Current address: Department of Biochemistry, Pir Mehr Ali Shah Arid Agriculture University Rawalpindi, Rawalpindi 46300, Pakistan
Kynurenine pathway metabolism in human blood–brain–barrier cells: implications for immune tolerance & neurotoxicity
Article first published online: 21 JAN 2008
© 2008 University of New South Wales
Journal of Neurochemistry
Volume 105, Issue 4, pages 1346–1357, May 2008
How to Cite
Owe-Young, R., Webster, N. L., Mukhtar, M., Pomerantz, R. J., Smythe, G., Walker, D., Armati, P. J., Crowe, S. M. and Brew, B. J. (2008), Kynurenine pathway metabolism in human blood–brain–barrier cells: implications for immune tolerance & neurotoxicity. Journal of Neurochemistry, 105: 1346–1357. doi: 10.1111/j.1471-4159.2008.05241.x
- Issue published online: 21 JAN 2008
- Article first published online: 21 JAN 2008
- Received August 20, 2007; revised manuscript received January 8, 2008; accepted January 11, 2008.
- Blood–brain barrier;
- immune tolerance;
- quinolinic acid;
- tryptophan metabolism
The catabolic pathway of l-tryptophan (l-trp), known as the kynurenine pathway (KP), has been implicated in the pathogenesis of a wide range of brain diseases through its ability to lead to immune tolerance and neurotoxicity. As endothelial cells (ECs) and pericytes of the blood–brain–barrier (BBB) are among the first brain-associated cells that a blood-borne pathogen would encounter, we sought to determine their expression of the KP. Using RT-PCR and HPLC/GC-MS, we show that BBB ECs and pericytes constitutively express components of the KP. BBB ECs constitutively synthesized kynurenic acid, and after immune activation, kynurenine (KYN), which is secreted basolaterally. BBB pericytes produced small amounts of picolinic acid and after immune activation, KYN. These results have significant implications for the pathogenesis of inflammatory brain diseases in general, particularly human immunodeficiency virus (HIV)-related brain disease. Kynurenine pathway activation at the BBB results in local immune tolerance and neurotoxicity: the basolateral secretion of excess KYN can be further metabolized by perivascular macrophages and microglia with synthesis of quinolinic acid. The results point to a mechanism whereby a systemic inflammatory signal can be transduced across an intact BBB to cause local neurotoxicity.
3-hydroxyanthranilic acid, 2-amino-3-hydroxybenzoic acid
glyceraldehyde phosphate dehydrogenase
gas chromatography-mass spectrometry
human immunodeficiency virus
high molecular weight-melanoma-associated antigen
macrophage inhibitory protein
tumor necrosis factor-alpha
von Willebrand Factor
There is increasing evidence of the importance of the KP in disease pathogenesis. Its activation has been associated with a variety of inflammatory brain diseases, especially AIDS Dementia Complex (Heyes et al. 1991). It is involved in the killing of brain pathogenic organisms (Feng and Walker 2000; Daubener et al. 2001; Schroten et al. 2001), the development of immune tolerance (Grohmann et al. 2003; Moffett and Namboodiri 2003) and tumor evasion (Friberg et al. 2002; Munn and Mellor 2004) as well as the pathogenesis of cerebral malaria in a murine model (Hunt et al. 2006).
Significant activation of the KP occurs with the interferons (IFNs), especially IFN-γ (Takikawa et al. 1988), interleukin-1 (Hu et al. 1995), lipopolysaccharide (Takikawa et al. 1999), pokeweed mitogen (Saito et al. 1992), CTLA-4 (CD28) via ligation of CD80 (B7-1)/CD86 (B7-2) (Mellor et al. 2004; Munn et al. 2004), the pro-inflammatory cytokines, such as platelet-activating factor, tumor necrosis factor-alpha (TNF-α), and human immunodeficiency virus (HIV) regulatory proteins tat and nef (Smith et al. 2001).
There are three dominant products of the KP: quinolinic acid (QUIN), kynurenic acid (KA) and kynurenine (KYN). Not all cells, however, have the full complement of enzymes to produce these and other KP products. The KP is fully expressed in monocytic lineage cells, such as macrophages and microglia (Heyes et al. 1992a). In a previous study, we showed that astrocytes lack a critical intermediate enzyme, kynurenine monooxygenase (KMO, EC 22.214.171.124), resulting in their ability to produce KA and KYN but not QUIN (Guillemin et al. 2001). The expression of the KP in other cells, especially those of the CNS, is unknown.
Given the importance of the KP in inflammatory brain diseases and that blood–brain barrier endothelial cell (BBB EC) and pericytes are the first brain-associated cells that a cerebral blood-borne pathogen would encounter, we sought to determine their complement of KP enzymes by examining human primary cultured brain microvascular EC and pericytes.
Methods and materials
Primary-cultured human brain microvascular EC (BBB EC) and pericytes were purchased from Cell Systems, Inc. (Kirkland, WA, USA) as frozen cultures. The frozen cells were rapidly thawed in a 37°C waterbath and seeded onto culture surfaces (Falcon, Becton Dickinson, North Ryde, NSW, Australia) that were pre-coated with Cell Attachment Factor (Cell Systems, Inc.) for 2 h at 37°C.
Blood–brain–barrier EC were maintained in Clonetics (Cambrex BioScience, Sydney, Australia) EBM-2 basal medium supplemented with low serum EGM-2 Bullet Kit (Clonetics, San Diego, CA, USA). Cells were fed every 48 h until ∼ 80% confluence was reached, at which point they were passaged at 1 : 3 split ratio using EDTA, trypsin and trypsin inhibitor (Passage Reagent Group, Cell Systems, Inc.).
