| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Gastroenterology Unit, Department of Clinical and Experimental Medicine, University Federico II, Naples, Italy;
Scientific Department, Sigma-Tau, Pomezia, Rome, Italy;
Institute of Genetics and Biophysics "A. Buzzati-Traverso", IGB-CNR and
Institute of Protein Biochemistry, IBP-CNR, Naples, Italy; and
|| Telethon Institute of Genetics and Medicine, Naples, Italy
1Correspondence: Institute of Protein Biochemistry, IBP-CNR, Via P.Castellino 111, Naples, Italy. E-mail: g.peluso{at}ibp.cnr.it
ABSTRACT
Carnitine transporters have recently been implicated in susceptibility to inflammatory bowel disease (IBD). Because carnitine is required for ß-oxidation, it was suggested that decreased carnitine transporters, and hence reduced carnitine uptake, could lead to impaired fatty acid oxidation in intestinal epithelial cells, and to cell injury. We investigated this issue by examining the expression of the carnitine transporters OCTN2 and ATB0+, and butyrate metabolism in colonocytes in a rat model of IBD induced by trinitrobenzene sulfonic acid (TNBS). We found that Octn2 and Atb0+ expression was decreased in inflammatory samples at translational and functional level. Butyrate oxidation, evaluated based on CO2 production and acetyl-coenzyme A synthesis, was deranged in colonocytes from TNBS-treated rats. Treatment with carnitine-loaded liposomes corrected the butyrate metabolic alterations in vitro and reduced the severity of colitis in vivo. These results suggest that carnitine depletion in colonocytes is associated with the inability of mitochondria to maintain normal butyrate ß-oxidation. Our data indicate that carnitine is a rate-limiting factor for the maintenance of physiological butyrate oxidation in colonic cells. This hypothesis could also explain the contradictory therapeutic efficacy of butyrate supplementation observed in clinical trials of IBD.—D’Argenio, G., Calvani, M., Casamassimi, A., Petillo, O., Margarucci, S., Rienzo, M., Peluso, I., Calvani, R., Ciccodicola, A., Caporaso, N., and Peluso, G. Experimental colitis: decreased Octn2 and Atb0+ expression in rat colonocytes induces carnitine depletion that is reversible by carnitine-loaded liposomes.
Key Words: butyrate • acetyl-CoA • mitochondria
THE ORGANIC CATION TRANSPORTER (OCTN) subfamily is involved in susceptibility to inflammatory bowel disease (IBD) (1)
and in carnitine transport. Because functionally important carnitine transporters are present in the bowel and because carnitine is crucial for ß-oxidation in bioenergetic metabolism (2)
, it has been suggested that a decrease in carnitine transporters, and hence reduced carnitine uptake, may lead to impaired fatty acid oxidation in intestinal epithelial cells and thus to cell injury (3)
. However, it is not known to what extent carnitine deficiency might predispose to IBD.
Carnitine (ß-hydroxy-
-trimethylaminobutyrate) is a prerequisite for lipid metabolism since it is a cofactor for the transport of acetyl groups and activated long-chain fatty acids into mitochondria for ß-oxidation and energy production (4)
. Carnitine, synthesized endogenously in humans in the liver and kidney (5)
, or absorbed in the intestinal tract from dietary sources (6)
, is delivered to tissues, such as skeletal and cardiac muscle, for which fatty acids are the main energy source. Although carnitine is not considered an essential nutrient in healthy adults, its biological importance for energy metabolism is illustrated by the clinical consequences of carnitine deficiency in a variety of genetic and acquired diseases (7)
that are characterized by defective membrane transport of carnitine into muscle, into renal tubular cells and/or into intestinal mucosa. Membrane transporters regulate the absorption, distribution, and elimination of carnitine, thereby maintaining carnitine homeostasis. Indeed, the cell-wide carnitine system consists of at least six proteins, three of which are encoded by the carnitine transporter subfamily (8)
. In the intact intestine, carnitine is absorbed mainly via an active mechanism (9)
that requires OCTN2 (encoded by the SLC22A5 gene), an organic cation/carnitine transporter of the OCTN family (10)
. OCTN2 functions as a Na+-dependent carnitine transporter and as a Na+-independent transporter for other organic cations. The low-affinity ATB0+ carnitine transporter, encoded by Slc6a14, is also involved in colon carnitine absorption (11)
. It has been postulated that carnitine absorption through ATB0+ prevails when OCTN2 is genetically or pharmacologically compromised and during therapeutic supplementation with oral doses of carnitine (11)
. Atb0+ is expressed in the colon and may therefore be important also for scavenging carnitine (although apparently not acetyl-carnitine) from the distal intestine, by virtue of its high concentrative capacity (11)
.
