Intestinal macrophages in health and inflammation

[摘要] The healthy large intestinal mucosa contains a vast pool of macrophages (mφ) that are close to the local bacterial flora and have several unique phenotypic and functional properties compared with other mφ populations. Although human colonic mφ retain some of the hallmark functions of mφ, such as the ability to phagocytose particulate material and exert bactericidal activity, they are unable to produce pro-inflammatory mediators. Thus it has been suggested that intestinal mφ are functionally adapted to the microbe-rich, immunostimulatory environment of the gut, where strong inflammatory responses to harmless commensal bacteria would lead to continuous inflammation and ultimately tissue pathology. Indeed, there is mounting evidence that mφ play an essential role in maintaining homeostasis and epithelial renewal in the normal intestine. In contrast, mφ from the intestine of patients with inflammatory bowel disease (IBD) differ markedly from those present physiologically, exhibiting heightened inflammatory and bactericidal activities, and contributing to the tissue damage. How these differing properties of colonic mφ are controlled and how this potentially dangerous population is kept quiescent under physiological conditions are important questions. Most existing information comes from either simple, observational studies of human tissue, or from work oncell culture systems which aim to reproduce the unusual phenotype of resident intestinal mφ in vitro. Importantly analogous experiments of resident and inflammatory mφ have not been carried out in murine systems, where it would be possible to characterise the cells fully and explore their origin, function and role in inflammatory processes.Therefore, the aims of this thesis were first to characterise mφ in the resting murine colon both functionally and phenotypically, focussing particularly on their expression of toll-like receptors (TLR) and responsiveness to TLR ligation and on their population dynamics in vivo. By comparing colonic mφ with other tissue mφ populations, I hoped to gain an understanding of how resident gut mφ might have adapted to their local environment. In the second part of my work, I examined how the properties of colonic mφ altered in inflammation, employing a well-established experimental model of colitis, with the aim of determining how resident and inflammatory mφ might relate to each other. Lastly, I explored the effects of the ES-62 parasite product, known to have potent anti-inflammatory effects on mφ, on experimental colitis in vivo.Experiments detailing my initial characterisation of the myeloid cells expressing the F4/80 mφ marker in the colon of normal mice are described in Chapter 3. These revealed that the F4/80+ population in the gut is extremely heterogeneous compared with other mφ populations in the body. Virtually all in vitro-differentiated BM mφ (BMM) and mφ from the resting peritoneum (PEC mφ) exhibited the conventional F4/80+CD11b+CD11c- phenotype of classical mφ and upregulated costimulatory molecules in response to TLR ligation. In stark contrast, the colon contained three F4/80+ subsets, one F4/80+CD11b+CD11cint, one F4/80+CD11b+CD11c- and a smaller population of F4/80+CD11b-CD11c- cells. None of these subsets expressed co-stimulatory molecules, even after LPS stimulation, but unlike other mφ, the majority of colonic mφ expressed high levels of class II MHC without stimulation.BMM and PEC mφ also produced several pro-inflammatory cytokines and chemokines following stimulation, whereas colonic mφ showed no mediator production under these conditions. Nevertheless, colonic mφ did retain avid endocytic and phagocytic activities, indicating that colonic mφ may engulf bacteria without initiating inflammation.In Chapter 4, I explored the unresponsiveness of resting colonic mφ to microbial stimuli in more detail and found that the TLR refractoriness is associated with reduced expression of TLR2, 3, 4 and 9. Apart from a small proportion of mφ that retained TLR2 expression, TLR expression was downregulated both at the protein level and to some extent also at the mRNA level; TLRs were not re-expressed following ex vivo culture of purified mφ. This global downregulation of TLRs could not be reproduced in BMM by treatment with TLR ligands, and was also present in colonic mφ taken from mice unable to signal via TLR2 or TLR4, suggesting it was not simply a form of endotoxin “tolerance”. However, the mechanism seemed to involve IL-10, as colonic mφ from IL-10-deficient animals displayed a heightened level of TLR expression and responsiveness, even prior to the onset of intestinal inflammation. In Chapter 5, I examined the phenotype and function of mφ during the experimental colitis induced by feeding dextran sodium sulphate (DSS). During inflammation, the absolute number of F4/80+ mφ increased 6-fold, the majority of which now expressed TLR, CD11b and low levels of CD11c. This new population of colitic mφ also expressed class II MHC, low levels of co-stimulatory molecules and produced large amounts of TNFα. In Chapter 6, I went on to examine the population dynamics of colonic mφ under resting conditions and during inflammation, showing that the overall turnover rate of the total mφ population was increased during colitis, as assessed by uptake of BrdU in vivo. The increased turnover was mainly due to the TLR-expressing, TNFα+ population of mφ and more detailed analysis showed that the small number of these cells present in resting colon had identical turnover rates to those found in colitis. In contrast, the TLR negative mφ had much lower turnover rates in resting and inflamed colon, suggesting that the TLR+ and TLR- subsets may represent distinct mφ populations with different population dynamics, and that during intestinal inflammation, the TLR+ subset may display a preferential recruitment into the gut. Indeed, proliferation in situ was minimal, indicating that the recently divided, TLR-expressing mφ proliferated outside the intestine before being recruited into the gut. My subsequent experiments suggested that this recruitment may involve the CCR2 chemokine receptor, which was expressed at high levels specifically by the TLR+ subset of mφ both in resting and inflamed colon.Finally in Chapter 7, I treated colitic mice with ES-62, a phosphorylcholine (PC)-containing glycoprotein secreted by the filarial nematode, Acanthocheilonema viteae, which has been shown to modulate pro-inflammatory cytokine production by mφ in vitro.ES-62 treatment had no significant effect on weight loss or pro-inflammatory cytokine production in the colon of mice with DSS colitis, although it slightly delayed the onset of the clinical signs of disease. Thus further studies of ES-62 as a modulator of mφ-dependent intestinal inflammation may be warranted.Taken together, my data suggest that under resting conditions, intestinal mφ are heterogeneous and adapt to their microenvironment by being non-inflammatory via active downregulation of TLR expression and function, which may be partly dependent on IL-10. During inflammation, large numbers of TLR expressing, fully responsive mφ appear, probably via CCR2-dependent recruitment of recently divided blood-derived monocytes. Interestingly, small numbers of these TLR-expressing, rapidly turning over mφ are also present in normal colon and my data suggest that these pro-inflammatory mφ may be quite distinct from the more sessile mφ which are the dominant “resident” population in normal gut. A delicate balance between these two mφ populations must ensure homeostasis and appropriate responses to inflammation.

[发布机构] University:University of Glasgow;Department:Institute of Infection Immunity and Inflammation

[关键词] intestine, macrophages, inflammation, toll-like receptor