点击显示 收起
1 From the Department of Nutritional Sciences (NA and JS), the Graduate Program in Nutrition (DK), the Department of Biochemistry and Molecular Biology (AM), and the University Health Services (GH), The Pennsylvania State University, University Park, PA.
2 Supported by research grants 96-35200-3132 and 2001-35200-10722 from the US Department of Agriculture and a grant from the National Cattlemen's Beef Association. 3 Reprints not available. Address correspondence to N Ahluwalia, S126 Henderson, Department of Nutritional Sciences, The Pennsylvania State University, University Park, PA 16802. E-mail: nxa7{at}psu.edu.
ABSTRACT
Background: Aging is often associated with a dysregulation of immune function. Iron deficiency may further impair immunity in older adults. Published reports on iron deficiency and immune response in humans are inconsistent. Most studies are focused on young children in developing countries and are often confounded by comorbid conditions, infections, and nutrient deficiencies.
Objective: Our objective was to determine the relation of iron status with immune function in homebound older women, who often have impairments in both iron status and immune response. The subjects were selected according to rigorous exclusion criteria for disease, infection, and deficiencies in key nutrients known to affect immunocompetence.
Design: Seventy-two homebound elderly women provided blood for comprehensive evaluation of iron status and cell-mediated and innate immunity. Women were classified as iron-deficient or iron-sufficient on the basis of multiple abnormal iron status test results. Groups were compared with respect to lymphocyte subsets, phagocytosis, oxidative burst capacity, and T cell proliferation upon stimulation with mitogens.
Results: In iron-deficient women, T cell proliferation upon stimulation with concanavalin A and phytohemagglutinin A was only 40-50% of that in iron-sufficient women. Phagocytosis did not differ significantly between the 2 groups, but respiratory burst was significantly less (by 28%) in iron-deficient women than in iron-sufficient women.
Conclusions: Iron deficiency is associated with impairments in cell-mediated and innate immunity and may render older adults more vulnerable to infections. Further prospective studies using similar exclusion criteria for disease, infection, and concomitant nutrient deficiencies are needed for simultaneous examination of the effects of iron deficiency on immune response and morbidity.
Key Words: Iron status • cell-mediated immunity • phagocytosis • oxidative burst capacity • elderly women
INTRODUCTION
Optimal nutrition is important for maintaining immunity (1, 2). Aging is often associated with a decline in immunocompetence, particularly in T cell-mediated functions (3, 4). Nutrient deficiencies can contribute to further impairments in immune function in older adults and thus render them even more vulnerable to acute and chronic infections and disease than they would be if they were nutrient sufficient (5-7). The role of various nutrients in maintaining immune function in the elderly has been the focus of several studies (5-10). However, the effect of iron deficiency on immune response in older adults has not been fully examined, despite indications of a fairly high prevalence of iron deficiency among frail, homebound, or institutionalized elderly (11-15). In a study examining the relation between nutritional status and response to influenza vaccine among institutionalized elderly, Fulop et al (10) reported that the vaccine-nonresponsive group had significantly lower hemoglobin, hematocrit, and serum iron concentrations than did the vaccine-responsive group, which suggested that iron-deficient older adults may have a low cell-mediated immune response.
The relation of iron status and immune function in humans has been addressed primarily in studies of infants, young children, and adults (16-26). Impaired cell-mediated immunity (CMI) and bactericidal function are generally noted in iron-deficient persons; however, the findings are inconsistent (18-27). The conflicting results with respect to iron deficiency and immunocompetence may be related to the different ages of the subjects, the presence of underlying inflammation or coexisting nutritional deficiencies, the various operational definitions for iron deficiency, and differences in the methods of assessment of immune function across studies. Thus, our interest was to comprehensively determine the relation of iron status to measures of CMI and innate immunity in homebound older women who are at risk of developing iron deficiency (11-15); the presence of confounding factors such as poor health, inflammation, and other nutrient deficiencies was adjusted for by the use of rigorous screening criteria (1, 5, 28).
SUBJECTS AND METHODS
Subjects
Homebound older women (>60 y old; n = 225) who were receiving services such as meals or assistance with activities of daily living from the Office on Aging and were free of chronic conditions known to affect immune response were first identified from the medical records of the Office on Aging in 3 rural counties (Blair, Centre, and Clearfield) in Pennsylvania. They were contacted via flyers and letters of invitation for participation in the study, and 119 women expressed interest in participating in the study. All of these women were contacted to obtain a verbal medical history to rule out the presence of any chronic medical conditions or acute or chronic infections and the current use of any medication known to affect immune response, according to the guidelines of the SENIEUR protocol (28). Eighty women met the health criteria of SENIEUR protocol, and 72 agreed to provide blood samples and to participate in the study. None of the women had taken iron supplements routinely in the 3 mo before enrollment into the study. Subjects provided written informed consent according to protocols approved by the Office for Regulatory Compliance at the Pennsylvania State University. Subjects received an honorarium of $50 at the completion of the study.