In some experiments, BBB EC (104/well) were cultured on collagen-coated polyester membrane transwell inserts (Cat. No. 3472, Corning, Inc., Lowell, MA, USA) and cultured to confluence (day 3). Transwells were coated with Bovine Collagen Type 1 (Cat no. 3442-100-01; Cultrex, Trevigen, Inc., Gaithersburg, MD, USA) at least 1 day prior to EC culture. Collagen (3 mg/mL in 0.1% acetic acid) was mixed 1 : 1 with absolute ethanol and 30 μL was added to each polyester membrane and left to air dry.
Pericytes were maintained in serum-free conditions using the CS-C serum-free kit (Cell Systems, Inc.), supplemented with 50 μg/mL gentamycin (Sigma–Aldrich, Castle Hill, NSW, Australia). Cells were fed every 48 h until ∼ 80% confluence was reached, when they were passaged at 1 : 5 split ratio using Passage Reagent Group as above.
All culture media were further supplemented with 2 mM (final concentration) l-glutamine (Invitrogen Australasia, Sydney, NSW, Australia). All cultures were maintained at 37°C in a humidified atmosphere of 5% CO2/95% air.
Immunofluorescence phenotyping of BBB cells
Cultures of cells were seeded at equivalent split ratio into appropriate attachment factor-coated, four-well glass chamber slides (Lab-Tek, Nalge Nunc, In Vitro Technologies, Sydney, Australia) and maintained as described above. At ∼ 80% confluence, cells were fixed for 8–10 min with 2% formaldehyde in phosphate-buffered saline (PBS) containing 1 mM Ca2+/Mg2+ (PBS + Ca/Mg). The cells were washed three times with PBS + Ca/Mg and permeabilized for 2 min on ice with 0.1% Nonidet P-40 (Sigma–Aldrich) in PBS + Ca/Mg (Mukhtar et al. 2002). After three washes in PBS + Ca/Mg, anti-human primary antibodies, diluted in antibody diluent (Pharmingen 559148, Becton Dickinson) were applied to the cells and incubated overnight at 4°C, after which the wells were washed three times with PBS + Ca/Mg. Secondary labeled antibodies, diluted 1 : 100 in antibody diluent, were incubated on the cells for 1 h at 20°C in a light-proof box. Negative control staining was performed on wells which had the primary antibody omitted. Primary antibodies, their specificities, sources and the dilution used are listed in Table 1.
|EC||vWF||Rabbit||polyclonal||1 : 60||Sigma-Aldrich|
|BBB EC||ZO-1||Mouse||IgG1||1 : 100||Zymed|
|BBB EC||GLUT-1||Rabbit||polyclonal||1 : 60||Chemicon|
|EC||CD146||Mouse||IgG1||1 : 100||Chemicon|
|Astrocytes||GFAP||Mouse||IgG1||1 : 100||Novocastra|
|Macrophages||CD68||Mouse||IgG1||1 : 100||Clone Ki-M1P|
|Neurons||MAP-2||Mouse||IgG1||1 : 200||Roche|
|Pericytes||HMW-MAA||Mouse||IgG1||1 : 10||Clone 225.28|
|Pericytes||PDGFBr||Mouse||IgG2B||1 : 100||Sigma|
Secondary FITC-labeled antibodies were goat anti-mouse IgG (H + L)-FITC and goat anti-rabbit IgG (H + L)-FITC (Southern Biotechnology Associates, Birmingham, AL, USA).
The cells were washed three times with PBS + Ca/Mg, then once in distilled water, and incubated in 2 μg/mL in water of 4′, 6-diamidino-2-phenylindole (D-9542, Sigma–Aldrich) for 5 min at 20°C to counter-stain cell nuclei. The cells were mounted in Fluoromount-G (Southern Biotechnology Associates) and visualized with an Olympus BX-60 microscope (Olympus Australia, Mt Wavetley, Vic., Australia) equipped with 60× oil immersion lens and illumination provided by a mercury vapor lamp providing excitation, through suitable filters, at 488 nm. Precision focusing was achieved by an electronically controlled microscope stage (Ludl Electronic Products, Hawthorne, NY, USA). Images were captured with a SensiCam CCD camera (PCO Imaging, Kelheim, Germany) interfaced to a Macintosh G4 computer (Apple Computer, Palo Alto, CA, USA); camera control was provided by IPLab v3.2 software (Scanalytics, BD Biosciences, Rockville, MD, USA). Post-capture image manipulation was performed in Adobe Photoshop v7.1 (Adobe Systems Pty. Ltd., Chatswood, NSW, Australia).
Mapping of the constitutive and inducible KP
Human cytokines IFN-γ and TNF-α, were obtained from R&D Systems (Bioscientific, Sydney, NSW, Australia) as sterile lyophilized powders. Stock solutions were made in sterile PBS containing 0.1% bovine serum albumin (Sigma–Aldrich) and aliquots were stored at −80°C. Cytokines were thawed and diluted to working concentrations in the appropriate culture medium immediately before use.
QUIN (2,3-pyridinedicarboxylic acid, Sigma–Aldrich) was dissolved in 4.8 μM NaOH to a stock concentration of 1 mM, sterile-filtered through 0.22 μm filters (Millipore-Waters, Lane Cove, NSW, Australia) and stored at 4°C. Working concentrations in culture medium were made just before use.
3-Hydroxyanthranilic acid (3HAA; 2-amino-3-hydroxybenzoic acid, Sigma–Aldrich) was dissolved in dimethyl sulphoxide (Sigma–Aldrich) to a stock concentration of 1 mM and stored at −80°C. Working concentrations were made in culture medium just before use and sterile-filtered through 0.22 μm.
To map the constitutive expression of KP enzymes, cells were seeded in triplicate wells of matrix-coated six-well plates at equivalent split ratio and fed every 48 h. At 80% confluence, they were re-fed with fresh medium and incubated for 24 h as described in Cell Culture.
