Adventures in the Gland of Plenty

ADIPO INSIGHTS

Once regarded as a simple storage depot for fat, adipose tissue is now recognized as an endocrine organ, secreting hormones on its own. "It was leptin that really changed things," says Alan Saltiel, director of the Life Sciences Institute at the University of Michigan, Ann Arbor, referring to the 1994 discovery.⁴ Later, researchers showed that adipocytes secrete tumor necrosis factor α (TNF-α), resistin, adipsin, and adiponectin, collectively known as adipokines.

Adiponectin was discovered independently by four groups in 1995 and 1996 (hence its wealth of aliases: acrp30, GBP28, adipoQ, apM1).² The peptide comprises 247 amino acids, with a collagenous domain at the N-terminus and a C-terminal globular domain. Identifying it was one thing; working out what it does was another. "For a number of years ... no one knew quite what to do with it," says Lodish, who leads one of the teams that identified the peptide.

Scherer, an erstwhile member of that group, says that "for no obvious reason," the protein proved tricky to knock out in mice. Also, the full-length ligand is an "extremely complex, supramolecular structure," he says. It forms trimers, hexamers, 12mers, and 18mers, "and it took, at least at our end, years to produce this in a recombinant form such that it would be bioactive." Then, in 1998, Scherer's group showed that the globular domain is structurally akin to TNF-α.⁵ That suggested a hormone-like action, says Lodish, "but lots of things look like hormones and are not."

More clues followed: Adiponectin expression appeared reduced in obese mice and humans, and in patients with type 2 diabetes and coronary heart disease. The implication that adiponectin expression has a role in developing insulin resistance prompted the three Hot Papers.

LEAN BURN

Lodish says that Fruebis did the crucial experiment showing that the globular domain injected into mice caused a reduction in free fatty acid elevation after a meal. "It was Tsu-Shuen [Tsao] in our lab who made the key observation that the only tissue that could do that would be muscle," Lodish says. Experiments showing increased fatty acid oxidation in isolated adiponectin-treated muscles bolstered the hypothesis.

The other two Hot Papers appeared six months later. Kadowaki and his team were building on their work with PPAR-γ knockout mice. Unexpectedly, says Kadowaki, they found that the mice were protected from obesity and insulin resistance induced by a high-fat diet.⁶ Their Hot Paper confirmed, in obese and diabetic murine models, adiponectin's effects on muscle. And, by injecting leptin (which was already known to be insulin-sensitizing) and/or adiponectin into insulin-resistant lipoatrophic mice, they showed that both were needed to ameliorate completely the resistant condition.² This paper also hinted that adiponectin affected the liver, reducing triglyceride content.

In the same issue of Nature Medicine, Scherer's group reported that an injection of adiponectin into wild-type mice reduces basal glucose, and that in ob/ob and diabetic mice the treatment transiently abolishes hyperglycemia. Furthermore, adiponectin increased insulin's ability to suppress glucose production in isolated hepatocytes.³

Kadowaki and Scherer's papers both reinforced the link between adiponectin and insulin resistance by showing that PPAR-γ agonistic, insulin-sensitizing drugs such as thiazolidinediones increase adiponectin levels in insulin-insensitive mice.

SPECIES DIVERSITY

But confusion remains over a rather unhormone-like characteristic of adiponectin. "What baffled a lot of people, and still does, is the fact that [adiponectin's] concentration in the serum is very high," says Lodish. There's a lot of inactive material of unknown function in the serum, he says, which may be a precursor for the signaling molecules.

CUT TO THE QUICK

The full-length adiponectin molecule may be processed through activation of a serum reductase and/or proteolytic cleavage. Effects in liver, muscle and adipocytes have been shown, and central effects in the brain have been postulated. AMPK (AMP-activated protein kinase) is a key target for activation by adiponectin.

Confusion over which chemical species cause the reported effects is compounded by the variety of ligands used by the different labs: a bacterially derived C-terminal globular fragment,^1,2 and full-length molecules derived from mammalian cells³ and bacteria.² The globular fragment, which forms only trimers, is more potent in its effects on muscle than the full-length ligand. But, to date, it remains undetected in vivo. This could, says Kadowaki, "be used as a therapeutic tool, but we doubt [its] physiological significance."

The identity of the physiologically active species remains in doubt, but its mode of action is becoming clearer. Insulin-sensitizing and glucose-lowering effects of AMP kinase (AMPK) have emerged in parallel with those of adiponectin. AMPK's effects are stimulated by exercise, for example, and the insulin-sensitizing biguanide drug, metformin. Several groups had made the AMPK-adiponectin connection, but Kadowaki's group was first to publish,⁷ showing that, in vitro and in vivo, AMPK is stimulated with globular and full-length adiponectin in skeletal muscle, but only with the full-length molecule in the liver.

"AMP kinase seems to be an insulin-independent means of increasing glucose uptake and ... fatty acid oxidation, and decreasing fatty acid synthesis," says Saltiel. "It mimics insulin's action on glucose uptake, but then it antagonizes insulin's action on lipid synthesis."

So, adiponectin's insulin-sensitizing effects may be mediated through activation of AMPK, but this is not necessarily its only path of action. Lodish has shown that the high-molecular weight species, "probably 12- and 18-mer and higher" (the most abundant forms in serum), activate the NF-κB pathway. "Perhaps their activation of NF-κB in tissues like vascular muscle or endothelial cells may be important," says Lodish, "but for the moment this is speculative."

MIXED RECEPTION

In June 2003, Kadowaki's lab published work using expression-cloning techniques to identify the genes for two adiponectin receptors.⁸ AdipoR1 is expressed ubiquitously, but predominantly in skeletal muscle, and AdipoR2 is most abundant in the liver. Both mediate adiponectin's activation of AMP kinase. "If we lose AdipoR1 and AdipoR2," says Kadowaki, "then cells cannot bind adiponectin. That is very solid data."

The findings are "wholly consistent with a dual mode of action," observes Scherer. One part of the protein, in its full-length form, targets the liver, he says, whereas the globular fragment, if it exists in vivo, works on muscles.

As well as shedding light on adiponectin's downstream effects, the receptors might prove fruitful targets for therapies for obesity-related conditions. But, not everyone is convinced. "It's still a little confusing to all of us," says Saltiel.

"There is a little bit of uncertainty whether these are the right receptors ... they don't look like receptors. They don't have features of any other proteins that have been identified as receptors.... I think a lot of people would like to see that pursued in a little bit more detail."

Stuart Blackman stuart.blackman@talk21.com is a freelance writer in Edinburgh, UK.