Although IBD is believed to result from genetic, metabolic, and immune factors, the finding that perturbation of carnitine transport is a predictive marker of IBD lends weight to the intriguing suggestion that the prime mover in ulcerative colitis might be the inability of colonocytes to oxidize fatty acid substrate adequately (12)
. However, this hypothesis has attracted little interest because there is no evidence that carnitine transporters are decreased in colonocytes in IBD. In addition, colonocytes preferentially use butyrate as respiratory fuel, butyrate being a short-chain fatty acid (SCFA) that does not require carnitine for its import into the mitochondrial matrix (13)
. Although the profile of circulating plasma carnitine has recently been reported to be altered in ulcerative colitis (14)
, a causal role for carnitine in IBD has yet to be demonstrated.
The aim of this study was to investigate whether perturbed butyrate metabolism in colonocytes from rats treated with the contact-sensitizing agent trinitrobenzene sulfonic acid (TNBS) was related to impaired expression of carnitine transporters and/or to a decreased carnitine content in colonocytes.
MATERIALS AND METHODS
Animals
Specific pathogen-free male Wistar rats weighing 200 ± 20 g were obtained from Charles River (Milan, Italy), housed in a clean, temperature-controlled environment with a 12-h light–dark cycle, and given free access to a regular laboratory chow diet and water for several days. All animals received humane care, and the study protocols were approved by the ethics committee of the Naples University Federico II School of Medicine.
Experimental colitis
To ensure a homogeneous distribution of lesions, chronic colitis was induced by 2, 4, 6-TNBS with a modified version of the method described by Morris (15)
. Briefly, rats (10 per group) were fasted the day before colitis induction and the colon was rinsed with tepid water enemas (
3 ml) 3 h before TNBS instillation. Each rat was then anesthetized with ether, and TNBS (12% TNBS in 50% ethanol; total vol, 0.25 ml) was instilled via a rubber catheter inserted 8 cm into the colon via the anus. TNBS was instilled through holes in the apical half of the catheter. The instillation procedure took only a few seconds. Intracolonic administration of TNBS resulted in long-lasting inflammation of the distal colon characterized by marked thickening of the colonic wall, infiltration of polymorphonuclear leukocytes, and granuloma formation.
Histological examination and colonocyte isolation
Two weeks after induction of colitis, rats were anesthetized with ether and exsanguinated, and the entire colon, from the ceco-colonic junction to the distal anal canal, was removed. The colon was flushed clear of luminal content with isotonic sodium chloride, opened longitudinally, weighed, and evaluated macroscopically according to the criteria described by Morris et al. (15)
(Table 1
). Several small biopsies were sampled from the macroscopically damaged area and from the intact mucosa. The specimens were fixed in 10% formalin and embedded in paraffin. The sections were analyzed by the pathologist, blind to the treatment of the animals, in random order. As in previous studies (16
, 17)
, four parameters, each scored on a 0–5 scale of severity, were considered, namely, extent of ulceration, submucosal infiltration, crypt abscesses, and wall thickness. The final total histological score (from 0 to 20) was determined by calculating the sum of the four individual scores, to give a rating of slight (1
2
3
4
5)
moderate (6
7
8
9
10)
or severe (11
12
13
14
15
16
17
18
19
20)
colonic inflammation (Table 2
).
|
|
Colonocytes were isolated with the procedure described by Roediger and Truelove (18)
, modified by Jorgensen et al. (19)
. Lactate dehydrogenase (LDH; EC 1.1.1.27) release was used to evaluate membrane integrity (20)
. Leakage of LDH into the medium at 90 min was <10%. Contaminating lymphocytes/macrophages were removed as reported elsewhere (21)
, by negative immunomagnetic bead selection using a mixture of the monoclonal antibodies anti-CD3 (Antigenix America Inc., NY, USA) and anti-ED-2 coated with magnetic beads (Biosciences Pharmigen, Milan, Italy).
Carnitine uptake
The cells and transport medium (125 mM NaCl, 4.8 mM KCl, 5.5 mM D-glucose, 1.2 mM CaCl2, 1.2 mM KH2PO4, 1.2 mM MgSO4, and 25 mM HEPES, pH 7.4) containing radiolabeled carnitine (0.5 µM, 0.5 mCi/ml) were preincubated separately and then mixed to start the carnitine uptake reaction. At appropriate times, the transport reaction was stopped by rapidly washing the cells four times with ice-cold 0.1 M MgCl2. The resultant pellets were solubilized in 3 N KOH and neutralized with HCl, and the associated radioactivity was measured in a liquid scintillation counter. Cellular protein content was determined with a protein assay kit (Bio-Rad, Milan, Italy) and BSA as the standard. When Na+-free conditions were required, the isolated cells were suspended in Na+-free medium, in which Na+ was replaced isotonically with N-methyl-D-glucamine or lithium. To verify the specificity of carnitine transporter(s) and to distinguish between OCTN2 and ATB0+ transporters, molecular competitors, i.e., unlabeled carnitine, acetyl-carnitine, tetraethylammonium (TEA) or glycine, were added to the reaction mixture. Data were analyzed according to Rytting and Audus (22)
.