To map the inducible expression of KP enzymes, parallel triplicate cultures of each cell type were established in six-well plates as above. At 80% confluence they were re-fed with fresh medium which included vehicle alone (unstimulated control), IFN-γ (100 IU/mL), TNF-α (100 IU/mL) and in some experiments IFN-γ + TNF-α (100 IU/mL each) and incubated for 24 h as described in Cell Culture.
At the end of the incubation period, the supernatant medium was aspirated from all wells and processed for HPLC and GC/M-S quantitation of KP metabolites as described below.
To assess the distribution of newly synthesized KYN between apical and basolateral surfaces of BBB EC, cells were cultured in triplicate on collagen-coated transwells for 3 days. After replacing culture media in the bottom well (600 μL) on day 3, BBB EC in the top well were cultured in 100 μL of EBM-2 media ± IFN-γ 100 IU/mL for 24 h. Culture media from the apical side (top well) and basolateral side (bottom well) was recovered for KYN assay by HPLC as described below.
After the supernatants were removed, to each well was added 1 mL/10 cm2 culture area RNA reagent Biotecx UltraSpec Total RNA isolation reagent (Fisher Scientific, Perth, WA, Australia) or Trizol (Invitrogen Australasia). The wells were incubated for 5 min on a rocking platform at 20°C to allow for solubilization of the cells. The cell lysates were dispensed into sterile, RNAse-free/DNAse-free microfuge tubes, and snap frozen in dry ice for storage at −80°C for later RNA isolation and cDNA synthesis.
To assess the downstream catabolism of the KP, cells were seeded in triplicate wells of matrix-coated six-well plates at equivalent split ratio and fed every 48 h. At 80% confluence they were re-fed with fresh medium which included an exogenous KP intermediate metabolite with or without cytokine stimulus:
3HAA (100 μM) or QUIN (1200 nM) without cytokine stimulation (control);
The cytokine treatments were as follows:
- 13HAA (100 μM) or QUIN (1200 nM) + IFN-γ 100 IU/mL
- 23HAA (100 μM) or QUIN (1200 nM) + TNF-α 100 IU/mL
- 33HAA (100 μM) or QUIN (1200 nM) + IFN-γ+TNF-α 100 IU/mL each
and incubated for 24 h, after which the supernatant medium was aspirated and processed for HPLC/GC-MS quantitation of KP metabolites. The cells were solubilized with RNA reagent as detailed above, for later RNA isolation and cDNA synthesis.
RNA isolation and cDNA synthesis for RT-PCR
Trizol-extracts of monocyte-derived macrophages (MdM) stimulated with IFN-γ were kindly supplied by Dr Gilles Guillemin, Centre for Immunology (Guillemin et al. 2001), for use as positive controls in KP enzyme RT-PCR, as the KP is fully expressed in these cells. The MdM were grown as previously described (Kerr et al. 1997).
The aliquots of frozen solubilized cells and MdM were thawed at 20°C, and RNA was isolated according to the manufacturer’s instructions.
The RNA pellets were dissolved in 50 μL Water For Injection (WFI, Baxter Healthcare, Old Toongabbie, NSW, Australia) and the concentration of a 1 : 50 dilution of the recovered RNA was quantitated by UV spectrophotometry at 260 nm; 2 μg total RNA from each sample was used for poly-A cDNA synthesis primed with 200 nM oligo-dT15 (Roche Diagnostics, Sydney, NSW, Australia), 0.5 U Avian Myoblastosis Virus reverse transcriptase and 2 U RNAsin RNAse inhibitor (both Promega Corp, Annandale, NSW, Australia) and 10 mM dinucleotide triphosphates (Roche Diagnostics) in a reaction volume of 50 μL. Reverse transcription was performed in a Techne (Cambridge, UK) Genius thermocycler, using one cycle of 60 min at 42°C for the reverse transcription reaction to proceed, followed by one cycle of 10 min at 65°C to heat-inactivate the reverse transcriptase enzyme, and then a final, untimed cycle at 4°C to keep the samples chilled until further storage at −20°C, preceding RT-PCR.
RT-PCR of KP enzymes in BBB cells
Primers, synthesized to HPLC purity by GeneWorks (Thebarton, SA, Australia) for each of the major KP enzymes indoleamine 2,3-dioxygenase (IDO, EC 126.96.36.199), kynurenine monooxygenase (KMO, EC 188.8.131.52), l-kynurenine hydrolase (KYNU, EC 184.108.40.206), 3-hydroxyanthranilate 3,4-dioxygenase (HAAO, EC 220.127.116.11) and quinolinate phosphoribosyltransferase (QPRTase, EC 18.104.22.168) were used in RT-PCR to map KP enzyme expression as previously described (Guillemin et al. 2001). Primers for reporter gene glyceraldehyde phosphate dehydrogenase (GAPDH, EC 22.214.171.124) were used as positive controls for the RT-PCR (Benveniste et al. 1996). The primer sequences are listed in Table 2.
|IDO||Forward Reverse||5′-ACCACAAGTCACAGCGCC-3′ 5′-CCCAGCAGGACGTCAAAG-3′|
|KMO||Forward Reverse||5′-GATGAGGAAGATAAGCTGAGGC-3′ 5′-CTTAAGGTTTCTTCCCCCTCTC-3′|
|KYNU||Forward Reverse||5′-GACTATTCCACCTAAGAACGGAGA-3′ 5′-ACAGGAAGACACAAACTAAGGTCG-3′|
|HAAO||Forward Reverse||5′-GAGGGCCTTCTTGGTGTG-3′ 5′-CATTGATTGGTTGTGAGTGG-3′|
|QPRTase||Forward Reverse||5′-GGTCACAGGAGCAGCAGG-3′ 5′-AAGCCAGAGGAGCTGCAC-3′|
|GAPDH||Forward Reverse||5′-ACCACCATGGAGAAGGCTGG-3′ 5′-CTCAGTGTAGCCCAGGATGC-3′|
Cycling conditions for KP enzyme RT-PCR were: 35 cycles at 92°C for 1 min, 60°C for 1 min, 72°C for 90 s. Cycling conditions for GAPDH were: 32 cycles at 95°c for 45 s, 60°C for 2 min, 72°C for 1 min.
