Metabolic diseases, including insulin resistance, type 2 diabetes, obesity, hypertension, and cardiovascular diseases associated with malnutrition and sedentary life are developing worldwide (Popkin et al., 2012). The Developmental Origins of Health and Disease hypothesis was formulated in the early 1990s (Hales and Barker, 1992). It suggested causal links between early nutrition and later chronic diseases, particularly components of metabolic syndrome. A large body of evidence has now accumulated to demonstrate that maternal and early postnatal nutritional deficiencies or excesses do program genes implicated in energy homeostasis, involve epigenetic modifications that modulate gene expression, and finally affect offspring phenotype (Canani et al., 2011; Sébert et al., 2011a). So-called “predictive adaptive responses” are made, in an evolutionary perspective, early during critical windows of development for future advantage of the species (Gluckman and Hanson, 2004). A mismatch between the environment predicted during the perinatal period and the actual environment encountered later in life is suggested to be causal in the increased disease risk in the progeny (Gluckman et al., 2005).
Most organs including adipose tissue, pancreas, kidney, skeletal muscle, and brain appear to be imprinted by early disturbances (Warner and Ozanne, 2010). However, the gastrointestinal tract (GIT) is rarely present in this picture. It is surprising given the central role of the GIT as a barrier between the outer environment and the body and in food digestion and nutrient absorption. Causal implication of GIT barrier in the development of obesity has been proposed recently (Cani and Delzenne, 2009). Indeed, bacterial lipopolysaccharide (LPS) leakage from the gut to the body increases with high fat diet consumption and leads causally to low grade (or metabolic) chronic inflammation and adipose tissue expansion in mice (Mus musculus). Besides nutrition, the influence of stress early in life is important to consider because it is now well established that it leads to long-lasting alterations in both intestinal and colonic mucosal function (Soderholm and Perdue, 2006; Gareau et al., 2008). Many studies on early programming have been conducted in rodents and sheep, and a recent literature survey suggests that swine is a good translational model to man (Guilloteau et al., 2010). Additionally, most basic neuroimmune regulatory mechanisms of gut physiology and barrier function involving, notably, corticotropin-releasing factor (CRF) and mast cells (see below) have been confirmed in pigs after weaning (Moeser et al., 2007a, b). The intestines are the main focus of this review because little information is available on the stomach (Sébert et al., 2011b).
Interactions between the microbiota and the host are being considered as potential players in the early programming of gut functions. Increasing evidence indicates that the gut microbiota program functions and metabolism of host organs, including the GIT (Bäckhed, 2011; Kaplan and Walker, 2012). In humans, early alterations in GIT microbial colonization, for example, by perinatal antibiotic treatment, are suspected to be responsible for increased disease risks in later life (Bedford Russell and Murch, 2006). Therefore, preservation of microbiota–host mutualism setup appears crucial (Willing et al., 2011). The involvement of the enteric microbiota is also reviewed as an important modulator of GIT programming.
Potentially all nutrients, vitamins and minerals provided deficiently or in excess to the mother during pregnancy and lactation or to the progeny during the neonatal period may play a role in offspring GIT programming. Nevertheless, only some nutrients, including the major nutrient sources (e.g., protein, fat) and specific methyl donors (e.g., betaine, choline, folic acid, vitamin B12) that are important in epigenetic modifications (see below) have been studied (Table 1).
