Impact of Childhood Malnutrition on Host Defense and Infection

The thymus is the primary lymphoid organ where bone marrow-derived lymphocytes undergo differentiation prior to migration to peripheral lymphoid tissues. Autopsy studies of malnourished children describe profound thymic atrophy, thymocyte depletion, and an alteration of the extracellular matrix (32). However, many of these children died from severe infection, itself a cause of acute thymic atrophy (33). Malnutrition- and infection-related thymocyte depletion is caused by the increased apoptosis of CD4- and CD8-double-positive (immature), -double-negative, and -single-positive thymocyte populations (34). Reduced thymocyte proliferation also contributes to thymic hypocellularity (35). Deficiencies in both dietary protein and zinc lead to thymocyte apoptosis (36, 37). Thymocyte apoptosis during malnutrition is driven by elevated levels of circulating glucocorticoids (38) and reduced leptin levels (37). Treatment of protein-deprived rats with leptin abrogated malnutrition-related thymocyte apoptosis (39). In a model of mild maternal protein deprivation during lactation, thymocytes in the offspring were protected from apoptosis by enhanced leptin activity (37). Alteration of the thymic microenvironment, including a reduced volume of the thymic epithelium, expansion of the extracellular matrix, and reduced thymic hormone production, is associated with thymocyte depletion (reviewed in reference 40).

The high rates of cell proliferation and self-renewal make bone marrow particularly vulnerable to the effects of nutrient deficiencies, especially protein-energy malnutrition and iron deficiency. Megaloblastic and dysplastic changes with erythroid-series hypoplasia were found in the bone marrow of children (n = 34) with marasmus (28.5%), kwashiorkor (50%), and marasmic kwashiorkor (30%) (41). In mice fed a protein-deficient diet, bone marrow atrophy with gelatinous degeneration, expansion of the extracellular matrix, and a loss of markers of cell proliferation was observed (42). Protein malnutrition suppressed the cell cycle progression of hematopoietic progenitor cells, with arrest in the G₀/G₁ phase (43, 44). This was associated with reduced levels of cell cycle-inducing proteins and increased levels of inhibitory proteins (44). The arrest of progenitor cells led to a reduction in myeloid and erythroid lineages (42). Altered erythropoiesis in protein-deficient mice occurred independently of iron or erythropoietin deficiency (45). Bone marrow granulocytic cells showed losses at all developmental stages, blunted maturation, an impaired blastic response to granulocyte colony-stimulating factor (G-CSF) (46), and reduced mobilization in response to lipopolysaccharide (LPS) (47). Lymphoid populations, which are relatively rare in bone marrow, were also reduced in malnourished mice (48).

Nonhematopoietic stromal cells play a role in the growth and maintenance of hematopoietic progenitor cells. The stroma of malnourished mice did not sustain CD34⁺ hematopoietic stem cell growth (42). Bone marrow mesenchymal stem cells in protein-deficient mice were found to differentiate into adipose cells, leading to an altered cytokine microenvironment and compromised hematopoiesis (49).

Malnourished children with bacterial infection showed no difference in total blood leukocyte counts or numbers of lymphocytes, granulocytes, or monocytes compared to well-nourished children with bacterial infection (50). Children with severe acute malnutrition had normal numbers of total mononuclear cells but reduced numbers of dendritic cells (DCs) in peripheral blood (51). Protein-malnourished mice were anemic and leukopenic, with reduced numbers of neutrophils, lymphocytes, and monocytes (49, 52).

The effect of malnutrition on secondary lymphoid tissues (spleen and lymph nodes) in children is unknown, but animal models suggest significant pathological changes. Mice fed a protein-deficient diet had a small, hypocellular spleen with a thickened capsule. There were reduced numbers of total splenocytes and splenic mononuclear cells (52, 53). Spleen cells showed reduced proliferation and were increasingly observed in the G₀/G₁ cell cycle phase (52). Similarly, malnutrition in weanling rats led to reduced proportions of cells in the S and G₂/M phases, with abnormal lengths of both the G₁ and S phases (54). Lactating malnourished mice showed increased splenocyte apoptosis (34). The splenic inflammatory milieu was altered in protein-malnourished mice. The production of interferon gamma (IFN-γ) and interleukin (IL-5) was unchanged, but IL-2 production was reduced and IL-10 production was increased in activated splenocytes from protein-malnourished mice (52). Activated STAT3 expression (involved in IL-10 production) was increased, but STAT1 expression (involved in IFN-γ responses) was reduced (52). There was a disorganization of the splenic white pulp in protein-malnourished mice, which was accentuated when malnourished mice were chronically infected with Leishmania infantum (55).

Lymph node cellularity is similarly affected by malnutrition. In a mouse model of moderate multinutrient deficiency (reduced zinc, iron, protein, and energy levels), the lymph node had fewer DCs, fibroblastic reticular cells, and macrophages. The reduction in myeloid cell populations (macrophages, DCs, and neutrophils) was amplified following challenge with the protozoan parasite Leishmania donovani (3 days postinfection), and the lymph node had an impaired capacity to act as a barrier to pathogen dissemination (56, 57). Trafficking of soluble antigens through the lymph node conduit system was also altered in this model (57).

Children with malnutrition had reduced numbers of cells positive for IgA in the jejunal mucosa, but other immunoglobulin subtypes were not affected (58). Reduced levels of secretory IgA were also found in the intestinal fluid (59). Extrapolating from the malnutrition-related hypocellularity of other lymphoid organs, one would expect the sizes of Peyer's patches to be reduced, but this and other analyses of gut-associated lymphoid tissue (GALT) in malnourished children have not been reported. Malnutrition-related low secretory IgA levels in protein-deprived mice were restored following supplementation with dietary protein (60). In the above-mentioned mouse model of chronic malnutrition (26), the typical histological and functional features of environmental enteropathy were reproduced by serial exposure to a diet poor in proteins and fat and a bacterial gavage of Bacteroidales species and E. coli. The deprived diet alone did not induce structural changes in the small intestinal mucosa but was associated with an increased number of intraepithelial lymphocytes, predominantly γδ CD8⁺ T cells, compared to those in mice fed a normal diet. Sequential exposure to the bacterial cocktail induced the flattening of mucosal villi and an influx of natural killer (NK) cells. Intraepithelial lymphocytes obtained from the duodenum of these mice secreted significantly higher levels of tumor necrosis factor alpha (TNF-α) and IFN-γ (26). Increased numbers of langerin-positive DCs were found in the gut lamina propria and mesenteric lymph nodes of vitamin A-deficient mice (61). Malnutrition of rat neonates during suckling reduced the numbers and delayed the maturation of B cells and T cells (including recent thymic emigrants) in Peyer's patches (62, 63). Following mucosal immunization with cholera toxin, specific IgG, IgA, and IgM antibody-forming cells were diminished in Peyer's patches and mesenteric lymph nodes of malnourished rats (63).

Alterations of the gastrointestinal mucosal barrier and GALT function suggest that the efficacy of oral vaccines would be reduced in malnourished children. Indeed, childhood malnutrition and environmental enteropathy are considered to be contributors to the so-called “tropical barrier,” referring to the phenomenon of a reduced efficacy of live oral vaccines in developing countries (10, 64, 65). The oral poliovirus, rotavirus, and cholera vaccines have shown reduced immunogenicity and efficacy in children in a number of developing countries (65–67). However, recent studies in children indicated that there was no effect of mild underweight (weight for age, ≤10th percentile) on vaccine responses (68). Thus, failures of oral vaccine-induced immunity may be limited to children with more severe malnutrition and are likely to have multiple contributing factors. In mice, PEM impaired the mucosal IgA response to rotavirus vaccine but not protective efficacy (69).

