AMP-activated protein kinase: function and dysfunction in health and disease

Rising levels of obesity and inactivity in many parts of the developed world are leading to pandemic proportions of so-called ‘metabolic syndrome’, with its accompanying symptoms of insulin resistance, hypertension, dyslipidaemia and elevated pro-inflammatory/thrombotic state. These symptoms in turn give rise to a significant increase in risk for a variety of disease states including type 2 diabetes mellitus, coronary heart disease, atherosclerosis and stroke. The widespread scale of metabolic syndrome has major healthcare cost implications for many nations that are currently driving national and international governmental health initiatives and guidelines aimed at minimizing future crises through disease prevention. It is estimated that up to 50 million people in the USA alone may have this syndrome and the prevalence in the UK has been put at around 25% of the population.

Whilst there appears to be a genetic component in a number of cases, the largest risk factor for metabolic syndrome – obesity – may be largely preventable. Pharmaceutical therapies are therefore combined with advice on altering diet and lifestyle. Unfortunately, whilst most people appear to be aware of the risk of overeating and inactivity, persuasion to alter lifestyle does not appear to have halted the increasing incidence of obesity with its associated morbidity and mortality. Much of this failure must be due to the particular pressures of modern living with its high demands on an individual's time and the prevalence of ‘convenience’ foods excessively rich in fats, sugar and salt. However, failure to curb increasing obesity may also be due to a lack of our understanding of how metabolism is sensed and controlled in individuals during health as well as disease. Greater understanding would clearly lead to an improvement in disease risk and, it would be hoped, disease prevention. Such hopes lie behind much of the research reviewed in this issue of The Journal of Physiology.

To sustain metabolism, intracellular ATP concentration must be regulated within an appropriate range. This is achieved both at the subcellular, biochemical level as well as at the systemic level, encompassing regulation of cellular anabolic and catabolic pathways, substrate storage and release as well as the hormonal and cytokine regulation of food intake and satiety. This Special Issue of The Journal of Physiology brings together a number of invited reviews that focus upon how this co-ordination may be achieved through the mechanisms and functions of AMP-activated protein kinase (AMPK). This kinase is typically, but not exclusively, activated by an increase in the cellular AMP: ATP ratio and, once activated, enables increased energy production whilst decreasing non-essential energy consuming pathways. AMPK can also modulate transcription of specific genes implicated in energy metabolism, thereby exerting longer-term metabolic control. It is thus apparent how a dysfunction in the AMPK signalling pathway might have sustained, deleterious effects at a systemic level. The growing realization that this pathway may be of significant clinical importance is revealed by a cursory search demonstrating that the annual number of papers published that have AMPK in their title alone has risen by more than 10-fold over the past 8 years.

The particular timeliness of this Special Issue lies, we believe, with the relatively recent developments in our understanding of precisely how this key, heterotrimeric enzyme complex might operate to control metabolism at both the cellular and organismal level. This is due in part to the development of transgenic rodent models that have enabled specific experiments to be performed that would not have been possible just a few years ago, and also through the increasingly higher throughput techniques of molecular and cell biology. The fascinating thought that AMPK might have evolved as this singular, energy conserving function from single cell eukaryotes through to more complex multicellular systems, involving hormonal and neuronal control, is explored in a number of reviews and is, to us, particularly captivating. An explicit emphasis in this issue of The Journal is on the potential role of AMPK in disease: as well as diseases associated with metabolic syndrome mentioned above, roles for AMPK in the aetiology and/or treatment of cancer, pulmonary hypertension and hepatic injury are also described. Although the series of reviews, taken together, form a valuable, integrated resource, it was not our intention to generate an interdependent sequence of papers and so each author was requested to provide a ‘stand-alone’ article that could be downloaded and read independently, if required. That said, there is a somewhat logical order to the sequence of papers and the committed reader is recommended to read all reviews in the order presented.

As with other emerging fields, caution is required when interpreting findings based upon incomplete knowledge of either pharmacological or genetic manipulation and it is noteworthy that many authors have provided these warnings to prevent possible misinterpretation of the data. In addition, it is becoming apparent that the widespread and various cellular and systemic functions of AMPK make its selective targeting in therapeutics a difficult one with, simultaneously, both advantageous and deleterious consequences possible. This theme is commented upon in a number of these reviews, but the matter remains unresolved.