|Factor||Species||Effects on the gut||Pathways/players||Reference|
|Protein restriction||Rodents||Reduced colonic expression [intrauterine growth restriction (IUGR)] of mucin 2 and mucin 4 genes||Fança-Berthon et al. (2009)|
|Rodents||Blunted jejunal interaction with dietary fat content,reduced goblet cell density||Intestinal alkaline phosphatase (IAP) activity [in transcription factors Kruppël-like factor 4 (Klf4) and caudal homeobox protein 2 (Cdx2)]||Lallès et al. (2012)|
|High protein||Pigs||Reduced ileal inflammatory responses ex vivo and nuclear factor-kappa B (NFkB) pathway activation||Reduced inflammatory cytokines||Chatelais et al. (2011)|
|High fat [n-3 poly-unsaturated fatty acids||Pigs||Increased intestinal expression of glucose transporters (PUFA)]||Gabler et al. (2007)|
|Rodents||Reduced colonic crypts; increased colonic inflammatory responses and increase in (n-3) PUFA||Early reduction in arachidonic acid||Innis et al. (2010)|
|High fiber||Rodents||Reduced colonic diverticulosis||Wess et al. (1996)|
|High methyl donors||Rodents||Increased colonic sensitivity to inflammation||Epigenetic modifications||Schaible et al. (2011)|
|High folic acid||Rodents||Reduced incidence of colorectal cancer||Epigenetic modifications||Sie et al. (2011)|
Nutrient restriction during pregnancy, leading to so-called intrauterine growth restriction (IUGR), is associated with an increased risk of developing inflammation, obesity, and metabolic syndrome in later life (Godfrey et al., 2011). Intrauterine growth restriction is also a production problem in farm animals (Wu et al., 2006). Intrauterine growth retarded rats displayed reduced colonic mucin 2 and mucin 4 gene and protein expressions after weaning (Fança-Berthon et al., 2009). We recently observed that among the variables investigated, intestinal alkaline phosphatase (IAP) (and aminopeptidase N to a lesser extent) in the jejunum was influenced by both mother protein malnutrition and adult offspring diet in IUGR rats submitted (or not) to a high-fat diet in adulthood (Lallès et al., 2012). Whereas IAP activity doubled in response to the high-fat diet in control adult rats, this response was blunted in adult IUGR offspring. High fat-fed IUGR rats also displayed lower goblet cell densities. Intestinal alkaline phosphatase is a crucial brush border enzyme involved in various homeostatic functions, including detoxification of bacterial pro-inflammatory components (e.g., LPS) and control of intestinal fatty acid absorption (Lallès, 2010). Our data suggest that adult IUGR offspring consuming a high-fat diet are at higher risk of gut (and systemic) inflammation. Kruppël-like factor 4 (Klf4), a major IAP transcription factor, was downregulated in IUGR individuals (Lallès et al., 2012). Kruppël-like factor 4 governs epithelial cell progenitor proliferation and terminal differentiation into enterocytes, goblet cells (producing mucins), and Paneth cells (producing antibacterial peptides). Mice with deleted intestinal Klf4 gene display lower IAP activity and goblet cells densities, associated with alterations in epithelial distribution of Paneth cells (Ghaleb et al., 2011). Therefore, Klf4 may be a target gene involved in early GIT programming.
Feeding a high-protein milk formula to normal birth-weight piglets led to increased ileal permeability in the neonatal period but not at 160 d of age (Chatelais et al., 2011). Ileal pro-inflammatory cytokine responses to LPS in cultured explants were lower in these young adult pigs that had been fed the high-protein formula, suggesting a long-term protective effect against inflammation of this formula on the ileum. In 4-wk-old IUGR pigs, a high-protein milk formula also increased ileal permeability in association with alterations in its nervous regulation (Boudry et al., 2011a). Long-term outcomes were not reported in this study.
Investigations in rats suggest that maternal fat consumption during gestation and lactation would predispose to obesity in adult progeny (Howie et al., 2009; White et al., 2009). In mice, 3-mo-old male offspring born to mothers fed a diet rich in (n-3) polyunsaturated fatty acids (PUFA; fish oil) during gestation and lactation showed reduced colonic crypt depth and enhanced inflammatory response to chemically induced colitis (Innis et al., 2010). It was concluded that early restriction in particular (n-6) fatty acids (e.g., arachidonic acid) may program long-term colonic proliferation and functional alterations.