Study protocol
Apparently healthy, homebound women (n = 72) provided fasting blood samples at home after a 15-min period of rest. Height and weight were recorded by using portable, standardized instruments, and these measurements were used to exclude women with body mass index (in kg/m2) <20 according to the SENIEUR protocol (28). Venous blood (35 mL) was collected by a certified phlebotomist into 3 kinds of evacuated tubes (Vacutainer; Becton Dickinson, Franklin Lakes, NJ) containing heparin, EDTA, or no additive (trace mineral-free tube) for screening of health and nutritional status by using laboratory tests.
Health screen
A complete blood count (CBC) with differential evaluation was carried out on a Coulter MAXM analyzer (Beckman Coulter Corporation, Miami). A clinical chemistry panel (Chem-24 profile) was performed by using a Roche Mira Plus random-access chemistry analyzer (Roche Boehringer Mannheim Corporation, Indianapolis). Results of these tests were reviewed by a physician (GH) to exclude subjects with infection, inflammation, liver disorders, kidney disorders, or bone marrow proliferative disorders or any combination of those conditions; 4 subjects were excluded.
To assess inflammatory status, the erythrocyte sedimentation rate was measured by using the Westergren method (Dispette72, Ulster Medical Products, Rio Rancho, NM). Serum 1-acid glycoprotein was determined by using radial immunodiffusion (Kent Laboratories, Redmond, WA). An elevated white blood cell count was also considered as an indicator of inflammatory status. The cutoffs for abnormal results were >30 mm/h for erythrocyte sedimentation rate, >11 x 103/mm3 for white blood cell count, and >1.4 g/L for 1-acid glycoprotein (29). Persons with abnormal results on =" BORDER="0">2 of these 3 tests of inflammatory status were excluded from analysis (n = 3).
Nutritional status screening criteria
Laboratory tests were performed to ensure the adequacy of protein, vitamin B-12, folic acid, and zinc status because the deficiency of these nutrients is associated with impairments in immune response (1, 2, 5, 8, 21, 24, 30-33). Serum total protein and serum albumin concentrations were measured as part of the clinical chemistry profile. Serum vitamin B-12 and folate concentrations were measured by using a commercial radioassay (ICN Pharmaceuticals, Orangeburg, NY). Plasma zinc was measured by using atomic absorption spectrophotometry.
Two subjects with serum total protein concentrations <60 g/L, serum albumin concentrations <35 g/L, or both were excluded from analysis (29). To exclude persons with underlying vitamin B-12 or folic acid deficiency, 2 subjects with serum vitamin B-12 concentrations <250 ng/L (34), serum folate concentrations <2 µg/L (29), or mean corpuscular volume >102 fL (35) were dropped from analysis. None of the subjects had suboptimal zinc status (plasma zinc <600 µg/L) (29). Two women had to be excluded from analysis because their blood samples were insufficient to allow the conduct of all of the screening and immune function tests.
Tests of iron status and characterization of subjects as iron-sufficient or iron-deficient
Of the 72 women enrolled into the study, 59 met the inclusion criteria, and they underwent a comprehensive assessment of iron status conducted with a panel of laboratory tests. Serum ferritin was assayed by using a radioimmunoassay (Diagnostic Products Corporation, Los Angeles). Serum transferrin receptors were assayed by using a commercial enzyme-linked radioimmuno-assay (Ramco Laboratories, Houston). Serum iron, total-iron-binding capacity, and transferrin saturation were analyzed by using colorimetric methods of the American Medical Laboratories (Chantilly, VA). Hemoglobin, hematocrit, and other red blood cell indexes were obtained from the CBC analysis. Women whose iron stores were depleted (serum ferritin concentration: <20 µg/L; 11, 362 other tests of iron status were considered iron-deficient (35).
Immune function
For sample volume considerations, immune function values were determined by using whole-blood assays under previously established protocols that had been validated in our laboratory (37).
Lymphocyte phenotypes
Lymphocyte subsets in 1 mL heparin-treated blood were estimated with the use of flow cytometry (EPICS XL; Beckman Coulter, Miami) and fluorescence-labeled monoclonal antibodies (Becton Dickinson Pharmingen, San Diego) that were specific for surface antigens—ie, anti-CD3 recognizes total T cells, anti-CD4 recognizes total helper T cells, anti-CD8 recognizes total cytotoxic T cells, anti-CD56,16 recognizes natural killer cells, and anti-CD19 recognizes B cells. The total lymphocyte number, obtained from the CBC with differential evaluation, was used to compute the absolute numbers of cells for various lymphocyte subsets.
