Effect of route and type of nutrition on intestine-derived inflammatory responses

Endothelial responses to injury 

Polymorphonuclear neutrophils (PMNs) generate nonspecific immune responses capable of controlling bacterial invasion with the risk of injuring or destroying otherwise viable tissue. While many factors modify this response, interaction between adhesion molecules on the vascular endothelium and ligands on circulating PMNs lead to PMN attachment, priming, and activation, which can affect their subsequent inflammatory response. Expression of the vascular adhesion molecules P-selectin, E-selectin, and intracellular adhesion molecule-1 (ICAM-1) are low on the vascular endothelium of normal healthy animals. Noxious events such as hemorrhagic shock, sepsis, and tissue injury result in their rapid upregulation and increased expression, which attracts PMNs to the vascular endothelium (“the first hit”) [8]. Initially, these selectins induce slowing, rolling and eventual adhesion of the PMNs to the vascular endothelium, which, in association with other factors, prime the PMNs [9]. Priming renders the cells capable of generating an augmented inflammatory response should another proinflammatory stimulation occur subsequently. If the magnitude of the initial insult is extreme, the primed cells induce significant tissue injury immediately. This immediate intense response is hypothesized to play a role in early multiple organ dysfunction if the patient survives the initial insult [8]. However, if the initial insult is rapidly controlled and resuscitation is successful, the primed PMNs are distributed throughout the body especially to the liver and lung. In these sites, current work suggests that a subsequent insult (“the second hit”) generates an augmented inflammatory response by these primed neutrophils.

The experimental evidence for these concepts have been elegantly reviewed recently [8]. Evidence implicates a series of intracellular events induced by the hypoxic insult as important factors in the post ischemic priming of PMNs. During low flow, ischemic states, the hypoxic tissue sequentially breaks adenosine triphosphate (ATP) to adenosine diphosphate (ADP), adenosine monophosphate (AMP), adenosine and hypoxanthine. Simultaneously, cellular levels of xanthine dehydrogenase are converted to xanthine oxidase through uncontrolled calcium accumulation in the hypoxic cells. As the tissues are reperfused and oxygenated, xanthine oxidase converts hypoxanthine to uric acid, which results in the production of reactive oxygen metabolites. Release of the reactive oxygen metabolites during reperfusion also activates cytosolic phospholipase-A2, which is found in very high concentrations in gut mucosa. Activated phospholipase A2 releases lysophospholipids and free fatty acids into the cell cytosol which act as substrate for eicosanoid production. These substances generate platelet-activating factor and other PMN chemoattractants. Simultaneously, adhesion molecules are expressed on the vascular endothelium. Preformed P-selectin expression on endothelial tissue increases almost immediately followed within 4 to 6 hours by expression of inducible E-selectin. Eventually, ICAM-1 is expressed to augment adherence of PMNs to the endothelial surface.

The intestinal tract appears to be an important priming site for PMNs and gut ischemia reperfusion (I/R) is capable of priming PMNs in vivo [8], [10], [11], [12]. Substances carried either in the portal blood or in postshock lymph [11] activate neutrophils and potentiate neutrophil-mediated endothelial injury. After a “second hit,” both the lung and the liver are important targets with increases in permeability. Diversion of intestinal lymph, but not portal blood, abrogates the lung injury implicating the mesenteric lymph in the generation of pulmonary injury. The role of the gut in PMN priming was shown in an elegant series of experiments in which PMNs harvested from the superior mesenteric artery were compared with those in the portal vein after I/R [10]. PMNs exiting the gut demonstrated evidence of significant activation and priming compared with those entering the splanchnic circulation through the mesenteric artery.

Additional work in rats demonstrated that PMNs primed by shock (identified by increased expression of CD11b/CD18 on their surfaces) were extremely sensitive to activating agents rendering and capable of producing high levels of reactive oxygen metabolites [13]. These activated cells increase lung permeability and increase mortality to a second hit of gut ischemia [10]. Blockade of the CD11b molecule with a specific monoclonal antibody prevented the pulmonary permeability changes [14]. Since many of these PMN changes have been shown to occur in severely injured trauma patients, Moore et al [8] proposed a “unifying hypothesis” that links PMN activation with subsequent multiple organ dysfunction. This has implicated the PMN and other circulating humoral factors as components of this clinical problem.

