Sirs, For a paper purporting to be a systematic review the recent manuscript by John et al. about adipose tissue and gastrointestinal diseases seems to have been remarkably selective in only reporting data supporting the authors’ currently unproven suggestions linking obesity and gastrointestinal mucosal inflammation.1
It is particularly disappointing that the authors uncritically accepted the idea of adiponectin as an important anti-inflammatory mediator. As adiponectin levels decline with obesity, this is clearly an attractive idea but not necessarily supported by evidence. Initial reports did show anti-inflammatory actions of adiponectin but these effects were seen at supraphysiological concentrations, greater than those typical of lean people.2 Data on adiponectin effects require careful consideration; adiponectin circulates in various molecular forms, a full-length form, a truncated (globular) form and several different multimers. These all appear to have different receptor binding characteristics and biological activities. For example, high-molecular weight and hexamer isoforms activated NF-κB in C2C12 cells but trimers did not, yet trimers and monomers were much more effective at activating muscle AMP-activated protein kinase.3 There are several papers showing that adiponectin concentrations in the physiological range have proinflammatory actions and globular adiponectin seems to be more potent in this regard. Adiponectin stimulated increased production of TNF-α and IL-6 by placental and adipose tissue explants4 and macrophages and THP-1 cells.5 Increased IL-8, MCP-1 and GM-CSF secretion by adiponectin-stimulated colonic HT-29 cells has recently been reported.6 A variety of enzymes, including leucocyte-produced elastase can cleave full-length adiponectin to the globular form,7 thus potentially altering the proinflammatory/anti-inflammatory balance within inflamed tissues in vivo. Therefore, it would not be wise to accept adiponectin as a purely anti-inflammatory agent on currently available data.
It was also surprising that the authors ignored a significant body of data showing that leptin is a growth factor for the gastrointestinal tract. Serum leptin levels are proportional to body mass index8 and leptin concentrations equivalent to those seen in obese patients have been shown to stimulate colonic proliferation in vivo and in vitro,9 inhibit colonic epithelial apoptosis10 and stimulate proliferation of MKN28 and HS746T gastric cancer cells.11 More recently leptin has been shown to stimulate proliferation and inhibit apoptosis in Barrett's adenocarcinoma cells via a complex COX-2-dependent mechanism.12 John et al. have identified that obesity seems to promote the development of oesophageal adenocarcinoma from Barrett's oesophagus rather than the initial development of columnar-lined oesophagus, and the growth factor effects of leptin would be expected to have an action at this point in the pathway. John et al. have stressed the effects of COX-2 and inflammation induced by obesity in promoting oesophageal carcinogenesis; the available data show that the upregulation of COX-2 in Barrett's oesophagus is not closely related to other markers of mucosal inflammation, suggesting that other factors are acting as important drivers in vivo.13 Leptin may be such a driver, and promote carcinogenesis without invoking the subclinical mucosal inflammation proposed by the authors.
We agree that the relationships between adipose tissue, adipokines and gastrointestinal diseases provide exciting new insights into the pathogenesis and treatment of several disorders but feel that the review by John et al. was not as systematic as claimed.