sign in sign up
<<
Home , Anaplerotic reactions, Citric acid cycle

Glucose-Regulated Anaplerosis and Cataplerosis in Pancreatic -Cells Possible Implication of a Pyruvate/Citrate Shuttle in Secretion Salah Farfari, Vera Schulz, Barbara Corkey, and Marc Prentki

The hypothesis proposing that anaplerosis and cata- plerosis play an important role in fuel signaling by providing mitochondrially derived coupling factors he stimulus-secretion coupling of calorigenic nutri- for stimulation of insulin secretion was tested. A rise ent secretagogues in the pancreatic -cell involves in citrate coincided with the initiation of insulin 2+ several pathways: the KATP- and Ca -dependent secretion in response to in INS-1 -cells. The 2+ pathway, the KATP-independent and Ca -dependent dose dependence of glucose-stimulated insulin T 2+ pathway, and possibly a K - and Ca -independent pathway release correlated closely with those of the cellular ATP contents of citrate, malate, and citrate-derived mal- (1–5). Little is known about the biochemical nature of the onyl-CoA. The glucose-induced elevations in citrate, second and third pathways. However, the hypothesis has been -ketoglutarate, malonyl-CoA, and the 3-[4,5- proposed that anaplerosis, the refilling of Krebs cycle inter- dimethylthiazol-2yl]-2,5-diphenyltetrazolium reduc- mediates, is implicated in the KATP-independent pathway(s) tion state, an index of -cell metabolic activity, were (1,6,7), perhaps via malonyl-CoA formation and esterifi- unaffected by the Ca2+ chelator EGTA. Glucose cation processes (8) or through a pyruvate/malate shuttle (9). induced a rise in both mitochondrial and cytosolic cit- The following -cell features are consistent with the view that rate and promoted efflux of citrate from the cells. anaplerosis is implicated in fuel signaling. 1) The anaplerotic The latter amounted to ~20% of glucose (PC) is expressed at very high entering the glycolytic pathway. Phenylacetic acid, a levels, amounting to 0.4% of total in islet (10), pyruvate carboxylase inhibitor, reduced the glucose- induced rise in citrate in INS-1 cells and insulin INS-1 cells (11), and purified -cells (7), but at low levels in secretion in both INS-1 cells and rat islets. The non– islet cells (7). The level of PC in the islet of Langerhans, results indicate the feasibility of a pyruvate/citrate which is not a gluconeogenic tissue (12), is similar to that of shuttle in INS-1 -cells, allowing the regeneration of (11), which contains ~5 times more mitochondria (13). NAD+ in the and the formation of cytosolic Malic enzyme (ME), another anaplerotic enzyme, is abundant acetyl-CoA, malonyl-CoA, and NADPH. The data sug- in islets (14) and INS cells (15). 2) Approximately 40% of the gest that anaplerosis and cataplerosis are early sig- glucose entering the cycle is carboxylated naling events in -cell activation that do not require in rat islets (16,17). This is a very high percentage for a non- a rise in Ca2+. It is proposed that citrate is a signal of gluconeogenic tissue (12) and a cell synthesizing at a low fuel abundance that contributes to -cell activation in rate (11). 3) The dose dependencies of citrate and malate both the mitochondrial and cytosolic compartments and that a major fate of anaplerotic glucose carbons accumulation in response to glucose correlate highly with is external citrate. Diabetes 49:718–726, 2000 secretion in INS-1 cells (7); only the or combination of nutrients that cause a rise in citrate and citrate-derived malonyl-CoA are secretagogues in HIT cells (8,11). 4) The anaplerosis of glucose carbons highly correlates with glucose- induced secretion in purified -cells (7). 5) Methyl-succinate is a potent secretagogue in intact -cells (18), and succinate From the Molecular Nutrition Unit (S.F., M.P.), Department of Nutrition, Uni- directly promotes the exocytotic release of insulin in perme- versity of Montreal, the Centre de Recherches du CHUM and Institut du Can- abilized pancreatic -cells (19). 6) Anaplerosis is exacerbated cer, Montreal, Quebec, Canada; and the Diabetes and Unit (V.S., B.C.), Boston University Medical School, Boston, Massachusetts. in INS cells showing high basal insulin secretion after long-term Address correspondence and reprint requests to Marc Prentki, PhD, exposure to elevated glucose (15); the expression level of PC CR-CHUM, Pavillon de Sève 4e, 1560 Sherbrooke Est, H3P3J7-Montrèal is markedly reduced in glucose incapacitated islets preincu- (PQ), Canada. E-mail: [email protected]. Received for publication 31 August 1999 and accepted in revised form bated at very low (1 mmol/l) glucose for 24 h (10,20). 7 January 2000. 7) Inhibitors of carbonic anhydrases that catalyze the – BSA, bovine serum albumin; IBMX, 3-isobutyl-1-methylxanthine; KRBH, reversible hydration of CO2 to HCO3 , a substrate for the PC Krebs-Ringer bicarbonate HEPES medium; ME, malic enzyme; MTT, reaction, inhibit glucose-induced insulin secretion; further- 3-[4,5-dimethylthiazol-2yl]-2,5-diphenyltetrazolium; PAA, phenylacetic acid; PBS, -buffered saline; PC, pyruvate carboxylase; TCA, more, islet -cells, but not -cells, show intense staining with trichloroacetic acid. an antibody reacting with carbonic anhydrase V (21).

