lumen of blood vessels in higher magnification (Fig. 1g), indicating that FABP4 was localized in the endothelial cells. FABP5 did not co-localize with BS-1 and did not express in endothelial cells (Fig. 1d, f).
Fig. 1

Localization of FABP4 and FABP5 in the thymus. a Expression of FABP4 and FABP5 in the mouse thymus examined by RT-PCR (left) and by western blotting (right). Adipose tissue and epidermis were used as control for FABP expression, and β-actin was used as an internal control gene. b Confocal laser micrographs showing the localization of FABP5 in the thymus at low magnification. Note that the fine meshwork of FABP5 (green) is distributed throughout the cortex, and a part of medullary stromal cells were also positive for FABP5. c Confocal laser micrographs showing the localization of FABP4 in the thymus at low magnification. Note that FABP4 (green) is distributed tubular structure and honeycomb-like islets in the cortex indicated by arrowheads. A part of medullary stromal cells were also positive for FABP4. Nuclei were stained by nuclear dye. Co thymic cortex, Me thymic medulla. Bar 50 μm. d, e Confocal micrographs showing the colocalization of FABP4 or FABP5 (red) and BS-1 (green) in the microvessels of the thymic cortex. The arrowheads indicate FABP4+/BS-1+ vessels, and asterisk indicates FABP4+ cortical stromal cells. Bar 20 μm. f, g High magnification view of a large blood vessels in the medulla showing FABP5 and FABP4 expression (green). An arrow indicates nucleous of endothelial cell. Bar 10 μm

To identify the cell type of the FABP4+ cells forming honeycomb-like structure in the cortex, we performed multiple immunofluorescence staining procedures with antibodies against FABP4 and various stromal cell-specific markers. FABP4+ cortical stromal cells were negative for CD11b, ER-TR7 and ER-TR4 (Fig. 2a–c), but positive for cytokeratin 8, MHC class II and NLDC145 (Fig. 2d–f). The FABP4+ cortical stromal cells were also positive for FABP5 (Fig. 2g–i). These findings indicate that the FABP4+ cortical stromal cell is not thymic macrophage or reticular cell/reticular fiber, but a population of cTEC.
Fig. 2

Expression of various cTEC markers in FABP4+ cortical stromal cells. Confocal laser micrographs of thymus sections stained for FABP4 (green), and various cell type-specific markers (red); a CD11b for macrophages, b ER-TR-7 for reticular fiber and reticulocytes, c ER-TR4 for type-1 TEC, d cytokeratin (CK) 8 for cTECs, e MHC class II for cTECs and mTECs, and f NLDC145 (DEC205) for cTECs. The arrowheads in panelsdf indicate FABP4+/cTEC marker+ cells (yellow). gi Confocal laser micrographs of a thymus section double stained with anti-FABP5 (green), anti-FABP4 (red). The arrowheads indicate FABP5+ FABP4+ cTEC. In panelsaf and i, the nuclei were stained with a nuclear dye (blue). BV blood vessel. Bars 20 μm

It has been shown that a subpopulation of cTEC can be defined as TNC based on its feature of enclosing many (2–200) thymocytes and further by their MHC molecule and cytokeratin expression in vitro and in vivo (Guyden and Pezzano 2003; Wekerle and Ketelsen 1980; Wekerle et al. 1980). To clarify whether the FABP4+ cTEC was TNC, we isolated TNC from freshly removed thymi by enzymatic digestion and sedimentation. By immunocytochemistry, isolated TNCs were positive for MHC (II) and NLDC145, and were also positive for both for FABP4 and FABP5 (Fig. 3a, b). The expression of FABP4 and FABP5 protein in the isolated TNCs was further confirmed by western blotting analysis (Fig. 3c). In addition, FABP4+ TNC enclosed CD4+CD8+ (double-positive; DP) thymocytes in the thymic cortex (Fig. 3d).
Fig. 3

Identification of FABP4 in isolated TNC. a Confocal micrographs of isolated TNC stained for FABP4 or FABP5 (left, green), NLDC145 (middle, red) and merged images. b Confocal micrographs of isolated TNC stained for FABP4 and FABP5 (left, green), and MHC class II (middle, red), and merged images (right, yellow). Bars 20 μm. c Expression of FABP4 and FABP5 protein examined by western blotting in isolated TNC and thymic stromal cell lines. Adipose tissue and epidermis were used as positive controls for FABP4 and FABP5 expression. β-Actin was used as a loading control. d Confocal micrographs stained for FABP4 (green), CD4 (red), CD8 (blue) showing the interaction of FABP4+ cTEC with DP thymocytes. BV blood vessel. Bar 10 μm. e Confocal laser micrographs showing the phagocytosis of Dex-induced apoptotic thymocytes (8 h) by FABP4+ cTECs (red). Arrows TUNEL+ thymocytes (green), arrowheads pyknotic thymocytes (blue). Nuclei were stained with a nuclear dye (blue). Bar 20 μm

A previous study showed that TNCs are involved in the removal of apoptotic thymocytes by phagocytosis through the recognition of apoptotic changes in cell membrane of thymocytes (Cao et al. 2004). Therefore, we examine the relationship between the apoptotic cells and FABP4+ TNC after the induction of thymocyte apoptosis by Dex. At 8 h after Dex injection, the FABP4+ TNCs had engulfed TUNEL+ cells and/or TUNEL- pyknotic thymocytes (Fig. 3e). In addition, FABP4+ TNCs enclosed DP thymocytes (Fig. 3d), consistent with the previous finding that TNCs interact with αβTCR+ DP thymocytes (Pezzano et al. 1996).