RT-PCR was performed in a final volume of 100 μL, using 2 μL cDNA, 10 mM dNTPs (Roche Diagnostics), 20 μM each forward and reverse primers, 0.5 U Taq DNA polymerase (Roche Diagnostics) in a Techne Genius thermocycler. PCR products were run out on 1.5% agarose/Tris/borate/EDTA gels (Amresco, Solon, OH, USA) with HaeIII-digested ϕX-174 DNA molecular weight standards (Promega) and visualized on a UV lightbox. The gels were photographed with an Olympus C-2040z digital camera (Olympus Optical Co., Tokyo, Japan).
HPLC and GC/MS quantitation of KP metabolites
l-Trp and KYN were measured by RP-HPLC in culture supernatants obtained from stimulation experiments, after precipitating protein with an equal volume of 10% TCA. KA was measured using HPLC with a fluorescence detector (excitation, 344 nm; emission, 398 nm) after isocratic elution from a polymeric reversed phase column. The limit of detection of KA was 0.3 pmol (Nicholls et al. 2001).
QUIN and PIC were assayed in TCA-extracted culture supernatants by GC–MS in electron-capture negative ionization mode. The on-column limit of detection for QUIN and PIC was < 1 fmol (Smythe et al. 2002).
Results of KP metabolite quantitation were analyzed for significance using Student’s unpaired T-test, using Prism v.4 software (GraphPad Software, San Diego, CA, USA).
Recovery of frozen cells upon thawing appeared satisfactory: cultured ECs grew well, necessitating passaging at least once weekly. The pericytes grew much more slowly, being maintained in the absence of serum. All cells were easily passaged with trypsin and the viability from these procedures appeared to be > 90%. EC were not used for these experiments beyond passage 6, as expression of vWF/FVIII:RAg was rapidly lost beyond this passage.
Immunofluorescence phenotyping of BBB cells
Blood–brain barrier EC (Fig. 2) expressed vWF/FVIII:RAg, GLUT-1 and ZO-1. They did not express any of the other markers tested, thus showing that they expressed at least two BBB-specific markers ZO-1 and GLUT-1. This also demonstrated that these cultures were homogeneous, and free from other contaminating neural cells. Expression of vWF/FVIII:RAg by EC was maintained until the seventh or eighth passage, beyond which expression was rapidly lost; cells were routinely used up to passage 6.
Pericytes were weakly positive for HMW-MAA (not shown) but negative for all other antibodies tested. Expression of the platelet-derived growth factor-B receptor was not detected.
RT-PCR mapping of the constitutive and inducible KP
All cDNAs from quiescent and cytokine-stimulated EC (Fig. 3a lanes 1–4) amplified for the control gene GAPDH, except for the PCR negative control (lane 5-water alone). Lane 6 is MdM.
In control, unstimulated EC (lane 1, Fig. 3b) there was no expression of IDO or KMO. KYNU was constitutively expressed, and following IFN-γ, TNF-α or IFN-γ+TNF-α stimulation.
Endothelial cells stimulated with IFN-γ (lane 2) showed IDO expression, with weak KMO expression. TNF-α stimulation (lane 3) induced IDO and KMO expression, with KMO expression being weak. Stimulation with IFN-γ+TNF-α (lane 4) induced IDO and KMO expression. In all PCRs, the MdM (positive control) was expressed (lane 6). Expression of QPRTase and HAAO were equivocal in control (unstimulated) cells and cells stimulated with IFN-γ, TNF-α or IFN-γ+TNF-α (not shown).
In pericytes (Fig. 4), all cDNAs amplified for the control gene GAPDH. IDO expression was induced by both IFN-γ, TNF-α and IFN-γ+TNF-α (lanes 2–4). There was no expression of KMO, constitutively or after cytokine treatment. KYNU expression was constitutive with apparent induction after TNF-α (lanes 1–4). HAAO and QPRTase were constitutively expressed and in all treatment conditions.
Table 3a and b summarizes the RT-PCR results for the constitutive and cytokine-activated expression of the KP in these cells, respectively.
|pericytes||+ ( + TNF-α)||(+)||X||(+)||(+)|
HPLC and GC/MS quantitation of KP metabolites
Production of KP metabolites in control, unstimulated- and cytokine-stimulated EC is shown in Fig. 5(a). In the unstimulated BBB EC control (n = 3), constitutive production of KP metabolites was: PIC 1.13 ± 0.66 μM, QUIN 0.09 ± 0.07 μM, KYN 1.04 ± 1.81 μM and KA 10.91 ± 1.28 nM. After IFN-γ treatment, there was no up-regulation of PIC (0.44 ± 0.34 μM, p > 0.05), nor QUIN (0.08 ± 0.07 μM, p > 0.05). KYN was significantly up-regulated (10.90 ± 1.10 μM, p = 0.0013), indicating expression of IDO. IFN-γ significantly increased KA expression (15.35 ±1.28 nM, p = 0.0132).
After TNF-α treatment, there was no change in PIC (0.44 ± 0.44 μM, p > 0.05), QUIN (0.08 ± 0.05 μM, p > 0.05), KYN (1.04 ± 1.81 μM, p > 0.05) or KA (10.91 ± 1.28 nM), indicating that TNF-α did not stimulate the KP in these cells.
When BBB EC were cultured on transwells, there was no detectable KYN in the supernatants above (i.e., luminally) in either control or IFN-γ stimulated cells. There was, however, 102.1 ± 40.7 nM KYN in the supernatant below (i.e., basolaterally) in IFN-γ stimulated cells. No specific KYN synthesis was detected in the supernatant below control cells.