This Special Issue begins with a review from the group that first correctly identified, and named, AMP-activated protein kinase as a multisubstrate enzyme. This report (Hardie et al. 2006) serves as an introduction to the development of the concept of AMPK as an ‘energy sensor’. As well as describing the key upstream kinases of AMPK, particularly the tumour suppressor kinase, LKB1 and a calmodulin-dependent protein kinase kinase, CaMKKβ, and its many downstream targets, the authors also provide a clear exposition of the multifarious ways in which a change in cellular energy status might occur. This review is recommended as a starting point for those new to this field as well as those keen to see the directions in which it might be heading. The authors point out that, whilst AMPK dysfunction does not, in itself, appear to induce type 2 diabetes, its low activation, particularly in obese and under-active individuals, might be a major contributory factor as exemplified by the beneficial use of certain classes of widely available antidiabetic drugs that are now known to activate AMPK.

The large range of energy turnover that is possible in skeletal muscle, between its resting and active states, requires significant variation in ATP synthesis that is itself dependent upon the uptake and cellular utilization of glucose and fatty acids. Jørgensen et al. (2006) provide a detailed account of experiments that have demonstrated activation of AMPK in both in vivo and in vitro models of exercise and provide an excellent summary of its key, regulatory effects upon glucose uptake and glycogen and fatty acid metabolism as well as on protein synthesis. In addition, these authors speculate on a role for AMPK in mediating the alteration in metabolic capacity of this tissue via specific expression of certain genes, for example those encoding GLUT-4 and hexokinase II, which might underlie the beneficial effects of exercise training. However, the AMPK signalling pathway is not the only one activated during exercise and the authors are mindful of the somewhat less than conclusive findings from some of the available mouse models, noting that ‘correlations do not necessarily imply causality’.

Decreased mitochondrial protein expression might underlie the confounding issues raised by Jørgensen et al. (2006), and the review by Reznick & Shulman (2006) describes the long-known association between exercise and mitochondrial content before commenting on recent studies utilizing, amongst other approaches, chronic pharmacological activation of AMPK via depletion of muscle phosphocreatine stores and the use of transgenic mice expressing a dominant negative mutant form of skeletal muscle AMPK. These studies suggest a role for AMPK in mitochondrial biogenesis that may be mediated via regulation of the PGC-1α–NRF pathway and this correlates well with the more physiological activation of AMPK via exercise. Activation of this pathway may also be responsible for the increased expression of GLUT-4 protein in the adaptation of muscle to exercise (see Jørgensen et al. 2006). Interestingly, Reznick & Shulman (2006) also report on human studies that demonstrate how mild mitochondrial dysfunction might play a role in insulin resistance and/or type 2 diabetes and, perhaps more intriguingly, in the process of ageing. As endurance exercise can partially reverse the defects in AMPK signalling and PGC-1α protein expression, the authors comment upon its potential therapeutic benefit in type 2 diabetes, but one is left wondering also about the reversal of the ageing process! However, the authors allude to the difficulty of ascribing cause and effect in such chronic studies and this caution should be heeded in similar protocols.

Further insight into possible therapeutic approaches to obesity and type 2 diabetes, via pharmacological manipulation of hepatic AMPK, is provided by Viollet et al. (2006). Additional, potentially therapeutic applications for AMPK manipulation are also highlighted by these authors, including use in the management of hepatic steatosis subsequent to chronic alcohol consumption and in ischaemic preconditioning prior to liver transplantation. The authors firstly provide, however, a detailed description of the more physiological function of hepatic metabolism in regulating systemic energy status, via alterations in lipid and glucose homeostasis and protein synthesis during periods of feeding and fasting, highlighting the role of AMPK in this regulation. They also introduce the concepts of circulating adipocyte-derived hormones and the potential for AMPK-related, anti-cell proliferation treatments for cancer – topics explored further by Daval et al. (2006) and Motoshima et al. (2006), respectively, in this Issue. As pointed out above, AMPK has been implicated in mediating long-term effects via alterations in gene expression in skeletal muscle. This also appears to be the case for glucose- and lipid-related gene expression in the liver, and Viollet et al. (2006) review the evidence for this, including their own studies utilizing adenoviral gene transfer techniques to induce short-term over expression of AMPK. However, as not all downstream targets of AMPK activation might prove beneficial to the diabetic patient, these authors end their review with a consideration of some of the potentially damaging consequences of uncontrolled AMPK-stimulated fatty acid oxidation and the inhibition of protein synthesis.