T cell proliferation upon stimulation with mitogens
Proliferation of T cells in response to several concentrations of phytohemagglutinin (PHA) and concanavalin A (Con A) were determined by measuring the incorporation of [3H]thymidine (6.7 Ci/mmol; 38). The PHA and ConA concentrations assayed were 5 and 10 mg/L and 3, 12, and 25 mg/L, respectively, based on our previous study (37). Heparinized blood was diluted (1:10) with RPMI 1640 medium (Cellgro; MediaTech, Herndon, VA) containing 10% fetal bovine serum, L-glutamine (2 mmol/L), penicillin (100 000 U/L), and streptomycin (100 mg/L). Diluted blood and mitogens (100 µL each) were added to wells in a 96-well, round-bottom microtiter plate. For each mitogen at each concentration, 6 replicate measurements per subject were carried out. Cells were incubated for 42 h (humidified, in 5% CO2, at 37 °C), and then 10 µL [3H]thymidine was added to each well, the cells were incubated for another 6 h, and then the cells were harvested onto glass fiber filters (Skatron 7025 Combi Cell Harvester; Skatron Inc, Sterling, VA). Incorporation of [3H]thymidine (cpm) into cellular DNA was determined by using a Wallac 1205 beta plate counter (EG&G Wallac Inc, Gaithersburg, MD). Results are presented after normalization for T cell concentration in blood and are expressed per 1000 T cells.
Phagocytes, phagocytosis, and oxidative burst capacity
Monocytes were estimated by using flow cytometry (EPICS XL; Beckman Coulter) with fluorescence-labeled monoclonal antibodies that were specific for surface antigens (eg, anti-CD14 recognizes monocytes). The total lymphocyte number, obtained from the CBC with differential evaluation, was used to compute the absolute number of monocytes. The granulocyte number was obtained from the CBC with differential evaluation.
Phagocytosis was determined by using flow cytometry and a commercial kit (PhagoTest; Orpegen Pharmaceuticals, Heidelberg, Germany). Heparinized blood was incubated with fluorescein isothiocyanate-labeled opsonized Escherichia coli for 10 min at 37 °C, which was followed by the quenching of fluorescence associated with uningested bacteria. An EPICS XL flow cytometer was used to determine the percentage of granulocytes containing fluorescence (ie, expressing phagocytosis) and the average fluorescence intensity per cell (ie, the number of bacteria engulfed per granulocyte).
Oxidative burst capacity of granulocytes, an indirect measure of bactericidal function, was also measured (BurstTest; Orpegen Pharmaceuticals). Heparinized blood was incubated with E. coli bacteria for 10 min at 37 °C, which was followed by the addition of dihydrorhodamine 123, which fluoresces on reaction with oxygen species produced during the oxidative burst. The percentage of fluorescent granulocytes (ie, cells expressing oxidative burst) and the fluorescence intensity per cell (ie, the degree of oxidative burst per granulocyte) were measured by using flow cytometry.
Statistical analyses
Statistical analyses were performed by using SAS software for WINDOWS (version 8.1; SAS Institute Inc, Cary, NC). The distributions for all variables were first examined for normality. Logarithmically transformed data for body mass index, serum ferritin concentrations, leukocyte subsets, and lymphocyte proliferation response to stimulation with PHA and Con A were used for statistical analysis because the transformed distributions were consistent with normality.
With the use of the iron status assessment criteria described above, healthy women (n = 59) who were free of underlying inflammation and nutritional deficiencies were categorized into 2 groups: iron-deficient (age: 76.1 ± 6.6 y; n = 12) and iron-sufficient (age: 76.9 ± 5.5 y; n = 47). Differences between iron-deficient and iron-sufficient women with respect to nutritional status variables and tests of inflammation were first evaluated by using Student's t test. Analysis of variance was carried out to examine the differences between groups in all immune function variables examined. Significance was set for all analyses at P < 0.05.
