The great metformin mystery–genomics, diabetes, and cancer


Galega officinalis (Goat’s Rue) From JoJan

Metformin (Bristol-Myers Squibb’s Glucophage, generics), an oral biguanide antidiabetic drug, is the most widely prescribed agent for treatment of type 2 diabetes. The drug mainly works by lowering glucose production by the liver, and thus lowering fasting blood glucose.

Although metformin–approved in the United States in 1994, and in Europe prior to that–has been used for many years, its mechanism of action is not well understood. In 2005, signal-transduction pioneer Lewis Cantley (Beth Israel Deaconess Cancer Center/Harvard Medical School, Boston MA), and his colleagues–including Reuben J. Shaw (now at the Howard Hughes Medical Institute, The Salk Institute for Biological Studies, La Jolla, CA)–published a report showing that metformin targets the adenosine monophosphate (AMP)-activated kinase (AMPK) pathway in the liver. We discussed this report and its implications in our 2007 Cambridge Healthtech Institute Insight Pharma Report, Diabetes and Its Complications.

AMPK is found in all eukaryotic organisms, and serves as a sensor of intracellular energy status. In mammals, it also is involved in maintaining whole-body energy balance, and helps regulate food intake and body weight. We  have discussed the potential role of AMPK in regulation of lifespan, and as a target in anti-aging medicine and in metabolic disease, in earlier articles on this blog. (See here and here.)

AMPK is activated by increases in the ratio of AMP to ATP, caused by energy stress. Under conditions of energy stress, AMP levels go up, and AMP binds to a specific site on the AMPK γ subunit. This induces a conformational change that exposes the activation loop of the α subunit. This allows an upstream serine/threonine kinase to phosphorylate this activation loop. In several mammalian cell types, including liver and skeletal muscle, that kinase is LKB1. Drs. Cantley and Shaw in 2005 showed that metformin targets the LKB1-AMPK pathway in the liver, and that metformin requires LKB1 to lower glucose production by the liver. However, neither LKB1 nor AMPK is the direct target of metformin, and as of 2005, that direct target was unknown.

A new genetic study that suggests that ATM kinase may affect the ability of patients to respond to metformin

Now–as of February 2011–comes a Nature Genetics paper that indicates that the serine/threonine kinase ATM (ataxia telangiectasia mutated) acts upstream of AMPK to mediate the therapeutic effects of metformin. ATM is a DNA repair protein that is recruited and activated by double-strand breaks in DNA. It initiates activation of the DNA damage checkpoint, leading to cell cycle arrest, followed by DNA repair or apoptosis. Thus the role of ATM in the AMPK pathway and in the therapeutic effects of metformin is surprising indeed.

In the study reported in the Nature Genetics paper, researchers of The GoDARTS and UKPDS Diabetes Pharmacogenetics Study Group and The Wellcome Trust Case Control Consortium 2 performed a genome-wide association study (GWAS) for glycemic response to metformin in type 2 diabetes patients in the U.K. In a population of nearly 4,000 patients, they identified a single-nucleotide polymorphism (SNP) designated rs11212617, which was associated with treatment success. This SNP occurs in a genetic locus that also contains the gene that encodes ATM. In a rat hepatoma cell line, inhibition of ATM by the specific inhibitor KU-55933 (KuDOS Pharmaceuticals, Cambridge, U.K., which was acquired by AstraZeneca in 2005) attenuated metformin-mediated phosphorylation and activation of AMPK.

The analysis by Morris Birnbaum and Reuben Shaw in the 17 February 2011 issue of Nature

The 17 February 2011 issue of Nature contained a Forum entitled “Genomics: Drugs, diabetes and cancer.” This consisted of two analyses of the implications of the Nature Genetics paper for metformin’s mechanism of action, and for understanding diabetes and the connections of the metformin-activated ATM/AMPK pathway with cancer. The first analysis was by Morris J. Birnbaum, M.D., Ph.D. (University of Pennsylvania Medical School, Philadelphia, PA), who does research on the role of AMPK and insulin in energy metabolism and in diabetes. The second analysis is by Dr. Reuben Shaw, mentioned earlier. Dr. Shaw’s research centers around LKB1 [also known as serine/threonine kinase 11 (STK11)]. LKB1, a serine/threonine kinase, is not only a regulator of hepatic glucose production via AMPK, but is also a tumor suppressor. Germline mutations in LKB1 are associated with the familial cancer Peutz-Jegher syndrome, and somatic mutations in LKB1 are present in various other cancers. In particular, the Lkb1 gene is one of the most frequently muted genes in human lung adenocarcinomas.

Dr. Birnbaum’s analysis

Dr. Birnbaum notes that the finding of a role for ATM in metformin responsiveness may be an important clue to the mechanism of action of this drug. However, it may also be a false lead, with ATM having only an indirect effect on metformin’s action. He cites recent evidence that metformin acts independently from LKB1 and AMPK and of transcriptional regulation in general. In these studies, genetic ablation of LKB1 and AMPK was used to show that these mediators are dispensable for metformin’s glucose-lowering activity. Instead, metformin appears to work by inhibiting mitochondrial production of ATP in the liver, thus reducing the level of liver glucose production via gluconeogenesis (which uses ATP). This is in apparent contradiction to the 2005 results of Dr Shaw and his colleagues. Nevertheless, metformin’s inhibition of mitochondrial ATP production increases the ratio of AMP to ATP, and thus activates AMPK. There are also other pathways by which inhibition of mitochondrial ATP production may inhibit gluconeogenesis. Thus the mechanisms by which metformin causes a decrease in glucose production by the liver appear to be very complex, and are not well understood.

