Address correspondence and reprint requests to Luc Pellerin, Institut de Physiologie, rue du Bugnon 7, 1005 Lausanne, Switzerland.E-mail: Luc.Pellerin@iphysiol.unil.ch
Regulation of the expression of MCT1 and MCT2, two isoforms of the monocarboxylate transporter (MCT) family, was investigated in primary cultures of mouse cortical neurons. Under basal conditions, both MCT immunoreactivities (IR) were found in the cell soma and dendrites, although IR for MCT1 appeared less bright than for MCT2. Treatment of cultured cortical neurons with 100 μm noradrenaline (NA) led, after a few hours, to a striking enhancement in fluorescence intensity associated with MCT2 IR in the cell soma as well as in dendrites. In contrast, MCT1 IR was not altered by NA treatment. Western blot experiments performed on cultured neurons treated with NA confirmed that MCT2 protein expression was increased. Forskolin and dBcAMP also enhanced MCT2 expression, suggesting the implication of a cAMP-mediated pathway in the effect of NA. Surprisingly, neither NA, dBcAMP nor forskolin affected MCT2 mRNA expression. Application of cycloheximide, a protein synthesis inhibitor, prevented the enhancement of MCT2 IR, while the mRNA synthesis inhibitor actinomycin D also blocked the effect of NA on MCT2 IR levels. These results suggest that regulation of MCT2 expression in neurons by NA occurs at the translational level despite the requirement for an as yet unknown transcriptional step.
Evidence for the regulation of MCT expression exists but characterized examples are rare and remain limited, essentially to skeletal muscle and heart (Bonen 2001). Thus, it was shown that MCT1 expression was strongly and reversibly increased by activity in fibres of the skeletal and cardiac muscle or after congestive heart failure (McCullagh et al. 1997; Baker et al. 1998; Bonen et al. 2000; Eydoux et al. 2000; Johannsson et al. 2001). Similar findings were obtained in humans in which case physical exercise was found to transiently enhance the expression of both MCT1 and MCT4 in skeletal muscle, together with alterations in their capacity to transport lactate (Bonen et al. 1998; Pilegaard et al. 1999; Dubouchaud et al. 2000; Evertsen et al. 2001; Green et al. 2002). MCT1 was shown to be up-regulated by butyrate in human colonic cell cultures (Cuff et al. 2002). In the central nervous system, developmental changes in expression have been described, including a transient increase of both MCT1 and MCT2 expression during the suckling period that could be related to a greater ketone bodies utilization by the brain during this period (Pellerin et al. 1998; Leino et al. 1999). Nevertheless, the observation that activity can trigger an enhancement of MCT expression in muscle suggests that a related phenomenon could also take place in the central nervous system. In this study, we sought to determine whether noradrenaline, a neurotransmitter known not only to modulate neuronal excitability but also to regulate brain energy metabolism (Magistretti and Morrison 1988; Magistretti et al. 1993), can alter the expression of MCT1 and/or MCT2 in mouse cultured cortical neurons.
Neuronal cultures and pharmacological treatments
Primary cultures of mouse cortical neurons were prepared from embryonic day 17 OF1 mice, as described previously (Debernardi et al. 2003). Briefly, they were cultured at a density of 500 cells/mm2 in Neurobasal-B27 medium (Brewer et al. 1993). Half of the medium was changed at day 3 and experiments were all performed on day 7 in vitro (DIV). Cultures contained approximately 1% of glial cells at this stage. Pharmacological agents [100 μm noradrenaline (NA), 100 μm forskolin (FK), 1 mm dibutyryl-cyclic-adenosine monophosphate (dBcAMP)] were added directly to the medium and cells were incubated for the indicated time in their presence. Transcription and translation inhibitors (5 μm actinomycin D and 10 μm cycloheximide) were added 30 min before pharmacological agents. All chemicals were purchased from Sigma (Buchs, Switzerland).