718 DIABETES, VOL. 49, MAY 2000 S. FARFARI AND ASSOCIATES

The fate of glucose carbon entering the via described above. The citrate output from the cells was calculated by subtracting anaplerosis is uncertain. It is particularly intriguing in view of the amount present at time 0 of the incubation periods. the magnitude of this flux in comparison to other cell types. Rapid cell fractionation procedure. Rapid cell fractionation through oil to mea- sure in the cytosolic and mitochondrial compartments has been car- Of the glucose carbon entering the cycle via anaplerosis, 25% ried out in isolated rat hepatocytes (28,29). We used the same protocol with sus- is channeled into synthesis in the -cell (7), but the pensions of INS-1 cells. After the incubation periods, cells were scraped from the fate of the remaining 75% is presently unknown. Nonetheless, dishes and resuspended in 2 ml of the same corresponding media. There was it should be emphasized that cataplerosis, the efflux of Krebs 200 µl of the cell suspension (2.5 107 cells) added to Eppendorf tubes containing (from bottom to top) 40 µl of 10% trichloroacetic acid (TCA), 300 µl of an oil (Dow- cycle intermediates and their metabolites from the mito- Corning, Acton, MA) mixture (80.75% heavy oil, fluid 550; 19.25% light oil, fluid chondrial to the cytosolic compartment (22), must be quan- 200), and 160 µl of the detergent saponin (1.12 mg/ml). After 20 s of further cell titatively similar to anaplerosis because the citric acid cycle incubation with saponin, the tubes were centrifuged at 16,000g for 1 min. The upper intermediates are not a sink for glucose carbons. Further- phase was collected for “cytosolic” citrate measurements. There was 360 µl of more, the sum of the Krebs cycle intermediate–derived car- bidistilled added to the remaining phases. The tubes were vortexed vigor- ously, and 2 phases were obtained after centrifugation at 16,000g for 1 min. The bons used for biosynthetic reactions within the mitochondria upper phase consisted of TCA soluble extracts of the cellular “particulate” frac- (de novo synthesis of proteins, lipids, , por- tion, including the mitochondrial citrate. The intermediate phase was the oil and phyrines, and other compounds) should be relatively minor the lower phase, i.e., the pellet, was used for cellular protein determinations. in comparison to those used in all other compartments. Cytosolic and particulate citrate were assayed after neutralization by 6 successive ether extractions of the samples to remove TCA. The validity of the method was In view of these considerations and the possibility that also assessed through demonstration that the total cellular citrate content of anaplerotic/cataplerotic fluxes may be at the heart of -cell unpermeabilized cells was identical to the sum of the particulate (mitochondrial) signaling and the generation of mitochondrial signals trig- and cytosolic citrate. gering the exocytotic release of insulin (1,2), we endeavored Glucose metabolism measurements. Glucose usage was determined radio- metrically as the production of 3H O from [5-3H]glucose (27). INS-1 cells were cul- to 1) further characterize anaplerosis and study the fate of 2 tured in 21 cm2 dishes, as described above, and subsequently incubated for 1 h anaplerotic carbons in INS-1 cells, 2) determine the possible at 2 or 20 mmol/l glucose in 3 ml of KRBH medium containing 0.15 µCi/µmol of route(s) or shuttle(s) implicated in the cataplerotic output of [5-3H]glucose (Amersham, Piscataway, NJ). Glucose oxidation was measured as 14 14 2 citric acid cycle intermediates, and 3) directly test the action CO2 production from [U- C]glucose (0.1 µCi/µmol) (27). Cells grown in 21 cm of phenylacetic acid (PAA), a PC inhibitor (23), on insulin Petri dishes were incubated for 1 h at 37°C in 3 ml of KRBH medium at 2 or 20 mmol/l glucose in the presence of 0.07% BSA. At the end of the incubations, secretion promoted by various secretagogues. media were collected and transferred to 25-ml Erlenmeyer flasks, which were cov- ered with septa caps. Media were acidified by injecting perchloric acid with a RESEARCH DESIGN AND METHODS syringe (6% final concentration). The liberated CO2 was trapped in a plastic cup, suspended to the septa cap, which contained 0.4 ml of methanolic benzethonium Cell culture and incubation conditions. INS-1 cells were seeded in 21 cm2 petri hydroxide. After overnight incubation at room temperature, the cups were dishes (1.4 106 cells/dish) and grown in regular RPMI medium containing 10% removed and the trapped 14CO was measured by liquid scintillation counting. heat-inactivated fetal calf serum as described previously (24). When cells reached 2 Pancreatic islet isolation and incubation. Islets of Langerhans from Wistar 80% confluence after ~7 days, they were washed twice with phosphate-buffered rats (weighing ~200 g) were isolated by collagenase digestion of pancreases, saline (PBS) and preincubated at 37°C for 2 days in culture medium containing separated from the bulk of digested organs by centrifugation in a discontinuous 5 mmol/l glucose. Cells were then washed twice with a Krebs-Ringer bicarbon- Ficoll gradient, and selected by morphological criteria from exocrine cells under ate medium containing 1 mmol/l CaCl , 5 mmol/l NaHCO , 25 mmol/l HEPES (pH 2 3 the microscope (11). They were then incubated in regular RPMI medium con- 7.4) (Krebs-Ringer bicarbonate HEPES [KRBH] medium), 2–4 mmol/l glucose, and taining 10% heat-inactivated fetal calf serum and 11.1 mmol/l glucose. Islets (10 0.07% bovine serum albumin (BSA) (Fraction V, free from Sigma), and per condition) were preincubated in 1 ml of KRBH medium containing 2 mmol/l subsequently preincubated in the same KRBH medium for a period of 30 min at glucose and 0.07% BSA for 1 h. Media were removed by suction with a Pasteur 37°C. Cells were then incubated with various concentrations of glucose and test pipette and islets were further incubated in 1 ml of the same fresh KRBH medium substances. Some experiments were carried out in the presence of the cAMP rais- at 2 or 20 mmol/l glucose without or with 5 mmol/l PAA for 1 h. Insulin was assayed ing agent 3-isobutyl-1-methylxanthine (IBMX) (0.1 mmol/l). Thus, we noticed in the media at the beginning (0.1 ml aliquot) and the end of the incubation peri- during the course of this study that this compound allowed a more robust glucose- ods. Secreted insulin was the difference between that measured at the end of the stimulated insulin release. incubation and that measured at time zero. Citrate, malate, -ketoglutarate, and malonyl-CoA measurements. Krebs Statistical analysis. All results are expressed per milligram protein as means ± cycle intermediates were determined as follows. Incubation media were discarded SE of the indicated number of experiments. Statistical significance was calculated or collected for insulin measurement, and 0.5 ml of 13% perchloric acid was added with the unpaired Student’s t test. to the cells. Cells were scraped from the dishes, precipitated proteins were Materials. Phenylacetic acid and phenylpropionic acid were purchased from removed by centrifugation, and the supernatants were neutralized by adding 2N Aldrich Chemicals (Milwaukee, WI). In water, 100-fold concentrated stock solu- KH CO . The precipitated potassium perchlorate salt was eliminated by centrifu- 2 3 tions were prepared, and the pH value was adjusted to 7.4 with 1 N NaOH. gation, and the resulting supernatants were stored at –20°C and subsequently used for citric acid cycle intermediate measurements. Citrate was assayed by coupling the citrate lyase/malate (Boehringer Mannheim, Mannheim, Ger- RESULTS many) reactions according to Williamson and Corkey (25). Malate and -ketoglu- Increased anaplerosis is an early process of -cell tarate were determined using the and glutamate dehydro- metabolic signaling that does not require a rise in Ca2+. genase (Boehringer Mannheim) reactions, respectively (25). Malonyl-CoA was extracted from cells with 1 ml of 10% trichloroacetic acid. After centrifugation of Figure 1 shows that glucose stimulation of INS-1 -cells precipitated proteins, cell extracts were brought to pH 5–6 by successive ether caused a >2-fold rise in cellular citrate, which coincided with extractions. Samples were lyophilized and store at –70°C. Malonyl-CoA was the initiation of insulin secretion and was sustained for at least assayed according to McGarry et al. (26) and Segall et al. (27). Total cellular pro- 30 min. By contrast, the rise in malate was markedly delayed teins were measured by the Bradford assay (Bio-Rad, Mississauga, ON, Canada) after relative to that of citrate and did not occur at the onset of dissolution of the trichloroacetic acid precipitated protein pellets in 1N NaOH. Insulin secretion and extracellular citrate measurements. At the beginning insulin secretion. A maximal 4-fold elevation in malate was and the end of the incubation periods, media were collected for insulin mea- observed after 10 min of high glucose stimulation. It should surements by radioimmunoassay using guinea pig anti-rat insulin antibody and rat be mentioned that in another -cell line (HIT), the glucose- insulin as standards (Linco Research, St. Louis, MO) (27). Secreted insulin was induced rise in citrate occurring at 2 min was found to pre- the difference between the amount of insulin measured at the end of the incubation periods and that at time 0. Extracellular citrate was measured using 1 ml of incu- cede the initiation of insulin secretion (11). bation medium after sedimentation of cells that might have detached from the The dose dependence of insulin secretion in response to var- dishes at 1,500g for 10 min. Supernatants were extracted with perchloric acid as ious glucose concentrations tightly correlated with the citrate