Total RNA extraction, reverse transcription, sequencing, and semiquantitative RT-PCR
Total RNA was isolated from isolated colonocytes with TrizolTM (Invitrogen), and reverse-transcribed with Superscript III (Invitrogen) using 1 µg total RNA and 100 ng random primers. To detect rat Octn2 (Slc22a5) by PCR, two primers located in the coding sequence of rat Octn2 (accession number NM_019269) were constructed. Oligonucleotide rOCTN2s (5'-TCCTGTGGCTGACCATATCA-3') corresponded to bases 1146–1165, oligonucleotide rOCTN2a (5'-CACCAAAGCTCTCTGGGAAG-3') corresponded to bases 1624–1643 of the rat Octn2 cDNA sequence. We used a glyceraldehyde-3-phosphate dehydrogenase (Gapd)-specific fragment to verify the integrity of the RNA preparation, (accession number NM_017008) with oligonucleotides GAPDH3 (5'-AACTACATGGTCTACATGTT-3') and GAPDH4 (5'-GTGGTGCAGGATGCATTGCT-3') corresponding to bases 145–164 and 466–485, respectively, of the rat Gapd sequence. We used 1 µl of the reverse-transcriptase reaction as a template to amplify specific fragments in a semiquantitative PCR reaction (25, 30, and 35 cycles: 1 min at 95°C, 1 min at 55°C, and 1 min at 72°C) using 2.5 U AmpliTaq-Gold (Applied Byosistems) and 100 pmol of each primer.
To detect rat Atb0+, two primers for RT-PCR were constructed that corresponded to bases 750–769 (5'-ACTTTGCCTTCTTCTAGCTT-3'; upper primer) and to bases 1282–1301 (5'-TGAGAATCCAGACCCAATGT-3'; lower primer) of the predicted mRNA (see Results). Semiquantitative RT-PCR was performed under the conditions described above. The intensity of the amplified bands was quantified by densitometry and referred to that obtained with Gapd (Quantity One, Bio-Rad). The oligonucleotides used to detect the 469-bp fragment of the rat Abcd1 transcript were: 5'-GGGCCTAAAGCAACAGTCTCA-3' and 5'-GGGCAACATACACAGACAGGAA-3'. PCR fragments were directly sequenced with an ABI Prism Dye Terminator sequencing kit and analyzed on an ABI PRISM automated sequencer (Applied Biosystems, Milan, Italy).
Real-time PCR
Real-time quantitative PCR was carried out on an ABI PRISM 7700 sequence detector (Applied Biosystems) to determine the relative expression levels of the two carnitine carriers in rat colon tissues. Forward and reverse primer sequences for the gene targets were: OCTN2-forward primer 5'-3': CGC CTT CCA CTA TCT TCG AT; OCTN2-reverse primer 5'-3': TGA TAT TTC GTG TTC GGA CC, ATB0+-forward primer 5'-3': ATT GGG ATA AAG TGA CGC; and ATB0+-reverse primer 5'-3': AGA AGA AGG CAA AGT GCT AAG. We used the following primers for the internal control cyclophillin A (CypA): forward primer 5' TAT CTG CAC TGC CAA GAC TGA GTG 3', reverse primer 5' CTT CTT GCT GGT CTT GCC ATT CC 3'. cDNA corresponding to 25 ng of reverse transcribed total RNA was amplified in a 25 µl vol reaction using the TaqMan universal PCR Mastermix (Applied Biosystems) in triplicate assays for the Octn2 and Atb0+ targets and the endogenous control CypA. Primers (Promega, Milan, Italy) were used at a concentration of 300 nM each, and the FAM/TAMRA fluorophore/quencher reporter probe (Eurogentec, Seraing, Belgium) at a concentration of 100 nM. Thermocycling conditions were as for standard TaqMan protocol. The primers and probes were validated for amplification efficiency by running the reaction at five different template dilutions and subtracting the results obtained with the control CypA from those obtained with the target. Amplification efficiency was found to be the same for all primer pairs. The results were analyzed with the C
method, where C is the cycle number at which the fluorescence of the sample crosses a given threshold. The expression levels of the full-length carnitine transporters in the liver for Octn2 and in the lung for Atb0+ were arbitrarily taken as the calibrators for all calculations, and colon tissues were normalized accordingly. Results are reported as expression of each transcript relative to the full-length liver or lung gene ± SD for each tissue of three animals.