MCT1 and MCT2 production and characterization have been described previously in Pierre et al. (2000). Briefly, for production, two peptides consisting of the 16 and 15 carboxyl-terminal amino acids of Chinese hamster MCT1 (CPQQNSSGDPAEEESPV) and rat MCT2 (CNTHNPPSDRDKESSI) were synthesized, with a cysteine added at the N-terminal site for optimal conjugation to Keyhole Limpet Hemocyanin. MCT1 and MCT2 antisera were obtained after subcutaneous injections into two respective rabbits. Specificity for both polyclonal MCT antibodies was demonstrated in previous studies where they were used for immunohistochemistry or for western blot experiments. Peptide antigens have been used as competitive inhibitors of anti-MCT antibodies, by pre-absorbing the MCT1 or MCT2 primary antibody with 10 μg/mL of the appropriate peptide antigen. Staining was absent on membranes or in sections that had been incubated in such solutions (see Pierre et al. 2000, 2002). In addition, specificity of our MCT2 antibody was further assessed by western blot using fractions from different mouse tissues (data not shown). As expected, our MCT2 antibody recognized a single band with a molecular weight of 40–43 kDa in brain, liver and kidney, but not in striated muscle, heart, lung or stomach, in accordance with previous studies (Jackson et al. 1997; Bergersen et al. 2001). Anti-microtubule associated protein-2 (MAP2) was a mouse monoclonal antibody from Sigma.
Immunocytochemistry and related quantification
After removal of the culture medium, cells were briefly rinsed in phosphate-buffered saline (PBS) at 37°C and directly post fixed in a paraformaldehyde fixative (4% in PBS, 15 min, at 20°C). They were then treated with casein (0.5% in PBS) for 1 h to block non-specific sites. For both single and double immunostainings, cultures were incubated in parallel overnight at 4°C, either in 50 μL taken from the same freshly prepared primary antibody solution (anti-MCT1 or anti-MCT2 diluted 1/400 in PBS containing 0.25% of bovine serum albumin), or in 50 μL taken from the same mixture of anti-MCT1 or anti-MCT2 (1/400) and anti-MAP2 (1/200) antibodies.
After careful rinsing in PBS, cultures were incubated in a solution containing goat FITC-conjugated anti-rabbit Igs (diluted 1/400, 2 h, room temperature) or a mixture of goat FITC-conjugated anti-rabbit Igs and goat Texas Red-conjugated anti-mouse Igs for single or double immunolabellings, respectively. After rinsing in PBS, cultures were mounted with Vectashield (Reactolab SA, Sigma) and examined with an Axioplan2 microscope (Zeiss, Germany), using epifluorescence with an appropriate filter.
Controls included omission of the primary serum or its substitution by non-immune rabbit serum. No specific staining was visible in such conditions.
To quantitatively assess the influence of different treatments on MCT2 protein expression, we carried out a quantitative analysis on images obtained by epifluorescence with a 20 × objective and acquired using a cooled CCD Camera (Axiocam, Zeiss, Germany) together with the 2.05 Axiovision software (Zeiss). Three fields were chosen randomly in each untreated or treated culture that contained at least 20 MCT2-labelled neurons per field. All pictures were acquired and presented as different levels of grey, with identical acquisition time for all. Pictures were then analysed using NIH software. We assessed the fluorescence intensity of eight isolated cells taken randomly in each of the three captured areas. The mean fluorescence intensity representing neuronal MCT2 expression in a given condition was thus obtained by an average of 24 counts. Counts were obtained in a blind fashion with the investigator unaware of the culture treatments. Data were statistically analyzed with anova followed by Dunnett's test.
Western blot and related quantification
Neuronal cultures were homogenized in a buffer containing 62.5 mm Tris-HCl, 50 mm dithiotreitol (DTT), 0.3% sodium dodecyl sulfate (SDS) and a mixture of protease inhibitors(1 mg/mL of each, anti-papain, leupeptin, pepstatin and E64, and 100 mm phenyl methylsulfonylfluoride, all purchased from Sigma). Lysates from three Petri dishes were pooled and protein concentration determined by the method of Bradford (Bradford 1976). Five micrograms of protein were heated at 95°C in SDS–PAGE sample buffer (62.5 mm Tris-HCl, 50 mm DTT, 2% SDS, 10% glycerol and 0.1% bromophenol blue) and loaded onto 10% polyacrylamide gels. After electrophoresis, samples were transferred to polyvinylidene difluoride membranes (NEN Life Science Products, Boston, MA, USA). Membranes were incubated in a blocking solution of Tris-buffered saline (TBS) containing 0.1% Tween, 10% non-fat milk and 1% BSA for 1 h. Membranes were incubated overnight at 4°C with the anti-MCT2, diluted at 1/5000 in TBS, 0.1% Tween, 1% milk (TBSTM). After four washes in TBSTM, membranes were incubated with the secondary antibody (diluted 1/250000) in TBSTM coupled to horseradish peroxidase (Boehringer, Mannheim, Germany) for 1 hour at room temperature, followed by four washes in TBSTM and two in TBST. Bound antibody was visualized by chemiluminescence (ECL Femto sensitive, Pierce, Rockford, IL, USA), detected with FluorS-Imager (Bio-Rad Laboratories, Hercules, CA, USA) and quantified with Quantity One software (Bio-Rad). Kodak AR films were then generally apposed for better image resolution. Differences in protein gel loading and blotting were assessed by re-incubating the membranes with an anti-β-actin antibody (diluted at 1/2000), after a stripping step at 50°C for 30 min in a solution of 62.5 mm Tris-HCl pH 6.8, 2% SDS and 0.7% β-mercaptoethanol. Incubation and washing conditions for the anti-β-actin antibody were identical to those for anti-MCT2.
