Toll-like receptors (TLR) are a family of proteins generally thought to recognize molecules of microbial origin and, subsequent this recognition, initiate an immune response. Related to the interleukin-1 receptors, TLRs stimulate production of the usual suspects of inflammatory cytokines, predominantly through transcription and activation of NF-κB. Interest in this receptor family increased when it was demonstrated that their activation by saturated fatty-acids (SFA) was the mechanism behind the so-called inflammatory effects of saturated fats. This effect has been shown in vitro (1) and suggested in transgenic experiments (2) that reduced weight gain from saturated fat through deletion of TLR4.
It’s worth pointing out that the mechanistic explanation for saturated fats causing inflammation must be particularly strong in order to overcome the question of why humans, like most mammals, preferentially produce these fatty acids. Indeed de novo lipogenesis is dominated by palmitic acid, a singularly counterproductive organismal action if this fatty-acid is recognized by said organism as invasive and dangerous. The search for the smoking gun of inflammatory actions by saturated fats continues and, this 2009 publication notwithstanding, activation of TLRs is a relevant discussion and research topic (3).
Erridge and Samani set out to confirm TLR dependent signaling by SFA using robust and varied techniques (4). HEK-293 (embryonic kidney) and RAW 264.7 (mouse leukemia) cell lines were treated with lauric, myristic, palmitic and stearic (12, 14, 16 and 18 carbons) acids in fatty-acid free bovine serum albumin (BSA). TLR signaling was measured by ELISA and Western Blot of cytokine production and other down-stream signaling molecules as well as a luciferase reporter assay. TLR activation by SFA was readily confirmed using BSA vehicle, but when fatty-acids alone were applied up to 500 uM no activation was detected. This led to the group testing BSA alone and discovering it as a potent activator of TLR dependent signaling through bacterial lipopolysaccharide (LPS), lipopeptide and flagellin contamination. None of the characteristic mediators of TLR activation, including induction of NF-κB and p38 MAP-kinase signaling cascades, inflammatory cytokine release or E-selectin expression in macrophages, endothelial cells, aortic smooth muscle, adipocytes or skeletal muscle cells were detectable from SFA treatment. Treatment of peripheral blood mononuclear cells with SFA induced expression of signaling molecules and cytokines but in a completely different profile as LPS treatment. SFA delivered through uncontaminated BSA did not activate TLR signaling in any cell line.
The investigators went on to show that polymyxin-B treatment, which sequesters and inactivates LPS, and the limulus reaction, used to show LPS contamination, are both ineffective against other bacterial contaminants such as di and triacyl lipopeptides and flagellin. Treatment of BSA with lipases proved more effective in blocking TLR signaling. Previous investigations into activators of TLR signaling showing positive results are therefore likely showing activation from bacterial contaminants rather than, or perhaps in addition to, the studied agent. In fact, as the author’s point out, the supposed TLR signaling induced by heat shock protein-60 (5) and endothelial immune activation by C-reactive protein (6) have both been shown to be artefacts of similar contamination.
As a post-script to the study the authors responded to criticism by Hwang et al. on the Arterioscler. Thromb. Vasc. Biol. journal website. The most salient point made was in response to the findings of Hwang et al. that SFA causes TLR signaling at concentrations of 75 to 100 uM in a similar potency as LPS at 200 ng/mL, which in humans corresponds to severe endotoxemia. If this were true then humans in a normal physiological state, in which circulating SFA is routinely above 150 uM, would regularly experience symptoms such as shock and multiple organ failure. The surface level implausibility of this, similar to SFA being at once preferentially synthesized de novo and treated as an invading molecule, does not occur to many authors or cause them to put their findings to the question. This highlights the danger of relying on molecular experiments that are only verifiable through other molecular experiments, with no grounding in whole organismal measurement and theory. Molecular biology best serves research when it is used as a tool subservient to more objective and visible traits. Molecular findings should be viewed skeptically when in conflict with macro-physiology, and never vice versa, as a model is built to the specifications of its subject and not the other way around.
- M. Milanski et al., J. Neurosci. 29, 359–370 (2009).
- J. E. Davis, N. K. Gabler, J. Walker-Daniels, M. E. Spurlock, Obesity (Silver Spring). 16, 1248–55 (2008).
- A. C. Ko, J. C. Bru, 22 (2011), doi:10.1016/j.tem.2010.08.007.
- C. Erridge, N. J. Samani, Arterioscler. Thromb. Vasc. Biol. 29, 1944–1949 (2009).
- B. Gao, M.-F. Tsan, J. Biol. Chem. 278, 22523–22529 (2003).
- K. E. Taylor, J. C. Giddings, C. W. Van Den Berg, Arterioscler. Thromb. Vasc. Biol. 25, 1225–1230 (2005).