In situ hybridization
To generate riboprobes, PCR fragments derived from amplification with oligonucleotides containing the same sequences used for semiquantitative RT-PCR of the Atb0+ and Octn2 cDNAs, as described above, and sequences recognized by T3 and T7 RNA polymerases were in vitro transcribed by standard methods with the Roche DIG RNA Labeling kit (Roche Diagnostics, Cat. No. 11175025910, Milan, Italy). Paraffin-embedded tissue sections, 10 µm thick, were floated onto silanated slides, which were then heated for 4 h at 60°C. Sections were deparaffinized by two washes (10 min each) in xylene and two washes (10 min each) in 100% ethanol, and then air dried. Slides were rehydratated through an ethanol series, fixed in 4% paraformadehyde in PBS, and acetylated in acetic anhydride in TEA buffer. After acetylation, slides were treated with 0.6% hydrogen peroxide in methanol, extensively washed with PBS, and treated with proteinase K solution (proteinase K, 50 µg/ml in 0.2 M Tris-HCl, pH 7.5; 2 mM MgCl2) for 20 min at room temperature. Prehybridization and hybridization were performed at 64°C in Ambion in situ hybridization buffer (Cat. No. B8807G). Posthybridization washes were carried out at 62°C in 5x saline-sodium citrate (SSC), 2x SSC 50% formamide, 1x SSC 50% formamide. Probes were detected with the Anti-Digoxigenin-activating protein, FAB fragments antibody (Ab) from Roche (Cat. No. 11093274910).
Butyrate metabolism assay
Cell suspensions (3 mg protein/ml) were placed in siliconized 25 ml Erlenmeyer flasks. Each flask contained the cell suspension, one or a mixture of metabolic substrates, a 14C-labeled butyrate (Amersham Bioscience, Milan, Italy) as substrate and KRB buffer (pH 7.4) with 5 g/L BSA (fraction V, essentially fatty acid free; Sigma, Milan, Italy) in a vol of 2 ml. All incubation media contained 5.6 mmol/l glucose (Glc), which represents physiological blood levels. Various amounts of butyrate were added to the medium to test for a concentration effect. The flasks were gassed with O2/ CO2 (95%:5%) and sealed by a rubber double-seal stopper with a center well. The flasks were incubated at 37°C in a water bath with shaking at 50 cycles/min for 60 min. The incubation was terminated by injection of 1.5 mol/L HClO4 into the incubation flask. Organic bases NCSII (Amersham Life Science, Milan, Italy) were injected into the hanging center well to trap CO2. After another 60 min of incubation, the center well with the alkali solution was transferred into a scintillation vial with 15 ml of a modified Bray’s solution (23)
. The medium in the incubation flask was neutralized with K2CO3 and centrifuged at 1000 g for 20 min. The supernatant was stored at –80°C. The bacterial contribution to metabolite formation was minimized by isolating and incubating cells with penicillin 1.8 x 105 U/l (Sigma) and streptomycin 180 g/l (Sigma). 14CO2 was quantified using a liquid scintillation counter (model LS-3801, Beckman Instruments, Naples, Italy). The background 14CO2 in flasks containing no cells represented the action of HClO4 on 14C-labeled substrate. Generation of 14CO2 (nmol) from 14C-labeled substrate was calculated by specific radioactivity (dpm/nmol) of 14C-labeled substrate in the incubation medium after the background had been subtracted. Data are expressed as nanomoles of 14CO2 produced per milligram protein per hour. The concentration of protein in the cell suspensions was assayed as described above.
Acetyl-coenzyme A assay in whole cell lysates and in mitochondrial fractions
The acetyl-coenzyme A content was measured in freshly isolated colonocytes that were incubated in medium with or without butyrate (see above). After incubation, the cells were centrifuged, the pellet was treated with 5 mmol/l HCl and then placed for 1 min in a boiling bath. After centrifugation, the acetyl-coenzyme A content was measured in clear supernatants with a cycling assay (24)
. To determine the intracellular distribution of acetyl-coenzyme A, the cells were broken by means of a nitrogen cavitation method using a Parr bomb (25)
, and pelleted by centrifuging at 750 g for 15 min. The supernatant was further centrifuged at 1 x 105 g for 1 h at 4°C using Beckman L8–55 M ultracentrifuge. The pellet was suspended in 3.0 ml of imidazole buffer (3 mM imidazole, 0.25 M sucrose; pH 7.4) and then fractionated into the plasma membrane/nucleus, cytosol, endosome/lysosome, and mitochondria. To check the reliability of the separation procedure, glutamate dehydrogenase and lactate dehydrogenase activities were determined in each fraction. We verified the results of the experiments by measuring cytosolic and mitochondrial acetyl-coenzyme A/coenzyme A with another procedure (26)
. The results of the two methods were essentially the same.
Preparation of carnitine-loaded liposomes
Liposomes were obtained and characterized as described previously (27)
. Briefly, a liposome drug formulation was prepared by simply injecting a lipid organic solution into a drug-containing aqueous solution (28)
. The liposome size distribution was determined by dynamic light scattering as a function of lipid concentration and composition. The injection of an ethanol solution of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC; Chemi S.p.A., Patrica, Frosinone, Italy) in pure water resulted in a monomodal size distribution of rather small vesicles (mean, 65 nm). When a POPC ethanol solution was injected in a carnitine and water solution, spontaneously formed liposomes were generally 30% larger.