DIABETES, VOL. 49, MAY 2000 719 CATAPLEROSIS AND INSULIN SECRETION

FIG. 2. The malonyl-CoA content of INS-1 cells correlates with the dose dependence of glucose-induced insulin release. Cells were incu- bated as described in the legend to Fig. 1 at various glucose concen- trations indicated in millimoles per liter (noted as numbers close to FIG. 1. Time course of citrate and malate accumulation in INS-1 cells. the values in the lower panel). Means ± SE of 5 experiments are Cells were preincubated for 30 min in a KRBH medium at 4 mmol/l glu- shown. cose and subsequently incubated for the indicated times at 4 mmol/l (G 4) or 20 mmol/l (G 20) glucose. Means ± SE of 5 experiments are that the initial increases in anaplerosis, malonyl-CoA, and shown. MTT reduction are all Ca2+ -independent processes. Consis- tent with this view, Fig. 4 shows that Ca2+ influx promoted by and malate content of INS cells after 30 min of incubation (7). elevated concentrations of KCl did not alter the content of cit- In fact, the correlation coefficients (y) of insulin secretion ver- rate and malate at either low or high glucose. sus citrate and malate were 0.93 and 0.96, respectively. To fur- Glucose promotes cataplerosis and citrate output from ther investigate whether anaplerosis/cataplerosis is tightly INS cells. To gain further insight into the fate of anaplerotic/ related to insulin secretion and may underlie the production cataplerotic carbons, citrate was measured in particulate of candidate metabolic coupling factors, malonyl-CoA and and cytosolic fractions after rapid cell fractionation of INS-1 insulin release were measured at various glucose concentra- cells previously incubated at low and high glucose. Saponin tions. It is apparent that citrate-derived malonyl-CoA levels treatment of insulinoma cells has been used before for intra- closely correlated with insulin secretion (Fig. 2). cellular Ca2+ (33) and mitochondrial metabolism To assess whether anaplerosis as well as the initial -cell (34) measurements; it is an optimal system for directly study- metabolic activation depends on glucose-induced Ca2+ influx, ing these complex functions in an environment with pre- the citric acid cycle intermediates, citrate and -ketoglu- served intracellular architecture but with the plasma mem- tarate, malonyl-CoA, and reduced 3-[4,5-dimethylthiazol-2yl]- brane bypassed. As far as citrate measurements are con- 2,5-diphenyltetrazolium (MTT) (an index of glucose oxidation cerned, its content in the particulate fraction should reflect rate and reducing equivalent production) (30,31), were mea- mostly the mitochondrial concentration of this . sured in the absence and presence of the Ca2+ chelator Thus, to our knowledge, citrate is concentrated only in this EGTA. Previous work showed that glucose does not pro- subcellular compartment and nuclear citrate is expected to mote a rise in cytosolic Ca2+ in islet cells and -cell lines in diffuse rapidly through nuclear pores into the cytosol and the presence of EGTA (32). The results in Fig. 3 demonstrate extracellular milieu of permeabilized cells. Similarly, citrate