The acute-phase response is a systemic response to infection or other causes of inflammation. It leads to appetite suppression and a negative energy balance. Energy expenditure is increased by 7 to 11% for each unit (degrees Celsius) increase in fever (72, 73). The acute-phase response is accompanied by proinflammatory cytokine production, which drives the catabolism of muscle protein and increased hepatic protein synthesis. Insulin resistance and hepatic glycogenolysis and gluconeogenesis contribute to increased plasma glucose levels during the acute-phase response. Owing at least in part to insulin resistance, there is also increased peripheral lipolysis and hepatic triglyceride and very-low-density lipoprotein (VLDL) synthesis but decreased cholesterol synthesis. All of these metabolic changes amplify growth faltering in children with insufficient nutrient intake. Children with severe malnutrition often have a blunted febrile response to infection. Consistent with this clinical observation, some studies have reported the reduced production of acute-phase proteins and proinflammatory cytokines (IL-1, IL-6, and TNF) in children with kwashiorkor and marasmus (74–77). Furthermore, the acute-phase proteins C-reactive protein (CRP) and procalcitonin were not reliable predictors of invasive bacterial infection in severely malnourished children (78). However, other studies demonstrated high levels of circulating TNF and increased cellular responsiveness to bacterial lipopolysaccharide in uninfected malnourished children (79, 80). This discordance may be due to differences in intestinal barrier function, bacterial translocation, and endotoxemia, which are common in severely malnourished children (17, 51, 81). Endotoxin tolerance may have a role in blunting the acute-phase response and the production of inflammatory mediators in severely malnourished children. Children with protein and energy deficits showed reduced levels and impaired activities of components of the complement system (82, 83).

There is a long-recognized need for noninvasive biomarkers to identify children at risk for growth faltering. Most studies have focused on markers of systemic inflammation, such as the cytokines and acute-phase proteins described above. More recently, markers of intestinal barrier disruption, bacterial translocation, and intestinal inflammation have been evaluated (84). Recent work from the MAL-ED Network demonstrated the interaction of inflammation, linear growth, and the growth hormone axis, suggesting that serum growth hormone, IGF-1, and IGF binding protein 3 (IGFBP-3) could be useful biomarkers of growth faltering (85). Fecal markers of inflammation have also been evaluated (21, 86, 87). To our knowledge, there has been no study of biomarkers that might identify impaired host defense and an increased risk of infection in malnourished children.

A number of clinical and experimental animal studies demonstrated reduced numbers of monocytes and macrophages in malnourished hosts. Infants with PEM had elevated expression levels of the apoptotic marker CD95 (Fas) in peripheral blood neutrophils, lymphocytes, and monocytes, which were decreased after nutritional rehabilitation (88). This suggests that the life span of monocytes is reduced in malnourished children. Protein-deficient mice had reduced numbers of circulating blood monocytes (49, 52). Acutely starved mice had decreased numbers of peritoneal macrophages, which were restored by refeeding (89). Polynutrient (protein, energy, zinc, and iron)-deficient mice had reduced numbers of resident (subcortical) and subcapsular sinus macrophages in their lymph nodes compared to those in nourished controls (57). Rats exposed to dietary protein restriction during lactation had fewer alveolar macrophages (90).

Macrophage effector function is also decreased in the malnourished host. Peritoneal macrophages from mice suffering from PEM showed impaired phagocytosis (91, 92) and diminished production of reactive oxygen and nitrogen intermediates (93). Peritoneal macrophages from protein-deficient mice exhibited dysregulated NF-κB activation, decreased TRAF-6 expression, dysregulated proinflammatory cytokine expression with low-level TNF-α production, and lower expression levels of the CD14 and TLR4/MD-2 receptors upon exposure to lipopolysaccharide (94–96). TNF-α-stimulated macrophages from protein-deficient mice showed lower expression levels of TNF-RI and reduced NF-κB phosphorylation together with the reduced production of IL-1β and IL-12 (97). NF-κB dysregulation was also found in a model of moderate polynutrient (protein, iron, and zinc) deficiency (93).

Surprisingly little is known about neutrophil function in childhood malnutrition. Neutrophil chemotaxis and microbicidal activity were impaired in children with PEM (98–100). Impaired synthesis of lysosomal enzymes and reduced glycolytic activity in neutrophils from malnourished children were reported (98, 99). Retinoic acid plays a critical role in neutrophil maturation. Neutrophils from vitamin A-deficient rats displayed impaired chemotaxis, phagocytosis, and generation of reactive oxygen species (101). A single dose of vitamin A supplementation enhanced the phagocytic capacity of neutrophils in 68 preschool children evaluated at a Venezuelan nutrition clinic (25% were vitamin A deficient) (102). The numbers of neutrophils in the skin-draining lymph nodes of mice deficient in protein, energy, iron, and zinc were reduced (57). Folate-deficient rats also had lower numbers of neutrophils and eosinophils (103). Conversely, zinc-deficient rats were shown to have increased circulating neutrophil counts, which were probably the result of increased corticosterone levels and enhanced release from the bone marrow (104, 105). Circulating granulocyte counts (and elevated corticosterone levels) returned to normal after 2 weeks of feeding a zinc-sufficient diet (105). Neutrophils from vitamin C (ascorbate)-deficient animals failed to undergo spontaneous apoptosis, resulting in reduced clearance (106).

Children 8 to 36 months of age with moderate or severe malnutrition showed no decrease in the number of circulating natural killer (NK) cells (107), but NK cell activity was depressed (108) and recovered with therapeutic nutritional intervention (109). NK cell numbers and cytotoxic activity were reduced in the lungs and spleen of energy-restricted mice in response to influenza virus infection (110). The number and activity of splenic NK cells were also reduced in vitamin A-deficient rats and returned to normal after vitamin A repletion (111). Total numbers of NK cells (103) and their cytotoxicity (112) were reduced in rats fed a folate-deficient diet.

Dendritic cells (DCs) bridge innate and adaptive immunity through the production of cytokines and the initiation of antigen presentation. Severely malnourished children from Zambia had reduced numbers of DCs that recovered after standard nutritional treatment (51). In a subpopulation of these children who had evidence of endotoxemia, DCs showed impaired maturation (failure to upregulate HLA-DR) and a reduced capacity to stimulate T cell proliferation (51). In a murine model of multinutrient deficiencies (protein, zinc, and iron deficiencies), a reduced number of lymph node-resident DCs was associated with the dysregulation of DC chemoattractants under inflammatory conditions (56, 57). The adoptive transfer of immortalized syngeneic DCs (but not CD3⁺ T cells) to protein-energy-deficient mice partially restored the impaired delayed-type hypersensitivity response (113). The effect of PEM on the DC antigen-presenting capacity and the induction of T cell activation was found to be nil (114) or impaired (115, 116). Differences in model systems, including the age of the mice and purity of DCs, probably account for these discrepancies. Studies of highly purified, defined DC subsets under inflammatory and noninflammatory conditions are needed to resolve this important issue. The critical role of vitamin A in DC differentiation (117) is discussed in the section on vitamin A, below.

Malnourished children hospitalized with bacterial infection showed no difference in numbers of peripheral CD8⁺ and CD4⁺ T cells (50) but had reduced numbers of CD4⁺ CD45RO⁺ memory T cells (120) and reduced numbers of effector T cell (CD4⁺ CD62L⁻ and CD8⁺ CD28⁻) subsets (121). Anemic children with vitamin A deficiency showed remarkable increases in the total numbers of CD4⁺ and CD8⁺ T cells after vitamin A supplementation (122). As noted above, malnutrition may impair antigen-presenting cell function, so altered adaptive T cell responses may not be due to an intrinsic change in T cell function. Peripheral blood mononuclear cells from malnourished children with bacterial infection had reduced levels of key cytokines required for both Th1 differentiation (IL-7, IL-12, IL-18, and IL-21) and function (IFN-γ and IL-2) (123, 124) and overexpression of the Th2 cytokines IL-4 and IL-10 (125). Increased apoptosis of CD3⁺ T cells, which was associated with decreased IL-7/IL-7 receptor alpha (IL-7Rα) and increased Fas (CD95) and PD-1 expression levels, was reported for children with severe acute malnutrition and respiratory and/or gastrointestinal infection (123). In mice, dietary protein restriction led to splenic atrophy but variable T cell numbers in the spleen (55, 126). Fasting for as few as 2 days decreased the numbers of T cells in the spleen (127, 128). PEM and zinc deficiency in rats caused a decreased level of production of immature CD4⁺ CD8⁺ cells due to enhanced thymocyte apoptosis and reduced lymphocyte proliferation (34).

