Overexpression of Monocarboxylate Transporter-1 (Slc16a1) in Mouse Pancreatic β-Cells Leads to Relative Hyperinsulinism During Exercise

The murine Mct1 coding sequence was amplified from pUC18-Mct1 (14), with the addition of a single COOH-terminal c-Myc epitope tag by PCR (primer sequences in Supplementary Table 1), and inserted between the NheI and EcoRV sites of plasmid pBI-L Tet (Clontech). This generated a plasmid with a bidirectional tetracycline-regulated promoter driving expression of both Mct1∷Myc and firefly luciferase.

The expression cassette was excised from the plasmid backbone by AatII and AseI digestion and transferred by pronuclear microinjection into C57Bl/6 CBA F1 hybrid mouse oocytes (Geneta, Leicester, U.K.). Three successful integrants were identified by PCR screening of DNA extracted from ear biopsies by the HotSHOT method (15) (primer sequences in Supplementary Table 2). Founders were backcrossed five times with C57Bl/6 mice to generate the Mct1-Luc strain. RIP7-rtTA mice on a C57Bl/6 background (16), expressing the reverse tetracycline transactivator under the control of the rat insulin promoter, were crossed with Mct1-Luc mice to permit β-cell–specific, tetracycline-inducible expression of Mct1∷Myc and luciferase. Of three founder lines, two produced transgenic offspring, one of which was propagated for subsequent experiments. Heterozygous Mct1-Luc mice were crossed with heterozygous RIP7-rtTA mice to produce littermates of four genotypes as follows: wild-type mice bearing no transgenes (−/−) and both single-transgenic (ST, RIP7-rtTA/− and −/Mct1-Luc), and double-transgenic (DT, RIP7-rtTA/Mct1-Luc) mice. All offspring were genotyped for both transgenes. Of these, only DT mice were expected to express the transgenes on induction with doxycycline. All procedures with animals received ethical approval and were compliant with both U.K. (Animals Scientific Procedures Act, 1986) and Danish legislation.

Pancreatic islets were isolated by collagenase digestion (17) from noninduced DT animals. Islets from each mouse were divided into two dishes containing approximately equal distributions of variously sized islets. Islets were cultured for 48 h in the presence or absence of 5 μg/mL doxycycline and then used for luciferase assay to measure transgene expression or used for insulin secretion assays. Luciferase activity was determined with the Luciferase Assay System (Promega) on a Lumat LB9507 luminometer (Berthold Technologies). Three batches of islets from each induction state were assayed from each of three mice. Insulin secretion from isolated islets was assessed essentially as described elsewhere (18).

Transgene expression was induced by administering 1 g/L doxycycline in autoclaved distilled water for at least 5 days. RNA was extracted from islets and other tissues from three DT mice after doxycycline treatment and from three control DT mice not treated with doxycycline. Mct1 mRNA was quantified relative to cyclophilin A by quantitative RT-PCR.

MCT1 expression in islets was assessed by immunofluorescence of pancreatic sections prepared from ST and DT mice after doxycycline treatment as described previously (19).

Littermates of ST (RIP7-rtTA/-) and DT (RIP7-rtTA/Mct1-Luc) animals were fasted overnight (16 h), and fasting blood glucose levels were measured in tail blood samples. Sodium pyruvate (0.5 g/kg) was administered by intraperitoneal injection, and blood glucose was measured at 15, 30, 60, 90, and 120 min after injection with the Accu-Chek Compact Plus glucometer (Roche). Transgene expression was induced by administering doxycycline in drinking water as before, and the pyruvate tolerance test was repeated on induced animals 1 week after the previous test. Doxycycline treatment was maintained for an additional week before glucose tolerance tests were performed (1 g/kg glucose, same procedure as pyruvate tolerance test).