Treatment of colonocytes with carnitine-loaded liposomes
The targeting efficiency of liposomes was determined by in vitro incubation of cells with carnitine-loaded liposome formulations. Briefly, carnitine (1 mM) in either the free form or the liposomally encapsulated form was added to each sample. Cells were incubated with the carnitine mixture at 37°C in a 5% CO2 environment. After incubation, cells were placed on ice, washed three times with 1 ml of ice-cold PBS containing calcium and magnesium to remove extracellular carnitine. Cells were then lysed with 2 ml of 5% Triton X-100, and the carnitine content of the lysed cell solution was measured (see "HPLC and mass spectrometry" section below). Aliquots of cell suspension were used for the butyrate metabolism assay.
Intracellular localization of carnitine
Liposomes loaded with 14C-labeled carnitine (
4x105 c.p.m./mg cell protein; Perkin Elmer, Milan, Italy) were used in these experiments. Colonocytes were incubated for 1 h at 37°C with lipid vesicles and then washed with NaCl/Pi followed by fixation with 1% paraformaldehyde solution. The intracellular distribution of labeled carnitine was determined with the procedure described for acetyl-coenzyme A (see above).
HPLC and mass spectrometry
Total carnitine content was measured using standard HPLC methods. Total, free, and acyl-carnitine fractions were determined by tandem mass spectrometry as described (29)
.
Treatment of colitis with carnitine-loaded liposomes
This experiment was designed to determine the effects of carnitine on histological activity in TNBS-induced experimental colitis. Thirty rats received an intracolonic injection of TNBS and were divided into three groups of 10. The first group was treated twice a day for 2 wk with 10 mmol/L carnitine-loaded liposomes administered intraluminally, the second group was treated with empty liposomes, and the third was treated with empty liposomes and carnitine admixed. Treatments were administered via a rubber catheter (diameter, 1.5 mm) inserted 8 cm into the colon via the anus; it was not necessary to anesthetize the animals. Ten rats that were not manipulated constituted the control group. At the end of treatments, rats were killed and the last 8 cm of the colon was stripped, gently washed with cold saline, longitudinally opened, and evaluated macroscopically. A central piece (0.5 cm) was used for histopathological evaluation.
Statistical analysis
Data distribution was evaluated with the Kolmogorov-Smirnov test. Differences between two experimental groups were tested by the Student’s t test for unpaired data. The Wilcoxon test was used for discontinuous data. Differences between means were considered significant at P < 0.05.
RESULTS
Macroscopic appearance and histology
Fourteen days after TNBS administration, the colon showed gross signs of inflammation, i.e., wall thickening, increased vascularization, increased wet wt, and ulcerations (data not shown). Histologically, TNBS-treated rats had typical IBD lesions, namely areas of ulcerations characterized by prominent inflammatory infiltrates at the deep and lateral margins of the lesions. Lymphocyte infiltrates were present in the crypt epithelium. There were several crypt abscesses and evidence of granulomas (data not shown).
Characteristics of isolated colonocytes
Cells isolated from control and colitic tissue were always more than 90% viable. The viability of control cells was not significantly decreased after 90 min of incubation with 5 mM butyrate as shown by trypan blue dye exclusion and confirmed by LDH assay. The two procedures showed that cells isolated from TNBS-treated rat tissue were 87% viable; this value remained unchanged after incubation with 5 mM butyrate for 90 min. Because 75% of cells of TNBS-treated rats were colonocytes, and the rest were inflammatory cells, we carried out negative immunomagnetic bead selection using a mixture of anti-CD3 and anti-ED-2 monoclonal antibodies. The purity of colonocytes from treated rats was similar to that of control rats (>90%).
Carnitine uptake in isolated colonocytes
In many tissues, carnitine transport is sodium-dependent, a characteristic mainly ascribed to carnitine transport by OCTN2 (30
, 31)
and ATB0+ (11)
. We examined the Na-dependence of [3H]-carnitine uptake in colonocytes from control rats and from TNBS-treated animals. As expected, control colonocytes accumulated a large amount of carnitine, but carnitine uptake was significantly decreased when sodium was replaced by N-methylglucamine (Fig. 1
).