720 DIABETES, VOL. 49, MAY 2000 S. FARFARI AND ASSOCIATES

FIG. 3. Glucose-induced anaplerosis, rise in malonyl-CoA, and production of accelerated reducing equivalents are Ca2+ independent. Cells were 2+ preincubated for 30 min at 4 mmol/l glucose in KRBH containing 1 mmol/l CaCl2. Cells were then washed twice with a Ca -free KRBH and sub- sequently incubated for 2 min in Ca2+-free KRBH in the absence or presence of 0.5 mmol/l EGTA. Means ± SE of 3–11 experiments are shown. OD, optical density. *P < 0.05. measured in the supernatant fraction should reflect its con- centration in the cytosol and the nucleus, of which the mem- brane is not a barrier to small and ions. Figure 5 indicates that elevated glucose promoted a rise in the content of this metabolite in both the particulate (mitochondrial) and cytosolic compartments. This finding demonstrates the occurrence of cataplerotic output of citrate from the mito- chondria, since citrate is not synthesized in the cytosol. The data also establish the feasibility of a pyruvate/citrate shuttle (see also Fig. 9). Citrate has been documented to enter a number of cell types, in particular, and intestine cells (35,36), via an Na+-dependent cotransporter and to exit some tissue, such as the intestine (35) and the heart (22,37), possibly via an anion exchanger (38). Therefore, we sought to investigate whether this phenomenon takes place in INS-1 cells and as such might represent a quantitatively important fate of anaplerotic car- bons. Interestingly, citrate output from INS cells occurred lin- early as a function of time at low glucose (Fig. 6). The citrate efflux rate at 20 mmol/l glucose was ~3 times that at low glu- cose (Fig. 6). Citrate efflux was not causally related to insulin granule release. Thus, the citrate content and insulin secre- tion rates in cells incubated with high KCl and epinephrine, which respectively stimulate and inhibit secretion indepen- dently of metabolic signaling, were dissociated. For instance, although 10 µmol/l epinephrine totally suppressed glucose- stimulated insulin release, it did not alter the rise in citrate pro- moted by sugar (data not shown).

Phenylacetic acid, a PC inhibitor, specifically reduces 2+ glucose-induced insulin secretion. A pharmacological FIG. 4. Accelerated Ca influx does not affect the citrate and malate content of INS-1 cells. Cells were incubated for 30 min at 4 or 20 mmol/l approach was used to directly test the hypothesis that glucose with 5 (Control) or 35 mmol/l KCl. Means ± SE of 8–10 exper- anaplerosis plays an important role in glucose-induced iments are shown. P < 0.001 for the values at high vs. low glucose.

DIABETES, VOL. 49, MAY 2000 721 CATAPLEROSIS AND INSULIN SECRETION

FIG. 5. Elevated glucose causes a rise in citrate in both the mito- FIG. 6. Glucose promotes citrate efflux from INS-1 cells. The figure chondrial and cytosolic compartments. Cells were incubated for 30 min shows the time dependence of citrate efflux from INS-1 cells at low at 4 or 20 mmol/l glucose. They were then scraped from the dishes, (3 mmol/l) and high (20 mmol/l) glucose. Cells were incubated as resuspended in KRBH, and incubated with saponin for 20 s. The par- described in the legend to Fig. 1, except that the basal glucose con- ticulate and cytosolic fractions were isolated by rapid cell fractiona- centration was 3 mmol/l instead of 4 mmol/l. Means ± SE of 3–6 exper- tion as detailed in RESEARCH DESIGN AND METHODS. Means + SE of 10 iments are shown. experiments are shown. *P < 0.05. acetyl-CoA (23). The results in Fig. 7 show that PAA dose insulin release. PAA, an inducer of tumor cell differentiation dependently inhibited both glucose-stimulated insulin secre- and a growth arrest agent (39,40), has recently been demon- tion and the rise in citrate. Insulin secretion as a function of strated to suppress both in vivo and in vitro liver gluconeo- inhibitor concentration closely correlated with the citrate genesis by inhibition of PC (23). In fact, its CoA derivative content of INS cells. Maximal effect occurred at ~10 mmol/l phenylacetyl-CoA is thought to mediate inhibition of the PAA. The specificity of the PAA effect is indicated by the enzyme by competing with the allosteric site activated by observation that PAA reduced secretion promoted by glucose

FIG. 7. The pyruvate carboxylase inhibitor phenylacetic acid reduces both glucose-induced insulin secretion and citrate accumulation in INS-1 cells. Cells were incu- bated for 45 min with 0.1 mmol/l IBMX at 3 or 20 mmol/l glucose (Glc) in the absence or presence of the indi- cated concentrations of PAA. Means ± SE of 5 experi- ments, except at 20 mmol/l PAA (n = 1), are shown. The numbers in parentheses in the lower panel indicate the tested concentrations of PAA (in mmol/l).