Piglets born to sows fed a diet enriched in linseed oil [rich in (n-3) PUFA] during gestation and lactation displayed, in the short term, increased ileal permeability associated with alterations in architecture and function of enteric nerves (De Quelen et al., 2011) together with reduced ileal barrier sensitivity to mast cell degranulation (Boudry et al., 2009). A gestation diet rich in (n-3) long chain (eicosapentaenoic; docosapentaenoic) PUFA also increased intestinal expression of glucose transporters [glucose transporter 2 (GLUT2) and sodium-dependent-D-glucose cotransporter 1 (SGLT-1)] and glucose absorption in weaned pigs (Gabler et al., 2007). Changes in lipid composition of cell membranes may mediate, at least partly, these functional changes. No data are available on long-term effects of PUFA in swine.
Few data are available on fiber. In rats, the incidence of colonic diverticulosis in adult offspring varied according to the fiber content of mothers’ diet (Wess et al., 1996). Diverticulosis was absent in offspring fed a high fiber diet and born to mothers also fed a high fiber diet. The incidence of diverticulosis was the highest (46%) when both mothers and offspring had been fed low fiber diets and intermediate when only offspring received the high fiber diet. The underlying mechanisms were not reported.
Methyl donors (e.g., betaine, choline, folic acid, vitamin B12) are important to consider as epigenetic modifications included genomic DNA methylation (see below). Maternal folic acid supplementation reduced but postweaning supplementation increased the occurrence of chemically induced colorectal cancer in adult rat progeny (Sie et al., 2011). In mice, maternal diet supplementation with a mixture of methyl donors increased colonic sensitivity to chemical inflammation in adult offspring (Schaible et al., 2011). No data are available on methyl donors provided in early life and GIT function in older pigs.
Various kinds of stress (e.g., maternal deprivation, water avoidance, animal transport and handling) when applied acutely or chronically profoundly affects gut mucosal function (Soderholm and Perdue, 2006; Gareau et al., 2008) (Table 2). Alterations include increases in ion and mucous secretion and in para- and transcellular permeability, ultimately leading to local inflammation. Both visceral and central nervous systems constitute the primary players in these disorders, involving cholinergic regulation and central and local CRF and growth factors [e.g., nerve growth factor (NGF)]. Mast cells appear to be central to these gut disorders. They control mucosal secretion and permeability through nervous stimulation and mediator release in CRF- and NGF-dependent manners (Gareau et al., 2008). These neuroimmune regulations of colonic function have been reported in man too (Wallon et al., 2008). Corticotropin-releasing factor, neurokinin [through neurokinin 1 (NK1) receptors], and mast cells are also involved in long-term hyperalsegia after early maternal separation (Schwetz et al., 2004; Barreau et al., 2008). Maternal deprivation increases nerve density and reduces nerve–mast cell distances specifically in rat colon (Barreau et al., 2008). It also reduces Klf4 expression and alters goblet and Paneth cell distribution CRF-dependently in rat jejunum (Estienne et al., 2010).
|Factor||Species||Effects on the gut||Pathways/players||Reference|
|Perinatal stress||Rodents||Increased ion and mucus secretion; corticotropin-releasing factor (CRF) (central and local), nerve growth factor (NGF) increased trans- and paracellular (local), neurokinin, and mast cell permeability; reduced nerve-mast transcription factor Kruppël-like factor 4 (Klf4) cell distance; visceral hyperalgesia||Gareau et al.(2008); Barreau et al. (2008); Estienne et al (2010)|
|Gut irritation||Rodents||Chronic visceral hyperalgesia||Central CRF, colonic NGF, colonic transient receptor potential vanilloid 1 (TRPV1) and transient receptor potential ankyrin 1 (TRPA1)||Al Chaer et al. (2000); Barreau et al. (2008); Christianson et al. (2010)|
|Contaminants [bisphenol A (BPA)]||Rodents||Decreased colonic permeability and increased colonic sensitivity to inflammation||Estrogen receptor β||Braniste et al. (2010)|