|
We evaluated the substrate specificity of carnitine transport in colonocytes in terms of inhibition of [3H]-carnitine uptake by unlabeled carnitine, acetyl-carnitine, and tetraethylammonium (TEA). Each of these compounds has been shown to inhibit Na-dependent carnitine transport in several cell types including enterocytes (32)
. We also determined whether the Atb0+ transporter is involved in carnitine uptake by isolated colonocytes, by examining the effect of glycine, which interacts with the Atb0+ transporter (11)
. As shown in Fig. 2
, all substrates tested significantly inhibited [3H]-carnitine uptake in colonocytes, although the inhibition exerted by the organic cation TEA was not as strong as that of carnitine and acetyl-carnitine and occurred only at high concentrations. The IC50 values are in the micromolar range for carnitine and acetyl-carnitine and in the millimolar range for the other two compounds. Glycine inhibited carnitine accumulation in colonocytes, but to a much lesser extent vs. the other substrates, and only at very high concentrations (Fig. 2)
. Interestingly, Na-dependent carnitine uptake was significantly lower in colonocytes from TNBS-treated rats than in control colonocytes (Fig. 1)
.
|
Analysis of Octn2 and Atb0+ expression
To gain insight into the mechanism by which colon inflammation prevents accumulation of carnitine in colonocytes, we measured the expression of two transporter genes, Octn2 (Slc22a5) and Atb0+ (Slc6a14), in rat colon samples by semiquantitative RT-PCR. Because the rat Atb0+ coding sequence has not been elucidated, we used a mouse mRNA sequence (Genebank accession AF320226) to obtain the rat homologue. We searched for the genomic rat sequence using the genome browser (http://genome.ucsc.edu, May 2003) and identified a predicted transcript of 1923 bp. We used this in silico mRNA to construct primers to verify the full transcript sequence on rat colon cDNA (the sequence obtained was deposited in the EMBL databank under accession number AJ969954) (Fig. 3
A). The predicted protein sequence showed an identity of 86% and 81% with the homologous mouse and human proteins, respectively (Fig. 3B
).
|
Fourteen days after colitis induction, Octn2 mRNA in colonic tissue was decreased nearly 5-fold, whereas Atb0+ expression was undetectable. Gapd mRNA level served as the internal standard control for densitometric quantification of the products obtained with 30 cycles of amplification (Fig. 4
A). Similar results were obtained with real-time PCR using the same RNAs (Fig. 4B, C
). In these experiments, we standardized the expression level of each target gene for each sample by subtracting the internal CypA standard, and we assigned a value of 1 to the expression levels of the Octn2 liver and Atb0+ lung transcripts (see Materials and Methods). To verify amplification efficiency, we performed serial template dilutions for each primer pair for each target gene. The quantitative PCR demonstrated that Octn2 was highly expressed in the small and large intestine of control rats, whereas it was significantly reduced in TNBS-treated rats (Fig. 4B
). The expression of the Atb0+ transcript was more restricted than that of the Octn2 transcript. Real-time RT-PCR analysis showed that Atb0+ was expressed in control colonocytes but not in the duodenum, jejunum, or ileum of the same animals. In addition, Atb0+ expression barely exceeded background levels in colonocytes from TNBS-treated rats (Fig. 4C
).
|
To verify that the effect of TNBS on carnitine transporters in colonocytes is specific and not a generalized defect in membrane transporters due to mucosal injury and inflammation, we measured Abcd1 expression in colonocytes from control and TNBS-treated rats. The product of this gene is involved mainly in the membrane transport of fatty acids and, as in the case of Atb0+, it is located on chromosome X and is highly expressed in colonocytes (Genbank ID 11666, primary source: MGI1349215). There was no difference in Abcd1 expression between colonocytes from control animals and those from TNBS-treated animals (Fig. 4A
). We next verified the decrease of Atb0+ and Octn2 expression directly in tissue specimens in an in situ hybridization experiment. The expression of both genes was greatly decreased in the epithelial cells that survived TNBS treatment (Fig. 5
).
|
Analysis of carnitine and carnitine derivatives in isolated colonocytes
As shown in Table 3
, carnitine was present in control colonocytes mainly in its free form (60%), with acetyl and propionyl carnitine accounting for the remaining 40%. Only traces of long-chain acyl-carnitine were detected. Carnitine and its derivatives, particularly acetyl-carnitine, were dramatically decreased in colonocytes from TNBS-treated rats. Although the free form and the short-chain acyl-carnitines predominated, there was only a slight increase in carnitine acylated in the form of long-chain acyl derivatives. The addition of carnitine significantly increased the level of carnitine in control cells (Table 3)
. The addition of carnitine to colonocytes from TNBS-treated rats did not increase carnitine accumulation to the level found in control colonocytes. Only after incubation of colonocytes with carnitine-loaded liposomes was there a dramatic increase in carnitine level.
|
CO2 production from butyrate in colonocytes
To investigate the effect of carnitine on butyrate metabolism, we measured the rate of 14CO2 production from 14C-labeled butyrate in colonocytes from control and from TNBS-treated rats. These experiments were carried out with cell suspensions from the same batch of colonocytes. We first examined butyrate uptake in all colonocyte cultures, and found that TNBS did not affect the uptake of butyrate by colonocytes (Fig. 6
). Nonisotopic butyrate reduced 14C butyrate uptake by 95% (namely, to 0.41±0.1 nmol/g protein). However, the production of 14CO2 from 1-14C-labeled butyrate in isolated colonocytes decreased significantly in samples from TNBS-treated animals vs. controls.