722 DIABETES, VOL. 49, MAY 2000 S. FARFARI AND ASSOCIATES

TABLE 1 Phenylacetate acid (10 mmol/l) does not impair glucose usage and oxidation in INS-1 cells

Glucose usage Glucose oxidation (nmol · mg–1 (nmol · mg–1 protein · h–1) protein · h–1) PAA– PAA+ PAA– PAA+

Glucose 2 mmol/l 15.8 ± 0.7 ND 3.5 ± 0.7 ND 20 mmol/l 88.3 ± 2.3 80.1 ± 1 22.0 ± 1.9 18.8 ± 1.1

Data are means ± SE of 6–8 independent determinations. ND, not determined.

potassium in INS-1 cells. It should be mentioned that PAA reduced liver from lactate/pyruvate but not from or aspartate (23). Furthermore, PAA did not alter glucose or palmitate oxidation in isolated rat hepatocytes (23). Consistent with the view that anaplerosis is an essential component of - is the fact that the dose depen- dence of insulin release highly correlates with the accumu- lation of citrate, malate, and citrate-derived malonyl-CoA. Furthermore, the rise in citrate and -ketoglutarate coin- cided with the initiation of insulin secretion and did not require a rise in Ca2+, indicating that anaplerosis is an early event of -cell activation. Cataplerosis occurs in INS-1 cells, at least in part, through mitochondrial citrate efflux as demonstrated using a rapid cell fractionation procedure and by the fact that citrate exits from the cells. The latter is quantitatively important. Thus, cit- FIG. 8. Phenylacetic acid specifically inhibits glucose-induced insulin –1 secretion in INS-1 cells and isolated rat islets. A: INS cells were incu- rate output from the cells at high glucose was 18 nmol · mg –1 bated with 0.1 mmol/l IBMX at 2 mmol/l glucose (G 2), without or with protein · h (means ± SE of 3 experiments) relative to methyl-succinate (Met-Succ) (10 mmol/l) or KCl (35 mmol/l), and at 88 nmol · mg–1 protein · h–1 of glucose usage. Thus, cata- 20 mmol/l glucose (G 20) in the absence or presence of 5 mmol/l PAA. plerotic citrate output is equivalent to ~20% of the flux of glu- Means ± SE of 10 experiments are shown. B: Islets were incubated for 1 h at 2 (G 2) or 20 (G 20) mmol/l glucose in the absence or presence cose carbons entering the glycolytic pathway. The citrate of 5 mmol/l PAA. Means ± SE of 10 experiments are shown. *P < 0.01. carbons escaping the cells are likely derived almost entirely from glucose. Thus, high glucose reduces oxida- tion (41), which occurs in part after input of amino acid car- but not by methyl-succinate, an anaplerotic fuel secreta- bons in the citric acid cycle via , such as gogue (1), or high KCl, a non- stimulus (Fig. 8A). that catalyzed by . Furthermore, Likewise, PAA did not affect glucose usage or glucose oxi- the cellular citrate content remained elevated at a constant dation (Table 1). Finally, PAA also markedly reduced glu- level at least from 2 to 30 min incubation, and citrate output cose-induced insulin release in isolated rat islets (Fig. 8B). was unrelated to insulin secretion, suggesting that secretory The action of phenylpropionic acid, another agent that granules do not contain appreciable amounts of citrate. It reduces hepatic gluconeogenesis, likely through PC inhibition remains to be determined whether this phenomenon is quan- (23), was tested. The following insulin secretion values were titatively as important in normal islet tissue. obtained: 2 mmol/l glucose, 18.3 ± 1.2; 20 mmol/l glucose, What might be the role of citrate in these 3 compartments 95.6 ± 6.1; 20 mmol/l glucose plus 10 mmol/l phenylpropionate, with respect to -cell function? The following possibilities can 51.7 ± 4.5 ng insulin · mg–1 protein · 30 min–1 (means ± SE, be envisaged. Mitochondrial citrate may be the precursor of n = 3). Thus, 2 PC inhibitors (PAA and phenylpropionic acid) cataplerotic signaling molecules such as malonyl-CoA and oth- reduce glucose-stimulated insulin secretion. ers (also called mitochondrial signals [1,19]) implicated in the KATP-independent pathway of secretion. Cytosolic citrate may DISCUSSION be an essential component of a pyruvate/citrate shuttle, gen- The results provide strong evidence that anaplerosis is impli- erating cytosolic acetyl-CoA, NAD+, and NADPH (Fig. 9). In cated in glucose signaling and insulin secretion. Thus, PAA, addition, it may regulate glycolytic oscillations through its a PC inhibitor, reduced glucose-stimulated insulin release in inhibitory action on 6-phosphofructo-1-kinase. Thus, this key both INS cells and pancreatic islets in association with a regulatory enzyme of is thought to control oscilla- reduction in the elevation in citrate promoted by sugar. The tions of -cell metabolism, KATP channel activity, and insulin specificity of PAA is supported by control experiments show- secretion (2,34). It cannot be discounted that citrate is a sig- ing that PAA neither altered glucose usage or oxidation nor nal directly influencing the exocytotic machinery. Extracel- impaired secretion in response to methyl-succinate and high lular citrate may simply reflect a sink for anaplerotic car-

