Introduction vitamin through plasma membrane transporters. Intracellu-

Brain Research 881 2000 144–151 www.elsevier.com locate bres Research report Sodium-ascorbate cotransport controls intracellular ascorbate concentration in primary astrocyte cultures expressing the SVCT2 transporter Jasminka Korcok, Raphael Yan, Ramin Siushansian, S. Jeffrey Dixon, John X. Wilson Department of Physiology , Faculty of Medicine and Dentistry, The University of Western Ontario, London, Ontario, Canada Accepted 8 August 2000 Abstract 1 Expression of the Na -ascorbate cotransporter, SVCT2, was detected in rat brain and in primary cultures of cerebral astrocytes by Northern blot analysis. SVCT2 expression in cultured astrocytes increased in response to the cyclic AMP analog, dibutyryl cyclic AMP. A 1 mathematical model of ascorbic acid transport was developed to evaluate the hypothesis that Na -ascorbate cotransport across the plasma membrane regulates the steady state intracellular concentration of ascorbic acid in these cells. The outcomes predicted by this model were compared to experimental observations obtained with primary cultures of rat cerebral astrocytes exposed to normal and pathologic conditions. Both cotransport activity and intracellular ascorbic acid concentration increased in astrocytes activated by dibutyryl cyclic 1 AMP. Conversely transport activity and ascorbic acid concentration were decreased by hyposmotic cell swelling, low extracellular Na 1 concentration, and depolarizing levels of extracellular K . In cells incubated for up to 3 h in medium having an ascorbic acid concentration typical of brain extracellular fluid, the changes in intracellular ascorbic acid concentration actually measured were not 1 significantly different from those predicted by modeling changes in Na -ascorbate cotransport activity. Thus, it was not necessary to specify alterations in vitamin C metabolism or efflux pathways in order to predict the steady state intracellular ascorbic acid concentration. These results establish that SVCT2 regulates intracellular ascorbic acid concentration in primary astrocyte cultures. They further indicate 1 that the intracellular-to-extracellular ratio of ascorbic acid concentration at steady state depends on the electrochemical gradients of Na and ascorbate across the plasma membrane.  2000 Elsevier Science B.V. All rights reserved. Theme : Neurotransmitters modulators transporters and receptors Topic : Uptake and transporters Keywords : Ascorbic acid; Transport; Mathematical model; Brain cell

1. Introduction vitamin through plasma membrane transporters. Intracellu-

lar AA functions as an enzyme cofactor and may contrib- Vitamin C occurs almost entirely in its reduced form, ute to antioxidant defense in brain cells. The concentration ascorbic acid AA, in normal brain [4,10,16,18,19]. While and redox state of vitamin C are crucial for these functions. the steady state concentration of AA is 200–400 mM in On the one hand, high concentrations of AA protect brain brain extracellular fluid, it is approximately 10-fold higher cells from ischemic, excitotoxic and oxidative injury in the cellular compartment. Brain cells cannot synthesize [9,15,16,23,35]. On the other hand, high concentrations of vitamin C from glucose de novo. Instead, they obtain the oxidized vitamin C dehydroascorbic acid, DHAA are cytotoxic [20] and especially neurotoxic [12]. Vitamin C transport systems have been characterized in Abbreviations: AA, ascorbic acid; DHAA, dehydroascorbic acid; an experimental model of brain cells, namely, primary dBcAMP, dibutyryl cyclic AMP; HPLC, high-performance liquid chroma- cultures of rat cerebral astrocytes. Cultured astrocytes tography achieve intracellular AA concentrations [AA] as high as i Corresponding author. Tel.: 11-519-661-3475; fax: 11-519-661- 8 mM when incubated with physiologic levels of extracel- 3827. E-mail address : john.wilsonmed.uwo.ca J.X. Wilson. lular AA for 3 h [24]. Longer incubation periods to do not 0006-8993 00 – see front matter  2000 Elsevier Science B.V. All rights reserved. P I I : S 0 0 0 6 - 8 9 9 3 0 0 0 2 8 2 9 - 8 J . Korcok et al. Brain Research 881 2000 144 –151 145 raise [AA] further [24]. Because the hydroxyl groups at were ubiquitous but had relatively high expression in i positions 2 and 3 ionize with pK values of 4.17 and 11.57, brain. In the present study, we report that astrocytes in most molecules of vitamin C exist as the monovalent anion primary culture express SVCT2 and that the expression ascorbate at physiologic pH. Simple diffusion of AA, level is