|
Before studying the effect of carnitine-loaded liposomes on butyrate metabolism, we determined the intracellular distribution of carnitine cells after exposure to carnitine-loaded liposomes. Approximately 38% and 17% of all cell-associated 14C-labeled carnitine was localized in the cytosolic and mitochondrial portions, respectively, of cells incubated with liposomes loaded with 14C-labeled carnitine (Fig. 6)
. The addition of carnitine-loaded liposomes restored the oxidation of butyrate in colonocytes from TNBS-treated rats. Empty liposomes, with or without carnitine admixed, did not affect butyrate metabolism in these cells. The addition of free or carnitine-loaded liposomes to control cells did not increase butyrate metabolism.
Effect of carnitine on acetyl-coenzyme A distribution in butyrate-treated colonocytes
To determine whether butyrate ± carnitine affects the intracellular distribution of acetyl-coenzyme A, we measured the level of acetyl-coenzyme A in the various compartments of colonocytes incubated with butyrate in the presence or absence of carnitine-loaded liposomes. In control cells, butyrate did not significantly affect mitochondrial acetyl-coenzyme A level, whereas it caused a 3-fold increase of cytosolic acetyl-coenzyme A (Fig. 7
A). The addition of carnitine to the cell suspension did not affect either total cellular acetyl-coenzyme A content or its cell distribution (Fig. 7B
). On the contrary, mitochondrial acetyl-coenzyme A content was significantly higher in colonocytes from TNBS-treated rats vs. control colonocytes (Fig. 7A
). The addition of butyrate did not affect either mitochondrial or cytosolic acetyl-coenzyme A level (Fig. 7A
). The addition of carnitine-loaded liposomes to these cells induced a dramatic increase in cytosolic acetyl-coenzyme A and a concomitant decrease in mitochondrial acetyl-coenzyme A (Fig. 7B
). Interestingly, the acetyl-coenzyme A level in whole cells was significantly higher in TNBS-treated colonocytes without butyrate than in control colonocytes because of the increased acetyl-coenzyme A content at mitochondrial level (Fig. 7A
).
|
Effect of carnitine-loaded liposomes on TNBS-induced colitis
To evaluate the in vivo effects of carnitine on colitis, we administered carnitine-loaded liposomes intracolonically to rats for two weeks beginning the day after TNBS administration vs. rats given empty liposomes or empty liposomes with carnitine admixed, and control rats. On completion of treatment, the total carnitine content in the colon mucosa of rats treated with carnitine-loaded liposomes was higher than in untreated rats and in animals treated with empty liposomes (Table 4
). The results of treatment with empty liposomes plus carnitine admixed were similar to those obtained with empty liposomes. Furthermore, based on anatomic and histological evaluations, treatment with carnitine-loaded liposomes improved the macroscopic appearance of the colon and significantly improved colitis (Table 4)
.
|
DISCUSSION
Colitis induced in rats by TNBS closely reproduces human gut inflammation (33)
. We used this model to address the question as to whether intestinal inflammation affects colonocyte carnitine uptake, and whether depletion of colonocyte carnitine affects butyrate oxidation, which is a fundamental energy pathway.
Studies performed with animals, animal intestinal preparations, human intestinal biopsy samples, and human intestinal epithelial cell lines have shown that carnitine is significantly absorbed only in the small intestine, without undergoing first-pass degradation, and in a dose-dependent manner presumably due to the involvement of saturable transport by carnitine carriers (34
, 35)
. On the contrary, carnitine uptake in the colon was reported to be negligible and restricted to a passive component (36)
. The results of our study modify the scenario depicted by these previous experimental studies. Indeed, our quantitative PCR data and functional analysis demonstrate that the Octn2 transcript is present and very abundant in the rat colon and small intestine. These results coincide with the original data obtained in rats by Northern blot analysis when this transporter was first cloned (36)
.
We also demonstrate that Atb0+, a low-affinity carnitine transporter expressed primarily in the mammary gland and lung (37)
, is expressed in rat colonocytes. Thus, rat colonocytes have a surplus rather than a deficiency of carnitine transporters. Because absorption of carnitine by the colon does not appear to be relevant for maintenance of systemic carnitine homeostasis in physiological conditions (35)
, we speculate that the two carriers are needed to supply colonocytes with carnitine. The abundant expression of carnitine transporters could reflect the fact that colonocytes must compete with colonic bacteria for the carnitine present in the intestinal lumen.