DIABETES, VOL. 49, MAY 2000 723 CATAPLEROSIS AND INSULIN SECRETION

rises after glucose addition, and malic enzyme is expressed at very high levels in rat islets (14) and INS-1 cells (15). Mac- Donald (43) provided experimental support for a role of a pyruvate/malate shuttle in islet tissue, but did not favor the view that a pyruvate/citrate shuttle is quantitatively important in the -cell. This decision was based on the observation that citrate output from isolated islet mitochondria was min- imal in comparison to malate efflux. Our results are at vari- ance with this previous study with respect to the pyru- vate/citrate shuttle, possibly because intact cells represent a more preserved system in comparison to isolated mitochon- dria, where the activity of a tricarboxylic carrier might be impaired. Furthermore, citrate was shown to rapidly escape mitochondria after pyruvate stimulation of permeabilized HIT -cells (44), and glucose promotes citrate (45,46) and mal- onyl-CoA (47) accumulation in rat islets. Nonetheless, the cur- rent study does not contradict the view that a pyruvate/ malate shuttle may also be important for -cell signaling. Finally, additional evidence in favor of a pyruvate citrate shuttle in the -cell stems from a recent report addressing the role of shuttles in insulin secretion. Eto et al. (48) showed that suppressing the activities of the phos- phate and malate/aspartate shuttles impaired secretion in response to glucose; surprisingly, glucose usage was unaf- fected under this condition. If both shuttles were essential contributors to redox traffic between the cytosolic and mito- chondrial compartments, one would have predicted a lack of reoxidation of cytosolic NAD and, consequently, impaired gly- FIG. 9. Model illustrating the pyruvate/citrate shuttle and the fate and colysis. Because the latter was not observed (48), the results possible signaling roles of citrate. The possible functions of the high- indicate the existence of other means of preserving a high lighted metabolites are 1) mitochondrial citrate, which serves as a car- + bon precursor for cataplerotic mitochondrial signals (Krebs cycle NAD /NADH ratio in cytosol. The pyruvate/citrate shuttle intermediates and derived compounds); 2) cytosolic citrate, which provides a possible explanation for the dichotomy reported may control glycolytic oscillations through PFK-1 inhibition, serves as in this study (48). a carbon precursor for malonyl-CoA formation and promotes ACC As exemplified in Fig. 9, the occurrence of a pyruvate/cit- activation; 3) external citrate, which may synchronize metabolism in rate shuttle would present several advantages for -cell islet cells and insulin secretion oscillations; 4) malonyl-CoA, which reduces oxidation for full glucose usage, controls lipid partition- metabolism and signaling. First, the occurrence allows the ing, and may act as a coupling factor; 5) NAD+ regeneration for high production of malonyl-CoA to cause a metabolic switch glycolytic flux; and 6) cytosolic NADPH, which reduces flux through from fat to glucose oxidation and possibly the generation of the pentose phosphate pathway, thus sparing glucose-6-phosphate lipid messenger molecules (8,42). Second, it reoxidizes carbons for glycolysis, and may act as a coupling factor. Citrate efflux from the occurs in exchange with malate possibly cytosolic NAD for the glyceraldehyde 3-phosphate dehy- formed during the operation of citrate/malate or malate aspartate drogenase reaction, thus allowing sustained elevated gly- shuttles. See text for discussion. ACC, acetyl-CoA carboxylase; colytic flux. Third, like the pyruvate/malate shuttle, it pro- Ac-CoA, acetyl-CoA; ACL, ATP-citrate lyase; CS, citrate ; Ext., vides cytosolic NADPH, a candidate coupling factor (9). Fur- extracellular; Mal-CoA, malonyl-CoA; MDH, malate dehydrogenase; thermore, an increase in the cytosolic NADPH/NADP ratio OAA, ; PDH, ; Pyr, pyruvate. will inhibit the pentose phosphate pathway, as demon- strated before for glucose (7), thus sparing glucose-6-phos- bons at times of fuel abundance. More interestingly, it may be phate carbons for high glycolytic flux. The latter 3 consid- speculated that exogenous citrate may act as a coordinator erations are compatible with the fact that cytosolic NAD synchronizing glucose metabolism and secretion in and NADP are respectively oxidized and reduced relative to individual islets. Consistent with this possibility is the obser- the state of these pyridine in the mitochondrion vation that exogenous citrate promotes secretion of insulin (28,29). Furthermore, glucose causes an initial oxidation of in INS-1 cells, rat islets, and purified -cells (S.F., C. Alcarcon, NAD in the cytosol of islet tissue (49). K. Moens, F. Schuit, C. Rhodes, M.P., unpublished data). The question arises as to the fate of cytosolic acetyl-CoA The following considerations deriving from this and pre- carbons if the pyruvate/citrate shuttle was to be highly vious work underscore the feasibility of a pyruvate/citrate active. One possibility is the rapid of mal- shuttle in the -cell (Fig. 9) and its potential role in insulin onyl-CoA by cytosolic malonyl-CoA decarboxylase or mito- secretion. Thus, PC is highly expressed in the -cell (7,10,11), chondrial malonyl-CoA decarboxylase (50) after the entry of and citrate increases in response to glucose in both mito- malonyl-CoA into the mitochondrion via putative malonyl- chondrial and cytosolic compartments. Citrate, malate, and carnitine (51) or malonate (52) shuttles. The possible malonyl-CoA levels in response to glucose tightly correlate using NADPH have been discussed before (9) and with insulin release; hydroxy-citrate, a citrate lyase inhibitor, may include reductase, nitric oxide synthase, impairs secretion promoted by sugar (42). Malate markedly and fatty acid synthase.