BBB EC + 3HAA
Production of KP metabolites in control, unstimulated- and cytokine-stimulated EC in the presence of the intermediate catabolite 3HAA is shown in Fig. 5(b). In the presence of 100 μM 3-HAA without stimulus (control) (n = 3) PIC was 0.27 ± 0.28 μM, QUIN was 0.91 ± 0.13 μΜ, KYN was 13.12 ± 1.92 μM and KA was 7.96 ± 1.28 nM.
After IFN-γ treatment, there was no change to PIC (0.27 ± 0.33 μM, p = 0.9756), QUIN (0.88 ± 0.10 μM, p = 0.7924) or KA (10.91 ± 2.26 nM, p = 0.1479). KYN was significantly up-regulated (23.11 ± 4.00 μM, p = 0.0175).
After TNF-α treatment, there was no change to PIC (0.25 ± 0.22 μM, p = 0.8939), QUIN (1.19 ± 0.19 μM, p = 0.1067) or KYN (12.01 μM, p = 0.3733). KA was significantly up-regulated (10.91 ± 1.28 μΜ, p =0.0474).
IFN-γ + TNF-α treatment was not done in this series.
BBB EC + QUIN
Production of KP metabolites in control, unstimulated- and cytokine-stimulated EC (n = 3) in the presence of the intermediate catabolite QUIN is shown in Fig. 5(c). In untreated control cells, PIC was 1.24 ± 0.18 μM, QUIN was 1.33 ± 0.14 μM, KYN was 1.04 ± 1.81 μM, KA was 13.13 ± 1.28 nM.
After IFN-γ treatment, there was no change to PIC (1.05 ± 0.39 μM, p = 0.4756), QUIN (1.31 ± 0.23 μM, p = 0.8930) or KA (13.13 ± 1.28 nM, p = 0.9999). KYN was significantly up-regulated (14.23 ± 2.94 μM, p =0.0027).
After TNF-α treatment, there was no change to PIC (1.36 ± 0.62 μM, p = 0.7761), QUIN (1.28 ± 0.08 μM, p = 0.6291) or KA (10.17 ± 2.56 μM, p = 0.1481). KYN was not detected.
IFN-γ + TNF-α treatment was not done in this series.
The functional expression of the KP in BBB EC is represented by the coloured enzyme names in Fig. 1.
In this series we noted the increased constitutive production of KYN by EC in the presence of the intermediate metabolite 3HAA (Fig. 5b), which is unprecedented; 3HAA is not a substrate for IDO, and levels of impurities in the pure compound as purchased from the manufacturer would not account for such production. We hypothesized that auto-oxidation products of 3HAA might serve as substrate for IDO. In a separate experiment (n = 3), we added superoxide dismutase (SOD) in the incubation medium (1, 10 & 100 U) to block basal induction of IDO by superoxide radicals however basal production of KYN was increased, shown in Fig. 6(a). In the control + 3HAA, KYN was 3.15 ± 0.08 μM; in the presence of 1U SOD, KYN was 5.80 ± 0.22 μM (p < 0.0001); in the presence of 10U SOD, KYN was 5.74 ± 0.52 μM; in the presence of 100U SOD, KYN was 5.35 ± 0.33 μM.
Production of KP metabolites in control, unstimulated- and cytokine-stimulated pericytes (n = 3) is shown in Fig. 7(a). In unstimulated (control) pericytes, constitutive production of KP metabolites was: PIC 1.38 ± 0.15 μM, QUIN 0.12 ± 0.09 μM, KYN was at baseline and KA was not detected under any condition.
With IFN-γ treatment, PIC was unchanged (1.54 ± 0.25 μM, p = 0.265), with QUIN and KYN at baseline.
With TNF-α treatment PIC was unchanged (1.65 ± 0.44 μM, p = 0.104), with QUIN and KYN at baseline.
With IFN-γ + TNF-α treatment PIC was 1.94 ± 0.20 μM (p = 0.018, just significant, compared to control). There was a large release of KYN (n = 2, mean = 30.14 nM).
Pericytes + 3HAA
Production of KP metabolites in control, unstimulated- and cytokine-stimulated pericytes (n = 3) in the presence of the intermediate catabolite 3HAA is shown in Fig. 7(b). In untreated control cells, PIC was 1.79 ± 0.48 μM, QUIN was not detected, KYN was 4.51 ± 4.1 μM, KA was not detected.
After IFN-γ treatment, there was no change to PIC (4.03 ± 0.36 μM, p = 0.3467), or KYN (4.03 ± 3.61 μM, p = 0.8864). Neither QUIN nor KA were detected.
After TNF-α treatment, there was no change to PIC (2.32 ± 0.46 μM, p = 0.2395) or KYN (4.30 ± 3.76 μM, p = 0.951). QUIN was 0.09 ± 0.03 μM. KA was not detected.
After IFN-γ + TNF-α treatment, PIC was unchanged (2.37 ± 0.43 μM, p = 0.1940). QUIN & KYN were not detected. KA was 17.57 ± 3.39 nM.
Pericytes + QUIN
Production of KP metabolites in control, unstimulated- and cytokine-stimulated pericytes (n = 3) in the presence of the intermediate catabolite QUIN is shown in Fig. 7(c). In untreated control cells, PIC was 2.14 ± 0.20 μM, QUIN was 1.68 ± 0.17 μM, KYN was 0.80 ± 0.69 μM, KA was not detected.
After IFN-γ treatment, there was no change to PIC (1.93 ± 0.11 μM, p = 0.1863), QUIN (1.62 ± 0.03 μΜ, p = 0.5796) or KYN (0.26 ± 0.46 μM, p = 0.3224). KA was not detected.