|Antibiotics||Rodents||Alterations in expression of genes||Various involved in gut development and barrier function||Schumann et al. (2005)|
|Rodents||Increased mast cell density and mast cell mediator concentrations||Mast cells and mast cell mediators||Nutten et al. (2007)|
|Pigs||Alterations in ileal paracellular permeability (in interaction with dietary fat content); reduced jejunal intestinal alkaline phosphatase (IAP) and increased dipeptidyl peptidase IV (DPP-IV) activities,transient reduction in intestinal heat shock proteins (HSP)||Boudry et al.(2011b); Mroz et al. (2011); Arnal et al. (2012)|
|Probiotics||Rodents||Restoration of gut barrier alterations||Induced by neonatal stress or antibiotics||Garcia Rodenas et al. (2006); Eutamene et al. (2007); Gareau et al. (2007a)|
Irritable bowel syndrome is a common disease characterized by chronic abdominal pain probably caused by multiple factors of various origins, including neonatal stress and involving the nervous system. Mechanical or chemical (e.g., acetic acid) irritation of the colon before weaning (but not in adulthood) resulted in chronic visceral hypersensitivity in adult rats (Al Chaer et al., 2000) (Table 2). This was associated with central neuron sensitization in the absence of colonic histopathology. The mechanisms involved the transient receptor potential vanilloid 1 (TRPV1) whose gene expression and protein production increased in colonic afferent neurons in dorsal root ganglia (Winston et al., 2007). Gene expression of growth factors (e.g., NGF) in distal colon was also increased transiently after birth. In a similar model in mice, a second receptor [transient receptor potential ankyrin 1 (TRPA1)] present on colonic afferent neurons was increased in adults following early colonic mechanical stress (Christianson et al., 2010). In rats, long-term nervous alterations were reported in the spinal cord (NGF; serotonin) and brain (CRF and CRF1-R) of adult progeny after perinatal stress (Chung et al., 2007; Ren et al., 2007; Tjong et al., 2010). Collectively, these studies point to the high sensitivity of the immature nervous system to early local and central conditioning. Such investigations remain to be conducted in swine.
Endocrine disruptors are hormone-like synthetic compounds that are present in our environment, including food. They are characterized largely by their adverse effects on fertility and development. One recent study reports programming of the gut with low doses of bisphenol A (BPA). Perinatal exposure to low doses of BPA (reduced colonic paracellular permeability, BPA dose—and estrogen receptor β—dependently but increased pro-inflammatory responses in female adult offspring (Braniste et al., 2010) (Table 2). Underlying mechanisms included increased tight junction protein [occludin, junctional adhesion molecule A (JAM-A)] production and enhanced immune cell sensitivity, respectively.
INVOLVEMENT OF THE GUT MICROBIOTA
Bacteria are determinant in the development of the gut and the maintenance of its homeostasis. Components of gut bacteria stimulate the production of tight junction proteins, participate in gut protection from injury (Yu et al. 2012), and promote intestinal homing of mast cells (Kunii et al., 2011). Various bacterial components, signaling pathways, and effects on gut epithelial cells have been identified (Lebeer et al., 2008; Lallès, 2009).
Neonatal stress induces long-lasting disorders in rodent colonic function and also affects gut bacterial colonization (Gareau et al., 2008). Mother-deprived offspring displayed 10- to 100-fold increases in bacterial adherence to colonic tissue and translocation to the spleen (Gareau et al., 2006). By contrast, colonic counts of lactobacilli were reduced 10-fold (Gareau et al., 2007b). Colonic functional disorders could be alleviated by oral administration of lactobacilli to offspring during the separation period. In IUGR rats, postnatal alteration in colonic maturation was associated with changes in bacterial composition and metabolic activity (Fança-Berthon et al., 2009, 2010). However, mechanisms linking the microbiota and gut function programming are still poorly understood.