All mice were acclimatized to treadmill exercise before experiments on days 1, 3, and 5 by being placed on the inactive treadmill for 10 min, followed by running for 5 min at 10 m/min and 4 min at 16 m/min (5% incline). Maximum running capacity was determined on day 7. Predoxycycline blood glucose levels were measured after a 4-h fast on day 10. Mice were administered with doxycycline as previously described for 2 weeks. Maximum running capacity after doxycycline treatment was determined 3 days before the exercise experiment. Because of limitations on the volume of blood taken in a single day, blood samples were taken from tail vein 2 days before exercise experiment after a 4-h fast to determine basal levels of plasma insulin, pyruvate, and lactate. On the experimental day, mice were fasted for 4 h and then exercised for 20 min at 80% maximum running speed at 5% incline. Blood glucose was determined by tail vein sampling 2 min before exercise and after 10- and 20-min periods of exercise. Blood was also collected directly after exercise to determine plasma levels of insulin, pyruvate, and lactate.

Insulin levels relative to glucose were calculated as homeostasis model assessment 2 (HOMA2) of β-cell function (HOMA2-B%) according to the HOMA2 model by means of HOMA Calculator software (20). The original HOMA1 model corrects for the nonproportional response of β-cells to glucose, according to the following equation:

The updated HOMA2 model (21) provides a further refinement of this calculation (which cannot be reduced to a simple formula).

Differences were determined by Student t test and one- or two-way ANOVA, with Dunnett or Bonferroni post hoc test as appropriate. A P value of <0.05 was considered statistically significant.

RIP7-rtTA mice have previously been used successfully to create a number of transgenic mouse strains in which β-cell–specific expression of transgenes is tightly regulated by doxycycline (22). In this experiment, we generated a Mct1-Luc mouse strain harboring the coding sequences of COOH-terminally epitope-tagged MCT1 and firefly luciferase under the control of a bidirectional, tetracycline-regulated promoter. Mct1-Luc mice were crossed with RIP7-rtTA mice producing offspring of four genotypes as follows: −/−, RIP7-rtTA/−, −/Mct1-Luc, and RIP7-rtTA/Mct1-Luc (Fig. 1A). Only doxycycline-treated DT mice (RIP7-rtTA/Mct1-Luc) were expected to express Mct1 and luciferase in the β-cell, so other genotypes were used as controls. In initial experiments, the two ST and wild-type genotypes all produced similar results, so only results from the RIP7-rtTA/− ST genotype are presented here as controls.

To assess the effectiveness of transgene regulation by the tetracycline analog doxycycline, luciferase expression was compared between noninduced and induced islets from the same mouse. Islets were isolated from DT mice and cultured in the presence or absence of 5 μg/mL doxycycline for 48 h. Batches of six size-matched islets were lysed, and luciferase activity was measured in the lysate (Fig. 1B). Luciferase activity increased more than tenfold on induction with doxycycline. This result also gives a good indication of the regulation of the Mct1 transgene, because both genes are expressed from the same promoter (22).

Overexpression of MCT1 in vivo was confirmed by immunofluorescence of pancreatic sections from doxycycline-treated mice. Strong staining with an anti-MCT1 antibody was observed in the islets of DT mice but was absent from islets of ST control mice (Fig. 1C). Quantification of the anti-MCT1 signal intensity within insulin-positive areas of the islets showed significant upregulation (Fig. 1D) in DT mice. An anti-MCT1 signal was observed in 68 ± 9% (mean ± SEM) of insulin-positive cells in DT islets (n = 3), and no MCT1 was detected in glucagon-positive cells (Supplementary Fig. 1).

The tissue-specificity of Mct1 overexpression was determined by quantitative RT-PCR in tissues extracted from DT mice treated or not with doxycycline in vivo (Fig. 1E). A significant increase in Mct1 expression was detected in islets, whereas doxycycline had no effect on Mct1 expression in any other tissue examined.