After TNF-α treatment, there was no change to PIC (2.14 ± 0.24 μM, p = 0.9999), QUIN (1.62 ± 0.08 μM, p = 0.6096) or KYN (0.86 ± 0.12 μM, p = 0.8892). KA was not detected.
After IFN-γ + TNF-α treatment, there was no change to PIC (1.94 ± 0.16 μM, p = 0.2476) or QUIN (1.44 ± 0.25 μM, p = 0.2411). KYN was 26.53 μM (because of insufficient sample, only one well of this triplicate was able to be analyzed), KA was 9.43 ± 2.22 nM.
The pericyte KP can be represented in Fig. 1, where the boxed distal portion is not functionally present.
Table 4a and b summarizes the constitutive and inducible KP metabolites in BBB EC and pericytes.
|PIC||X||+(IFNγ + TNFα)|
|KYN||+||+(IFNγ + TNFα)|
These results show that BBB EC constitutively synthesize KA, and that after IFN-γ (but not TNF-α) treatment they additionally synthesize kynurenine (KYN), which is secreted basolaterally, that is abluminally. In the presence of 3HAA, there were no increases in downstream metabolites PIC or QUIN, indicating that 3HAA was not a substrate for the KP in EC. Furthermore, there was no downstream metabolism of exogenous QUIN. These data show that the KP in BBB EC is functionally expressed to allow only KYN and KA synthesis.
The inclusion of SOD in the incubation medium paradoxically increased constitutive KYN production in the presence of 3HAA. However, 3HAA is known to undergo auto-oxidation (Goldstein et al. 2000) to produce hydrogen peroxide, an inhibitor of IDO (Poljak et al. 2006), and superoxide radicals (Dykens et al. 1987), which are substrates for IDO. SOD enhances the rate of auto-oxidation of 3HAA (Dykens et al. 1987; Ishii et al. 1990) by up to fourfold, probably by reducing back reactions between superoxide and either the anthranilyl radical or the quinoneimine formed during the initial step of auto-oxidation, thus providing more free superoxide for use as a substrate for IDO (Fig. 6b). SOD removes the superoxide radical, generated by the oxidation of 3HAA, thereby promoting the forward reaction; the superoxide radical is then used by IDO as substrate. Thus, the increase in basal KYN production in the presence of 3HAA intermediate arises as a result of its auto-oxidation supplying free radicals for use by IDO as a substrate, and inclusion of SOD increases the rate of 3HAA auto-oxidation.
To date, there are no similar published data on the expression of the KP beyond KYN in human EC. The biological significance of KYN production at the BBB is not known. The BBB is permeable to KYN via the large amino acid transporter and the possibility that peripherally produced KYN can be a substrate for further CNS KP metabolism has already been proposed (Stone 1993). The data presented here demonstrate that KYN produced at the BBB by cytokine stimulation is secreted abluminally, or possibly secreted luminally and taken up rapidly by large amino acid transporter to be passed through BBB, providing substrate for further processing by the KP of CNS-resident monocytic-lineage cells.
Animal data show that KA can be produced constitutively by EC in conduit vessels: rat aorta (Stazka et al. 2002) and bovine aorta (Wejksza et al. 2004), but its role in these vascular beds is unknown. The production of KA may be a constitutive and inducible protective response of the BBB to immune activation, as KA is an antagonist of NMDA receptors at the glycine subunit (Stone and Addae 2002). The significance of this is that activation of NMDA receptors can induce neurocytoxicity (Kerr et al. 1995, 1998). KA is also is a non-competitive inhibitor of α7 nicotinic acetylcholine pre-synaptic receptors (Nemeth et al. 2005). KA thus has anticonvulsant and neuroprotectant properties (Stone 2000). However, in a variety of inflammatory diseases such KA production is significantly less than QUIN that is produced by monocytic lineage cells (Heyes et al. 1992b). In addition, the absence of the QUIN-degrading enzyme QPRTase in EC would be expected to augment the potential toxicity of QUIN in the neurocellular microenvironment, where it may exceed the low QPRTase activity which we have already demonstrated in human astrocytes (Guillemin et al. 2001).
Pericytes constitutively produce PIC; with IFN-γ + TNF-α treatment, PIC was significantly increased, as was KYN. In the presence of either 3HAA or QUIN, constitutive PIC production was maintained, but both KYN and KA were produced in response to IFN-γ + TNF-α treatment. There was no metabolism of exogenous QUIN. These data indicate that, despite undetectable RT-PCR KMO expression, pericytes are able to synthesize PIC, a distal KP metabolite. Additionally, in comparison to BBB EC, pericyte IDO is differentially regulated by the combination of IFN-γ + TNF-α.
PIC is an unselective metal ion chelator (Aggett et al. 1989) and activates macrophages via induction of macrophage inhibitory protein (MIP)-1α and MIP-1β (Bosco et al. 2000), which is potentiated by simultaneous IFN-γ treatment (Pais and Appelberg 2000). It possesses both extracellular and intramacrophage anti-microbial activity against Mycobacterium avium (Cai et al. 2006; Shimizu and Tomioka 2006), Candida albicans (Abe et al. 2004) as well as antiviral/apoptotic activity against HIV-1- and Herpes simplex virus-2-infected cells (Fernandez-Pol et al. 2001). Further, it blocks the neurotoxic (but not the neuroexcitant) effects of quinolinic acid (Beninger et al. 1994; Kalisch et al. 1994), possibly by zinc chelation (Jhamandas et al. 1998). In the context of the present experiments, the constitutive pericyte synthesis of PIC may contribute to brain homeostasis by countering the neurotoxic effects of QUIN, until the constitutive synthesis is overwhelmed by de novo QUIN produced by inflammatory insults.