The ability of T cells to respond to inflammatory stimuli is also negatively affected by malnutrition. In response to DNA vaccination (ovalbumin expression plasmid), protein-deficient mice exhibited an impaired antigen-specific T cell response (decreased numbers of ova-specific CD8⁺ T cells and lower-level IL-2 production by CD4⁺ T cells) but an unaltered antigen-specific antibody response (129). Similar to what was described for malnourished children, malnourished mice showed enhanced Th2 cytokine polarization and skewing of the Th1-Th2 balance (130). Mice fed a very-low-protein diet and infected with lymphocytic choriomeningitis virus (LCMV) showed fewer activated (CD44^hi) virus-specific CD8⁺ T cells in the spleen and reduced virus clearance (131). Virus-specific CD8⁺ T cells from protein-deficient mice showed effective T cell activation when transferred into normally nourished LCMV-infected mice. This suggests that protein deficiency does not lead to intrinsic defects in T cells, but rather, the malnourished environment does not effectively support T cell activation (131). Acute malnutrition inhibits glucose metabolism-dependent T cell activation (proliferation and cytokine production) (128, 132, 133). The in vitro activation of T cells from mice fasted for 48 h showed an impaired production of the Th1 cytokines IL-2 and IFN-γ that was rescued by exogenous leptin (128).

Malnourished children with respiratory or gastrointestinal bacterial infection had reduced numbers of B cells compared to those of infected well-nourished controls (50). B lymphocyte function generally appears to be maintained in PEM, although specific antibody-mediated immune responses may be affected. Levels of Th2-type immunoglobulins (IgG1 and IgE) are increased, whereas levels of Th1-type immunoglobulins (IgG2a and IgG3) are unaltered (134). Numbers of IgA-secreting cells and secretory IgA concentrations are reduced (58, 59). However, oral administration of the probiotic Lactobacillus pentosus to protein-deficient mice restored the levels of intestinal IgA and numbers of splenic B and Th2 cells to the levels of normal controls (135). This suggests that malnutrition mediates its effect on mucosal immunity by affecting the intestinal microbiota (see below). Folate-deficient rats had lower numbers of B and T cells than those in the controls (103). Vitamin A-deficient mice produced a poor IgG response that was restored with vitamin A repletion (136). Vitamin A-deficient rats had reduced numbers of IgA⁺ plasma cells and CD4⁺ cells and increased numbers of CD8⁺ cells in their Peyer's patches (137). Zinc deficiency depleted immature and mature cells of the B cell lineage in bone marrow (138).

Dietary iron exists as heme iron and nonheme iron, with the former being found exclusively in animal foods and the latter being found in both animal and plant foods. The efficiency of intestinal heme iron absorption is much higher than the efficiency of absorption of nonheme iron. Iron deficiency is the world's most widespread micronutrient disorder. Anemia affects over 1.6 billion people worldwide, one-quarter of the world's population (151), and half of these anemia cases are associated with iron deficiency (152). Worldwide, nearly 47% of preschool children, 42% of pregnant women, 30% of nonpregnant women, and 12.7% of men are anemic (151). The prevalence of iron deficiency in the poorest populations is attributed to cereal-based diets that lack heme iron and contain low levels of nonheme iron and high levels of inhibitors of iron absorption (153). Severe anemia in children is associated with fatigue and may result in developmental delays and behavioral problems. Iron is critically important for both innate and adaptive immunity (154, 155). Intracellular iron has been shown to activate NF-κB via promoting the release of reactive oxygen species (156, 157). Hypoxia-inducible factor-1 alpha (HIF-1α), an iron-dependent transcription factor, promotes the production of antimicrobial peptides by macrophages (158). Peripheral blood mononuclear cells from iron-deficient patients showed increased TNF-α, IL-6, and IL-10 mRNA expression levels after the administration of iron (159). Mitogen-activated spleen cells from iron-deficient mice showed reduced IFN-γ production (160). Transferrin receptor 1 (TfR1)-deficient mice, which have reduced cellular iron uptake, exhibited impaired T cell development and fewer mature B cells than wild-type mice (161). The proliferation of human B and T lymphocytes was also reduced by TfR1-blocking antibodies (155). Mice with a conditional deletion of ferritin H in their bone marrow had fewer mature B and T cell populations in lymphoid tissues (162). On the other hand, too much iron is detrimental to host defense. Macrophages from Hfe^−/− mice, which have enhanced iron absorption that leads to iron overload, produced low levels of inflammatory cytokines (IL-6 and TNF-α) in response to Salmonella infection (163). Similarly, children with low levels of the cellular iron transporter ferroportin, which leads to reduced iron efflux and increased accumulation of intracellular iron, had low levels of circulating TNF-α (164). Collectively, these findings indicate that an alteration of iron homeostasis, whether resulting from too much or too little iron uptake, impairs innate and adaptive responses.

The impact of iron deficiency on susceptibility to infection is difficult to dissect because free iron is essential for the growth of many pathogens (reviewed in reference 165). Some human and animal studies demonstrated that iron deficiency increased the risk of infection (155), but other studies observed that iron supplementation increased susceptibility to malaria and tuberculosis (TB) (166, 167). Host cells may harness pathways involved in iron homeostasis as an antimicrobial defense system. Upon infection, reticuloendothelial cells sequester iron from the blood and phagocytes by the release of lactoferrin. Lactoferrin binds iron more avidly (specifically at low pH) (168) than do bacterial siderophores, with a consequent deprivation of iron required for the replication of the pathogen (165). Therefore, iron deficiency results in the impaired killing of bacteria by phagocytes but may also lead to impaired pathogen replication. Clearly, iron deficiency leading to anemia is a major public health problem, but further research is needed to determine optimal iron levels and the impacts of iron repletion on maximizing host defense and minimizing pathogen replication and virulence.

Zinc deficiency affects one-fifth of the world's population and is responsible for the deaths of nearly 450,000 children under the age of 5 years annually (5, 169). Zinc deficiency often accompanies childhood PEM (170–172), and a protein-deficient diet led to zinc deficiency in experimental animals (173). Foods of animal origin (e.g., meat, shellfish, and organs such as liver) are the richest sources of zinc, and the bioavailability of this mineral from animal sources is higher than that of zinc found in plant sources. Animal-derived foods rich in both protein and zinc are severely limited in the diets of children whose families have inadequate resources. Zinc is a cofactor for more than 200 enzymatic reactions and thus has profound effects on cellular function and is critical to proper childhood growth and sexual maturation. It plays critical roles in the structure and functioning of biomembranes and in stabilizing DNA, RNA, and ribosomal structures (174). Zinc also regulates a wide range of immune functions (reviewed in references 153 and 175). It is important for the activity of thymic hormone (176–178), which regulates T cell maturation. Zinc promotes Th1 cell differentiation and Th1 cell responses by increasing IL-2, IFN-γ, and IL-12Rbβ2 expression levels (179, 180). Additionally, zinc regulates the release of proinflammatory cytokines such as IL-1β, IL-6, and TNF-α by innate immune cells (181–183). It regulates neutrophil function by modulating the oxidative burst (184, 185). As a result, zinc deficiency leads to thymic atrophy, lymphopenia, a reduced CD4/CD8 ratio, and a reduced synthesis of Th1 cytokines. It is also associated with impaired NK cell function and impaired phagocytosis by macrophages (174, 175, 186, 187). Zinc deficiency may impair mucosal immune function through altered epithelial homeostasis.