Changes in mucosal immunity induced by nutritional manipulation reproduce this PMN priming effect [12], [15], [16]. Consistent with human data demonstrating an augmented inflammatory response to endotoxin after parenteral feeding [17] (ie, a non-fed gut), changes induced by gut “starvation” and parenteral feeding alter the intestinal tract so that it responds in a manner similar to the postischemic gut, ie, alterations in gut cytokines, adhesion molecule expression and myeloid cell activation induced by lack of enteral feeding constitutes a “first insult.”

Influence of route of nutrition on mucosal immunity 

The impact of route and type of nutrition on mucosal immunity was extensively reviewed recently [7] and the reader is referred to that article for greater detail (Fig. 1). In brief, approximately 50% of total body immunity resides in the lamina propria of the intestinal and respiratory tract where interaction between T and B cells generate production of IgA for immediate transport across the mucosal surfaces by secretory component. This IgA binds to specific bacterial and viral antigens to prevent their attachment to mucosal surfaces. Infection does not occur without attachment. The primary site for T and B cell sensitization resides in the Peyer patches of the small intestine and, at least in rodents, within the nasal-associated lymphoid tissue (NALT). The latter may be analogous to Waldmeyer’s ring in humans. The α4/β7 and L-selectin on naive B and T cells interact with their ligand, MAdCAM-1, located on the high endothelial venules in the initial steps necessary for migration into the NALT or Peyer’s patches. After sensitization to antigens processed in these sites, the T and B cells are distributed via the circulation or nasal lymphatics to the lamina propria of the intestinal tract, respiratory tract, and other extraintestinal sites. Within the lamina propria, T cells produce Th2 IgA stimulating cytokines (IL-4, IL-5, IL-6, and IL-10), which stimulate IgA production by fully differentiated, mature B cells. This effect is counterbalanced by levels of Th1 type IgA inhibiting cytokines (IFN γ and lymphotoxin). The size and function of the murine mucosal immune system is exquisitely sensitive to the presence or absence of enteral stimulation.

View Large Image Download to PowerPoint

Fig. 1. The common mucosal immune hypothesis: naive T and B cells migrate into the Peyer’s patches (PP) via modified MAdCAM-1 located on the high endothelial venules. After sensitization within the PP, they are directed into the thoracic duct via mesenteric lymph nodes (MLN) and into the blood stream for distribution for to the gut-associated lymphoid tissue (GALT) and extraintestinal sites such as the mammary glands and genitourinary tract where they produce specific IgA to bind the antigen. They also migrate through the nasal-associated lymphoid tissue (NALT) in rodent species, migrate to cervical lymph nodes (CLN) and are delivered via the blood stream as well. Peripheral lymph node addressin (PNa-D) directs these cells into effector sites within the nasal passages. Unmodified MAdCAM-1 is found throughout the body. (Previously published in Am J Surg 2002;183:390–8.)

Since lethal malnutrition occurs with just 3 days of murine starvation, parenteral feeding permits study of the effects of route and type of nutrition without the confounding variable of severe protein-calorie malnutrition and weight loss. Within hours of instituting parenteral feeding, MAdCAM-1 expression in the Peyer’s patches drops, reaching a nadir by 48 hours. Simultaneously, lymphocyte cell counts in the Peyer’s patches, lamina propria, and intraepithelial spaces decrease to 50% to 60% of normal. Within the lamina propria, changes in T cell profile result in a decrease in the CD4:CD8 ratio from approximately 2:1 to 1:1. This profile change is related to decreased levels of the Th2 cytokines, IL-4 and IL-10, which drop in proportion to intestinal IgA levels. Levels of the Th1 IgA-inhibiting cytokine, IFN-γ remain unchanged [18].

These alterations create functional defects in established antiviral [19] and antibacterial [20] defenses at the mucosal surfaces presumably through reduced IgA. The drops in IL-4 and IL-10 appear to affect the gut vasculature as well, since their decreases are associated with increased gut endothelial expression of the adhesion molecules ICAM-1, P-selectin, and E-selectin (Fig. 2) [9], [21].

View Large Image Download to PowerPoint

Fig. 2. Cytokines and gut-associated lymphoid tissue (GALT) function: within the lamina propria of the small intestine, interleukin (IL)-4 and IL-10 (as well as other Th2 type cytokines) stimulate immunoglobulin A (IgA) production on the mucosal surface and inhibit intracellular adhesion molecule-1 (ICAM-1) expression on the vascular endothelium. Interferon-γ (IFN-γ) inhibits IgA secretion and increases the expression of ICAM-1 on the vascular endothelium. Parenteral feeding (intravenous total parenteral nutrition [IV-TPN]) results in decreases in IL-4 and IL-10 that correlate with decreases in IgA and increases in ICAM-1.