The gut microbiota is known to be influenced by diet composition. For example, a high-fat diet decreased total bacterial numbers and increased the Bacteroides to Clostridia ratio in rats (de La Serre et al., 2010). More importantly, the density of enterobacteria was higher in obesity-prone rats compared to obesity-resistant rats. It was concluded that gut inflammation is a primary factor in obesity development and changes in microbiota composition are secondary. Interactions between mother’s diet and bacteria on rat offspring gut architecture and function were reported recently (Fåk et al., 2012). Maternal high-fat diet led to early decreases in maltase and sucrase in the progeny. Addition of Escherichia coli to dams that were fed a high-fat diet during pregnancy and lactation increased offspring intestinal permeability, systemic inflammation, and obesity (Fåk et al., 2012). Maternal administration of E. coli had long-term effects on weight gain and gut microbiota in the progeny (Karlsson et al., 2011), but specific effects on gut function were not reported. Other studies indicate that maternal administration of antibiotics in rat dams increased gut permeability and systemic inflammation in offspring (Fåk et al., 2008a). Also, supplementation of diets for mothers and offspring with Lactobacillus plantarum stimulated offspring intestinal tissue growth and restored intestinal permeability (Fåk et al., 2008c). Such responses were restricted to rats supplemented with this probiotic between 3 and 10 d of life (Fåk et al., 2008b), emphasizing the importance of the early window of action. Bacterial strain was also shown to be determinant (Eutamene et al., 2007). Neonatal administration of antibiotics was shown to alter the expression of genes involved in GIT development and intestinal barrier (Schumann et al., 2005). Importantly, this treatment also increased intestinal mast cell tissue density and mediator (e.g., protease) concentrations (Nutten et al., 2007). Conversely, probiotic bacteria used alone or with specific dietary ingredients (e.g., oligosaccharides, PUFA) were able to alleviate gut barrier disorders observed in young and adult rodents submitted to neonatal stress (Garcia-Rodenas et al., 2006; Eutamene et al., 2007; Gareau et al., 2007a) (Table 2).
We recently developed a swine model of perinatal disturbance in bacterial colonization by oral antibiotic administration to sows around farrowing. We observed transient short-term (day 14) and long-term (month 6) alterations in paracellular permeability (Boudry et al., 2011b) and reductions in IAP activity and gene expression in offspring born to antibiotic-treated mothers (Mroz et al., 2011). Conversely, the activity of peptidase [dipeptidyl peptidase IV (DPP-IV)] was doubled in the jejunum of the offspring (Table 2). A high intestinal DPP-IV activity was shown to specifically deteriorate glucose tolerance and reduce circulating insulin through incretin breakdown and generation of deleterious dipeptides in mice (Waget et al., 2011).
Heat shock proteins (HSP) regulate various facets of cell biology and intracellular protein trafficking. In the GIT, inducible HSP (HSP27, HSP70) are important contributors of health maintenance (Petrof et al., 2004; Otaka et al., 2006; Lallès, 2009). They induce subsequent GIT protection following a second stressor, leading to the concept of “memory to stress.” The microbiota is essential for intestinal production of protective HSP because germfree rodents or those treated with antibiotics display low HSP levels (Arvans et al., 2005). By contrast, bacterial pro-inflammatory components (LPS), metabolites (butyrate), and various soluble factors lead to increased production of protective HSP in intestinal epithelial cells (Lebeer at al., 2008; Lallès, 2009). We hypothesized that perinatal alteration of bacterial colonization programs HSP basal expression along the GIT. Following perinatal antibiotic treatment of sows, we observed a reduced expression of HSP70 (but not HSP27) in the jejunum and ileum of offspring at 28 (weaning) and 42 d of age (Arnal et al., 2012). However, no change in HSP was noted in 6-m-old offspring. Transcription factor heat shock factor 1 (HSF1) protein production was unchanged. Our data support the idea that gut HSP are not submitted to imprinting. This agrees with data on colonic smooth muscle HSP of mother-deprived adult rats (Lopes et al., 2008). We are currently working on swine offspring colonic HSP.
Increasing evidence indicates that the effects of early nutrition and environment on chronic diseases in offspring may be mediated by epigenetic mechanisms throughout life (McKay and Mathers, 2011). Epigenetic modifications include DNA methylation of gene promoter regions, multiple histone modifications (e.g., acetylation, methylation, phosphorylation, ubiquitination), and noncoding micro-RNA expression, all leading to finely tuned modulation of gene transcription. Epigenetic modifications of colonic mucosa continue during life and may contribute to age-associated increase in colon inflammation (Kellermayer et al., 2010).