Dietary zinc supplementation has been widely studied for its effects on childhood growth and mortality (reviewed in reference 188). Its effects on immune function and the risk of infection are somewhat controversial, but the general consensus is that zinc supplementation reduces the risk of diarrheal disease and pneumonia. Gender-related differences in response to zinc supplementation may contribute to some of the conflicting results of clinical trials (189). A double-blind, randomized, placebo-controlled study of daily zinc supplementation in a cohort of children aged 6 to 30 months in a New Delhi, India, slum demonstrated reduced frequency, duration, and severity of diarrheal disease in the zinc-supplemented group (190). In the same cohort, zinc supplementation had no effect on the rate of acute lower respiratory tract infection (LRTI) but was associated with a significant decline in the incidence of pneumonia (190). This large-cohort trial confirmed the results of previous smaller studies that demonstrated a benefit of dietary zinc for diarrheal disease and respiratory infections (191–193). A trial of zinc supplementation showed no effect on the incidence and morbidity of malaria but showed a reduction in the prevalence of diarrhea (194). Zinc reduced biofilm formation, adherence to epithelial cells, and virulence factor expression of enteroaggregative Escherichia coli (EAEC) (195). Zinc supplementation in deficient mice reduced EAEC stool shedding and abrogated infection-related growth stunting (195).

Selenium deficiency usually accompanies PEM in geographic regions with soil deficient in selenium (174). It is more pronounced in kwashiorkor than marasmus (196). Selenium plays a pivotal role in major metabolic pathways (197, 198) and contributes to antioxidant activity via selenoproteins (199). It has major anti-inflammatory effects through the mitogen-activated protein kinase (MAPK)-, NF-κB-, and peroxisome proliferator-activated receptor γ (PPARγ)-dependent regulation of proinflammatory mediators (200, 201). The genetic deletion of the whole family of selenoproteins by a knockout of the selenocysteine tRNA gene in mice resulted in fewer functional T cells, impaired T cell-dependent antibody responses, and an impaired migration of macrophages (202, 203). The genetic deletion of selenoprotein K did not alter the numbers of immune cells but resulted in impaired T cell responses, neutrophil migration, and phagocyte oxidative burst through the alteration of cellular calcium flux (204). Dietary selenium deficiency leads to several immune deficits (205), including reduced CD4⁺ T cell proliferation and function (reduced NFAT [nuclear factor of activated T cells] activation, IL-2 production, and IL-2 receptor expression and impaired calcium mobilization) (202, 205, 206). Supplementation with selenium along with vitamin A, the vitamin B complex, vitamin C, and vitamin E increased CD3⁺ and CD4⁺ T cell counts but did not augment the antituberculous T cell response in patients with active tuberculosis (207, 208).

Vitamin A, or retinol, is acquired exclusively through the diet, absorbed by enterocytes, and stored in the liver. Vitamin A deficiency is a global health problem that affects 100 million to 140 million children, with 4.4 million having xerophthalmia (209). Indeed, vitamin A deficiency is the leading cause of childhood blindness worldwide. PEM compounds vitamin A deficiency due to inadequate amino acid availability in the liver, which is required for the synthesis of vitamin A transport proteins such as retinol binding protein. Vitamin A, through its primary active metabolite retinoic acid, plays key roles in the proper differentiation of epithelial cells in skin; the cornea of the eye; and mucosal surfaces of the gastrointestinal, respiratory, and urogenital tracts. The lack of adequate epithelial barrier function makes pathogenic bacterial and viral invasion more easily accomplished. Retinoic acid is also involved in the regulation of a number of innate and adaptive immune functions (reviewed in references 210 and 211) (Fig. 3). Retinoic acid production is highly enriched in the intestinal tract, where it modulates intestinal immune homeostasis and defense. Its effects are highly cell specific and influenced by whether the tissue microenvironment is homeostatic or inflammatory. Maternal vitamin A intake plays a critical role in secondary lymphoid development in utero through the regulation of prenatal innate lymphoid cells, which determine the size of lymphoid organs in adult life (212). Retinoic acid has an essential role in mucosal immunity (213) through the regulation of mucin gene expression (214), the production of IgA (215, 216), the regulation of innate lymphoid cell development (217), and the regulation of DC and T cell differentiation in the lamina propria and gut-associated lymphoid tissue (117, 218). Retinoic acid acts on, and is secreted by, mucosal DCs and macrophages. It regulates specific DC subpopulations, most notably CD11b⁺ CD103⁺ DCs in the intestine (219). It enhances DC migration to draining lymph nodes (220) and, by doing so, regulates T cell differentiation and activation. It also promotes the activation of IFN-γ signaling through STAT1 and interferon regulatory factor 1 (IRF1) activation in lung epithelial cells (221). The effects of retinoic acid on T cell differentiation and function appear to be context dependent. Under homeostatic conditions, retinoic acid promotes (with the help of transforming growth factor β [TGF-β]) the conversion of naive CD4⁺ T cells into regulatory T cells and inhibits the development of Th17 cells. Both processes promote immune tolerance against commensal bacteria (222, 223). Retinoic acid regulates small intestine inflammation via the generation of regulatory and gut-homing IL-10-producing T cells (218, 224, 225). DC-induced T cell recruitment is mediated by the retinoic acid-induced expression of the gut-homing molecules α4β7 and CCR9 on CD4⁺ T cells (226, 227). Under inflammatory conditions, retinoic acid promotes CD4⁺ and CD8⁺ effector T cell responses (228–231) and in particular favors the development of Th2 over Th1 responses (231–233). Retinoic acid treatment of M. tuberculosis-infected rats led to reduced bacterial burdens in the lung and spleen, which were associated with the increased accumulation of CD4⁺ and CD8⁺ T cells, NK cells, and CD163⁺ macrophages at the site of lung infection (234). It also enhanced the proinflammatory response to and killing of tubercle bacilli by alveolar macrophages. Retinoic acid secreted by DCs and alveolar macrophages enhances the differentiation of T cells to regulatory T cells (222). Previous studies demonstrated that retinoic acid enhances the ability of regulatory T cells and gut-homing T cells to suppress acute small intestinal inflammation after adoptive transfer in mice (218, 225).

In light of the above-mentioned role of retinoic acid, it is not surprising that vitamin A-deficient mice possess altered innate and adaptive immunity. Vitamin A deficiency leads to a marked reduction in the number of type 3 innate lymphoid cells (ILC3s), leading to reduced IL-17 and IL-22 levels and increased susceptibility to acute enteric bacterial infection (217). At the same time, vitamin A-deficient mice exhibited an expansion of the IL-13-producing ILC2 population with consequent increases in the amount of intestinal mucus, goblet cell hyperplasia, and resistance to intestinal helminthes (217). This effect was dependent on signaling through the retinoic acid receptor (RARα). Thus, dietary vitamin A regulates intestinal barrier immunity by regulating the balance between these two subsets of ILCs. This enhances one arm of innate immunity to defend against nutrient-depleting worms at the expense of increased susceptibility to enteric bacterial pathogens. The numbers and functions of natural killer T (NKT) cells and NK cells are also modulated by the availability of retinoic acid (235, 236).

Regarding adaptive immunity, vitamin A deficiency altered homeostatic DC maintenance and differentiation in the gut-associated lymphoid tissue (61, 237). Gestational vitamin A deficiency in rats also decreased the numbers of CD11c⁺ DCs in Peyer's patches of offspring (238). CD4⁺ (Th1, Th2, and Th17) and CD8⁺ T cell numbers in the intestinal lamina propria were also altered (217, 226, 228, 239). Vitamin A deficiency promotes the differentiation of T cells toward Th2 cells and increases the ratio of Th2 to Th1 cytokines by suppressing the Th1 immune response (218, 240). This explains, at least in part, why vitamin A deficiency is associated with reduced effector T cell responses, suboptimal immune responses to some vaccines (241), and an increased risk for certain infections. A large number of clinical trials of vitamin A supplementation have been conducted, collectively involving several hundred thousand participants. Most of these trials have shown a reduction in all-cause mortality (20 to 30%) and reductions in the incidences and severities of diarrheal disease and measles but not lower respiratory tract infections (reviewed in references 242–245).