increased by cyclic AMP. Subsequently, we ascorbate and DHAA through cell membranes is a slow develop a mathematical model to evaluate the hypothesis 1 1 process [19]. However, a much faster, Na -dependent that Na -ascorbate cotransport activity regulates [AA] . i mechanism of ascorbate uptake has been observed in the Actual values for transport rates and intracellular con- plasma membrane of rat and mouse astrocytes. This centrations were obtained using primary cultures of cere- mechanism is an example of electrogenic, secondary active bral astrocytes exposed to normal and pathologic con- 1 transport and it is driven by the membrane potential and ditions. When measured rates for Na -dependent AA 1 1 transmembrane Na gradient. The rate of Na -ascorbate uptake were entered as parameters in the model, the cotransport is stimulated in astrocytes that have been predicted values for [AA] were found to conform closely i depleted of intracellular AA [33] as well as in astrocytes to experimental observations. Furthermore, changes in 1 that have been activated by dibutyryl cyclic AMP Na -ascorbate cotransport activity were sufficient to ex- dBcAMP or forskolin [26,31,32]. Conversely, this co- plain the sustained influences of several pathologic con- transport activity is inhibited by incubation in media ditions on [AA] . i 1 having low Na concentration [31], low osmolality [25] or 1 depolarizing K levels [34]. Furthermore, the rate of 1 Na -dependent ascorbate influx slows as [AA] rises after 2. Materials and methods i reintroduction of a physiologic AA concentration to the medium, with most of this downregulation completed 2.1. Cell cultures 1 within 3 h [33]. In addition to Na -ascorbate cotransport, astrocytes also take up DHAA through facilitative hexose The experimental protocols were approved by the transporters and reduce it to AA intracellularly [26]. This University of Western Ontario Council on Animal Care. recycling of DHAA to AA depends on the rate of Primary cultures of astrocytes were prepared from the formation of DHAA in the extracellular fluid, is inhibited neopallium of 1-day-old Wistar rats according to our by physiologic concentrations of glucose and is not published procedure [26]. All subsequent treatments were 1 influenced by Na [26]. The only pathway by which AA carried out at 378C unless otherwise noted. The astrocytes has been demonstrated to exit astrocytes is through vol- were grown to confluence in horse serum-supplemented, ume-sensitive organic anion channels [25]. These channels minimum essential medium that did not contain detectable become permeant to ascorbate in response to hyposmotic AA ,1 mM [25]. Like astrocytes in situ, these cells are cell swelling. However, this response is only transitory and coupled by connexin43-positive gap junctions and contain the plasma membrane permeability to ascorbate returns to glial fibrillary acidic protein [11]. They were used for normal within 3 min even though regulatory volume transport experiments after 3–4 weeks in culture. decrease is incomplete [25]. It is possible that [AA] is influenced by several 2.2. Experimental procedures i mechanisms, namely, ascorbate uptake, DHAA uptake and recycling to AA, and ascorbate efflux. However, the effects Northern blot analysis was performed according to a of some of these mechanisms may be transient and not modification of a published procedure [22]. Total RNA determinant for [AA] at steady state. Although several was isolated from whole brain of neonatal rat and primary i transport systems and metabolic pathways have been astrocyte cultures using TRIzol GIBCO. RNA was shown to influence AA levels in cells, their relative separated on 0.6 formaldehyde-agarose gel, transferred importance has remained controversial [19]. Indeed, their to nylon membrane and cross-linked by UV irradiation. respective roles may vary between tissues. For instance, The amount of RNA loaded in each lane was 15 mg. The the brain differs from many other organs by having an blot was hybridized overnight with a 1 kb rat SVCT2 extracellular AA concentration that is several-fold higher cDNA 32P-labelled probe, in ULTRAhyb Ambion at than plasma levels. As a result, the extracellular AA 428C. Following hybridization, membrane was washed concentration in brain 200–400 mM is sufficient to 2315 min in 0.13SSC, 0.1 S.D.S. and 2330 min in 1 saturate high affinity Na -ascorbate cotransport systems 0.13SSC, 0.1 S.D.S. at 428C. The signal was detected [31]. by exposing the membrane to X-ray film. The blots were Recently, the transporters SVCT1 and SVCT2 