RESULTS
All subjects included in the analyses were considered healthy, free of inflammation, and generally well-nourished according to the criteria used (Table 1). The results of tests of nutritional status and inflammation were in normal ranges for both iron-sufficient and iron-deficient groups, and the 2 groups did not differ significantly with respect to any of these variables (Table 1). As expected, the iron-deficient group had significantly lower iron status than did the iron-sufficient group (Table 2). The iron-deficient group had significantly (P < 0.05) lower values for all body iron compartments—namely, iron stores (lower serum ferritin), tissue iron (elevated transferrin receptor), transport iron (lower serum iron and transferrin saturation), and red blood cell iron (lower hemoglobin and hematocrit)—than did the iron-sufficient group. Approximately one-half of the iron-deficient subjects were anemic, and the remainder had iron-deficient erythropoiesis. No differences were seen between the ironsufficient and iron-deficient groups with respect to total lymphocytes or their subpopulations, granulocytes, or monocytes, when they were expressed as absolute counts or percentages (P > 0.10; data not shown); these values fell within normal ranges in both groups and were similar to those reported previously (37). T cell proliferation in response to stimulation with various concentrations of Con A and PHA was significantly less in iron-deficient women than in their iron-sufficient counterparts (Table 3). Phagocytosis, the extent to which granulocytes ingested E. coli bacteria, did not differ significantly by iron status. The number of granulocytes expressing respiratory burst on ingestion of E. coli did not differ significantly between the iron-deficient and iron-sufficient groups. However, the magnitude of oxidative burst expressed by granulocytes was significantly (28%) less in iron-deficient women than in iron-sufficient women, which potentially suggested a reduced capacity in the iron-deficient group for killing bacteria (Table 4).
View this table:
TABLE 1. Indicators of nutritional status and inflammation by group1
View this table:
TABLE 2. Laboratory measures of iron status by group
View this table:
TABLE 3. Lymphocyte proliferation response by group1
View this table:
TABLE 4. Phagocytosis and oxidative burst function of granulocytes by group1
DISCUSSION
The role of iron in maintaining immune response is considered important at biochemical and cellular concentrations (2, 26, 39). For example, iron is a component of the enzyme ribonucleotide reductase (EC 1.17.4.1) involved in DNA production and cell division, and iron is necessary to the activity of myeloperoxidase (EC 1.11.1.8), which is involved in killing bacteria by neutrophils (26). Iron is also important in the activation of protein kinase C (EC 2.7.1) that phosphorylates various factors regulating cell proliferation (39). Reports on iron deficiency and immune dysfunction in humans are often confounded by the presence of other nutritional deficiencies and subclinical infection or inflammation, which can perturb both tests of iron status and immune function (5, 11, 25-27, 33). Very few studies have examined the effect of iron deficiency on immune function after accounting for such potential confounders, and the results of such studies are mixed (23, 25, 27, 31). Thus, we conducted a carefully designed study, first, to rule out the presence of confounding factors (inflammation and other nutrient deficiencies) and, second, to examine the differences between iron-deficient and iron-sufficient older women with respect to immune function. We used a whole-blood assay that had been validated in our laboratory against an assay based on isolated mononuclear cells (37). The advantages of the whole-blood assay are its use of smaller blood volumes, the simplicity of sample preparation, and its closer mirroring of the in vivo situation.
We examined the relation of iron status to immune response in an older cohort for several reasons. Iron deficiency is quite common in homebound and institutionalized elderly (11-15). Moreover, few studies have examined iron status and immunity in older adults, despite the fact that the immune systems of this population may already be compromised because of the age-associated decline in certain T cell functions (3, 4, 37).
Most previous studies suggested that iron deficiency may have a greater effect on innate immunity and T cell-mediated functions than on humoral immunity (20, 25, 26, 40); therefore, we focused on those aspects in the current study. Iron deficiency was defined as the presence of depleted iron stores (11, 35) in conjunction with abnormal results on 2 other tests from a comprehensive panel of tests spanning tissue iron deficiency to anemia.
The results from this carefully designed cross-sectional study, in which factors such as disease, inflammation, and deficiency of key nutrients such as protein, zinc, vitamin B-12, and folic acid (1, 2, 5, 8, 24, 30-33) that may impair immune response were ruled out, suggest that iron deficiency is associated with significant impairments in both CMI and innate immunity.
It is interesting that, in the current study, the numbers and percentages (data not shown) of lymphocytes and their subsets, granulocytes, and monocytes did not differ significantly between the iron-sufficient and iron-deficient women, which is consistent with the findings of Thibault et al (18). In contrast, others have reported a decline in the lymphocyte count in iron-deficient anemic children (19) and adults (41). Declines in T cells (24, 40, 42) and helper T cells (19, 24, 42) in iron-deficient anemic persons have also been reported. The presence of comorbid conditions, infection, or other nutrient deficiencies and differences in the methods used to estimate lymphocyte subsets and in the severity of iron deficiency in subjects across the studies may partly explain differential findings.