Daval et al. (2006) focus upon the vital role of activated AMPK in adipose tissue during exercise, fasting and in the presence of hypoglycaemic drugs. These issues are considered both in relation to this tissue's role in energy storage as well as its less-well recognized role as an endocrine organ secreting the anorexigenic cytokine, leptin, and the insulin-sensitizing hormone, adiponectin. With regard to its energy storage role, the authors describe how a similar functional role for AMPK in adipocytes (as described above for hepatocytes) might be expected during energy depleted states but, utilizing evidence both from overexpression of constitutively active and from dominant negative forms of AMPK in adipocytes, these authors forward an interesting proposal, suggesting that as well as being activated by increased lipolysis, AMPK might itself limit lipolysis in adipocytes as an important feedback mechanism to conserve energy. As with many studies in this field, however, the findings are not equivocal and may depend, for example, upon species and/or selective cell lines but it does raise the beguiling question of how the varying requirements for energy conservation might be best served in the organism during extended periods of high demand or shortage.

Ultimately, cellular metabolism requires the continued delivery of oxidisable substrate. Two papers in this Special Issue describe the integrative, regulatory role played by AMPK in the hypothalamus in the control of feeding and/or the counter-regulatory hormonal responses to hypoglycaemia. Xue & Kahn (2006) and Ramamurthy & Ronnett (2006) describe how anorexigenic stimuli, including leptin and insulin as well as the orexigenic hormones, grehlin and cannibinoids, might, respectively, decrease or increase AMPK activity in specific, feeding-related nuclei of the hypothalamus. A reduction in hypothalamic AMPK activity, induced by an decrease in the cellular AMP: ATP ratio could thus generate weight loss through a reduction in food intake. Similar anorexigenic effects are observed in mice by modulation of fatty acid synthase and carnitine palmitoyltransferase by the drug C75 and by compound C, a newly available inhibitor of AMPK which can also prevent the hormonal-based recovery from hypoglycaemia. These actions may alter the set point of energy balance towards a more satiated state in the whole animal via alteration in neuropeptide gene expression. Ramamurthy & Ronnett (2006) do caution, however, that these particular findings may be of more relevance in extreme conditions of chronic fasting or overfeeding. Xue & Kahn (2006) focus, additionally, upon the tissue-specific, peripheral actions of adipokines and other hormones to develop the theme of a co-ordinated peripheral and central (hypothalamic) regulation of energy homeostasis and suggest that many of the peripheral effects of AMPK might, in fact, be mediated by the CNS. These authors also describe how a failure to regulate AMPK in both hypothalamic neurones and peripheral tissue might play a part in the leptin resistance observed in diet-induced obesity. Together with Ramamurthy & Ronnett (2006), they recommend further studies to determine the regulation and identity of upstream kinases and the nature of the downstream AMPK targets in the control of feeding.