were then stripped and reprobed for rat 18S ribosomal RNA cloned from both rat and human and shown to induce under the same conditions to normalize RNA loading. The 1 Na -dependent uptake of L -ascorbate when expressed intensities of SVCT2 and 18S rRNA autoradiographic heterologously [2,14,27–29]. Northern blot analysis of images were quantified by scanning densitometry and were mammalian tissues detected transcripts encoding SVCT1 expressed as the ratio, SVCT2 mRNA 18S rRNA. 1 in intestine, kidney and liver, whereas SVCT2 transcripts Isoosmotic normal-Na transport medium consisted of 146 J in mM: 134 NaCl, 5.4 KCl, 1.8 CaCl , 0.8 MgSO , 10 ANOVA and Dunnett’s test were used to evaluate the 2 4 1 glucose and 20 Hepes pH 7.3. The final Na con- effect of dBcAMP on SVCT2 expression levels. A ran- centration of this medium was 138 mM and its osmolality domised block factorial design was employed to evaluate was 300 mOsm. Other transport media were made by the effects of treatments on observed AA concentrations by changing the NaCl or KCl concentrations. The specific two-way ANOVA. Additionally, the paired t-test was used 1 changes were as follows: i Isoosmotic low-Na transport to compare the observed AA concentrations with the 1 media containing either 35 or 80 mM Na were prepared theoretical values. Finally, the effects of treatments on 14 by partial replacement of NaCl with the chloride salt of [ C]AA efflux rates were analysed using one-way 1 N-methyl- D -glucamine . ii Hyposmotic transport ANOVA. P,0.05 was considered to be significant for all medium was prepared by omission of 58 mM NaCl, so that statistical tests. 1 the final Na concentration was 80 mM and the osmolality 1 was 206 mOsm. iii Isoosmotic high-K transport medium was prepared by substitution of 50 mM KCl for 3. Results 1 an equal concentration of NaCl, thus raising the K 1 concentration to 55.4 mM. iv Hyperosmotic normal-K SVCT2 was detected by Northern blot analysis in whole transport medium was prepared by addition of sufficient rat brain and primary astrocyte cultures Fig. 1. DBcAMP N-methyl- D -glucamine chloride to increase the osmolality increased astrocytic expression of SVCT2 after a latent 1 to 400 mOsm. iv Hyperosmotic high-K transport medium was prepared by addition of 50 mM KCl to 1 increase both the K concentration and the osmolality. 14 Stock solutions of AA and L -[1- C]AA 8 mCi mmol, Dupont Canada were prepared at 48C and contained homocysteine to prevent oxidation of the vitamin [24]. These stock solutions were diluted in prewarmed transport media immediately before addition to cell cultures. Initial rate of AA uptake was measured by incubating 1 astrocytes in isoosmotic normal-Na transport medium 14 containing [ C]AA 5 mM, 378C. Uptake was linear with time for at least 90 s. To study intracellular accumulation of AA in the presence of a physiologic concentration of extracellular AA, the astrocytes were incubated for the indicated periods in transport media containing 200 mM AA with 160 mM homocysteine to prevent AA oxidation. Washing the cultures with ice-cold Tris-sucrose solution pH 7.3 terminated uptake. [AA] was determined by i acidic extraction and HPLC-based electrochemical assay assay sensitivity was 2 pmol AA [25]. Total cell protein was determined by the Lowry assay. For experiments investigating AA efflux, astrocytes were first loaded with the radiolabeled vitamin by incuba- 1 tion for 60 min in isoosmotic normal-Na transport 14 medium containing 200 mM [ C]AA with 160 mM homocysteine. Next, the cells were washed and incubated in the indicated transport media containing 200 mM unlabeled AA. Aliquots of media were collected at timed intervals and then the cells were harvested. The radioactive contents of the media and cells were measured by liquid scintillation counting. 2.3. Statistics Fig. 1. Northern blot analysis of SVCT2 expression in total RNA from rat brain and primary cultures of cerebral astrocytes. Confluent primary Data are presented as mean6S.E.M. values from n cultures of rat astrocytes were incubated with dBcAMP 0.25 mM or number of independent experiments, with triplicate replica- vechicle for the incubated periods. Shown is a representative blot panel tions i.e., 3 culture dishes per treatment in each experi- A and the mean6S.E.M. values for SVCT2 18S RNA ratios from 3 ment. In the figures, error bars are omitted where the experiments panel B. P ,0.05 compared to vehicle-treated control S.E.M. is less than the size of the symbol. One way astrocytes. J . Korcok et al. Brain Research 881 2000 144 –151 147 period of 12 h Fig. 1. The initial rates of AA uptake after 24 h of treatment were 6962 nmol g protein min in vehicle control and 10464 nmol g protein min in dBcAMP-treated astrocytes n57 experiments, P,0.05. We next developed a mathematical model to evaluate 1 the hypothesis that Na -ascorbate cotransport activity regulates [AA] . The mechanisms addressed by the mathe- i matical model are illustrated in Fig. 2 and the details are presented in the DISCUSSION below. Uptake rate through 1 the Na -ascorbate co-transport system is designated as J . 