Dietary Restriction Impairs Neutrophil Exudation by Reducing CD11b/CD18 Expression and Chemokine Production FREE

DESPITE MANY advances in nutritional therapy, the incidence of malnutrition in patients admitted to surgical services remains high.¹ Moreover, during hospitalization, because of multiple diagnostic examinations or underlying diseases, many patients are subjected to oral intake restriction, which can worsen their nutritional status.² Preoperative dietary restriction is associated with a high incidence of infectious complications in patients undergoing surgery.³ Polymorphonuclear neutrophils (PMNs) represent the first line of host defense, acting to eliminate invading bacteria at the site of infection.⁴ In response to an inflammatory stimulus, activated PMNs marginate, adhere to endothelium, and migrate into local inflammatory sites early in the inflammatory process.⁵ Exudative rather than circulating PMNs play the most important role in host defense at local sites. Recent studies^5- 6 have revealed that adhesion molecules on both PMNs and the endothelium, as well as chemokines at local inflammatory sites, are required for PMN recruitment.

The reported effects of malnutrition on PMN exudation into local inflammatory sites and their functional alteration are contradictory.^7- 8 Moreover, there have been no studies investigating the effects of dietary restriction on the expression of adhesion molecules on circulating PMNs or on chemokine production at local inflammatory sites. Therefore, we evaluated the effects of short-term dietary restriction on the exudation of PMNs into the peritoneal cavity, adhesion molecule expression on circulating PMNs, and peritoneal cavity chemokine levels in a murine peritonitis model. The effect of short-term dietary restriction on the function of exudated PMNs was also evaluated. Furthermore, we investigated whether refeeding can restore inhibited PMN exudation and PMN functional alterations induced by dietary restriction.

For this purpose, we used a glycogen-induced peritonitis model. Adhesion molecules and chemokines are reportedly required for peritoneal PMN recruitment in this model.^9- 11 Although bacterial peritonitis models may have more clinical relevance than this chemically induced model, the doses of bacteria affect the fate of PMNs, ie, necrosis or apoptosis, changing both PMN number and function.¹² Therefore, we believe our model is adequate for investigating adhesion molecule expression on PMNs, the chemokine production at local inflammatory sites, and subsequent PMN exudation into the peritoneal cavity.

Specific pathogen-free, 7- to 11-week-old, male C57BL/6J mice (Japan SLC, Hamamatsu, Japan) were used for the experiments. The mice were kept in animal facilities for 1 week before experiment initiation to allow acclimation. They were exposed to constant temperature (24°C) and humidity (60%) and were fed standard mouse chow (Oriental Koubo, Tokyo, Japan). Regular mouse chow contains protein, fat, carbohydrate, cellulose, minerals, and a vitamin mix (24.6, 5.6, 6.4, 3.1, 3.5, and 0.4 g per 100-g diet, respectively). All studies were performed in accordance with the Guide for Animal Experimentation, Faculty of Medicine, The University of Tokyo. Our institutional review board approved the protocol.

Mice (n = 174) were randomly assigned to 3 groups. The ad libitum, moderate, and severe diet-restricted groups received mouse chow ad libitum, 132 g/kg (2008 kJ/kg [480 kcal/kg]), 66 g/kg (1004 kJ/kg [240 kcal/kg]), and 33 g/kg (502 kJ/kg [120 kcal/kg]) daily for 7 days, respectively. Our preliminary experiment revealed that the average chow consumption by mice with free access to chow was 132 g/kg per day.