To explore the function of FABPs in cTEC, we established two thymic stromal cell lines stably expressing either FABP4 (TSt-4/DLL1-FABP4) or FABP5 (TSt-4/DLL1-FABP5), and confirmed over-expression of their transcript and protein (Fig. 4a). In this experiment, we examined the gene expression of IL-7 and IL-18, because they are produced from TEC and important for survival and expansion of thymocytes and for maintenance of thymic microenvironment (Hare et al. 2000; Ito et al. 2006; Kim et al. 1990). IL-7 and IL-18 gene expressions were significantly increased in both cell lines (approximately 2.3-fold for IL-7 and 2.1-fold for IL-18 in TSt-4/DLL1-FABP4 cells; 1.5-fold for IL-7 and 1.7-fold for IL-18 in TSt-4/DLL1-FABP5 cells versus mock control, Fig. 4b). Such alterations in the cytokine gene expression were further confirmed by the DNA array analysis (approximately 5-fold for IL-7 and approx. 11.8-fold for IL-18 in TSt-4/DLL1-FABP4 cells, and 6.9-fold for IL-7 and 4.9-fold for IL-18 in TSt-4/DLL1-FABP5 cells, see a Supplementary Table 3).
Fig. 4

Altered gene expression profiles and effect of stearic acid on IL-7 gene expression in FABP over-expressing cells. a Confirmation of over-expressed transcripts and protein of FABP4 and FABP5. The results of qRT-PCR and western blot analysis showed increased amount of transcripts and proteins. β-Actin was used as reference gene for normalization of gene expression and loading control of proteins. b Altered cytokine gene expression of FABP gene-transfected TSt-4/DLL1 cells (mock, TSt-4/DLL1-FABP4 and TSt-4/DLL1-FABP5). RT-PCR results showing the cytokine gene expressions in mock-transfected cells (left). Bar graphs showing the relative gene expression levels of IL-7 (left), IL-18 (right) by qRT-PCR. Quantitative data are shown as mean ± SEM (n = 3), and values in mock cells were defined as 1.0. c Effects of stearic acid on IL-7 gene expression. The IL-7 gene expression levels in mock cells (black bars), TSt-4/DLL1-FABP4 cells (gray bars) and Tst-4 DLL1-FABP5 cells (white bars) were examined by qRT-PCR after stimulation with stearic acid (18C0,) at 0, 50 and 100 μM. The data are indicated as mean ± SEM (n = 3). The values in mock cells without FA stimulation (0 μM) are defined as 1.0. β-Actin was used as reference gene for normalization of gene expression. Asterisks indicate statistically significant differences (P < 0.05)

It has been shown that the fatty acid environment of immune cells can affect their function, including cytokine production (Pompos and Fritsche 2002), and FABP has been shown to serve as a regulator of fatty acid-mediated immune cell control (Kitanaka et al. 2006). Thus, we examined whether the FABP over-expressing thymic stromal cells change IL-7 production in response to different cellular fatty acid environment. Interestingly, when 100 μM of stearic acids (18C0) was added into cultures, significant increase of IL-7 gene expression was observed in mock cells (approx. fourfold, Fig. 4c black bars) and such increase was more enhanced in TSt-4/DLL1-FABP4 (approx. sixfold, Fig. 4c grey bars), while TSt-4/DLL1-FABP5 did not show the enhancement of the gene expression (approx 1.5-fold, Fig. 4c white bars). On the other hand, arachidonic acids (20C4, n-6) or docosahexaenoic acids (22C6, n-3) did not enhance the increase of IL-7 gene expression significantly neither in TSt-4/DLL1-FABP4 nor TSt-4/DLL1-FABP5 (data not shown). These results suggest the possibility that cellular environment of saturated FA of TEC contributes to the IL-7 gene expression and that FABP4 may serve as a modulator of this process.


The production of functional T lymphocytes, which is largely dependent on the thymus, is crucial for adaptive immunity. The thymic stromal cells, including TEC, dendritic cells and macrophages, have been shown to regulate the maturation of thymocytes. The TEC produce many kinds of humoral factors, such as cytokines for induction of proliferation, differentiation and apoptosis of thymocytes (Chouaib et al. 1985; McCormack et al. 1991; Oida et al. 1995). In this study, we have shown that two FABPs, FABP4 and FABP5, are expressed in mouse cTEC with a spatial difference. Furthermore, we propose the possible involvement of FABP in the regulation of cytokine expressions in TNC.

