LKB1 couples glucose metabolism to insulin secretion in mice
- 1.5k Downloads
Precise regulation of insulin secretion by the pancreatic beta cell is essential for the maintenance of glucose homeostasis. Insulin secretory activity is initiated by the stepwise breakdown of ambient glucose to increase cellular ATP via glycolysis and mitochondrial respiration. Knockout of Lkb1, the gene encoding liver kinase B1 (LKB1) from the beta cell in mice enhances insulin secretory activity by an undefined mechanism. Here, we sought to determine the molecular basis for how deletion of Lkb1 promotes insulin secretion.
To explore the role of LKB1 on individual steps in the insulin secretion pathway, we used mitochondrial functional analyses, electrophysiology and metabolic tracing coupled with by gas chromatography and mass spectrometry.
Beta cells lacking LKB1 surprisingly display impaired mitochondrial metabolism and lower ATP levels following glucose stimulation, yet compensate for this by upregulating both uptake and synthesis of glutamine, leading to increased production of citrate. Furthermore, under low glucose conditions, Lkb1 −/− beta cells fail to inhibit acetyl-CoA carboxylase 1 (ACC1), the rate-limiting enzyme in lipid synthesis, and consequently accumulate NEFA and display increased membrane excitability.
Taken together, our data show that LKB1 plays a critical role in coupling glucose metabolism to insulin secretion, and factors in addition to ATP act as coupling intermediates between feeding cues and secretion. Our data suggest that beta cells lacking LKB1 could be used as a system to identify additional molecular events that connect metabolism to cellular excitation in the insulin secretion pathway.
KeywordsACC1 AMPK Glucose metabolism Glutamine metabolism Insulin secretion LKB1 Pancreatic beta cells
Acetyl-CoA carboxylase 1
AMP-activated protein kinase
Extracellular acidification rate
Electron transport chain
Green fluorescent protein
Glucose-stimulated insulin secretion
ATP-sensitive potassium channel
ATP-sensitive inward rectifier potassium channel
Lkb1 adult beta cell knockout
Liver kinase B1
Oxygen consumption rate
Readily releasable pool
Short hairpin RNA
Tetramethylrhodamine ethyl ester
The initiating steps in the glucose-stimulated insulin secretion (GSIS) pathway involve the stepwise breakdown of glucose through glycolysis and oxidation of the glycolytic product, pyruvate, in mitochondria. ATP-sensitive potassium channels (KATP channels) on the plasma membrane sense the increase in ATP and consequently close [1, 2, 3, 4], triggering membrane depolarisation and opening of voltage-sensitive L-type Ca2+ channels [4, 5]. The secretory response is biphasic; within minutes a readily releasable pool (RRP) of granules docked at the plasma membrane is released [6, 7, 8]. A second phase, which requires recruitment and trafficking of insulin granules to the plasma membrane, follows . The central role of the KATP channel in beta cell function is highlighted by deficiencies in both first and second phase GSIS in islets from sulfonylurea 1 (SUR1)−/− and ATP-sensitive inward rectifier potassium channel (KIR6.2)−/− mice . In humans, inappropriate inhibition or activation of KATP channel activity arising from mutations in KIR6.2 and SUR1 underlie congenital hyperinsulinaemia or neonatal diabetes, respectively [11, 12, 13].
Our understanding of precisely how glucose is coupled to insulin output is still incomplete. Several lines of evidence have indicated that the secretory response is more intricate than the classical model outlined above would suggest. For example, an increase in cytosolic ATP and closure of KATP channels does not always correlate with insulin output in GSIS . Beta cells can also secrete insulin independently of KATP channel closure by an unclear mechanism that still requires calcium influx and membrane depolarisation [10, 14]. Furthermore, KATP channel activity can be recorded from on-cell patches on beta cells exposed to glucose-free solutions, despite the fact that the intracellular ATP concentration under similar conditions would predict the channels to be closed . Finally, the presence of depolarising agents and the allosteric KATP channel activator diazoxide can further enhance insulin secretion in response to glucose . As a result, additional factors that drive metabolism-dependent secretion, such as mitochondrial metabolites (NADPH, GTP, phosphoenol pyruvate [PEP], α-ketoglutarate [α-KG]), and other secondary glucose-derived metabolites such as glutamate [17, 18] must exist.
We previously demonstrated that Lkb1 adult beta cell knockout (LABKO) mice are resistant to hyperglycaemia induced by a high-fat diet, owing in large part to a doubling of beta cell mass and enhanced insulin stores via mTOR activation . However, targeting liver kinase B1 (LKB1) in the beta cell for the treatment of diabetes would be undesirable as knockout of Lkb1 in adult mice has a deleterious effect on haematopoietic stem cell maturation [20, 21, 22]. We and others have identified non-overlapping functions in beta cells for LKB1 targets, including MAP/microtubule affinity-regulating kinase 2 (MARK2), salt-inducible kinase 2 (SIK2), brain-specific kinase 2 (BRSK2) as well as AMP-activated protein kinase (AMPK) itself [19, 23, 24, 25, 26, 27, 28], underscoring the potential for identifying the minimal target(s) of LKB1 responsible for enhancing insulin secretion. Here, we systematically interrogated insulin secretion pathway components to identify the signals responsible for enhanced insulin secretion in Lkb1 knockout islets.