After 7 days of dietary restriction, all mice in the 3 groups were intraperitoneally administered 2 mL of a 1% glycogen solution (Sigma-Aldrich, St Louis, Mo). Mice were killed before (0 hour) or after glycogen installation. A heparinized whole blood sample was obtained via cardiac puncture. The total numbers of leukocytes were measured by hemocytometer (Celltac, MEK-6258; Nihon Kouden, Tokyo). Differential counts of leukocytes were performed with Wright-Giemsa staining (Muto Pure Chemicals, Tokyo). Whole blood samples were centrifuged at 400g for 5 minutes to remove plasma and then stored in 1-mL quantities with RPMI-1640 (Nikken Biomedical Laboratory, Kyoto, Japan) supplemented with 1% fetal calf serum at 4°C until sample measurement. Peritoneal exudative cells (PECs) were recovered by lavaging the peritoneal cavity with 5 mL of phosphate-buffered saline (PBS; Nikken Biomedical Laboratory) without Ca²⁺ or Mg²⁺. Supernatants of the peritoneal lavage fluid (PLF) were collected and stored at −70°C until the cytokine assay. Contaminating erythrocytes were lysed with distilled water. The PECs were then resuspended in RPMI-1640 supplemented with 1% fetal calf serum. Live PECs were counted by hemocytometer and adjusted to 1 × 10⁶/mL. Cytocentrifuged PECs were fixed in methanol. The differential cell counts were performed by the Wright-Giemsa staining technique.

Adhesion molecule expressions on circulating and exudative PMNs were analyzed at 0, 2, 4, and 8 hours after glycogen installation. The monoclonal antibodies used were R-phycoerythrin–conjugated rat antimouse CD11b (integrin α_M β₂ chain), fluorescein isothiocyanate–conjugated rat antimouse CD18 (integrin β₂ chain), and CD62L monoclonal antibodies (L-selectin) (Pharmingen, San Diego, Calif). One hundred microliters of the PEC suspension and washed whole blood samples were incubated with saturating amounts of anti-CD11b, anti-CD18, anti-CD62L monoclonal antibodies, or the related isotype antibodies for 30 minutes at 4°C. At the end of incubation, the cells were washed twice with cold PBS followed by fixation with 0.5 mL of 1% paraformaldehyde (PFA). Each whole blood sample underwent erythrocyte lysis with 2 mL of FACSlysing solution (Beckton Dickinson Immunocytometry Systems, San Jose, Calif) for 10 minutes at room temperature before washing with PBS and stored in 1% PFA at 4°C until the flow cytometric analysis.

Cytokine levels in PLF were evaluated at 2 and 4 hours after glycogen installation. The levels of tumor necrosis factor α (TNF-α); interleukin (IL) 6; IL-10; and a C-X-C (PMN-specific) chemokine, macrophage inflammatory protein 2 (MIP-2), in the supernatants were measured using commercially available enzyme-linked immunosorbent assay kits, all from Genzyme Co, Cambridge, Mass (TNF, a mouse TNF-α immunoassay; IL-6, a mouse IL-6 immunoassay; IL-10, a mouse IL-10 immunoassay; MIP-2, a mouse MIP-2 immunoassay).

Opsonic activity in PLF was evaluated as the phagocytic index of PLF according to Deitch et al¹³ at 2 and 4 hours after glycogen installation. The PECs used in the experiment were harvested from other ad libitum–fed mice (n = 12) and pooled, at 4 hours after 2 mL of 1% glycogen solution had been instilled intraperitoneally, in the same way as described in the "Sample Preparation" section. Isolated PECs (1 × 10⁶/100 µL), together with the PLF (80 µL) samples from which they had been harvested, were incubated with fluorescent latex beads (Fluoresbrite, 1.0 µm in diameter; Polysciences Inc, Warrington, Pa) for 60 minutes at 37°C in a shaking water bath. The PECs-beads ratio was 1:50. After incubation, the PECs were washed 3 times with PBS and fixed with ice-cold 1% PFA before flow cytometric analysis.