Vitamin C is an essential water-soluble vitamin important for metabolic function and antioxidant activity (246), and it increases the absorption of nonheme iron when coingested in the same meal. Vitamin C deficiency affects approximately 10% of adults in the industrialized world (247, 248). It occurs more frequently in impoverished populations, but there is little information on its prevalence in children in the developing world. Its potential role in leukocyte function is suggested by the ascorbic acid (reduced form of vitamin C) content in leukocytes being severalfold higher than that in plasma (249). Vitamin C blunts the inflammatory cytokine response to LPS in peripheral blood mononuclear cells from adult human subjects (250) but paradoxically enhances inflammatory cytokine responses in neonatal cord blood leukocytes (251). Vitamin C regulates apoptosis in monocytes/macrophages, neutrophils, and B cells (106, 252–254). DCs cultured in the presence of vitamin C showed upregulations of the costimulatory molecules CD80, CD86, and major histocompatibility complex class II (MHC-II) (255) and increased CD8⁺ T cell expansion when cocultured with T cells (256). In vivo and in vitro experiments demonstrated that vitamin C regulated the isotype switching of mouse B cells (254). Vitamin C deficiency exaggerated inflammation and impaired its resolution in a murine model of sterile inflammation (257). Vitamin C administration attenuated acute lung, kidney, and liver injury in murine models of lethal LPS administration and intra-abdominal sepsis (257–259). The attenuated lung injury was accompanied by a reduced proinflammatory response, enhanced epithelial barrier function, increased alveolar fluid clearance, and reduced coagulopathy (257–259). The underlying mechanisms of this protective effect were attributed to reduced neutrophil NF-κB activation, endoplasmic reticulum stress, the induction of autophagy, and the generation of neutrophil extracellular traps (NETosis) (253). A phase 1 trial of intravenous ascorbic acid in adults with severe sepsis showed no evidence of ascorbic acid-induced toxicity and significantly reduced levels of biomarkers of both inflammation (CRP and procalcitonin) and vascular endothelial injury (thrombomodulin) (260). Subjects who received high-dose ascorbic acid also showed an attenuation of organ failure scores (260). There are no studies of the influence of vitamin C status on resistance or susceptibility to sepsis in malnourished children.

The primary role of vitamin D is in calcium homeostasis and bone metabolism, but it also has a number of effects that impact host defense. 25-Hydroxyvitamin D [25(OH)VD₃] is the major circulating form and is metabolized by 25-hydroxyvitamin D-1α-hydroxylase (CYP27B1) to the primary active form 1,25-dihydroxyvitamin D [1,25(OH)₂VD₃], which induces signaling when it binds to its cognate nuclear receptor, the vitamin D receptor (VDR). Genetic variation in the VDR may modify the associations of vitamin D with human health and the interpretation of data from clinical studies (261). The optimal level of serum vitamin D has been fiercely debated. Individuals are considered to be vitamin D deficient when the serum 25(OH)VD₃ level is <25 nmol/liter and vitamin D insufficient when the serum 25(OH)VD₃ level is <50 to 75 nmol/liter (262). Vitamin D deficiency is estimated to affect 1 billion people worldwide. More than 40% of the elderly in the United States and Europe and more than 50% of postmenopausal women suffer from vitamin D deficiency (263). Vitamin D deficiency may also be common in children and young adults (264). There are few foods that are naturally rich in vitamin D, and therefore, its synthesis in the skin via exposure to UV light is of critical importance. A lack of adequate sun exposure is a common cause of vitamin D deficiency. Children with darker skin, which contains more of the pigment melanin, which blocks the effects of UV radiation, are at a greater risk for deficiency.

The effect of vitamin D on immunity and host defense is complex, having roles in both proinflammatory antimicrobial effector function and anti-inflammatory suppressive activity (Fig. 4). The role of vitamin D in innate immunity was recently reviewed (265, 266). A seminal observation by Liu et al. (267) identified Mycobacterium tuberculosis as a trigger for the TLR2-mediated induction of CYP27B1 and VDR in monocytes. Signaling through the TLR4/NF-κB and IFN-γ receptor (IFN-γR)/STAT1 pathways also induced the expression of CYP27B1 and VDR (268–270). The IFN-γ-mediated induction of CYP27B1 in human monocytes and macrophages was dependent on STAT1 and the induction of IL-15 and, in the presence of sufficient vitamin D, led to an antibacterial effect via the induction of autophagy, autophagolysosomal fusion, and the generation of the antimicrobial peptides cathelicidin (LL37) and β-defensin-2 (270). Mycobacterial killing was abrogated in the presence of vitamin D-deficient serum (270). Vitamin D-induced antituberculous autophagy was driven by cathelicidin and dependent on TLR1/2 signaling (271, 272). 1,25-Dihydroxyvitamin D enhanced the M. tuberculosis-induced expression of proinflammatory cytokines and chemokines in a human macrophage cell line via the NLRP3/caspase-1 inflammasome (273). In this in vitro model, augmented IL-1β secretion led to increased antimycobacterial activity in cocultured lung epithelial cells via the production of antimicrobial peptides (273). Other studies also identified a critical role for VDR signaling in the production of the antimicrobial peptides cathelicidin (LL37) and β-defensin-2, which mediate the growth restriction of M. tuberculosis in macrophages (267, 274–277). In addition to the IFN-γ-induced production of cathelicidin, IFN-γ/TNF-independent production via TLR signaling has been proposed (267, 276).

Clinical studies have investigated the role of vitamin D in tuberculosis. Most of these studies included primarily adult subjects. The seasonality of the prevalence of tuberculosis has long been known. Recent studies associated this with seasonal variations in vitamin D levels, presumably related to sun exposure, in individuals in South Africa and Peru (278, 279). Vitamin D deficiency was associated with an increased risk of active tuberculosis in a large number of studies (recently reviewed in references 278 and 280). The risk was influenced by polymorphisms in the VDR and vitamin D binding protein (281, 282). Vitamin D insufficiency was also associated with an increased risk of relapse following antituberculous therapy in both HIV-uninfected and -coinfected patients (283). Vitamin D was used to treat tuberculosis in the preantibiotic era (284), but recent clinical trials of adjunctive vitamin D therapy for active tuberculosis have reported conflicting results in clinical, bacteriological, and/or immunological outcomes (285–288). Vitamin D supplementation accelerated treatment-induced sputum smear conversion (285, 289), the resolution of lymphopenia and monocytosis, and the normalization of increased levels of serum inflammatory cytokines and chemokines (285). Significant clinical benefit may be achieved by the accelerated resolution of inflammation, which is clearly associated with increased tuberculosis mortality (290). In a multicenter, randomized, placebo-controlled trial of adjunctive vitamin D treatment for sputum smear-positive pulmonary tuberculosis patients in London, vitamin D₃ (VD₃) (cholecalciferol; three doses of 2.5 mg each) significantly improved the time to sputum conversion only in subjects that had the tt genotype of the TaqI vitamin D receptor polymorphism (286). However, a lower dose of oral cholecalciferol (100,000 IU) given 0, 5, and 8 months after the initiation of antituberculous treatment did not lead to improved sputum conversion, clinical outcomes, or 12-month mortality in adults with pulmonary tuberculosis compared to placebo (288). In contrast, two doses of 600,000 IU of intramuscular vitamin D₃ accelerated clinical and radiographic improvement 12 weeks after the start of antituberculous therapy compared to placebo (291). In a study of children, most of whom had extrapulmonary tuberculosis, adjunctive vitamin D therapy improved clinical and radiological features (292).