After confirmation of transgene expression, we next studied its effects on insulin secretion from isolated islets. It is noteworthy that the inducible transgenic model developed here allowed islets from a single animal to be compared with and without transgene induction, thus excluding interanimal variation. Islets isolated from a single animal were divided and cultured in the presence or absence of doxycycline for 48 h before being subjected to insulin secretion assays (Fig. 2). MCT1 overexpression exerted no clear effect on insulin secretion in response to glucose at either 8 or 17 mmol/L. This suggests that the expected increased permeability of the β-cell plasma membrane to pyruvate in response to MCT1 induction does not lead to any loss of glycolysis-generated pyruvate from the cell. Whereas the addition of 1 mmol/L or 10 mmol/L pyruvate exerted no effect on noninduced islets, the monocarboxylate significantly and substantially stimulated insulin secretion from islets overexpressing Mct1. Although the effect of 0.1 mmol/L pyruvate was not statistically significant, a tendency toward increase insulin secretion was observed (Fig. 2). These effects were specific to pyruvate, with lactate having no significant stimulatory effect at the same concentrations, consistent with previous findings (11). Having confirmed that Mct1 expression renders insulin secretion specifically sensitive to pyruvate in vitro, we went on to study its effects on glucose homeostasis in the context of the whole animal.

Without doxycycline treatment, blood glucose levels after an overnight fast were similar between ST and DT genotypes (Fig. 3A). After doxycycline treatment, however, DT animals displayed significantly lower fasting blood glucose by an average of 1.5 ± 0.3 mmol/L relative to ST controls (P < 0.001) (Fig. 3A). It is noteworthy that lowered fasting blood glucose has also been observed in human subjects affected by EIHI (1). Although the fasting plasma insulin levels were not significantly different between mouse genotypes, there was a nonsignificant trend toward lower insulin levels in DT animals. Because there were no significant differences between males and females, data from both sexes were combined for this assessment and for pyruvate tolerance tests. After 3 weeks of doxycycline treatment, there were no differences in behavior or body weight (Supplementary Fig. 2) between ST and DT animals.

EIHI is characterized by the ability of pyruvate to stimulate insulin secretion, lowering blood glucose (1). To test whether induced Mct1 expression in β-cells may be sufficient to cause this effect in vivo, blood glucose was measured after intraperitoneal injection of pyruvate. Under these conditions, pyruvate is likely to be metabolized largely by the liver, serving as a gluconeogenic substrate in control animals and leading to glucose output and an increase in blood glucose levels (23). However, should pyruvate also stimulate insulin secretion, hepatic glucose output is likely to be lowered and the excursion in blood glucose diminished.

Initially, a relatively low dose (0.5 g/kg) of pyruvate was administered to avoid increasing blood glucose levels excessively. Pyruvate tolerance was first tested in mice before treatment with doxycycline (Fig. 3C and D). The drug was then administered in drinking water for 1 week, and a second pyruvate tolerance test was then performed on the induced animals (Fig. 3C and D). Before doxycycline treatment, the glucose excursion was similar in both genotypes, with no significant differences detected. After transgene induction, however, the DT animals displayed a slower initial increase in blood glucose in response to pyruvate administration, resulting in significantly lower levels at the 15- and 30-min time points. At 120 min, blood glucose levels had dropped significantly lower in DT animals than in ST animals. These effects are consistent with pyruvate stimulation of insulin secretion lowering blood glucose in DT animals relative to ST controls.

We next attempted directly to detect an increase in plasma insulin in response to injection of this level of pyruvate. Low basal plasma insulin levels after overnight fast approached the assay’s detection limit, however, meaning that any increase was too small to be detected over the sample-to-sample variation. To address this, we increased the level of pyruvate administered to 2 g/kg, a level routinely used in pyruvate tolerance tests (24,25). Intraperitoneal injection at this dose led to substantial increases in both plasma pyruvate and lactate, and these changes were not affected by MCT1 overexpression (Fig. 3F and G). The increased plasma pyruvate stimulated a significant and rapid increase in plasma insulin only in DT animals (Fig. 3E), confirming that MCT1 expression renders insulin secretion sensitive to pyruvate in vivo.

The effect of MCT1 overexpression on glucose tolerance was next studied in doxycycline-treated ST and DT mice. Because differences in glucose tolerance were observed between males and females, data from each sex were analyzed separately. Whereas ST males were less glucose tolerant than corresponding females, there was no significant effect of MCT1 expression in either case (Fig. 3H and I).