MCT2 cDNA fragment and radiolabelled probe generation
A MCT2 cDNA fragment located at 1297–1881 bp in the coding region of the mouse MCT2 cDNA sequence (Koehler-Stec et al. 1998) was generated by RT-PCR, inserted in a pGEMTeasy plasmid (Promega. Madison, WI, USA) and amplified in Escherichia coli cultures. Plasmids were then extracted and sequenced, confirming the expected sequence. The radiolabelled cDNA probe was generated by random primed nucleotide labeling (Roche Molecular Biochemicals, Indianapolis, IN, USA) with 200 ng of the RT-PCR product and 32P-dATP (3000 Ci/mmol, Hartmann Analytic GmbH, Braunschweig, Germany) at 37°C for 45 min. The radiolabelled cDNA probe was subsequently purified on column (Nucleospin Extraction kit, Macherey Nagel, Duren, Germany) and denatured.
Sample preparation and hybridization
Total RNA was extracted using the Trizol reagent procedure (Invitrogen, Carlsbad, CA, USA). Poly(A)+ RNA was obtained by passing total RNA through an oligo(dT)-cellulose spin column (Amersham Pharmacia Biotech, Piscataway, NJ, USA). Four micrograms of poly(A)+ mRNA were electrophoresed on a denaturing 1.3% agarose gel containing 2 m formaldehyde, and were transferred onto nylon membrane (Hybond N+ membrane, Amersham Pharmacia Biotech). Hybridization was performed overnight at 65°C in a solution containing EDTA 1 mm, NaH2PO4 0.2 m, Na2HPO4 0.3 m, SDS 7% and the denaturated 32P-cDNA probe specific for MCT2. Filters were then washed under high-stringency conditions (twice with 2 × SSC/0.1% SDS at 65°C for 20 min and once with 0.1 × SSC/0.1% SDS at 65°C for 15 min, where sodium saline citrate SSC 20 × = 3 m NaCl and 0.3 m Na citrate) and sealed. Bound radiolabelled probe was detected with a phosphoimager (Bio-Rad), quantified with Quantity One software (Bio-Rad) and subsequently apposed to Kodak AR film at −80°C with an intensifying screen in order to achieve better image resolution. Differences in RNA gel loading and blotting were assessed by rehybridizing the filters with a 32P-antisense actin riboprobe (Pellerin et al. 1998), after membrane stripping for 30 min at 95°C in a 0.1 × SSC/1% SDS solution. Hybridization and washing conditions for β-actin were identical to those for MCT2.
Double immunocytochemical labellings performed with MCT2 and the neuronal marker MAP2 showed that MCT2 immunoreactivity (IR) was present in all cultured cortical neurons (Fig. 1ai–iii). MCT2 IR was particularly intense and was found in the cell soma as well as in neurites that were often surrounded by several puncta (Fig. 1ai, inset). Treatment of cultured cortical neurons with 100 μm noradrenaline (NA) led, after 9 h, to a striking enhancement in fluorescence intensity corresponding to higher levels of MCT2 IR (Fig. 1bi vs. Fig. 1ai). This increased expression of MCT2 IR was observed in virtually all neurons in the culture and it was so remarkable that, in the overlay, the green fluorescence signal associated with MCT2 IR became predominant in all neurons (compare aiii and biii of Fig. 1). Moreover, at the subcellular level, enhanced MCT2 IR was observed in the cytoplasm and in neurites as well (Fig. 1bi, inset). No change in MAP2 IR was observed under the same conditions (Fig. 1bii vs. Fig. 1aii). In parallel, double immunocytochemical labellings with MCT1 and MAP2 revealed that many neurons expressed MCT1 IR (Fig. 1ci–iii). In contrast to MCT2, exposure of cortical neurons to 100 μm NA did not alter the intensity of MCT1 IR after the same time period (Fig. 1di–iii).