Phagocytosis by circulating and exudative PMNs was determined according to the method of Dunn and Tyrer¹⁴ at 2, 4, and 8 hours after glycogen installation. Briefly, isolated PECs (140 µL) and washed whole blood samples (70 µL) were incubated with RPMI-1640 containing fluorescent latex beads (20 µL in PEC samples, 10 µL in whole blood samples) and 20% pooled mouse serum in a metabolic shaker at 37°C for 60 minutes. After the incubation, red blood cells in whole blood samples were lysed with 2 mL of FACS-lysing solution for 10 minutes at room temperature, and both PEC and whole blood samples were washed 3 times with PBS and fixed with ice-cold 1% PFA before flow cytometric analysis.

Intracellular reactive oxygen intermediate (ROI) productions by circulating and exudative PMNs were determined, with a modification of previously described protocols,¹⁵ at 2, 4, and 8 hours after glycogen installation. Isolated PECs and washed whole blood samples were stimulated with 100 ng/mL of phorbol myristate acetate. After a 20-minute incubation at 37°C, the reaction was stopped by placing the samples on ice. The samples were then incubated with 7.5-µmol/L dihydrorhodamine 123 (Molecular Probes Inc, Eugene, Ore) for 10 minutes at 37°C, and the reaction was stopped by placing the samples on ice. Red blood cells in whole blood samples were lysed with 2 mL of FACS-lysing solution for 10 minutes at room temperature, and both PEC and whole blood samples were washed twice with PBS and fixed with ice-cold 1% PFA before flow cytometric analysis.

The flow cytometric analysis was performed using FACScan (Becton Dickinson Immunocytometry Systems). In each PEC sample, 10 000 leukocytes and, in whole blood samples, 20 000 leukocytes were counted. The PMNs were gated using morphologic characteristics displayed on a dot plot of forward light scatter vs side scatter. The results of adhesion molecule expression, phagocytosis, ROI production, and opsonic activity were expressed as mean channel fluorescence intensity (MCF).

The mice (n = 30) were divided into the 3 groups as described in experiment 1. The chow used in experiment 2 was the same as in experiment 1. After 7 days of dietary restriction, all the groups were fed ad libitum for 1 day. After 1 day of refeeding, all mice were injected intraperitoneally with 2 mL of 1% glycogen solution. At 4 hours after glycogen installation, all mice were killed. The numbers of PMNs, CD11b, CD18, and CD62L expressions on PMNs, phagocytosis, and ROI production by PMNs were measured in both circulating and exudative cells. Opsonic activities in the PLF were also evaluated.

Results are presented as means ± SEMs. One-way analysis of variance, followed by Fisher adjustment, was used for statistical comparisons. Linear regression analysis was also performed. Differences were defined as statistically significant at P<.05.

Body weight percentages (body weight at death, predietary restriction) were significantly lower in both diet-restricted groups than in the ad libitum group (76.5% ± 0.7% in the severe diet-restricted group, 89.8% ± 0.8% in the moderate diet-restricted group, and 107.1% ± 0.7% in the ad libitum group, respectively; P<.01). Body weight percentages were also significantly lower in the severe than in the moderate diet-restricted group (P<.01).

The numbers of circulating PMNs in all 3 groups peaked at 2 hours after glycogen installation (Figure 1). The number of exudative PMNs in the ad libitum group peaked at 2 hours after glycogen installation, whereas numbers of exudative PMNs peaked at 4 hours after glycogen installation in both of the diet-restricted groups (Figure 2). Moreover, the numbers of exudative PMNs were significantly lower in both diet-restricted groups than in the ad libitum group at 2 hours after glycogen installation. At 4 hours after glycogen installation, the number of exudative PMNs was also significantly lower in the severe diet-restricted than in the ad libitum group.