In the present study, the FABP4+ cells in the mouse thymus were positive for both cytokeratin 8 and MHC class II, and they were found to enclose many DP thymocytes. TNCs were initially reported as a unique multicellular complex that is positive for cTEC markers, including cytokeratin and MHC class II (Wekerle and Ketelsen 1980), and thereafter they were revealed to enclose the DP thymocytes (Li et al. 1992; Philp et al. 1993). Furthermore, Ezaki et al. found that 15–30 % of TNC isolated from rat thymus contained macrophages, and suggested their involvement in the clearance of apoptotic thymocytes (Ezaki et al. 1991), which is consistent with the present finding that CD11b+ macrophages were located adjacent to the FABP4+ TNCs in the thymic cortex. Although the distinct characteristic and/or function of TNCs are largely unknown, FABP4 can be a useful marker for TNCs and be involved in the control of the process of apoptotic cell clearance.

Several tissue-specific endothelial expression of FABP isoforms has been reported; FABP3 in endothelium of aorta and heart microvessels (Antohe et al. 1998), FABP4 in endothelium of spleen microvessels and arterioles (Abdelwahab et al. 2007), and FABP5 in tissue microvascular endothelial cells (Masouye et al. 1997). In this study, we showed that FABP4+ tubular structure was endothelial cells as revealed by its co-localization with FITC-conjugated lectin-positive staining. On the other hand, ER-TR7+ reticular cells which constitute perivascular space of intrathymic vasculatures were FABP4. Such spatial differences of FABP expression may indicate the distinct features of FA demands and/or homeostasis of given tissues, however the detail should be elucidated by further studies.

Recently it has been shown that FABP4 and FABP5 orchestrate the cellular lipid homeostasis in the adipocytes (Makowski and Hotamisligil 2005), thereby regulating the gene expression and/or signal transduction. Among the various FABPs, FABP4 and FABP5 are closely related to one another in their molecular structures and ligand-binding affinities (Hanhoff et al. 2002), while FABP4 exhibits a two- to threefold higher affinity for saturated and mono-saturated FAs than FABP5 (Simpson et al. 1999). In this study, FABP4 was localized exclusively to the TNC, a subpopulation of cTEC, in contrast to the wide distribution of FABP5 throughout cTEC. To explore the function of FABPs in TEC including TNC, we established two cell lines stably over-expressing FABP4 or FABP5 (TSt-4/DLL1-FABP4 or -FABP5, respectively) as an in vitro model system in this study. Comparative gene expression analysis using a DNA array and/or qRT-PCR revealed that the expression level of IL-7 and IL-18 were significantly increased by over-expression of FABP4 or FABP5. Furthermore, the increased gene expression of IL-7 in TSt-4/DLL1-FABP4 was significantly enhanced in response to saturated FA (stearic acid), but not to n-3 and n-6 PUFA. IL-7 is a pleiotropic cytokine that is produced by TEC, which has been shown as an essential cytokine for early differentiation of CD4CD8 (double negative; DN) thymocyte (Kim et al. 1998), for differentiation of DP thymocytes to CD8 single-positive (SP) thymocytes (Yu et al. 2003) and for intrathymic expansion of mature CD4 or CD8 SP thymocytes (Hare et al. 2000). Deletion or over-expression of IL-7 gene results in thymic atrophy through a reduction of the number of DP thymocytes caused by the inhibition of early thymocyte differentiation at the stages of DN1, DN2 and DN3 (El Kassar et al. 2004; Moore et al. 1996; von Freeden-Jeffry et al. 1995, 1997). Taken together, FABP4 may deliver stearic acid as a ligand for an intracellular receptor molecule in TNC, and regulate the T lymphocyte homeostasis by TNC though the control of IL-7 expression.


We cordially thank Dr. Willem V. Ewijk (Leiden University Medical Center, Leiden, The Netherlands) and Dr. Takashi Amagai for providing ER-TR4 antibody. We grateful thank to Dr. Hiroshi Kawamoto (RIKEN Research Center for Allergy and Immunology, Yokohama, Japan) for his useful comments and providing TSt-4 and TSt-4/DLL-1 cells, and Dr. Tsunetoshi Itoh (Tohoku University Graduate School of Medicine, Sendai, Japan) for providing IT-79MTNC cells. This work was supported by Grant-in-Aid for Young Scientists (B) to Y.A. (No. 19790152) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, from the New Frontier Project Research Fund of Yamaguchi University Graduate School of Medicine to Y.A. and from the Yamaguchi University Research Project on STRESS.

Supplementary material

418_2012_963_MOESM1_ESM.doc (41 kb)
Supplementary material 1 (DOC 41 kb)
418_2012_963_MOESM2_ESM.doc (31 kb)
Supplementary material 2 (DOC 31 kb)
418_2012_963_MOESM3_ESM.doc (30 kb)
Supplementary material 3 (DOC 29 kb)

Copyright information

© Springer-Verlag 2012