Turning to specific tissues, Dyck & Lopaschuk, 2006) describe the ‘Jekyll and Hyde’ actions of AMPK upon cardiac tissue during myocardial ischaemia, apoptosis and cardiac hypertrophy. Like Viollet et al. (2006) in this Issue, these authors also raise the complex but important question of how AMPK activation or inhibition might best be applied therapeutically. Despite the heart being a highly metabolic tissue, Dyck & Lopaschuk (2006) cast doubt as to whether cellular [AMP] might play any role in activating cardiac AMPK during aerobic metabolism, or indeed if AMPK activation is involved at all in physiological, exercise-induced increases in cardiac work. However, they leave little doubt of its activation during metabolic stress, when the rate of glycolysis accelerates in order to generate ATP, and following such stress (e.g. during reperfusion) when fatty acid oxidation is promoted by AMPK at the expense of glucose oxidation, leading to decreased cardiac function and efficiency. In addition, these authors describe data that suggest a causal link between adiponectin stimulation and pressure overload hypertrophy that may be associated with the cardiomyopathy of obesity and diabetes. It is clear, however, that the many in vitro models used in these studies cannot easily replicate the complete in vivo condition and thus the authors are currently unable to state definitively whether AMPK deficiency might be beneficial or harmful during ischaemia-reperfusion. They are, however, able to suggest ways in which future studies might best be conducted to gain further insight into the critical role of AMPK upon cardiac function. Similarly, Ramamurthy & Ronnett (2006), in this Issue, describe recent evidence to suggest an interaction between over-production of nitric oxide or peroxynitrite and AMPK activation in stroke, but comment on how AMPK activation might be either neuroprotective or pro-apoptotic depending upon the specific stoke model utilized.

Evans (2006) provides a departure from the somewhat ‘classical’ downstream targets of AMPK activation by suggesting a specific role for this kinase as an O₂ sensor in tissues of the pulmonary vasculature and the carotid body. This is particularly intriguing as it suggests an even more complex role for AMPK in systemic energy balance, in this case by regulation of O₂ supply. The author points out the evolutionary conservation of this chemotransducing function in lampreys, fish and amphibia. In this review, Evans (2006) compares and contrasts the O₂ sensing mechanisms of pulmonary artery smooth muscle cells (PASMCs) with that of the type I cells of the carotid body and attempts to provide a unifying hypothesis to link the hypoxia mediated increase in the AMP: ATP ratio in these two cell types with cell-specific elevations in intracellular Ca²⁺ which are fundamental to both tissues' responses to hypoxia. Thus, whilst AMPK activation in PASMCs, either by pharmacological intervention or by hypoxia, induced intracellular Ca²⁺ release from sarcoplasmic reticulum via a cADPR-dependant effect upon ryanodine receptors, in carotid body type I cells, the same interventions induced Ca²⁺ entry via voltage gated Ca²⁺ channels subsequent to inactivation of plasmalemmal K⁺ channels. The author speculates that the presence of significantly higher levels of the α1-catalytic subunit isoform might confer acute O₂ sensing to tissues whilst the α2-catalytic subunit isoform may be involved in longer-term gene expression mediated via the transcription factor, HIF-1α. This interesting possibility requires further experimental proof but could provide a novel therapeutic target for conditions ranging from pulmonary hypertension to obesity-related sleep apnoea.

Given the presence of tumour suppressor elements both upstream and downstream in the AMPK signalling pathway, it is not surprising that AMPK has been targeted as a potential therapeutic agent in disease states related to cell proliferation, including atherosclerosis and cancer. Motoshima et al. (2006) in the final paper of this issue provide a concise and clear overview of cell proliferation both in health and disease and describe how AMPK might arrest cell cycling at the G₁ phase in smooth muscle cells, by action of downstream p53 and/or upstream LKB1 proteins, thereby overcoming endogenous growth-promoting signals. Additionally, these authors describe an inhibitory action of AMPK upon mTOR regulated protein synthesis and translations thus reducing, potentially, cell growth and division at times of substrate limitation. Interestingly, Motoshima et al. (2006) also make clear the possible anticancer effects of circulating plasma adiponectin, as well as the antiproliferative effects of phenformin – both presumably acting via AMPK activation – and conclude by suggesting future avenues for research including more temporal studies to determine the long-term effects of AMPK activation.

Our aim for this Special Issue of The Journal of Physiology is to integrate the biochemical, pharmacological and clinical aspects of current AMPK research and thereby to promote its importance in physiological research. Thanks to the excellent (and punctual!) efforts of the contributing authors, this issue provides a concise, yet wide-ranging introduction to many of the key aspects of AMP-activated protein kinase. It is our hope that it will stimulate further interest amongst physiologists in this fascinating protein.