Methylation of DNA is a common epigenetic modification that can be altered by maternal diet supplementation with methyl donors, leading to persistent phenotypic modifications in the progeny (Waterland and Jirtle, 2003). It was later hypothesized that early nutrition affects the gut epigenome, causing metabolic imprinting of GIT structure and function (Waterland, 2006). However, data are still scarce. The reduced occurrence of chemically induced colorectal cancer in rat offspring following maternal folate supplementation was associated with higher colorectal global DNA methylation, but postweaning intervention had opposite effects (Sie et al., 2011) (Table 2). Maternal methyl donor supplementation-associated increase in mouse offspring colitis was accompanied with changes in colonic mucosa DNA methylation and gene expression (Schaible et al., 2011). Genotype- and gender-dependent reduction in genomic DNA methylation in the small intestine of murine adult offspring was observed following perinatal folate depletion (McKay et al., 2011a, b).
The microbiota is involved in the modulation of the host epigenome at the GIT level. Intestinal epithelial cells are hyporesponsive to LPS, and this is mediated by epigenetic modification of the Toll-like receptor 4 (TLR4) gene (Vamadevan et al., 2010; Takahashi et al., 2011). Cytosine nucleotide-phosphate-guanine nucleotide (CpG) methylation of this gene was lower in the small intestine than in the colon of conventional mice and lower in the colon of germfree than conventional mice (Takahashi et al., 2011). Histone deacetylation at the 5′ region of the TLR4 gene was also reported (Takahashi et al., 2009). Butyrate regulates IAP expression epigenetically (Hinnebusch et al., 2003). Finally, maternal methyl-donor supplementation persistently influenced offspring colonic mucosa microbiota, suggesting bacterial involvement in epigenomic reprogramming of gut function in mice (Schaible et al., 2011). No data is available on gut epigenetic modifications in swine.
CONCLUSIONS AND PERSPECTIVES
Increasing evidence supports the long-term influences of early nutrition and environment (stress) on GIT architecture and function. A number of macronutrients, vitamins, and minerals provided in restriction or excess during pregnancy or lactation or both in mothers may have long-lasting and often negative gender-specific influences on the progeny. Maternal preconception nutrition and environment should be considered too because the former affects the GIT of offspring (Mortensen et al., 2010). Independent models suggest that perinatal disturbances (either nutritional or bacterial) are responsible for the early programming of various intestinal key functions, some of which (e.g., IAP, Klf4) seem to respond to a broad range of stimuli (Figure 1). This supports the notion that early programming of common adult metabolic diseases resulting from perinatal disturbances targets a core of specific genes or gene pathways (housekeepers) regardless of the nature of the initial trigger (McMullen et al., 2012). Literature evidence discussed in this review support the view that the GIT contributes specifically to metabolic disorders in adult offspring.
Gastrointestinal tract disturbances often involve both the enteric and central nervous systems. Precisely how these are programmed and the specific contribution of the gut microbiota need further investigations (Bercik et al., 2012). Gastrointestinal tract responses are also region specific and in that respect the stomach has been understudied. The persistent shaping of GIT architecture and function involves a complex cross talk between the microbiome and the host. Underlying nongenetic mechanisms are being unraveled but more work is needed to get a clearer picture of epigenetic changes, including the actual role of gut bacteria. Finally, paternal influences on offspring programming as well as the transgenerational transmission of epigenetic markings are emerging as important factors leading to pancreas alterations, metabolic disorders, and obesity (Ng et al., 2010; Curley et al., 2011; Dunn and Bale, 2011). The GIT and its microbiome are probably contributing organs and should, therefore, be further investigated. In swine, future studies should evaluate long-term and intergenerational outcomes of nutrition and management of animals kept for breeding and neonates on offspring GIT function and health at critical periods of life.