Control LoxP/LoxP (L/L) and LABKO mice were obtained from mating homozygous Lkb1 floxed mice to Pdx1-CreER tam mice all on an FVB/n background as described previously  (electronic supplementary material [ESM] Methods).
Antibodies and short hairpin RNA sequences
Antibodies and short hairpin RNA (shRNA) sequences are included in ESM Methods.
GSIS and calcium experiments
Electron microscopy and insulin immunogold labelling
Mitochondrial membrane potential and content measurements
Mitochondrial inner membrane potential was measured in dispersed islet cells using tetramethylrhodamine ethyl ester (TMRE) and normalised to Mitotracker green (MTG; Invitrogen, Burlington, ON, Canada) intensity (ESM Methods).
NAD(P)H autofluorescence imaging
Islets were incubated in 2 mmol/l glucose for 1 h and loaded into microfluidic devices and placed in a heating chamber on a Zeiss LSM710 microscope (Zeiss Canada, Toronto, ON, Canada) and imaged as described previously .
NEFA quantification in islets
NEFA levels in chloroform extracts from 50 islets per mouse were determined using a colorimetric assay (Abcam, Toronto, ON, Canada).
Membrane potential recordings and whole cell patch-clamp
106 MIN6 cells per condition were infected with lentivirus expressing non-silencing control or Lkb1 shRNA. At 72 h post-infection, cells were starved for 1 h in KRBH + 1 mmol/l glucose prior to pulsing. Cells were pulsed with 10 mmol/l [U-13C6]glucose or 2 mmol/l [U-13C5]glutamine for the indicated times and compared with starved and unlabelled carbon source samples. GC-MS analysis was performed as described previously  (ESM Methods).
Loss of LKB1 in mouse beta cells increases insulin granule size and plasma membrane docking
LKB1 regulates glucose sensing
We previously observed an enhanced rate of calcium influx in response to glucose in LKB1 knockdown (LKB1i) MIN6 cells, raising the possibility that LKB1 regulates sensitivity to feeding cues. To determine if glucose responsiveness is enhanced in the absence of LKB1, we performed islet perifusion experiments. A glucose ramp (2.8 to 16.7 mmol/l) revealed an increase in maximal release at near saturating glucose concentrations in the GSIS response curve, with the response from LABKO islets to 8 mmol/l glucose equal to that of L/L control islets to 16.7 mmol/l glucose (Fig. 1c). Taken together, these data suggest that, in addition to having an increase in insulin cargo per cell, LABKO beta cells are also more sensitive to glucose.
LKB1 depletion increases sensitivity to the anti-diabetic drug tolbutamide
Enhanced membrane depolarisation and calcium influx in LABKO beta cells
Consistent with the increased tolbutamide-induced membrane depolarisation, maximal calcium response to tolbutamide in MIN6 cells was achieved at a tenfold lower concentration after silencing LKB1 (Fig. 3e), while calcium levels under basal conditions were unaffected (ESM Fig. 3a). A similar effect was observed in dispersed LABKO beta cells (ESM Fig. 4a). Treatment with the L-type specific calcium channel inhibitor diltiazem blocked this influx, indicating that the extracellular milieu and not intracellular stores was the source of calcium ions (ESM Fig. 4a, b). Taken together, these data reveal that LABKO beta cells secrete more insulin in part due to sensitisation of calcium influx.
LKB1 is required for mitochondrial respiration in beta cells
The reduction in medium ECAR in LABKO cells (ESM Fig. 5b) suggested that impaired generation of a metabolic intermediate may underlie the defect in mitochondrial respiration. Indeed, we observed a reduced ATP/ADP ratio in MIN6 cells (Fig. 4g) as well as lower ATP levels and mitochondrial membrane potential (ΔΨ) in LABKO islets (Fig. 4h; ESM Fig. 6). These changes occurred without a reduction in mitochondrial mass (ESM Fig. 7). These data indicate that if provided with sufficient fuel, the electron transport chain (ETC) in LKB1-deficient cells can support electron transport.