There was no difference in CD11b or CD18 expression on circulating PMNs among the 3 groups before glycogen installation (Table 1). At 2 hours after glycogen installation, neither CD11b nor CD18 expression on circulating PMNs differed significantly among groups, but both expressions tended to be lower in diet-restricted groups than in the ad libitum group. At 4 hours after glycogen installation, CD18 expression on circulating PMNs was significantly lower in both diet-restricted groups than in the ad libitum group. CD11b expression on circulating PMNs was also significantly lower in both diet-restricted groups than in the ad libitum group at 8 hours after glycogen installation. At 2 hours after glycogen installation, both CD11b and CD18 expression on circulating PMNs correlated significantly with the numbers of exudative PMNs (CD11b: n = 22, r = 0.54, P = .008, CD18: n = 22, r = 0.54, P = .01) (Figure 3). In contrast, CD62L expression on circulating PMNs was not affected by dietary restriction. Expressions of CD11b, CD18, and CD62L on exudative PMNs did not differ among the 3 groups before and after glycogen installation.

At 2 hours after glycogen installation, levels of TNF-α in PLF were significantly lower in the moderate diet-restricted group than in the ad libitum group, and levels of TNF-α in PLF tended to be lower in the severe diet-restricted than in the ad libitum group (P = .05 vs ad libitum) (Table 2). Levels of IL-6 and IL-10 in PLF were significantly lower in both diet-restricted groups than in the ad libitum group at 2 hours after glycogen installation. Moreover, PLF levels of MIP-2, a murine chemokine, were significantly lower in both diet-restricted groups than in the ad libitum group at 2 hours after glycogen installation. There was a significant positive correlation between the number of exudative PMNs and levels of MIP-2 in PLF at 2 hours after glycogen installation (n = 72, r = 0.49, P<.001, 95% confidence intervals: 0.015 to 0.042 for the line and 7.347 to 18.604 for the data). At 4 hours after glycogen installation, there was no difference in the PLF levels of cytokines and MIP-2 among the 3 groups. Opsonic activity in PLF at 4 hours after glycogen installation was significantly lower in the severe diet-restricted than in the ad libitum group.

Phagocytosis by circulating PMNs was unaffected by dietary restriction at 2, 4, and 8 hours after glycogen installation (data not shown). However, phagocytosis by exudative PMNs was significantly lower in both diet-restricted groups than in the ad libitum group at 4 hours after glycogen installation (Table 3).

Intracellular ROI production by circulating PMNs was also unaffected by dietary restriction at 2, 4, and 8 hours after glycogen installation (data not shown). However, ROI production by exudative PMNs, at 4 hours after glycogen installation, was significantly higher in the severe diet-restricted group than in the moderate diet-restricted and ad libitum groups (Table 3).

After 1 day of refeeding, the body weight percentages in both diet-restricted groups had recovered to nearly the same value as in the ad libitum group. There was no difference in the number of circulating PMNs between the ad libitum and the severe diet-restricted groups at 4 hours after glycogen installation (ad libitum: 15.3 ± 3.2 × 10⁵/mL; moderate: 13.0 ± 2.6 × 10⁵/mL; severe: 12.7 ± 2.4 × 10⁵/mL). Moreover, after 1 day of refeeding, at 4 hours after glycogen installation, there was no difference in the number of exudative PMNs between the ad libitum and the severe diet-restricted groups (ad libitum: 20.0 ± 3.9 × 10⁵/mouse; moderate: 41.9 ± 8.9 × 10⁵/mouse; severe: 23.6 ± 5.8 × 10⁵/mouse).

The difference in CD18 expression on circulating PMNs between the severe diet-restricted and ad libitum groups, at 4 hours after glycogen installation, disappeared with 1 day of refeeding (Table 4). After 1 day of refeeding, at 4 hours after glycogen installation, there was no significant difference in phagocytosis by exudative PMNs (ad libitum: 15.8 ± 2.2 MCF; moderate: 10.7 ± 1.4 MCF; severe: 13.7 ± 1.9 MCF), in ROI production by exudative PMNs (ad libitum: 30.7 ± 2.6 MCF; moderate: 24.4 ± 2.5 MCF; severe: 37.3 ± 3.6 MCF), or in PLF opsonic activities (ad libitum: 11.1 ± 3.1 MCF; moderate: 14.7 ± 2.3 MCF; severe: 10.3 ± 1.2 MCF) among the 3 groups.