The key characteristic of EIHI is inappropriate stimulation of insulin secretion during physical exercise. To test whether forced expression of MCT1 in β-cells is sufficient to replicate this phenotype, we investigated the effect of treadmill exercise on the transgenic mice. The running capacities of male DT mice and their ST littermate controls were measured both before and after doxycycline treatment. The drug exerted no significant effect on the running capacity of mice of either genotype (Fig. 4A). In contrast, doxycycline induced a significant decrease in blood glucose level after a 4-h fast in DT mice (Fig. 4B). One exercise bout at 80% of maximal running speed for 20 min induced a decrease in blood glucose level in both genotypes (Fig. 4C). The magnitude of the decrease (ST 1.82 ± 0.61 vs. DT 1.89 ± 0.30 mmol/L) was similar between genotypes; however, because the initial level was lower in the DT mice, the level fell to 3.3 ± 0.6 mmol/L relative to 5.4 ± 0.6 mmol/L in ST controls (Fig. 4C). In ST controls, the drop in blood glucose was accompanied by a significant fall in plasma insulin (Fig. 4D). Moreover, there was no significant change in HOMA2-B%, a measure of insulin relative to glucose (21,26), before and after exercise (93 ± 11 vs. 103 ± 22%) (Fig. 4E) in these mice.

DT mice displayed a nonsignificant trend toward higher fasting HOMA2-B% values than seen in ST mice (118 ± 28 vs. 93 ± 11%), suggestive of a higher level of insulin normalized to glucose. In response to exercise, no decrease in plasma insulin was observed despite a drop in glucose levels. This stable insulin level in the face of falling glucose in DT mice suggests that a stimulatory mechanism compensated for the expected exercise-mediated inhibition of insulin secretion. In supporting of this view and in contrast to ST controls, we observed a sharp increase in the HOMA2-B% value after exercise in DT mice (DT 486 ± 167 vs. ST 103 ± 22%; P < 0.05) (Fig. 4E).

To assess whether the imposed level of exercise might raise circulating pyruvate levels, we compared basal lactate and pyruvate levels with those immediately after exercise (Fig. 4F and G). Exercise produced a significant increase in plasma lactate, which was similar (ST 1.05 ± 0.35 vs. DT 1.46 ± 0.35 mmol/L) in the two genotypes. The expected lower levels of pyruvate rendered measurements of this parameter more difficult (Fig. 4G), and here no clear changes or differences between genotypes were observed. Providing confirmation that altered stress levels did not contribute to the results observed, no significant differences were observed between plasma corticosterone concentrations in ST and DT mice after exercise (Supplementary Fig. 3).

The principal aim of this study was to determine whether controlled overexpression of MCT1 selectively in pancreatic β-cells recapitulates the key features of EIHI, notably pyruvate- and exercise-induced insulin secretion and hypoglycemia. This has seemed important in view of the fact that mutations linked to the disease are likely to increase the expression of MCT1 in all tissues (2), raising the possibility that multiple tissue interactions may be involved in altered glucose homeostasis.

Although pyruvate and lactate are reportedly metabolized by islets (27), it seems likely that this occurs largely in the non–β-cells comprising 20–30% of the islet mass, as well as in endothelial and possibly contaminating acinar cells (see Ishihara et al. [11] for discussion). Previous studies (11) have demonstrated that pyruvate-induced insulin secretion can be unmasked by adenovirus-mediated overexpression of MCT1 in isolated islets. It is noteworthy that in these earlier studies, lactate failed to stimulate insulin release even after MCT1 overexpression, consistent with the absence of significant levels of LDHA activity in the β-cell. These earlier in vitro findings, in which the overexpression is likely only to have affected peripheral islet cells (28), were recapitulated here by means of transgene induction after MCT-1 overexpression throughout the islet β-cell mass with the small-molecule regulator doxycycline.

It is noteworthy that we now extend these findings to the in vivo case and show that MCT1 induction in the β-cell markedly improves pyruvate-induced excursions in blood glucose as a result of enhanced insulin release. By contrast, neither the studies reported by Ishihara et al. (11) nor our findings revealed any differences in glucose-induced insulin secretion in vitro or in glucose tolerance in vivo after MCT1 induction. This observation indicates that strict MCT1 suppression is not essential for the normal triggering of secretion by glucose. We note that although the mouse model that we used was generated by random insertion of the transgene, an off-target effect resulting from disruption of a nontargeted gene seems unlikely given that the metabolic phenotype was only apparent after gene induction with doxycycline.