A prospective cohort study showed a significant inverse association between vitamin D levels and the incidence of active tuberculosis disease among contacts of patients with pulmonary tuberculosis (293, 294). Vitamin D supplementation also reduced the incidence of latent tuberculosis infection (identified by tuberculin skin test conversion or a positive interferon gamma release assay) in contacts of patients with pulmonary tuberculosis (295, 296). In a double-blind, randomized, controlled trial with healthy adult tuberculosis contacts (94% of whom were either vitamin D deficient or insufficient), a single oral dose of vitamin D (ergocalciferol; 2.5 mg) enhanced the growth restriction of Mycobacterium bovis BCG in an ex vivo whole-blood assay (287). Collectively, data from these studies indicate that vitamin D modulates immune and inflammatory mechanisms that can enhance the control of infection and tissue damage. However, a beneficial effect has not been consistently demonstrated in clinical trials, possibly because the optimal dose and frequency of vitamin D supplementation remain to be determined. There is a need for further investigation of vitamin D in the management of children with tuberculosis.

The prophylactic or therapeutic effect of vitamin D supplementation on acute respiratory tract infection (ARI) was recently reviewed (297). A number of observational and cross-sectional studies have demonstrated an association of vitamin D deficiency with increased susceptibility to ARI, but randomized, controlled studies have inconsistently shown a benefit of vitamin D supplementation. This lack of consensus may arise from the variability in vitamin D dosing regimens, the variable prevalence of vitamin D deficiency in the study population, the failure to achieve or test for an effect on vitamin D levels, the use of endpoints that involved self-reported symptoms, the inclusion of diverse and unknown etiologies of ARI, and suboptimal power for subset analyses. In a randomized, controlled, double-blind trial of vitamin D-deficient school-age children in Mongolia in the winter, supplementation with vitamin D₃-fortified milk (300 IU/day) versus nonfortified milk significantly reduced the frequency of ARI reported by mothers (rate ratio = 0.52) (298). In a randomized, placebo-controlled trial, 100,000 IU (2.5 mg) of vitamin D₃ administered every 3 months for 18 months did not reduce the incidence of pneumonia in Afghan infants (299). The intermittent high dose of vitamin D used to achieve supraphysiological peaks followed by deficiency-level troughs may not be optimal (300). Indeed, high concentrations of vitamin D can impair adaptive immunity (301). In a large trial of adults (median age, 63 years) in Norway, vitamin D supplementation did not reduce the risk of influenza-like illness during a 6-month period, but vitamin D levels were not determined before or after the intervention (302). Similarly, in a randomized, controlled trial in New Zealand, vitamin D supplementation did not reduce the frequency or duration of upper respiratory tract symptoms (303).

In addition to the role of vitamin D in the activation of antimicrobial host defense, it has important anti-inflammatory activities. Vitamin D suppresses the proliferation and differentiation of B cells and blocks immunoglobulin secretion (304, 305). Through paracrine action, vitamin D leads to decreased expression levels of MHC class II on DCs, with consequent reductions in DC maturation, antigen presentation (306), and T cell priming (307). This is regulated by the balance of the activating (CYP27B1) and inactivating (CYP24A1) vitamin D hydroxylases and the consequent availability of active 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃] (308, 309). Vitamin D suppresses chronic T cell activation (reviewed in reference 310) and promotes Th2 and regulatory T cell expansion while blocking Th1 polarization (311–313). It also directly promotes the expression of the key transcription factor FoxP3 in regulatory T cells (314). In monocytes/macrophages, it leads to the decreased production of the proinflammatory mediators IL-1β, IL-6, IL-8, and TNF (315–317) and the increased production of anti-inflammatory mediators such as IL-10 (318). Not surprisingly, vitamin D has a significant effect on modulating the host inflammatory response to pathogens. In human airway epithelial cells, vitamin D restrained respiratory syncytial virus (RSV)-induced NF-κB-dependent inflammatory cytokine and chemokine production (319) and the activation of STAT1 and its downstream targets IRF1 and IRF7 (320), without compromising the antiviral effect. The Fok I polymorphism in the VDR, which predisposes patients to severe RSV bronchiolitis, was found to abrogate vitamin D-induced anti-inflammatory signaling (320). A similar anti-inflammatory effect was noted for influenza virus-infected lung epithelial cells (321). 1,25(OH)₂D₃ inhibited Th17 cytokine production in patients with severe asthma (322). In a murine model of cerebral malaria, vitamin D administration led to reduced neuropathology and improved survival. This was accompanied by reduced DC activation and pathogenic T cell infiltration but expanded regulatory T cells and IL-10 production (323). The vitamin D-dependent anti-inflammatory activity could also be detrimental for the control of some intracellular pathogens. The ablation of vitamin D signaling through receptor knockout or a block of vitamin D metabolism to the active form by CYP27B1 deletion led to an increased resistance of mice to the intracellular protozoan pathogen Leishmania major (324).

In a case-control study of children aged 0 to 10 years from indigenous people in Venezuela, malnutrition (WHZ or HAZ of <−2 for children aged <5 years and a body mass index [BMI]-for-age z score or a HAZ of <−2 for children aged 5 to 10 years) was significantly associated with increased nasopharyngeal or oropharyngeal colonization by Streptococcus pneumoniae (335). Alteration of upper respiratory mucosal immune function and/or the mucosal microbiota may facilitate this increased pathogen carriage, but this has not been investigated. No clinical studies have directly identified malnutrition as a risk factor for acute LRTI due to S. pneumoniae, but bacterial colonization is associated with a risk of acute LRTI. In mice, dietary protein deprivation was associated with more severe S. pneumoniae infection and impaired innate immune responses, including reduced leukocyte infiltration in the lung, impaired bactericidal activity of phagocytes, and diminished antipneumococcal mucosal IgA levels (336). The recovery of innate immune function and resistance to pneumococcal infection were accelerated in malnourished mice given a protein-replete diet supplemented with a Lactobacillus probiotic (336, 337).

In a longitudinal cohort study, children in the Philippines (median age, 1.8 months) who were moderately underweight (WAZ of <−2) at the time of their first immunization, or who had reduced weight gain between their first and third immunizations, were at a higher risk of severe RSV infection (338). In a longitudinal study of a birth cohort conducted in rural Kenya across three successive RSV epidemics, stunting (HAZ of <−2) was a risk factor for all-cause LRTI and LRTI due to RSV, as were household crowding and the number of siblings (339). In a prospective study of a birth cohort in the Netherlands, neonates who had a cord blood 25-hydroxyvitamin D level of <50 nmol/liter had a 6-fold increased risk of RSV lower respiratory tract infection in their first year of life compared to neonates who had a vitamin D level of >75 nmol/liter (340). Surprisingly, there are no reported studies of the risk of influenza virus infection in malnourished children, but data from experimental animal studies suggest that protein and energy deficits increase the risk of influenza virus infection. Mice fed a low-protein diet (2%) and infected with influenza virus showed increased viral burdens, lung disease, and mortality. This was associated with impaired virus-specific antibody and CD8⁺ T cell responses (126). A 40% energy (calorie) deficit in mice resulted in an increased risk of severe influenza that was associated with impaired virus-specific type I interferon and NK cell responses (110).