1 Volume-insensitive AA efflux from astrocytes is J ; this 2 efflux may occur by reversal of the co-transport system or by other volume-insensitive pathways. The AA efflux pathway that is activated transiently by cell swelling is J . 3 DHAA formation, uptake and intracellular reduction to AA are shown as discrete steps in Fig. 2 but are designated altogether as J in the model. The differential equation 4 describing the rate of change of [AA] as a function of i time was solved numerically and solutions are presented as the theoretical AA concentrations plotted in Figs. 3 and 4. The results of the HPLC-based electrochemical assay showed that the astrocytes cultured in vitamin C-free medium did not contain intracellular AA initially. How- ever, the cells gradually accumulated the vitamin when incubated with physiologic concentrations of AA 200 1 mM and Na 138 mM Fig. 3. Activation of astrocytes by dBcAMP increased the AA uptake rate and steady state [AA] . The [AA] observed at each time point were i i Fig. 3. Accumulation of AA in control and dBcAMP-activated as- compared to values predicted by the mathematical model. trocytes. Primary cultures of rat astrocytes were grown to confluence in For both the vehicle control and dBcAMP-activated as- vitamin C-free medium. Subsequently, dBcAMP-activated 0.25 mM, top trocytes, observed [AA] was not significantly different panel and control bottom panel astrocytes were incubated with 200 mM i AA for the indicated times in isoosmotic medium containing 138 mM from predicted values Fig. 3. This finding demonstrates 1 Na before being harvested and assayed for intracellular AA con- that the model is capable of predicting [AA] using values i centration. Plotted as open circles are mean6S.E.M. values for cell AA for J calculated from previously published V and K 1 max m concentration from 4 experiments. Curves represent the theoretical values 1 values for Na -ascorbate cotransport by control and predicted by the mathematical model that is described in the text. dBcAMP-activated astrocytes [31]. Furthermore, the effect of dBcAMP on steady state [AA] was modeled by i changing J , without altering J , J or J J 50.012[AA] , 1 2 3 4 2 i J 50, and J 50. 3 4 1 Incubation of astrocytes in media that decreased Na - dependent ascorbate uptake rates led to decreased steady state [AA] Fig. 4. The mathematical model was used to i calculate theoretical values for steady state [AA] using i values for J calculated from transport parameters obtained 1 in earlier studies as presented in the Legend of Fig. 4. When the other fluxes were kept constant J 50.012[AA] , 2 i J 50, and J 50, the [AA] values predicted were not 3 4 i significantly different from the concentrations actually measured Fig. 4. The equations modeled the effects of treatments on [AA] by varying the AA uptake rate and leaving the efflux i Fig. 2. Model of vitamin C recycling by cultured astrocytes. AA uptake rate unchanged. To further test the validity of the assump- 1 through the Na -AA cotransporter is designated as J , efflux through 1 tion that efflux rates do not change, we measured these volume-insensitive pathways as J , and efflux through volume-activated 2 rates under various experimental conditions. Table 1 shows pathways as J . DHAA formation, uptake and intracellular reduction to 3 the efflux rates obtained when astrocytes were incubated in AA are shown as discrete steps in this illustration but are designated 1 1 altogether as J in the mathematical model. hyperosmotic normal-K and high-K media. No effects 4 148 J Fig. 4. Accumulation of AA in astrocytes incubated under conditions of changed ionic gradients. Primary astrocyte cultures were incubated with 1 1 200 mM AA for 3 h in normal, low Na , hyposmotic or high K medium before being harvested and assayed for intracellular AA concentration. Bars show mean6S.E.M. values from 4 experiments. The theoretical values were generated from the equation described in the Discussion using the following parameters: J o 50.39 mmol g protein min for the