A novel finding of this study is that experimental colitis results in a simultaneous decrease in the expression of carnitine transporters and in the level of carnitine in colonocytes. This condition paralleled the impaired ability of colonocytes to oxidize butyrate, whereas butyrate uptake was unchanged. Therefore, while carnitine does not appear to be required for butyrate import into the mitochondrial matrix, it is required for butyrate metabolism.
It is generally agreed that SCFAs play a central role in colonocyte metabolism. Among SCFAs, butyrate is the most important energy source for colonocytes, because it provides them with 70% of their energy (38)
. Butyrate is almost never activated in the extramitochondrial space because it crosses the double-mitochondrial membrane very rapidly and, unlike long-chain fatty acids, does not require carnitine (13)
. In the mitochondrial matrix, butyrate undergoes ß-oxidation, which results in acetyl-coenzyme A production. Because of the excess of acetyl-coenzyme A produced by butyrate and the limited capacity of the Krebs cycle, part of the acetyl-coenzyme A is redirected to the synthesis of fatty acids in the cytosol (39
, 40)
. The acetyl-coenzyme A used in fatty acid production is transported from the mitochondria to the cytosol by a complicated transfer mechanism involving carnitine. This mechanism also serves to maintain a normal level of intramitochondrial free coenzyme A (CoA), which is thus available for further metabolic reactions (41
, 42)
.
Although apparently uncontrolled, the metabolism of butyrate can be subject to several intramitochondrial controls. For example, in case of free carnitine deficiency, impaired transformation of acetyl-coenzyme A esters into their carnitine esters would result in the breakdown of the mitochondrial oxidative metabolism of butyrate. This assumption is supported by our finding that the increase in total acetyl-coenzyme A level in butyrate-treated colonocytes results from an increased metabolic flow of butyrate into mitochondria with a consequent export of acetate groups from mitochondria to the cytosol. Indeed, in control colonocytes, where acetate can be efficiently transported from mitochondria to the cytosol through the carnitine acetyltransferase pathway, the increase of acetyl-coenzyme A after butyrate treatment occurs mainly in the cytoplasm. The finding that carnitine supplementation did not affect acetyl-coenzyme A distribution in control cells indicates that carnitine-dependent acetate transport is well supported by endogenous carnitine under physiological conditions. In the case of carnitine depletion, the finding that the increase in acetyl-coenzyme A was confined to the mitochondrial compartment of colonocytes is additional evidence that carnitine plays an important role in the regulation of cell metabolic fluxes. The dramatic increase in cytosolic acetyl-coenzyme A content induced by carnitine supplementation in colonocytes from TNBS-treated rats indicates that substrate flow at mitochondrial level was markedly decreased by Octn2/Atb0+ down-regulation consequent to decreased carnitine uptake and intracellular carnitine depletion. Octn2/Atb0+ down-regulation was specifically induced by inflammation, because it was not accompanied by decreased expression of the cell membrane transporter gene Abcd1.
The results of our experiments are compatible with a causal relationship between colonocyte carnitine depletion, which results in accumulation of acetyl-carnitine and a decrease in free carnitine, and the inability of mitochondria to maintain normal butyrate ß-oxidation. We also demonstrate that supplementation with carnitine-loaded liposomes restores butyrate metabolism in colonocytes from rats with colitis. These observations are supported by a significant inverse correlation between in vivo carnitine treatment and histological indicators of mucosal damage induced by TNBS treatment.
In conclusion, our data support the concept that carnitine is a rate-limiting factor for the maintenance of physiological butyrate oxidation in colonocytes. Thus, in agreement with Roediger (43)
, we suggest that it is not decreased butyrate uptake but rather deranged butyrate utilization that amplifies colonocyte damage in experimental colitis. This hypothesis would explain the contradictory therapeutic efficacy of butyrate supplementation observed in clinical trials of IBD (44)
.
Received for publication March 9, 2006. Accepted for publication July 6, 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  S. R. Srinivas, P. D. Prasad, N. S. Umapathy, V. Ganapathy, and P. S. Shekhawat Transport of butyryl-L-carnitine, a potential prodrug, via the carnitine transporter OCTN2 and the amino acid transporter ATB0,+ Am J Physiol Gastrointest Liver Physiol, November 1, 2007; 293(5): G1046 - G1053. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP |
), acetyl-carnitine (
), TEA (
), and glycine (
). Inhibitors were premixed with 0.5 µM [3H]-carnitine in buffer solution. Results of the 30-min experiments at 37°C are presented as percent of the control situation (0.5 µM [3H]-carnitine uptake in the absence of inhibitor). Each point and error bar represents the average and SD for 3–4 determinations of uptake corrected for protein content. *P < 0.001 vs. controls; 
), colonocytes from rats with TNBS-induced colitis; (•), colonocytes from TNBS-treated animals incubated with carnitine-loaded liposome. Values are mean ± SD from four experiments. * P < 0.001 vs. respective controls.