Effect of route and type of nutrition on gut endothelial adhesion marker expression 

The dual radio-labeled monoclonal (mAb) antibody technique allows accurate quantification of individual adhesion molecule expression on intact endothelial cells in vivo [22]. Radiolabeled specific binding mAbs directed against ICAM-1, E-selectin, P-selectin, or other molecules of interest are administered to mice to measure in vivo expression of that particular molecule. A nonbinding mAb labeled with 131I is injected intravenously to allow correction for nonspecific antibody binding of a second 125I-labeled binding mAb directed against a specific target molecule. After the two mAbs circulate for 5 minutes, animals are exsanguinated while perfusing the vascular system with buffered solutions to remove the blood. Organs are harvested for measurement of radioactivity, which allows calculation of the expression of the specific adhesion molecule in question. Expression is proportional to the accumulated radioactivity of the binding relative to the nonbinding mAb. The technique eliminates difficulty in quantifying expression inherent in immunohistochemical staining and allows direct measurement of protein expression not capable through measurement with mRNA techniques.

Role of cytokines and nutrition on endothelial expression and PMN accumulation 

Lack of enteral feeding alters levels of cytokines IL-4 and IL-10 within the lamina propria with no effect on the Th1 cytokine IFN-γ [18]. Under normal conditions, IL-4 and IL-10 inhibit endothelial expression of ICAM-1 while IFN-γ is a potent stimulator of its expression. Earlier work had identified increased levels of myeloperoxidase (as a marked of PMN accumulation) in the intestine of parenterally fed rats suggesting a potential inflammatory response to lack of enteral stimulation [23]. Use of the dual mAb technique allowed further study of this phenomena in vivo after altering route and type of nutrition in our mouse model.

Five days of gut “starvation” with parenteral feeding (to prevent malnutrition) increased adhesion molecule expression in the intestine, lung, or liver. Expression of P-selectin, a constituently expressed adhesion molecule important in initial rolling of PMNs, significantly increased in the small intestine of parenterally fed mice, with no changes in other organs [9]. E-selectin, an inducible molecule that appears several hours after in insult, was selectively increased in the pulmonary vasculature but not in other sites, although a trend existed toward increase in intestinal expression [9]. The most impressive effect occurred on ICAM-1 expression, which was significantly upregulated in the intestine and lungs of parenterally fed mice [21]. This increase in ICAM-1 in the parenterally fed mice occurred in conjunction with increased intestinal levels of myeloperoxidase. These ICAM-1 changes were reversible with reinstitution of enteral feeding. Five days of refeeding the parenterally fed mice with chow returned ICAM-1 levels to normal as myeloperoxidase levels disappeared. Further work supported the hypothesis that this increase in endothelial adhesion molecule expression and PMN accumulation in the unfed gut with was a diet-induced “first hit.”

Diet-induced susceptibility to intestinal ischemia/reperfusion 

Mice are more sensitive to intestinal I/R than rats. While 45 minutes of superior mesenteric artery occlusion with gut ischemic is well tolerated in rats, it is rapidly fatal in mice so that ischemic time must be limited to 15 to 30 minutes. Dietary manipulation and the concomitant intestinal altercations affected the response to the gut I/R within this timeframe. Animals underwent 15 or 30 minutes of gut I/R after pretreatment with parenteral feeding or enteral feeding with either chow, intragastric parenteral formula (IG-PN), or an isocaloric, isonitrogenous complex enteral diet containing whole protein, complex carbohydrates, and fat. Animals were also studied for permeability changes using radiolabeled albumin 3 hours after the ischemic event.

Both the type of diet and the duration of insult altered mortality [12]. After 15 minutes of ischemia, survival at 72 hours significantly dropped from 90% to 95% in animals fed chow, the complex enteral diet, or IG-PN to approximately 60% with parenteral feeding. Three hours after the ischemic event, significant increases in permeability to radiolabeled albumin were noted only in the lungs and liver of parenterally fed mice with no permeability changes in any of the enterally fed groups. When the ischemic insult was prolonged to 30 minutes, mortality of chow and CED-fed mice increased slightly to approximately 10% to 15% at 72 hours. However, mortality increased to approximately 80% in both the parenterally fed mice and those receiving IG-PN. Thus, whatever protective effects were adequate to protect IG-PN fed mice for 15 minutes of gut I/R were inadequate to protect them against the 30-minute insult.