Thus, it is likely that at the BBB, in disease states where there is induction of IDO, there will be depletion of microenvironment l-Trp with resultant restriction of certain viral and bacterial pathogens (Adams et al. 2004; Adam et al. 2005), thereby preventing their ingress into the brain. In conjunction with the constitutive pericyte production of PIC, this may constitute an effective brain homeostatic microbicidal and antiviral mechanism which dampens the immune response that such pathogens would otherwise stimulate. This may limit immune activation to minimize neural damage, and may contribute to the brain’s immunologically privileged state. Furthermore, l-Trp depletion and production of other KP catabolites can lead to apoptosis of (mainly CD4 + ) T cells, B cells and NK cells (Frumento et al. 2002; Moffett and Namboodiri 2003), and induction of a regulatory T-cell phenotype via down-regulation of T-cell receptor ζ-chain (Fallarino et al. 2006).
Kynurenine, which can cross the intact BBB could be processed by adjacent perivascular macrophages to QUIN, further limiting the immune response by mediating apoptosis of T cells. EC KYN production thus may be a factor in limiting CNS immune cell activity through l-Trp depletion. KYN could also be processed by adjacent BBB astrocytes to KA, thereby limiting the potential neuronal toxicity of QUIN. This provides a unique mechanism whereby an inflammatory signal, chiefly in the form of IFN-γ, can be transduced across an otherwise intact BBB into the brain parenchyma.
The data herein show that the human BBB has limited expression of the KP, which can be activated by immune stimulation to produce intermediate metabolites which could be further processed by brain-resident monocytic-lineage cells to neurotoxic KP metabolites, in particular QUIN, which are relevant to AIDS Dementia Complex pathogenesis. Moreover, the KP-mediated immune paresis particularly affecting T cells at the BBB is likely to facilitate selection of macrophage tropic viral strains and contribute to persistence of viral infection in perivascular macrophages.
This work was supported by a NSW Department of Health Infrastructure grant, and a Program Grant from the Commonwealth National Health and Medical Research Council.
- 2004) Augmented inhibition of Candida albicans growth by murine neutrophils in the presence of a tryptophan metabolite, picolinic acid. J Infect Chemother 10, 181–184. , , , and (
- 2005) Role of human brain microvascular endothelial cells during central nervous system infection. Significance of indoleamine 2,3-dioxygenase in antimicrobial defence and immunoregulation. Thromb. Haemost. 94, 341–346. , , , , , and (
- 2004) Inhibition of human herpes simplex virus type 2 by interferon gamma and tumor necrosis factor alpha is mediated by indoleamine 2,3-dioxygenase. Microbes Infect 6, 806–812. , , , , and (
- 1989) An in vitro study of the effect of picolinic acid on metal translocation across lipid bilayers. J. Nutr. 119, 1432–1437. , and (
- 1994) Picolinic acid blocks the neurotoxic but not the neuroexcitant properties of quinolinic acid in the rat brain: evidence from turning behaviour and tyrosine hydroxylase immunohistochemistry. Neuroscience 61, 603–612. , , , and (
- 1996) Cytokine mRNA levels in unmanipulated (ex vivo) and in vitro stimulated monkey PBMCs using a semi-quantitative RT-PCR and high sensitivity fluorescence-based detection strategy. Cytokine 8, 32–41. , , , , and (
- 2000) The tryptophan catabolite picolinic acid selectively induces the chemokines macrophage inflammatory protein-1 alpha and -1 beta in macrophages. J. Immunol. 164, 3283–3291. , , , , and (
- 2006) Antimicrobial activity of picolinic acid against extracellular and intracellular Mycobacterium avium complex and its combined activity with clarithromycin, rifampicin and fluoroquinolones. J. Antimicrob. Chemother. 57, 85–93. , , , , , and (
- 2001) Restriction of Toxoplasma gondii growth in human brain microvascular endothelial cells by activation of indoleamine 2,3-dioxygenase. Infect. Immun. 69, 6527–6531. , , , , , and (
- 1987) Oxidative reactivity of the tryptophan metabolites 3-hydroxyanthranilate, cinnabarinate, quinolinate and picolinate. Biochem. Pharmacol. 36, 211–217. , and (
- 2006) The combined effects of tryptophan starvation and tryptophan catabolites down-regulate T cell receptor zeta-chain and induce a regulatory phenotype in naive T cells. J. Immunol. 176, 6752–6761. , , , et al. (
- 2000) Mechanisms of intracellular killing of Rickettsia conorii in infected human endothelial cells, hepatocytes, and macrophages. Infect. Immun. 68, 6729–6736. and (
- 2001) Antiviral, cytotoxic and apoptotic activities of picolinic acid on human immunodeficiency virus-1 and human herpes simplex virus-2 infected cells. Anticancer Res. 21, 3773–3776. , and (
- 2002) Indoleamine 2,3-dioxygenase contributes to tumor cell evasion of T cell-mediated rejection. Int. J. Cancer 101, 151–155. , , , , , , , and (
- 2002) Tryptophan-derived catabolites are responsible for inhibition of T and natural killer cell proliferation induced by indoleamine 2,3-dioxygenase. J. Exp. Med. 196, 459–468. , , , , and (
- 2000) 3-Hydroxykynurenine and 3-hydroxyanthranilic acid generate hydrogen peroxide and promote alpha-crystallin cross-linking by metal ion reduction. Biochemistry 39, 7266–7275. , , , et al. (
- 2003) Tolerance, DCs and tryptophan: much ado about IDO. Trends Immunol 24, 242–248. , and (