Change in the levels of MCT2 IR induced by NA was studied as a function of time. It was observed that NA at a concentration of 100 μm caused a gradual increase of MCT2 IR in cortical neurons that reached more than 400% of control level 9 h after the addition of the neurotransmitter (Figs 2a and b). Levels of MCT2 IR rapidly decreased thereafter to come back to control levels by 24 h. Together with NA, the effect of dBcAMP and forskolin was quantitatively assessed on both MCT2 IR and protein expression. All three substances significantly enhanced MCT2 IR as observed by immunocytochemistry (Fig. 3a) and MCT2 protein expression as determined by western blot analysis (Fig. 3b). No specific differences in the distribution of MCT2 IR, both at the cellular and subcellular levels, could be observed following treatment with one or other substance, although forskolin led to a higher protein expression level than NA. Results shown were obtained 6 h after addition of each agent, but similar results were observed after 9 h (data not shown).
The effect of NA on MCT2 mRNA expression was investigated in the same preparation by northern blot analysis. Surprisingly, NA at 100 μm did not alter levels of MCT2 mRNA in cultured cortical neurons exposed for times between 3 and 24 h (Fig. 4a). In addition, neither dBcAMP nor forskolin applied to cultured cortical neurons for 3 h were able to alter MCT2 mRNA expression (Fig. 4b). A parallel investigation with quantitative RT-PCR confirmed that none of these treatments had an effect on MCT2 mRNA expression (data not shown). The implication of translational and/or transcriptional steps in the effect of NA on MCT2 IR levels was further explored using specific inhibitors. Application of cycloheximide, a protein synthesis inhibitor, prior to NA stimulation, prevented the enhancement of MCT2 IR (Fig. 5). Finally, and quite unexpectedly, the mRNA synthesis inhibitor actinomycin D also blocked the effect of NA on MCT2 IR levels (Fig. 5).
Despite our increasing knowledge about MCT distribution in various tissues, including the central nervous system, relatively few examples describing regulations of MCT expression have been reported so far. In the cerebral cortex, MCT1 and MCT2 were shown to be expressed by various cellular elements (Gerhart et al. 1997, 1998; Hanu et al. 2000; Pierre et al. 2000, 2002), and it was observed that their levels of expression transiently increased during postnatal development (Pellerin et al. 1998; Leino et al. 1999). It was proposed that this transient enhancement in expression could be caused by circulating ketone bodies whose concentrations rise during the suckling period. This hypothesis was supported by a study showing that, in adult rat, a ketogenic diet induced brain MCT1 expression (Leino et al. 2001). Recently, it was shown that both MCT1 and MCT2 are expressed by mouse cultured cortical neurons. Interestingly, the level of MCT2 expression was correlated with synaptogenesis in these cultured cells, suggesting that its expression could be regulated by factors unrelated to the diet but rather in relation with synaptic activity (Debernardi et al. 2003). Our results further support the idea that MCT1 and MCT2 are regulated by distinct type of signals, related either to diet (for MCT1) or neuronal activity (for MCT2).
Noradrenaline is an important neurotransmitter in the central nervous system which modifies neuronal excitability, and it has been suggested previously that it participates in the regulation of brain energy metabolism (Magistretti and Morrison 1988). Indeed, it was shown to enhance both glycogenolysis and glycolysis in brain slices and cultured astrocytes (Magistretti et al. 1993; Tsacopoulos and Magistretti 1996). Results presented here clearly demonstrate that NA selectively increases the expression of MCT2 protein in mouse cultured cortical neurons, as determined by both immunocytochemistry and western blotting, while it had no effect on MCT1 IR levels in the same preparation. A similar enhancement of MCT2 expression was obtained following dBcAMP or forskolin exposure. As NA is known to induce cAMP formation in cultured neurons via the activation of β-adrenergic receptors (Atkinson and Minneman 1991), these results suggest that the effect of NA could be mediated via cAMP formation and protein kinase A activation. In contrast to protein expression, no effect of either NA, dBcAMP or forskolin was observed on MCT2 mRNA expression. This observation would be consistent with a regulation of expression at the translational rather than transcriptional level. Inhibition of NA-induced MCT2 protein expression by cycloheximide, an inhibitor of protein synthesis, supports this view as it also excludes the possibility of other protein synthesis-independent effects such as unmasking epitopes as a consequence of a disruption of protein–protein interactions or alteration in protein turnover.