Malnutrition increases the risk of diarrheal diseases caused by some, but not all, enteropathogens. Impaired immune defenses, compromised gut integrity (81), and an altered intestinal microbiota (see below) are likely to influence defense against intestinal pathogens in the malnourished host. The Global Enteric Multicenter Study (GEMS), a large, 3-year, case-control study of moderate to severe diarrhea identified rotavirus, Cryptosporidium, enterotoxigenic Escherichia coli (ETEC), and Shigella as the most common pathogens across seven sites in sub-Saharan Africa and south Asia (343). Infection with each of these pathogens was associated with childhood malnutrition. In a prospective study of urban Bangladeshi children aged 2 to 5 years, a population that inherently has a high rate of exposure to enteric pathogens, only ETEC, Cryptosporidium sp., and Entamoeba histolytica were significantly more prevalent in malnourished (WHZ of <−2) children (344). Bacterial enteropathogens have also been studied in experimental models of malnutrition. Mice infected with enteroaggregative Escherichia coli (EAEC) demonstrated impaired growth that was proportional to the intensity of infection, and conversely, protein-malnourished mice showed an enhanced susceptibility to infection that further impaired their growth velocity (345). Similar results were found for mice fed a “regional basic diet” low in protein, fat, and micronutrients (346). Oral zinc supplementation in zinc-deficient mice resulted in improved weight gain and reduced bacterial shedding following challenge with EAEC (195). Interestingly, low zinc levels appeared to enhance EAEC virulence properties as well as alter the host inflammatory response. Vitamin A-deficient rats showed increased intestinal pathology following infection with Salmonella enterica serovar Typhimurium that was accompanied by increased numbers of mucosal DCs and the dysregulation of IL-12 and IFN-γ production (347).

It has long been held that malnutrition is a major contributor to the high mortality rates from viral gastroenteritis in low-income countries (343, 348). Surprisingly, there are few clinical studies that directly support this (344), and experimental animal studies have reported conflicting results. In the GEMS study noted above, diarrhea caused by rotavirus was associated with malnutrition (343). Protein-malnourished mice had increased weight loss, reduced antiviral mucosal IgA levels, high viral loads, and a delayed clearance of norovirus compared to normal controls (349). In contrast, malnourished mice infected with rotavirus or immunized with a rotavirus vaccine showed no deficit in virus-specific mucosal IgA, and disease severity or vaccine efficacy was not altered (69). Vitamin A supplementation reduced the prevalence of norovirus-associated diarrhea but increased the duration of viral shedding (350).

Intestinal infection with Cryptosporidium, Giardia, and Entamoeba histolytica is associated with growth faltering in children (344, 351–353). Limited human data also suggest that malnutrition has a role in increasing the risk or severity of infection by intestinal protozoa (344, 354). The GEMS study determined that moderate to severe diarrhea caused by Cryptosporidium was associated with childhood malnutrition (343). A community-based, prospective cohort study of infants from birth to 18 months of age determined that weight-for-age z scores at 6 months of age were inversely related to the risk of symptomatic giardiasis (354). Childhood malnutrition was also found to be strongly associated with intestinal amebiasis (344). Diarrheal illness due to Entamoeba histolytica within the preceding 3 years was associated with stunting in Bangladeshi children (353). In this cohort, antigen-induced IFN-γ production was linked to nutritional status and was associated with a reduced risk of subsequent infection by E. histolytica (355). Leptin signaling, which has a profound influence on innate and adaptive immunity (356), is involved in the mucosal defense against E. histolytica. A specific allelic amino acid substitution in the cytokine receptor homology domain of the leptin receptor was associated with increased risks of intestinal amebiasis in children and amebic liver abscess in adults in Bangladesh (357). Mice carrying a copy of this allele were also more susceptible to intestinal amebiasis (357).

Malnutrition as a risk factor for, and a consequence of, intestinal protozoal infection is well established through experimental animal studies. Infection of mice with Giardia led to impaired growth (358). Protein malnutrition in mice led to an increased severity of Giardia infection with evidence of increased mucous production, shortened intestinal villi, and a blunted host immune response (absence of crypt hyperplasia, decreased mucosal IL-4 and IL-5 levels, and reduced numbers of B cells in the lamina propria) (358, 359). Infection of protein-malnourished weanling mice with Cryptosporidium parvum oocysts led to increased weight loss and fecal shedding; increased attachment of the parasite to the ileum, cecum, and colon; a reduced ratio of villous height to crypt depth, and reduced proinflammatory signaling and levels of Th1 cytokines (351, 360–362). Supplementation with l-arginine, from which the antimicrobial molecule nitric oxide is generated by the action of inducible nitric oxide synthase (NOS2), in Cryptosporidium-infected protein-malnourished mice led to improved mucosal histology, weight gain, and decreased parasite burdens. The effect of l-arginine in this model was reversed by the inhibition of NOS2 (362).

Clinical studies have repeatedly demonstrated that chronic infection with helminths such as Schistosoma and soil-transmitted intestinal helminths contributes to underweight or stunting (363–368). Therefore, human studies to investigate nutritional deficiency as a cause of an increased risk or severity of gastrointestinal helminth infection are challenging. Plasma zinc levels were negatively correlated with the intensity of Trichuris trichiura infection in Jamaican children (369), but in Guatemalan children, the reinfection rate following antihelminthic treatment of zinc-deficient children was not reduced by zinc supplementation (10 mg Zn formulated as an amino acid chelate) (370). Experimental animal studies also demonstrated that zinc and protein deficiencies impair the host response and lead to more severe intestinal helminth infection (reviewed in reference 371). Protein-malnourished mice infected with Heligmosomoides polygyrus had an increased worm burden, which was associated with lower serum IgE levels, reduced numbers of intestinal eosinophils and activated mast cells, and reduced gut Th2 effector responses (372). Neonatal malnutrition, a consequence of a low level of protein in the diet of dams during lactation, led to increased egg output and liver damage in mice infected with Schistosoma mansoni (373).

Bacterial bloodstream infections are among the most dangerous complications of SAM. WHO guidelines for the management of malnourished children include evaluation for possible sepsis as part of the initial assessment. However, clinical assessment is poorly predictive of serious infections in these patients (374–376), and microbiology services are generally not available in low-resource health care facilities (377–380). Thus, routine observational data underestimate the true prevalence of bacteremia among malnourished patients. Epidemiological data regarding the prevalence of systemic bacterial infections in children with types of malnutrition other than complicated SAM are rare.

Reliable epidemiological data come from a few ad hoc research surveys in sub-Saharan Africa. In a study of 1,000 children treated as outpatients in Niger, 4% were bacteremic due to unspecified pathogens (381). However, most data come from hospitalized children with SAM. Studies conducted in Kenya, Ghana, and Mozambique found that hospitalized children with SAM had 2 to 3 times the risk of being bacteremic at admission compared with well-nourished children (380, 382–384). The prevalence of bacteremia in children admitted with complicated SAM ranged from 8% to 17%. Gram-negative enteric bacteria, particularly nontyphoidal Salmonella (NTS), E. coli, and Klebsiella spp., are common invasive pathogens in children with SAM, independent of HIV prevalence or the predominant type of malnutrition (marasmus versus kwashiorkor). A study of hospitalized children from Ghana found that underweight (low weight for age) but not stunting (low length/height for age) was associated with an increased risk of bacteremia (385). However, in a recent analysis of pooled data from 10 countries in Asia, Africa, and South America, severe stunting was associated with a 3-fold increased risk of death from sepsis, unspecified acute febrile illness, tuberculosis, or meningitis (386). Mixed infections are not rare, and the case-fatality rate ranges from 10% to 28% (374–376, 382).

A large study conducted in Uganda between 2003 and 2004 found that 17% of the 445 children admitted with SAM had bacteremia: of these children, 58% had at least one Gram-negative bacterium detected, with Salmonella Typhimurium and Salmonella enterica serovar Enteritidis being the most common species, followed by E. coli and Haemophilus influenzae. Among Gram-positive bacteria, Staphylococcus aureus was the most common pathogen, followed by Streptococcus pneumoniae. The odds ratio of being bacteremic was increased for children with oral thrush or hypoalbuminemia but not for children with HIV infection (diagnosed in 36% of the tested patients), focal infection, malaria, or severe anemia. The case-fatality rates were 28.9% overall and 43.5% in children with HIV (387).