Dr. Birnbaum therefore speculates that ATM may affect blood glucose levels via pathways that are parallel to, but not the same as, those modulated by metformin. However, the effects of these other pathways may be synergistic with those modulated by metformin when patients are treated with the drug. Dr. Birnbaum notes that 40 years ago, it was found that patients with ataxia telangiectasia often display a type 2-diabetes-like condition, including insulin resistance. Ataxia telangiectasia is a familial disease caused by germline mutations in ATM. This suggests that  ATM may act to counteract hyperglycemia and insulin resistance.

Dr. Birnbaum concluded that biochemical and cell biology studies should be conducted to determine the nature of the interaction of ATM and the antidiabetic effects of metformin. Key to these endeavors is to determine whether there are any biomolecules other than AMPK that both are influenced by ATM and control metabolism.

Dr. Shaw’s analysis

Dr. Shaw first discusses several animal studies that help elucidate the role in glucose regulation of the biomolecules involved in the putative ATM-LKB1-AMPK pathway. He notes notes that deletion of the Lkb1 gene in mouse liver results in loss of AMPK activity in that organ, and to the development of hyperglycemia and hepatic steatosis–two conditions that are seen in type 2 diabetes. Dr. Shaw also cites the 40-year-old finding about the connection between  ataxia telangiectasia and insulin resistance and diabetes. But as he also mentions the more recent (2006) finding that mice with defective ATM activity show increased insulin resistance and abnormal glucose regulation.

Dr. Shaw then speculates as to how ATM might work to modulate patients’ antidiabetic responses to metformin. He notes that ATM is known to phosphorylate LKB1, which is the key activator of AMPK in the liver. Alternatively, ATM might also regulate AMPK independently of LKB1, and might affect responsiveness of patients to metformin by regulating other relevant targets, independently of AMPK. In this context, ATM is known to phosphorylate other, LKB1 and AMPK-independent components of the insulin signaling pathway.

In the light of these considerations, Dr. Shaw says that it is important to determine whether the rs11212617 genetic variant results in modulation of ATM activity toward AMPK activation or toward other targets relevant to glucose regulation, or indeed whether this SNP affects ATM activity at all.

Dr. Shaw then focuses on the potential relevance of metformin to cancer therapy. Researchers have found, in retrospective studies, that diabetes patients who take metformin have a lower risk of developing cancer than those treated with other antidiabetic medications. Animal studies confirm the anticancer effects of metformin, but–as discussed in a 2010 review by Dr. Michael Pollak (McGill University, Montreal, Quebec, Canada)–they indicate that the anticancer effects of this drug are mechanistically complex. Dr. Shaw asks whether metformin is a general activator of ATM (and/or its targets) in the DNA damage-response pathway, or whether its specific effects on LKB1 and/or AMPK might be responsible for the apparent beneficial effects of metformin on cancer risk.

Dr. Shaw concludes with the statement that future studies of the relationship between metformin action, ATM, LKB1, and AMPK should shed light on the relationship between metformin’s antidiabetic effects and its apparent anticancer effects.

Our conclusions

The finding, based on a genome-wide association study, which suggests that ATM, a kinase best known for its involvement in DNA repair pathways, may also be involved in diabetics’ response to metformin is surprising and intriguing. It may eventually be important in unraveling metformin’s mechanism of action in inhibition of liver gluconeogenesis, and in other antidiabetic activities. This finding indicates a connection between pathways by which metformin exerts its antidiabetic activities, and pathways that are involved in cancer.

Nevertheless, the elucidation of metformin’s mechanism(s) of action in diabetes remains a work in progress. This situation is an example of how science works in the real world (as opposed to textbooks or much of science journalism)–generating more questions than answers.

A drug like metformin, with its complex and still poorly understood mechanism of action, could not have been discovered by modern, post-genomics drug discovery strategies. Metformin was discovered via research on natural products derived from the plant Galega officinalis (known as the French lilac, goat’s rue, and by various other names), which had been known by herbalists for centuries. It is fortunate that researchers were able to study the effects of extracts of this plant, and ultimately to develop metformin, well in advance of the modern era of drug discovery. Diabetics and their physicians now have access to metformin as an inexpensive generic drug.

The continued study of the antidiabetic mechanism(s) of action of metformin may yield additional insights into control of gluconeogenesis and other metabolic pathways. Some of the findings of these studies might be relevant to drug discovery and development, for example the development and use of AMPK activators in metabolic disease and in anti-aging medicine.

Continued study of the mechanism(s) of action of metformin may also be relevant to developing new therapies for cancer. As suggested by Dr. Pollak, although metformin is off-patent and is thus not an attractive agent for development as an oncology drug by pharmaceutical or biotechnology companies, other biguanides or related compounds might be better anticancer compounds, and would be patentable. In addition to identifying such compounds, it will be important to determine and define which groups of cancer patients could best benefit from them (perhaps via biomarkers). It will then be important to conduct personalized medicine hypothesis-testing clinical trials (as discussed in an earlier blog post) designed to obtain proof-of-concept that such compounds can indeed benefit specific groups of patients.


As the producers of this blog, and as consultants to the biotechnology and pharmaceutical industry, Haberman Associates would like to hear from you. If you are in a biotech or pharmaceutical company, and would like a 15-20-minute, no-obligation telephone discussion of issues raised by this or other blog articles, or of other issues that are important to  your company, please click here. We also welcome your comments on this or any other article on this blog.

Visited 1577 Times, 1 Visit today