PMC

Changes in relative beta cell mass, which might have explained the marked decrease in insulin release in βAMPKdKO animals, were not observed in βAMPKdKO mice, as assessed by optical projection tomography (OPT) of whole pancreata [] (; videos betaAMPKhet and betaAMKPdKO) or through analysis of pancreatic slices (). However, the distribution of islet sizes between heterozygous and βAMPKdKO animals, as assessed by OPT, revealed a significant ~40% decrease in the average volume of individual islets ().

Since mice with global homozygous deletion of AMPKα2 display abnormal insulin secretion in vivo and elevated catecholamine levels [], we determined whether the latter parameter may contribute to abnormal insulin secretion in βAMPKdKO mice. Indicating that this was not the case, we detected no alterations in the levels of urinary catecholamines (), and abnormal glucose tolerance was still observed, though diminished in extent, in the presence of the α-adrenoreceptor blocker, phentolamine ()[].

By contrast, no differences were apparent in the ratio of alpha to beta cells within individual islets, nor with the relative disposition of the two cell types (). Similarly, despite proposed roles for AMPK and the upstream kinase LKB1 in the control beta cell polarity [-](Sun et al, unpublished), we observed no abnormalities in the formation of adherens (anti-E-cadherin antibodies) () or tight (anti-zona occludins-1, ZO-1, antibody) junctions () and microfilament and microtubule structure was unchanged in islets from βAMPKdKO mice (not shown).

Maintained on normal chow, AMPK.CA and AMPK.DN transgenic displayed normal body weight increases (). Whereas male AMPK.CA Tg mice displayed abnormal glucose tolerance at three (), but not six () months, no abnormalities were seen in AMPK.DN Tg animals () nor in female AMPK.CA mice (not shown). The abnormalities in the AMPK.CA Tg mice were not associated with any alterations in insulin sensitivity () but with decreased fasting and stimulated plasma insulin levels () and a 25% decrease in beta cell area (0.09±0.02% in wild-type versus 0.068±0.04% in transgenic littermates) (data not shown). No significant changes in these parameters were observed in AMPK.DN Tg mice.

Male βAMPKdKO mice displayed entirely normal growth and normal food intake () whilst females displayed a small reduction in body weight up to ten weeks of age (). Assessed at three () and six months (not shown) of age, βAMPKdKO mice of either sex, but not mice deleted for either AMPKα1 [] or α2 () alone, displayed markedly elevated plasma glucose and decreased plasma insulin levels ; ). Correspondingly, double knockout mice displayed abnormal glucose tolerance and insulin release in vivo, despite increased insulin sensitivity (; ). No differences in glucose tolerance, insulin release or sensitivity were observed between AMPKα1^+/+,α2^+/+,Cre⁺ and AMPK,α1^+/−,α2^fl/+,Cre⁺ control mice (not shown).

An expression vector containing the RIP2 promoter fragment (600bp), c-myc-tagged rat AMPKα1³¹².T172D (CA) or AMPKα2.D157A (DN) cDNA and an SV40 poly (A) cassette () was excised with BssHII and microinjected into the male pronucleus of fertilized C57BL/6 oocytes. The injected zygotes were re-implanted into pseudo-pregnant female C57BL/6 mice (GenOway, France). We obtained three AMPK.CA (named C1, C2 and C10) and two AMPK.DN (D1 and D2) founder mice that stably transferred the corresponding transgenes to their offspring. Founder mice were crossed with wild type C57BL/6 mice to achieve F1 generation. Distributions of genotypes in the offspring followed a Mendelian pattern. All AMPK transgenic mice were kept heterozygous. F3 and later generations and their littermate wild type controls were used for experiments. All lines were maintained on a pure C57BL/6 background.

To determine whether increases in AMPK imposed selectively in beta cells may affect insulin secretion in vivo, we generated transgenic mice in which the constitutively active enzyme (“AMPK.CA”) [] was expressed under the control of the insulin promoter (). Of two founder mouse lines generated (), we examined one line carrying two transgene copies in detail. Over-expression of the mRNA was clearly evident in islets (), and under the control of glucose ex vivo (as expected for expression under the insulin promoter; ) but barely (<0.001% of the islet level) in the hypothalamus or other tissues (), reflecting the more restricted expression of the RIP2 transgene to the beta cell in adult mice [].

mRNAs encoding both AMPKα1 and α2 subunits were present in highly purified [,] wild-type mouse beta cells (). AMPKα1 mRNA was ~15-fold more abundant than that encoding the α2 subunit, in line with previous AMPK activity measurements in clonal beta cells [] and with distinct roles for each isoform in these cells []. Since mice deleted globally for either subunit display essentially normal insulin release in vitro [], we generated trigenic mice inactivated for both α1 and α2 subunits selectively in beta cells and a small population of hypothalamic neurons. Mice globally inactivated for AMPKα [] were first crossed with animals bearing flox’d AMPKα2 gene alleles (, left). Crossing with RIP2.Cre mice [] led to a selective loss of the catalytic domain of the AMPKα2 subunit (aa189-260) from islets and hypothalamus (, right). Consistent with the abundant expression in islets of the Cre transgene (, left), and deletion of the both catalytic subunits selectively from beta cells [], the crossing of AMPKα2 flox’d and RIP.Cre mice resulted in a decrease in islet AMPKα2 mRNA of 60-70% (, right). Assessed at low glucose concentrations to near-maximally stimulate the enzyme [], total islet AMPK activity was decreased by 93% in AMPKα1^−/− versus AMPKα1^+/− (α2^fl/fl.Cre^- in each case) mice, and by 95% in βAMPKdKO mice versus AMPKα1^+/− ().

We next determined whether AMPK present in beta cells and RIP.Cre neurons may contribute to the deleterious effects of a high fat diet (HFD) on insulin release and glucose homeostasis []. No differences in body weight gain were apparent between heterozygous and βAMPKdKO mice maintained on HFD (). However, after six weeks on HFD, sufficient to cause profound abnormalities in glucose-induced insulin secretion in C57BL/6 mice [], the differences in glycemia observed between heterozygote and βAMPKdKO mice observed on a normal diet () were abolished (). This change reflected a more dramatic increase in glucose levels in the heterozygous mice on HFD versus normal chow (compare and ). Likewise, the difference in glucose tolerance between βAMPKdKO and heterozygous mice observed on normal chow (, ~5 mmol/l at 60 min) was substantially reduced when the comparison between genotypes was performed in mice maintained on HFD (; ~ 2 mmol/l at 60 min.). By contrast, the enhanced insulin sensitivity of βAMPKdKO mice (; ~20 % of the initial glucose level 30 min. post insulin injection) was maintained or enhanced on HFD (; ~ 30 % of initial glucose). Strikingly, the 3.5-fold decrease in glucose-stimulated insulin secretion observed in islets from heterozygous mice maintained on HFD versus normal chow (from ~0.7 to 0.2 %/30 min; vs ) was reduced in βAMPKdKO mouse islets to ~0.6 fold (from ~1.3 to 0.8%/30 min.). Hence, βAMPKdKO mice were still susceptible to the effects of HFD, albeit to a lesser extent than controls.