724 DIABETES, VOL. 49, MAY 2000 S. FARFARI AND ASSOCIATES

In conclusion, the data suggest that anaplerosis via the PC 20. MacDonald MJ: Estimates of glycolysis, pyruvate (de), pentose reaction and cataplerotic output of citrate from the mito- phosphate pathway, and methyl succinate metabolism in incapacitated pan- chondria are important metabolic transducing events in creatic islets. Arch Biochem Biophys 305:205–214, 1993 21. Parkkila AK, Scarim AL, Parkkila S, Waheed A, Corbett JA, Sly WS: Expres- insulin secretion. The results also favor the view that a pyru- sion of carbonic anhydrase V in pancreatic beta cells suggests role for mito- vate/citrate shuttle is implicated in -cell activation and that chondrial carbonic anhydrase in insulin secretion. J Biol Chem 273:24620– citrate is not only a mitochondrial signal of nutrient availability 24623, 1998 (1), but also a cytosolic and extracellular signal of glucose 22. Comte B, Vincent G, Bouchard B, Jette M, Cordeau S, DesRosiers C: A 13C abundance in the -cell. mass isotopomer study of anaplerotic pyruvate carboxylation in perfused rat hearts. J Biol Chem 272:26125–26131, 1997 23. Bahl JJ, Matsuda M, DeFronzo RA, Bressler R: In vitro and in vivo suppres- ACKNOWLEDGMENTS sion of gluconeogenesis by inhibition of pyruvate carboxylase. Biochem This work was supported by grants from the Medical Pharmacol 53:67–74, 1997 Research Council of Canada, the Canadian Diabetes Associ- 24. Asfari M, Janjic D, Meda P, Li G, Halban PA, Wollheim CB: Establishment of 2-mercaptoethanol-dependent differentiated insulin-secreting cell lines. ation, and the Juvenile Diabetes Foundation (grant number Endocrinology 130:167–178, 1992 197047) (to M.P.), and National Institutes of Health Grant 25. Williamson JR, Corkey BE: Assay of citric acid cycle intermediates and DK35914 and DK50662 (to B.E.C.). M.P. is a Medical related compounds. Methods Enzymol 13:434–513, 1969 Research Council of Canada Scientist. 26. McGarry JD, Mannaerts GP, Foster DW: A possible role for malonyl-CoA in the We thank Dr. Christine DesRosiers for helpful discussions. regulation of hepatic fatty acid oxidation and ketogenesis. J Clin Invest 60:265–270, 1977 27. Segall L, Lameloise N, Assimacopoulos-Jeannet F, Roche E, Corkey P, REFERENCES Thumelin S, Corkey BE, Prentki M: Lipid rather than glucose metabolism is 1. Prentki M: New insights into pancreatic beta-cell metabolic signaling in implicated in altered insulin secretion caused by oleate in INS-1 cells. Am J insulin secretion. Eur J Endocrinol 134:272–286, 1996 Physiol 277:E521–E528, 1999 2. Prentki M, Corkey BE: Are the beta-cell signaling molecules malonyl-CoA and 28. Siess EA, Brocks DG, Wieland OH: Distribution of metabolites between the cytosolic long-chain acyl-CoA implicated in multiple tissue defects of obesity cytosolic and mitochondrial compartments of hepatocytes isolated from fed and NIDDM? Diabetes 45:273–283, 1996 rats. Hoppe Seylers Z Physiol Chem 359:785–798, 1978 3. Newgard CB, McGarry JD: Metabolic coupling factors in pancreatic beta-cell 29. Akerboom TPM, Bookelman H, Zuurendonk PF, Van der Meer R, Tager JM: signal transduction. Annu Rev Biochem 64:689–719, 1995 Intramitochondrial and extramitochondrial concentration of adenine 4. Komatsu M, Schermerhorn T, Noda M, Straub SG, Aizawa T, Sharp GW: Aug- nucleotides and inorganic phosphate in isolated hepatocytes from fasted mentation of insulin release by glucose in the absence of extracellular Ca2+: rats. Eur J Biochem 84:413–420, 1978 new insights into stimulus-secretion coupling. Diabetes 46:1928–1938, 1997 30. Janjic D, Wollheim CB: Islet cell metabolism is reflected by the MTT (tetra- 5. Gembal M, Gilon P, Henquin JC: Evidence that glucose can control insulin zolium) colorimetric assay. Diabetologia 35:482–485, 1992 release independently from its action on ATP-sensitive K+ channels in mouse 31. Roche E, Assimacopoulos-Jeannet F, Witters LA, Perruchoud B, Corkey BE, cells. J Clin Invest 89:1288–1295, 1992 Asfari M, Prentki M: Induction by glucose of genes coding for glycolytic 6. Corkey BE, Glennon MC, Chen KS, Deeney JT, Matschinsky FM, Prentki M: enzymes in a pancreatic -cell line (INS-1). J Biol Chem 272:3091–3098, 1997 2+ A role for malonyl-CoA in glucose-stimulated insulin secretion from clonal pan- 32. Prentki M, Matschinsky FM: Ca , cAMP, and phospholipid-derived messen- creatic beta-cells. J Biol Chem 264:21608–21612, 1989 gers in coupling mechanisms of insulin secretion. Physiol Rev 67:1185–1248, 7. Schuit F, De Vos A, Farfari S, Moens K, Pipeleers D, Brun T, Prentki M: Meta- 1987 bolic fate of glucose in purified islet cells: glucose-regulated anaplerosis in 33. Prentki M, Corkey BE, Matschinsky FM: Inositol 1,4,5-trisphosphate and the 2+ -cells. J Biol Chem 272:18572–18579, 1997 endoplasmic reticulum Ca cycle of a rat insulinoma cell line. J Biol Chem 8. Prentki M, Vischer S, Glennon MC, Regazzi R, Deeney JT, Corkey BE: Malonyl- 260:9185–9190, 1985 CoA and long chain acyl-CoA esters as metabolic coupling factors in nutrient- 34. Civelek VN, Deeney JT, Fusonie GE, Corkey BE, Tornheim K: Oscillations in induced insulin secretion. J Biol Chem 267:5802–5810, 1992 oxygen consumption by permeabilized clonal pancreatic beta-cells (HIT) 2+ 9. MacDonald MJ: Feasibility of a mitochondrial pyruvate malate shuttle in pan- incubated in an oscillatory glycolyzing muscle extract: roles of free Ca , creatic islets: further implication of cytosolic NADPH in insulin secretion. substrates, and the ATP/ADP ratio. Diabetes 46:51–56, 1997 J Biol Chem 270:20051–20058, 1995 35. Pajor AM: Citrate transport by the kidney and intestine. Semin Nephrol 19: 10. MacDonald MJ: Influence of glucose on pyruvate carboxylase expression in 195–200, 1999 pancreatic islets. Arch Biochem Biophys 319:128–132, 1995 36. Pajor AM: Sodium-coupled transporters for Krebs cycle intermediates. Annu 11. Brun T, Roche E, Assimacopoulos-Jeannet F, Corkey BE, Kim KH, Prentki M: Rev Physiol 61:663–682, 1999 Evidence for an anaplerotic/malonyl-CoA pathway in pancreatic beta-cell 37. Comte B, Vincent G, Bouchard B, Des Rosiers C: Probing the origin of acetyl- nutrient signaling. Diabetes 45:190–198, 1996 CoA and oxaloacetate entering the citric acid cycle from the 13C labeling of 12. MacDonald MJ, McKenzie DI, Walker TM, Kaysen JH: Lack of glyconeogen- citrate released by perfused rat hearts. J Biol Chem 272:26117–26124, 1997 esis in pancreatic islets: expression of gluconeogenic enzyme genes in islets. 38. Wolffram S, Unternahrer R, Grenacher B, Scharrer E: Transport of citrate Horm Metab Res 24:158–160, 1992 across the brush border and basolateral membrane of rat small intestine. Comp 13. Dean PM: Ultrastructural morphometry of the pancreatic -cell. Diabetologia Biochem Physiol Physiol 109:39–52, 1994 9:115–119, 1973 39. Prasanna P, Shack S, Wilson VL, Samid D: Phenylacetate in chemoprevention: 14. Ashcroft SJH, Randle PJ: Enzymes of glucose metabolism in normal mouse in vitro and in vivo suppression of 5-aza-2-deoxycytidine-induced carcino- pancreatic islets. Biochem J 119:5–15, 1970 genesis. Clin Res 1:865–871, 1995 15. Roche E, Farfari S, Witters LA, Assimacopoulos-Jeannet F, Thumelin S, Brun 40. Samid D, Hudgins WR, Shack S, Liu L, Prasanna P, Myers CE: Phenylacetate T, Corkey BE, Saha AK, Prentki M: Long-term exposure of -INS cells to high and phenylbutyrate as novel, nontoxic differentiation inducers (Review). glucose increases anaplerosis, and lipogenic gene expression. Adv Exp Med Biol 400A:501–505, 1997 Diabetes 47:1086–1094, 1998 41. Malaisse WJ, Sener A, Carpinelli AR, Anjaneyulu K, Lebrun P, Herchuelz A, 16. MacDonald MJ: Glucose enters mitochondrial metabolism via both carboxy- Christophe J: The stimulus-secretion coupling of glucose-induced insulin lation and decarboxylation of pyruvate in pancreatic islets. Metabolism release. XLVI. Physiological role of L-glutamine as a fuel for pancreatic islets. 42:1229–1231, 1993 Mol Cell Endocrinol 20:171–189, 1980 17. Khan A, Ling ZC, Landau BR: Quantifying the carboxylation of pyruvate in pan- 42. Chen S, Ogawa A, Ohneda M, Unger RH, Foster DW, McGarry JD: More direct creatic islets. J Biol Chem 271:2539–2542, 1996 evidence for a malonyl-CoA-carnitine palmitoyltransferase I interaction as a 18. MacDonald MJ, Fahien LA, Mertz RJ, Rana RS: Effect of esters of key event in pancreatic beta-cell signaling. Diabetes 43:878–883, 1994 and other citric acid cycle intermediates on insulin release and inositol phos- 43. MacDonald MJ: Evidence for the malate aspartate shuttle in pancreatic islets. phate formation by pancreatic islets. Arch Biochem Biophys 269:400–406, 1989 Arch Biochem Biophys 213:643–649, 1982 19. Maechler P, Kennedy ED, Pozzan T, Wollheim CB: Mitochondrial activation 44. Civelek VN, Deeney JT, Shalosky NJ, Tornheim K, Hansford RG, Prentki M, directly triggers the exocytosis of insulin in permeabilized pancreatic -cells. Corkey BE: Regulation of pancreatic beta-cell mitochondrial metabolism: EMBO J 16:3833–3841, 1997 influence of Ca2+, substrate and ADP. Biochem J 318:615–621, 1996