The proliferative capacity of T lymphocytes cultured with mitogens, which mimics the response of T cells to antigens in vivo, was significantly compromised in iron-deficient women. The proliferation of T cells upon stimulation with Con A in iron-deficient women ranged from 38% (with 25 mg/L) to 46% (with 3 mg/L) of that in iron-sufficient women. Similarly, the lymphocyte proliferation response to PHA in the iron-deficient women was 42% (with 5 mg/L) and 60% (with 10 mg/L) of that in the iron-sufficient women (Table 3). Such a dramatic reduction in the proliferative capacity of T cells in response to mitogens was reported previously in studies with iron-deficient anemic children (20-22) and adults (43, 44); however, those subjects may also have had other nutrient deficiencies or subclinical infections. It is interesting that 2 studies in which folic acid and vitamin B-12 were examined concomitantly reported contrary findings (23, 31), and these studies deserve further comment. Kulapongs et al (23) did not find a reduction in T cell proliferation upon stimulation with PHA in their small sample of 8 young, severely iron-deficient Thai children who did not have folic acid and vitamin B-12 deficiencies. The study of Gross et al (31) in Africa also had few iron-deficient subjects (n = 5). In these 2 studies, however, details on protein and zinc status, both of which were likely to be low in these cohorts, were not provided for the entire study cohort (31) or the control group (23), which makes it difficult to interpret these studies' findings.
Few studies have examined phagocytosis and bactericidal function in adults. In the present study, phagocytosis did not differ significantly between iron-sufficient and iron-deficient women; this is consistent with previous studies of children (16, 17, 20, 45, 46). Our finding of reduced oxidative burst is also consistent with the findings of several groups working with children in developing countries (16, 17, 20, 22, 45, 46) who may, however, also have had other subclinical nutrient deficiencies and infection. In contrast, in a small study, Kulapongs et al (23), using a more rigorous design ensuring adequacy of folic acid and vitamin B-12 status, did not find significant differences in the bactericidal activity between iron-deficient (n = 8) and iron-sufficient (n = 20) children; however, those authors did not provide details on protein and zinc status, which possibly could have confounded the study findings. Thus the current study suggests that reduced respiratory burst by granulocytes may place iron-deficient older adults at greater risk of bacterial infections. This speculation is consistent with our finding in a randomized controlled study of young children in Sri Lanka (47) that improvement in iron status due to iron supplementation was accompanied by a reduction in the incidence and duration of upper respiratory tract infections.
The strengths of the present study include its design to exclude subjects with disease, inflammation, and underlying nutritional deficits (eg, undernutrition or deficiency of protein, vitamin B-12, folic acid, or zinc); the comprehensive evaluation of iron status on the basis of tests spanning all stages of iron deficiency (35); and the comprehensive assessment of CMI and innate immunity. Because of the cross-sectional nature of this study, however, our findings remain correlational. Experimental studies are needed to establish whether there is a causal relation between iron deficiency and immune response. Furthermore, the causes of iron deficiency, including dietary intake, hypochlorhydria, or blood loss, should also be examined in future investigations. In an ongoing study, we are investigating the effect of iron supplementation on immune function in iron-deficient older women.
In summary, by using strict criteria to establish the absence of underlying disease, inflammation, and deficiencies of key nutrients that may affect immune function, this cross-sectional study shows clear associative relations between impaired iron status and a decline in measures of innate immunity and CMI in older women. These findings are of public health significance, and they underscore the importance of diagnosing, treating, and preventing iron deficiency. The present study suggests that reduced immune function in iron-deficient cohorts may contribute to increased prevalence of infections. Future experimental studies using similarly rigorous subject-selection procedures should simultaneously examine the relation of iron status to immune function and infection.
ACKNOWLEDGMENTS
We thank the subjects for their participation in the study. We acknowledge the assistance of Cindy Pfitzenmaier in data collection and of Veronika Weaver in laboratory analyses. We are grateful to Juergen Erhardt (Institute of Biological Chemistry and Nutrition, University of Hohenheim, Stuttgart, Germany) for conducting the plasma zinc analyses.
NA was responsible for developing the study design; AM and GH contributed to the process. NA supervised the data collection by DK and JS. GH contributed to medical decisions pertaining to subject selection for the study. NA was responsible for the supervision of tests of clinical and nutritional status, inflammatory status, and phagocytosis and oxidative burst assays. AM was responsible for the supervision of lymphocyte phenotypes and proliferation assays and assisted with data collection for phagocytosis and oxidative burst assays. Data analysis was conducted by NA, JS, and DK. NA wrote the manuscript, and all other coauthors contributed to the manuscript development process and to the discussion section. None of the authors had any conflict of interest.
REFERENCES