Increased glutamine flux in beta cells lacking LKB1
Interestingly, levels of glucose-derived glutamine were elevated in LKB1 knockdown cells (Fig. 5b), which can be used as an alternative cellular carbon source for anabolic processes. To identify the fate of glutamine, we pulsed MIN6 cells with 2 mmol/l [13C5]glutamine for 10 and 60 min and measured incorporation into several intermediate metabolites using GC-MS (Fig. 5c). In cells silenced for Lkb1, there was a trend towards increased glutamine uptake and levels of glutamate, as well as a significant early increase in α-KG levels (Fig. 5d). Furthermore, while pyruvate levels were unaffected, citrate production from glutamine was significantly increased at all time points (Fig. 5d and data not shown). Taken together, these data indicate that in beta cells lacking LKB1, the breakdown of glucose is enhanced, resulting in an increase in citrate and glutamine abundance. That this occurs in the absence of a corresponding increase in glucose-derived TCA intermediates indicates that citrate is being redirected away from the TCA cycle and used for other biosynthetic pathways. Consistent with this observation, labelled carbons from glutamine contribute less to α-KG than to citrate, suggesting that this pool of citrate is shunted away from α-KG production.
Enhanced acetyl-CoA carboxylase 1 activity is required for insulin in LABKO islets
LKB1 depletion uncouples glucose metabolism from calcium influx and insulin secretion
We and others previously reported that deletion of Lkb1 in adult mouse beta cells leads to increased insulin secretion [19, 26]. Here, we report that LKB1 acts to attenuate KATP channel closure and calcium influx. This is evidenced by an enhanced closure of KATP channels, calcium influx and insulin secretion in response to tolbutamide in LABKO beta cells, even in the absence of glucose. Given the reduction in energy status in LABKO beta cells under both basal and stimulated conditions, the enhancement in insulin secretion is paradoxical. Taken together, our data are consistent with the notion that LKB1 regulates both exocytotic events as well as metabolic amplification of calcium influx through an alternate coupling factor that is not ATP and involves citrate and ACC1.
Metabolic rewiring in LABKO beta cells
That impaired mitochondrial respiration in LABKO islets can be restored with constitutively active AMPK indicates the LKB1–AMPK pathway may directly control mitochondrial bioenergetics. While provision of additional substrate can restore full ETC activity in beta cells lacking LKB1, following glucose stimulation there is a reduction by half in their bioenergetic capacity. Furthermore, in the absence of LKB1, glucose usage favours citrate and glutamine production, and glucose-derived glutamine is used to generate citrate, contributing to higher citrate levels. This metabolic shift towards glutaminolysis may help to maintain both the higher basal mitochondrial membrane potential and NADPH levels in LABKO β cells in which respiration is attenuated. Glutamine itself is known to amplify GSIS and enhance KATP channel-independent release , and may serve as a coupling factor to amplify metabolism-secretion coupling in LABKO beta cells. This idea is consistent with recent work using beta cells in which Lkb1 is deleted (using the Ins1–Cre system), wherein loss of LKB1 or AMPK results in a metabolic switch favouring glutamine metabolism .
Taken together, it appears that in cells lacking LKB1, fuel is redirected away from the TCA cycle to favour biosynthetic pathways such as amino acid synthesis, or NEFA and lipid generation for de novo granule generation, or perhaps for use in a signalling role. As ACC1 has been shown to be required for pyruvate cycling and generation of TCA intermediates [41, 42], we conclude that enhanced insulin secretion from LABKO beta cells is due at least in part to enhanced ACC1 activity and increased citrate production from glutamine. Taken together, these observations support the idea that activating ACC1 or elevating citrate levels may be beneficial for enhancing insulin secretion in settings of hypoinsulinaemia.
We would like to acknowledge members of the Screaton laboratory for helpful discussions and commentary on the manuscript, T. Alves and R. Kibbey (Department of Internal Medicine, Yale University, New Haven, CT, USA) for assistance with metabolomics studies and helpful discussions, P. Rippstein (University of Ottawa Heart Institute, Ottawa, Canada) for assistance with Electron Microscopy, J. McBane (University of Ottawa, Ottawa, ON, Canada) for assistance with islet isolations. Adenovirus expressing AMPK-CA was a kind gift of J. Dyck (Department of Pediatrics, University of Alberta, Edmonton, AB, Canada).
This work was supported by grants to RAS from the Canadian Institutes of Health Research (CIHR; MOP #111186), Canadian Foundation for Innovation (CFI), and by a Studentship from the Canadian Diabetes Association to AF. RAS acknowledges the Canada Research Chair program for support.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
AF researched data and wrote the manuscript. KR, BF and ABH researched data and edited the manuscript. CR, XQD, KSS, SO and JVR researched data and revised the article. OYAD, MBW and PEM designed methods, supervised data acquisition and edited the manuscript. RJ contributed to project design, supervised data acquisition and edited the manuscript. RAS designed the project and wrote the manuscript. All authors approved the final version. RAS is the guarantor of this work.
- 38.Metallo CM, Gameiro PA, Bell EL et al (2012) Reductive glutamine metabolism by IDH1 mediates lipogenesis under hypoxia. Nature 481:380–384Google Scholar