Discrepancies in the distribution of MCT2 mRNA and protein in different organs including the brain had previously led to the suggestion that MCT2 expression could be regulated at the level of translation (Jackson et al. 1997; Lin et al. 1998; Halestrap and Price 1999). In addition, the presence of multiple mRNA transcripts for MCT2 in the central nervous system (Koehler-Stec et al. 1998; Pellerin et al. 1998) gives rise to the possibility that tissue-specific, post-transcriptional regulation of MCT2 expression may occur through alternative splicing within the 5′- or 3′-untranslated regions, leading to differences in translation efficiency (Lin et al. 1998). An important body of work on the control of the translation machinery has recently uncovered some of the mechanisms leading to gene-specific regulation of translation (Dever 2002). One important control of translation is exerted by phosphorylation of a series of proteins called eukaryotic translation initiation factors (eIFs) involving different kinases (Gingras et al. 1999). Several examples whereby cAMP elevations (Lawrence et al. 1997; Morgan and Beinlich 1997), protein kinase A activation (Diggle et al. 2001) or even stimulation of β-adrenergic receptors (McLeod et al. 2001; Freeman et al. 2002) led to phosphorylation of some of these eIFs and alteration in protein synthesis have already been reported. Whether the specific phosphorylation of one or several of these initiation factors could explain the effect of NA on MCT2 expression in cultured neurons, however, is an attractive possibility that remains to be established.
A paradoxical observation that was made is the fact that actinomycin D, a general transcription inhibitor, can prevent NA-induced enhancement in MCT2 protein expression, despite the fact that MCT2 mRNA expression was unaltered by NA treatment. These data would suggest that a transcriptional step is required in the mechanism leading to enhancement in MCT2 protein expression by NA, although it does not involve MCT2 mRNA itself. Two hypotheses could be formulated to account for this observation. First, as described above, a number of eIFs are necessary for the control of the translation machinery. It is possible that the synthesis of some of these factors might be necessary for the initiation of translation and requires a transcriptionally regulated step (Dever 2002). The second hypothesis refers to a newly discovered class of non-coding RNA or ncRNA that is involved, among other functions, in the regulation of mRNA stability and translation (Storz 2002). Recently, the presence of one of these ncRNAs was identified in neurons of the central nervous system (Wang et al. 2002). BC1 RNA, as it is called, was found to be present in dendrites and demonstrated that it regulated local translation initiation. The possibility that MCT2 mRNA could be also found in dendrites and translated locally under the control of such factors is particularly interesting as there is already some evidence that the MCT2 transporter is expressed in post-synaptic terminals, both in the cortex (Pierre et al. 2002) and in the cerebellum (Bergersen et al. 2001, 2002).
The observation that NA can enhance MCT2 expression in neurons is particularly relevant in the context of the regulation of brain energy metabolism and neuron–glia interactions. There is a number of evidence suggesting enhanced lactate formation during increased synaptic activity as well as the existence of a net lactate transfer from astrocytes to neurons (reviewed in Bouzier-Sore et al. 2002). Lactate produced and released by astrocytes would become available to serve as an additional energy substrate for active neurons via its uptake through MCT2 transporters that were shown to be widely expressed on neurons (Pierre et al. 2002). As mentioned above, noradrenergic neurotransmission is known to have important effects on energy metabolism. By inducing both glycogenolysis and glycolysis in astrocytes (Sorg and Magistretti 1991; Yu et al. 1993), it would contribute to provide further energy substrates, and more specifically lactate, to neurons. Furthermore, through enhancement of MCT2 expression in neurons, it would facilitate their use of lactate. Consistent with the “enabling” actions of the noradrenergic system during wakefulness involving changes in gene expression (Cirelli and Tononi 2000), these processes might represent essential adaptative mechanisms to allow the brain to face important energy demands linked to enhanced level of activity.
The authors would like to thank Dr Sylvain Lengacher for his help with quantitative real time RT-PCR. This work was supported by a grant from Fonds National Suisse de la Recherche grant 31-56930-99 (to PJM and LP).