The present study demonstrated that 7 days of dietary restriction decreased CD11b/CD18 expression on circulating PMNs, MIP-2 levels in PLF, and subsequent PMN exudation into local inflammatory sites at an early phase of inflammation in glycogen-induced peritonitis. Seven days of dietary restriction also impaired phagocytosis but up-regulated ROI production by exudative PMNs. Only 1 day of ad libitum refeeding improved depressed PMN exudation into the peritoneal cavity, depressed phagocytosis, and the up-regulated ROI production by exudative PMNs in the diet-restricted groups.

The effects of dietary restriction on the number of exudative PMNs at local inflammatory sites have been controversial. The total numbers of leukocytes mobilized into skin abrasions were similar in kwashiorkor and well-nourished control patients.⁷ On the other hand, numbers of leukocytes exudated into the peritoneal cavity in response to glycogen installation are lower in mice fed protein-depleted chow than in well-nourished mice.⁸ However, these studies did not differentiate exudative PMNs from leukocytes. Our results demonstrated that 7 days of dietary restriction delayed the phase and decreased the magnitude of PMN influx into the peritoneal cavity in this murine peritonitis model.

Exudation of PMN into local inflammatory sites is controlled by the adhesion molecules on both circulating PMNs and the endothelium, by the migratory capacity of PMNs, and by the presence of chemotactic agents at local inflammatory sites.⁵ Several adhesion molecules and cytokines are reportedly required for PMN recruitment in glycogen peritonitis models.^9- 11

Adhesion molecules, such as CD11b, CD18, and CD62L, play important roles in PMN recruitment during glycogen-induced peritonitis.^9,11 In this study, dietary restriction reduced CD18 expression on circulating PMNs at 4 hours after glycogen installation. In addition, at an earlier stage, ie, 2 hours after glycogen installation, there was a significant positive correlation between the numbers of exudative PMNs entering the peritoneal cavity and CD11b/CD18 expressions on circulating PMNs. The results are consistent with our preliminary report that 1 week of severe dietary restriction significantly reduced CD11b and CD18 expressions on circulating PMNs at 4 hours after intraperitoneal glycogen installation.¹⁶ Thus, the reduced CD11b/CD18 expression on circulating PMNs may, at least in part, be responsible for the decreased PMN exudation at an early phase of inflammation in our murine model.

The specific cause of the reduced CD11b/CD18 expression on circulating PMNs in this study is not known at present. A possible cause is stress hormone alteration secondary to dietary restriction. One day of starvation reportedly decreases circulating insulinlike growth factor 1 levels in rats.¹⁷ Insulinlike growth factor 1 is a powerful primer for the decreased CD11b/CD18 expression on circulating PMNs.¹⁸ On the other hand, energy intake restriction elevates circulating free cortisol levels.¹⁹ Cortisol is a potent down-regulator of CD18 expression on circulating PMNs.²⁰ Neither insulinlike growth factor 1 nor free cortisol levels were measured in our study. Moreover, the effects of dietary restriction on both the intracellular storage of CD11b/CD18 and the translocation of these adhesion molecules are still unclear. Further study is needed to clarify the precise cause and mechanisms of reduced CD11b/CD18 expression on circulating PMNs with 1 week of dietary restriction.

Among various cytokines needed for PMN recruitment into the peritoneal cavity in glycogen-induced peritonitis, MIP-2, a murine C-X-C (PMN-specific) chemokine and a homologue of human IL-8, is important because of its strong chemotactic capacity.^6,21 In our study, levels of MIP-2 in PLF were significantly lower in both diet-restricted groups than in the ad libitum group at 2 hours after glycogen installation. This is the first report demonstrating dietary restriction to significantly reduce chemokine levels at local inflammatory sites. Moreover, there was a significant positive correlation between PLF MIP-2 levels and the number of exudative PMNs in the peritoneal cavity at 2 hours after glycogen installation. Thus, decreased MIP-2 production in the peritoneal cavity may be another important factor that accounts for reduced PMN exudation at an early phase of inflammation in this murine model.