One of the common characteristics of human patients with EIHI and the doxycycline-induced DT mice described here is a significantly lowered fasting blood glucose level. In both cases, this seems likely to be a reflection of the fact that normal (low) levels of circulating pyruvate constantly augment insulin secretion, thus suppressing hepatic glucose output and stimulating glucose clearance by other tissues. In the mouse model, however, fasting insulin levels (Figs. 3B and 4D) were not significantly different between ST and DT mice, although a slight trend toward lowered insulin in the DT mice was observed. Although this result may appear surprising, two points should be made. First, the observed lowering in insulin was proportionately less than the lowering in glucose level, such that β-cell function (as gauged in our calculated value of HOMA2-B%) was enhanced in the DT relative to the ST mice (Fig. 4E). Second, our measurements from tail-vein samples reflect systemic insulin levels. In mice of either genotype, concentrations of the hormone in the portal vein are likely to be higher than those in the general circulation, reflecting significant hepatic extraction (29). Indeed, portal vein insulin concentrations may conceivably be higher in DT mice than in ST mice after induction and might exert a more powerful inhibitory effect on hepatic glucose output in the former group.

One of the main differences between our rodent model relative to human EIHI was the degree to which physical exercise raised lactate and pyruvate levels. Moderately strenuous bicycle exercise for 10 min in human patients with EIHI and control subjects caused lactate levels to rise from a basal value of <1 mmol/L to ∼6 and 7 mmol/L, respectively (1). In contrast, exercising the mice at 80% of their maximum running speed for 20 min caused a far more modest increase, from 3 to 4.1–4.5 mmol/L. Although this change was still significant, it was clearly far less than that seen in the human subjects. Moreover, the bicycle exercise increased pyruvate levels in humans in a similar manner, from ∼50 to ∼200 μmol/L (1). Although the basal pyruvate level in the mice was similar at ∼50 μmol/L, the combination of a low concentration of pyruvate in blood and the considerably smaller volumes of blood we were able to obtain from mice relative to human subjects introduced considerable variability into our pyruvate measurements. Nonetheless, we anticipated that that a significant increase in plasma lactate would be accompanied by an increase in plasma pyruvate broadly similar in magnitude, as was the case with human exercise. Given the relatively modest increase in lactate, the expected similarly modest increase in pyruvate may explain why glucose levels did not drop further in DT mice in response to exercise. Although it is possible that we would have observed higher plasma lactate and pyruvate levels if we had run the mice at higher speeds, we note that the mice were already running at 80% of their maximum speed, and an increased exercise level thus would have been difficult to maintain for more than a few minutes without causing considerable stress to the animals.

Despite the relatively modest increase in lactate (and presumably pyruvate) in response to exercise, it appears that this increase was sufficient to affect β-cell function in DT mice. Thus, although blood glucose dropped in both genotypes during exercise, only in DT mice did the insulin level remain stable. We conclude that pyruvate stimulation of insulin secretion allowed by MCT1 expression compensates in DT mice for the exercise-induced inhibition of insulin secretion observed in ST mice. Thus, overexpression of MCT1 in β-cells is sufficient to replicate the key symptom of EIHI, highlighting the importance of maintaining the β-cell–specific block on the expression of this gene.

This study was supported by grants from the Wellcome Trust (Programme 081958/Z/07/Z) and from the Danish Medical Research Council. G.A.R. is a Royal Society Wolfson Research Merit Award holder.

No potential conflicts of interest relevant to this article were reported.

T.J.P. researched data and wrote the manuscript. L.S., G.S., and E.A.R. researched data. A.P.H. provided reagents and reviewed the manuscript. G.A.R. wrote the manuscript, coordinated the study, and is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented as a poster at the 71st Scientific Sessions of the American Diabetes Association, San Diego, California, 24–28 June 2011.

The authors thank Jorge Ferrer (Institut d’Investigacions Biomèdiques August Pi i Sunyer, Barcelona, Spain) for providing DNA constructs and Gerhard Christofori (University of Basel, Basel, Switzerland) for supplying the RIP7-rtTA mice.