DIABETES, VOL. 49, MAY 2000 725 CATAPLEROSIS AND INSULIN SECRETION

45. Matschinsky FM, Rutherford CR, Ellerman JE: Accumulation of citrate in pan- 985, 1999 creatic islets of obese hyperglycemic mice. Biochem Biophys Res Commun 49. Matschinsky FM, Ghosh AK, Meglasson MD, Prentki M, June V, von Allman 33:855–862, 1968 D: Metabolic concomitants in pure, pancreatic beta cells during glucose-stim- 46. Hellman B, Idahl LA, Danielsson A: levels of mam- ulated insulin secretion. J Biol Chem 261:14057–14061, 1986 malian pancreatic -cell after stimulation with glucose and hypoglycemic 50. Voilley N, Roduit R, Vicaretti R, Bonny C, Waeber G, Dyck JB, Lopaschuk GD, sulfonylureas. Diabetes 18:509–516, 1969 Prentki M: Cloning and expression of rat pancreatic -cell malonyl-CoA 47. Liang Y, Matschinsky FM: Content of CoA-esters in perifused rat islets stim- decarboxylase. Biochem J 340:213–217, 1999 ulated by glucose and other fuels. Diabetes 40:327–333, 1991 51. Yano S, Sweetman L, Thorburn DR, Mofidi S, Williams JC: A new case of mal- 48. Eto K, Tsubamoto Y, Terauchi Y, Sugiyama T, Kishimoto T, Takahashi N, onyl decarboxylase deficiency presenting with cardiomyopa- Yamauchi N, Kubota N, Murayama S, Aizawa T, Akanuma Y, Aizawa S, Kasai thy. Eur J Pediatr 156:382–383, 1997 H, Yazaki Y, Kadowaki T: Role of NADH shuttle system in glucose-induced acti- 52. Koeppen AH, Riley KM: Effect of free malonate on the utilization of glutamate vation of mitochondrial metabolism and insulin secretion. Science 283:981– by rat brain mitochondria. J Neurochem 48:1509–1515, 1987

726 DIABETES, VOL. 49, MAY 2000