Another study conducted in Niger between 2007 and 2008 found similar results: 17% of 311 children admitted with complicated SAM had a bloodstream infection. Sixty-eight percent of the bacteremic children had at least one Gram-negative bacterium detected in the blood, with NTS, Salmonella enterica serovar Typhi, E. coli, and Klebsiella pneumoniae being the most common. S. aureus and S. pneumoniae were the most frequently isolated Gram-positive microorganisms. Eight percent of the children had mixed infections. Another 7% of the admitted children were found to have blood cultures positive for opportunistic bacteria, such as Leuconostoc, Streptococcus equinus, Streptococcus infantarius, Staphylococcus epidermidis, or Enterococcus spp. The case-fatality rate was 16% in this study, which was carried out in a hospital with a dedicated research team and with a very low prevalence of HIV (1%). The only clinical sign associated with bacteremia was oral thrush (374). Fever was present only in 27.5% of children with bacteremia.

Septic shock in children with acute severe malnutrition is associated with high mortality rates and presents unique management challenges, especially in facilities where close hemodynamic monitoring is not possible (378). Children with SAM and septic shock often have severely impaired renal and cardiac function and multiple electrolyte derangements, which make them susceptible to fluid overload and congestive heart failure. The 2013 revised WHO guidelines suggest the use of 5% dextrose plus either half-strength Darrow's solution, Ringer lactate, or half-strength saline for the initial expansion of the circulating volume, in repeatable boluses of 10 to 15 ml/kg of body weight (378). However, the quality of evidence to support this recommendation is low. A large clinical trial aimed at defining the optimal fluid management for children presenting to African hospitals with severe infections (FEAST trial) found that the use of fluid boluses was associated with increased mortality. Severely malnourished children, however, were specifically excluded from the study (388, 389). The optimal management of septic shock in children with SAM therefore remains an area in great need of further research.

The high risk of invasive bacterial infections in children with malnutrition plausibly arises from a combination of factors. Beside the multiple immune cell dysfunctions discussed above, the relatively increased frequencies of mixed infections and sepsis due to enteric pathogens suggest that microbial translocation across a defective mucosal barrier is likely to play an important role. The structural and functional changes associated with environmental enteropathy (EE) are addressed above. In a murine model of EE, mice fed a low-fat, low-protein diet and infected with Salmonella Typhimurium had a significantly increased burden of bacteria in the jejunum compared with mice fed a normal diet (26). The increased load of potentially invasive bacteria colonizing permeable mucosae could explain the higher rates of bacteremia in malnourished children (390). A similar association between mucosal dysfunction and invasive infections is conceivable at the level of the respiratory system. Children in developing countries typically live in settings where crowding, inadequate room ventilation, and increased exposure to smoke from biomass fuel are common. Such risk factors have been associated with increased epithelial inflammation, pneumonia, and chronic lung disease (391). At the same time, epidemiological studies also showed that children in low-resource settings have high rates of nasopharyngeal carriage of S. pneumoniae and H. influenzae and a different immune response toward these bacteria than that of children living in industrialized countries (392). It is plausible that a disruption of the respiratory mucosal barrier coupled with an increased density of colonization facilitate invasive infections.

The impacts of micronutrient and vitamin deficiencies on tuberculosis are discussed above. The association of malnutrition with tuberculosis and early tuberculosis-related mortality is well established in adults (393–395) and was recently reviewed (396). TB is also a common cause of pneumonia in children with acute malnutrition (397), and WHO guidelines for the management of malnourished children as well as guidelines for national TB programs highlight the importance of the association of malnutrition and TB in children (378, 398). However, few studies of children have been completed, and a causal role of malnutrition as a risk factor for TB is not definitively established. Active TB itself often causes wasting, so data from retrospective studies are difficult to interpret. Furthermore, diagnostic tests for TB in children lack sensitivity, so diagnosis is often based on clinical findings without bacteriological confirmation. Therefore, we are left to conclude that malnutrition places a child at risk for TB through inference and extrapolation from data from studies of adults, BCG-vaccinated children, children with latent TB infection, and animal models (reviewed in reference 329). Severe malnutrition is associated with a reduced rate of tuberculin skin test positivity in children who received BCG vaccination (399), suggesting impaired cellular immune function and an increased risk for developing active disease. Accordingly, adults with coexisting latent TB infection and low BMI had a reduced protective cytokine response (e.g., IFN-γ and TNF-α) and increased regulatory cytokine (e.g., IL-10, IL-13, and TGF-β) production compared to individuals with latent TB infection and normal BMI (400). Severely underweight children were more likely to acquire TB infection following exposure to a household contact with pulmonary TB (401), but no studies have specifically addressed malnutrition as a risk factor for TB disease in children. Studies of M. tuberculosis-challenged guinea pigs (402–404) and mice (405) demonstrated that protein malnutrition compromised resistance to TB through defects in both innate and adaptive immune functions. Compromised host defense was associated with impaired T cell trafficking and proliferation, reduced production of protective cytokines (IFN-γ and TNF-α), impaired granuloma maturation, and reduced macrophage effector function (e.g., generation of nitric oxide) (329). Undernourished mice challenged with BCG also had reduced IFN-γ and TNF-α production and an increased risk of bacterial dissemination (406), and BCG vaccination failed to protect protein-malnourished guinea pigs against M. tuberculosis challenge (404, 407).

Placental malaria is a major contributor to fetal malnutrition (408). Intermittent preventive antimalarial treatment during pregnancy can reduce the risk of low birth weight and increase the length of the infant at 4 weeks of age (409–411). Repeated episodes of clinical malaria in children can also cause underweight or growth faltering, which can be prevented by measures to reduce malaria transmission (412–414). On the other hand, there has been considerable controversy over the impact of malnutrition on the outcome of malaria infection. Unfortunately, most studies were not designed to determine causality (415). Prospective cohort studies found an insignificant increase in the malaria incidence in children with moderate to severe undernutrition (416). Stunting and underweight increase the risk of malaria mortality (5), and recent data show that deficiencies in protein-energy, zinc, and vitamin A contribute significantly to malaria morbidity and mortality (reviewed in reference 417). Data pooled from several large-cohort studies identified relative risks of malaria mortality of 9.5, 4.5, and 2.1 for severe (weight-for-age z score of <−3), moderate (z score of between −2 and −3), and mild (z score of between −1 and −2) undernutrition, respectively, compared to children with a z score of >−1 (416). Based on these relative risks and prevalence data, the fractions of clinical malaria episodes and deaths from malaria in children under 5 years of age that were attributable to malnutrition (underweight, zinc deficiency, or vitamin A deficiency) were substantial (416).

The majority of people who are infected with the visceralizing Leishmania species Leishmania chagasi/L. infantum or L. donovani develop chronic latent infection without clinical disease. Both the innate immune response that occurs within a few days of infection and the long-term adaptive immune response play critical roles in preventing the development of active disease. Epidemiological studies have documented a greatly increased risk for visceral leishmaniasis (VL) in malnourished hosts (418–422). Malnutrition was identified as a risk factor for severe disease (419) and death from VL in both children (WFH value, <60%; odds ratio, 5.0) and adults (BMI, <13; odds ratio, 11.0) (423). Malnutrition-related VL is particularly evident in displaced and impoverished populations (424, 425), and the recently described movement of transmission into periurban slums is likely to lead to increased malnutrition-infection synergism (426). A murine model of polynutrient deficiency (deficient protein, energy, zinc, and iron) (93, 427, 428), which mimics moderate human malnutrition (429), recapitulated the epidemiological observations by demonstrating that polynutrient deficiency led to an increased rate of early dissemination following cutaneous infection with L. donovani (427). Subsequent studies using the polynutrient-deficient mouse model identified a defect in lymph node barrier function, likely related to fewer myeloid phagocytes and dysregulated cell trafficking, as a contributor to increased parasite dissemination (56, 57, 427). In a hamster model of progressive VL from L. infantum infection, protein malnutrition led to leukocyte depletion, impaired lymphoproliferation, and increased parasite burdens compared to well-nourished controls (430).