Reportedly, MIP-2 was produced mainly by peritoneal macrophages.²² The decreased MIP-2 levels in the PLF in the diet-restricted groups may be induced by a decrease in the number of resident peritoneal macrophages, decreased cytokine stimulation of macrophages, and/or decreased MIP-2 production within macrophages. The numbers of peritoneal resident macrophages, before glycogen installation, were significantly lower in both diet-restricted groups than in the ad libitum group in this study. The number of peritoneal resident macrophages in the moderate diet-restricted group was 45% of that in the ad libitum group (our unpublished data). The number in the severe diet-restricted group was only 35% of that in the ad libitum group.

Production of MIP-2 is known to be up-regulated by TNF-α²³ and down-regulated by IL-10.²⁴ In our study, levels of TNF-α in PLF were lower in the diet-restricted groups than in the ad libitum group at 2 hours after glycogen installation. Thus, it is possible that decreased TNF-α levels in the diet-restricted groups induced the decrease in PMN exudation by reducing MIP-2 production in our model. On the other hand, we found that levels of IL-6 and IL-10 in PLF were also lower in both diet-restricted groups than in the ad libitum group at 2 hours after glycogen installation. Chemotactic capacities of IL-6 have not been reported. In contrast, IL-10 reportedly inhibited MIP-2 production from peritoneal macrophages.²⁴ However, the diet-restricted groups had reduced PLF levels of both IL-10 and MIP-2. Thus, we can rule out the possibility that IL-10 down-regulates MIP-2 production in an early phase of the murine peritonitis described herein.

A recent study²⁵ revealed decreased intracellular signaling within macrophages to be a possible reason for decreased TNF-α and IL-6 production from peritoneal macrophages. It is suggested that low levels of MIP-2 in the diet-restricted groups in our study could be secondary to decreased intracellular signaling within peritoneal macrophages. Taken together, the decreased number of resident macrophages, the decreased production of TNF-α by resident macrophages in the peritoneal cavity, and possibly impaired intracellular signaling within peritoneal macrophages may be responsible for the decreased MIP-2 production in PLF observed in our study.

The complements, especially C5a, have a chemotactic effect on PMNs.²⁶ Our study also demonstrated that 1 week of dietary restriction significantly reduced PLF opsonic activity. Serum opsonin levels are well documented to be decreased in malnutrition.²⁷ However, the effects of dietary restriction on levels of opsonic activity at local inflammatory sites are unclear. The major opsonins include specific antibodies and complement components. The mice used in this experiment were not given fluorescent beads beforehand. Therefore, opsonic activity measured in our study may reflect mainly levels of complements in PLF. Thus, decreased levels of complements in the peritoneal cavity in the severe diet-restricted group, at 4 hours after glycogen installation, may also decrease exudation of PMNs into the peritoneal cavity in our model.

Reported effects of malnutrition on PMN phagocytosis are contradictory.^28- 29 Moreover, investigations concerning the adverse effects of malnutrition on PMNs are based primarily on circulating PMNs.^28- 29 Exudative, rather than circulating, PMNs play the most important role in host defense at the local site. Our results demonstrated that 7 days of dietary restriction significantly reduced phagocytosis by exudative PMNs at 4 hours after glycogen installation. The beads used in our experiment were opsonized with pooled serum from mice fed ad libitum in all 3 groups. Therefore, factors influencing phagocytosis by exudative PMNs in our study depend mainly on factors related to the PMN itself. Reportedly, PMN phagocytosis is influenced by the Fc receptor (CD16) and complement receptors on PMN.¹⁸ However, there are no reports that suggest that dietary restriction affects either complement or Fc receptor expression on exudative PMNs. Moreover, CD11b/CD18 (CR3) expression on exudative PMNs was not influenced by 7 days of dietary restriction in our study. Therefore, additional studies must be done to elucidate the effects of dietary restriction on complement and Fc receptors on PMNs.