- 2001) Kynurenine pathway metabolism in human astrocytes: a paradox for neuronal protection. J. Neurochem. 78, 842–853. , , , , , , and (
- 1991) Quinolinic acid in cerebrospinal fluid and serum in HIV-1 infection: relationship to clinical and neurological status. Ann. Neurol. 29, 202–209. , , , , , , , , and (
- 1992) Human macrophages convert L-tryptophan into the neurotoxin quinolinic acid. Biochem. J. 283, 633–635. , and (
- 1992) Relationship of neurologic status in macaques infected with the simian immunodeficiency virus to cerebrospinal fluid quinolinic acid and kynurenic acid. Brain Res. 570, 237–250. , , , , , , and (
- 1995) Interleukin-1 enhances indoleamine 2,3-dioxygenase activity by increasing specific mRNA expression in human mononuclear phagocytes. J. Interferon Cytokine Res. 15, 617–624. , and (
- 2006) Immunopathogenesis of cerebral malaria. Int. J. Parasitol. 36, 569–582. , , , , , , , and (
- 1990) Superoxide dismutases enhance the rate of autoxidation of 3-hydroxyanthranilic acid. Arch. Biochem. Biophys. 276, 248–250. , , , and (
- 1998) Role of zinc in blockade of excitotoxic action of quinolinic acid by picolinic acid. Amino Acids 14, 257–261. , , and (
- 1994) Picolinic acid protects against quinolinic acid-induced depletion of NADPH diaphorase containing neurons in the rat striatum. Brain Res. 668, 1–8. , , and (
- 1995) Neurocytotoxity of quinolinic acid in human brain cultures. J. Neurovirol. 1, 375–380. , and (
- 1998) Chronic exposure of human neurons to quinolinic acid results in neuronal changes consistent with AIDS dementia complex. AIDS 12, 355–363. , , and (
- 1997) Kynurenine pathway inhibition reduces neurotoxicity of HIV-1-infected macrophages. Neurology 49, 1671–1681. , , , , and (
- 2004) Specific subsets of murine dendritic cells acquire potent T cell regulatory functions following CTLA4-mediated induction of indoleamine 2,3 dioxygenase. Int. Immunol. 16, 1391–1401. , , , , , , , , and (
- 2003) Tryptophan and the immune response. Immunol. Cell Biol. 81, 247–265. Direct Link: and (
- 2002) Primary isolated human brain microvascular endothelial cells express diverse HIV/SIV-associated chemokine coreceptors and DC-SIGN and L-SIGN. Virology 297, 78–88. , , , , , and (
- 2004) IDO and tolerance to tumors. Trends Mol Med 10, 15–18. and (
- 2004) Ligation of B7-1/B7-2 by human CD4+ T cells triggers indoleamine 2,3-dioxygenase activity in dendritic cells. J. Immunol. 172, 4100–4110. , and (
- 2005) Role of kynurenines in the central and peripheral nervous systems. Curr Neurovasc Res 2, 249–260. , and (
- 2001) Kynurenine production and catabolism in fetal sheep with embolized or nonembolized placentas. Am. J. Obstet. Gynecol. 185, 988–995. , , and (
- 2000) Macrophage control of mycobacterial growth induced by picolinic acid is dependent on host cell apoptosis. J. Immunol. 164, 389–397. and (
- 2006) Inhibition of indoleamine 2,3 dioxygenase activity by H2O2. Arch. Biochem. Biophys. 450, 9–19. , , , , , , , , , and (
- 1992) Effects of immune activation on quinolinic acid and neuroactive kynurenines in the mouse. Neuroscience 51, 25–39. , and (
- 1990) Expression of the high molecular weight melanoma-associated antigen by pericytes during angiogenesis in tumors and in healing wounds. Am. J. Pathol. 136, 1393–405. , , , and (
- 2001) Potential role of human brain microvascular endothelial cells in the pathogenesis of brain abscess: inhibition of Staphylococcus aureus by activation of indoleamine 2,3-dioxygenase. Neuropediatrics 32, 206–210. , , , , , and (
- 2006) Activity of picolinic acid in combination with the antiprotozoal drug quinacrine against Mycobacterium avium complex. Antimicrob. Agents Chemother. 50, 3186–3188. and (
- 2001) Quinolinic acid is produced by macrophages stimulated by platelet activating factor, Nef and Tat. J. Neurovirol. 7, 56–60. , , , , , and (
- 2002) Concurrent quantification of quinolinic, picolinic, and nicotinic acids using electron-capture negative-ion gas chromatography-mass spectrometry. Anal. Biochem. 301, 21–26. , , , , , and (
- 2002) Endothelium-dependent production and liberation of kynurenic acid by rat aortic rings exposed to L-kynurenine. Eur. J. Pharmacol. 448, 133–137. , , , and (
- 1993) Neuropharmacology of quinolinic and kynurenic acids. Pharmacol Rev 45, 309–379. (
- 2000) Development and therapeutic potential of kynurenic acid and kynurenine derivatives for neuroprotection. Trends Pharmacol Sci 21, 149–154. (
- 2002) The pharmacological manipulation of glutamate receptors and neuroprotection. Eur. J. Pharmacol. 447, 285–296. and (
- 2002) Endogenous kynurenines as targets for drug discovery and development. Nat Rev Drug Discov 1, 609–620. and (
- 1988) Mechanism of interferon-gamma action. Characterization of indoleamine 2,3-dioxygenase in cultured human cells induced by interferon-gamma and evaluation of the enzyme-mediated tryptophan degradation in its anticellular activity. J. Biol. Chem. 263, 2041–2048. , , and (
- 1999) Interferon-gamma-dependent/independent expression of indoleamine 2,3-dioxygenase. Studies with interferon-gamma-knockout mice. Adv. Exp. Med. Biol. 467, 553–557. , , , and (
- 1990) A new pan-macrophage antibody Ki-M1P stains plasmacytoid cells in paraffin sections of lymph nodes. Verh Dtsch Ges Pathol 74, 159–164. , , , and (
- 2004) Kynurenic acid production in cultured bovine aortic endothelial cells. Homocysteine is a potent inhibitor. Naunyn Schmiedebergs Arch Pharmacol 369, 300–304. , , , , , , , , and (