Diet-induced PMN activation implication gut ischemia/reperfusion response 

Although the increased expression of P- and E-selectins and ICAM-1 influenced PMN accumulation in the gut, this did not necessarily establish their role in the survival or permeability changes noted after gut I/R. This was addressed by subjecting parenterally and enterally (chow and IG-PN) fed mice to 15 minutes of gut I/R followed by quantification of ICAM-1 expression in intestinal and extraintestinal sites 3 hours later using the dual mAb technique [16]. In parallel experiments, peripheral blood samples were obtained either with or without gut I/R after the same dietary manipulation to measure expression of the PMN activation markers, CD11a and CD11b. Lungs were also harvested 3 hours after gut I/R and stained for CD18, a marker of myeloid cell activation, using immunohistochemistry.

Parenteral feeding significantly increased ICAM-1 expression within the intestine and liver 3 hours after reperfusion compared with enteral feeding. No differences in lung, colon, stomach, pancreas, or muscle ICAM-1 expression were noted following gut IR. Although CD11a or CD11b expression were similar on the circulating myeloid cells of all groups prior to gut I/R, 3 hours of reperfusion increased CD11b expression (but not CD11a expression) of circulating cells only in the parenterally fed group. The upregulation of CD11b with no CD11a change is consistent with the literature. Only the lungs of parenterally fed animals exposed to gut I/R had evidence of CD18 positive leukocytes on immunohistochemical staining. The increase in CD11b expression in circulating cells and the pulmonary CD18 positive myeloid cells in only the parenterally fed group is consistent with our hypothesis that lack of enteral feeding creates a “first hit” and renders the animal susceptible to a subsequent ischemic insult.

Glutamine: an adjuvant to parenteral feeding 

There is a marked increase in the production of glutamine by skeletal muscle during time of stress and sepsis [24], [25], [26]. Glutamine is a respiratory fuel for both enterocytes as well as for proliferating lymphocytes—particularly T lymphocytes. Clinically, glutamine supplementation of parenteral feeding normalizes microbial columnization patterns within the gastrointestinal tract and reduces bacterial influence in bone marrow transplant patients [27]. Experimentally, glutamine exerts beneficial effects upon the mucosal and in our work has significant effects upon the GALT [28]. A 2% glutamine supplemented TPN solution normalizes lymphocyte counts within the lamina propria and Peyer’s patches. Within the gut and in isolated lamina propria cells, glutamine preserves IL-4, but not IL-10 [29], [30]. Glutamine also improves intestinal and respiratory tract IgA levels [31] and partially, but not completely, reverses immunological defects in established antiviral [28] and antibacterial defenses[30] induced by parenteral feeding without glutamine. It also alters the inflammatory response to gut I/R.

The replacement of 2% of the amino acids in TPN solution with glutamine significantly improved ICAM-1 expression in the small intestine of parenteral fed animals reducing its expression nearly to the level of chow fed mice [32]. However, increased pulmonary ICAM-1 expression was not affected. When animals were subjected to 30 minutes of gut I/R, the glutamine supplemented TPN improved survival from 20% in TPN fed mice to 65%, approaching the 85% survival of chow fed mice [33]. The reduction in ICAM-1 expression together with an improved survival to gut I/R provided further evidence that the cytokine and cellular changes noted with parenteral feeding modifies the gut and systemic inflammatory responses to subsequent insult while decreasing gut tolerance to hypoxia.

Summary 

The initial experiments from Sheldon’s laboratory in the late 1970s and early 1980s fostered a new direction for clinical and laboratory research. Subsequent investigations into gut barrier function, bacterial translocation and mucosal immunity can be traced to these early experiments. It is increasingly clear that benefits are gained when complex nutrients are delivered via the gastrointestinal tract that are not gained when similar nutrients are delivered intravenously. Lack of enteral feeding impairs the coordinated system of sensitization, distribution, and interaction of T and B cells important in the production of IgA, in the maintenance of normal gut cytokines, and in the regulation of endothelial markers of inflammation. As not all patients can be fed enterally, definition of these defects may lead to novel approaches to reverse the defects in gut barrier function and mucosal immunity that occur in parenterally fed patients. These changes in mucosal immunity first suggested by Sheldon’s research are consistent with the clinically observed increases in infectious complications and inflammation noted in nonenterally fed patients.