Eight weeks of protein depletion decreased superoxide production by peritoneal resident macrophages.³⁰ Contrary to our expectation, in this model, 1 week of severe dietary restriction significantly up-regulated phorbol myristate acetate–stimulated ROI production by peritoneal PMNs at 4 and 8 hours after glycogen installation. Factors influencing this enhanced ROI production by exudative PMNs are not precisely known. Possible mechanisms of increased ROI production by exudative PMNs include reduced antioxidant components or up-regulated neuroendocrine responses with 7 days of severe dietary restriction. Three days of starvation reportedly lead to the depletion of liver antioxidant stores with accelerated release of hepatic oxygen free radicals in a rat model.³¹ On the other hand, plasma catecholamine levels were increased after starvation.³² High catecholamine levels reportedly induce PMN superoxide generation.³³ Therefore, 7 days of dietary restriction may have up-regulated ROI production by decreasing antioxidant stores in peritoneal PMNs and/or by increasing local catecholamine levels. Nevertheless, the up-regulated ROI production by exudative PMNs may compensate for the decreased numbers of PMNs exudated into the peritoneal cavity in response to 7 days of dietary restriction.

We evaluated the effects of refeeding at 4 hours after glycogen installation in our refeeding experiment. At this time point, differences in the numbers of exudative PMNs, the functions of PMNs, and opsonic activity levels in PLF between the ad libitum and the diet-restricted groups were most prominent. In this experiment, 1 day of refeeding normalized decreased exudation of PMNs into the peritoneal cavity. Moreover, 1 day ad libitum refeeding restored CD11b/CD18 expression on circulating PMNs and phagocytosis, reversed up-regulated ROI production by exudative PMNs, and normalized depressed PLF opsonic activity in the severe diet-restricted group. Refeeding improves impaired cell-mediated and nonspecific immunity induced by malnutrition. Nutritional recovery reportedly occurred in in vitro cytokine production from monocytes after 3 weeks of refeeding.³⁴ However, increased susceptibility to Escherichia coli infection by malnourished mice was reversed by refeeding for only 1 day.³⁵ Our results appear to support rapid recovery of impaired host defense mechanisms induced by 7 days of dietary restriction. We cannot determine which component is responsible for the rapid recovery, because we examined protein and energy (calorie) malnutrition, as well as protein-energy refeeding, in the present study. Nonetheless, the changes are more likely related to carbohydrate provision than to restoration of lean body mass or correction of any deficits in fat or protein, since changes in body composition cannot occur over such a short time.

The results obtained from our study may account for observations made in patients with severe dietary restriction, although the relevance of the model used herein to clinical practice must be considered. Our recent in vitro investigation revealed that in preoperative patients malnutrition depresses adhesion molecule expression on circulating PMNs, inhibits PMN adhesion to cultured human umbilical vein endothelial cells, and reduces PMN migration beneath human umbilical vein endothelial cell monolayers. In addition, in patients with leukocyte adhesion deficiency 1 syndrome, which is characterized by a deficiency of CD11b/CD18 molecules on PMNs, there is impaired PMN exudation, resulting in recurrent bacterial infection.³⁶ Taken together, the results of this study imply that surgical patients subjected to dietary restriction may develop postoperative infectious complications in part via impaired PMN exudation into local inflammatory sites secondary to inhibited adhesion molecule expression on circulating PMNs and by decreased chemokine production at local sites. Nutritional replenishment even for a short period in diet-restricted patients may improve host defense via restoration of these functional abnormalities of PMNs and chemokine production at local inflammatory sites.

Corresponding author and reprints: Hideaki Saito, MD, Surgical Center, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan (e-mail: saito-ope@h.u-tokyo.ac.jp).