- DOHaD
developmental origins of health and disease
- FABP
fatty acid-binding protein
- UCP
uncoupling protein
Developmental origins hypothesis
Programming is defined as a process through which exposure to environmental stimuli or insults during critical phases of development brings about permanent changes to the physiology or metabolism of the organism. The embryonic, fetal and early postnatal periods represent critical stages of development during which programming of body systems can occur(Reference Symonds, Stephenson and Gardner1). Thus, there is great potential for the in utero and early postnatal environment to impact on future physiological and biochemical systems, including adipose tissue.
The link between birth weight and adult disease came initially from epidemiology highlighting that rates of CVD are highest in ‘poorer’ more-socially-deprived locations and in lower-income groups in the UK(Reference Barker and Osmond2). However, variations in adulthood lifestyle choices such as diet and smoking could not explain these differences in CVD. Investigation of the rates of infant mortality 60 years previously in the same geographical locations has demonstrated the highest infant mortality in the areas that 50 or 60 years later show the highest rates of CVD(Reference Barker and Osmond2, Reference Barker, Bull and Osmond3). By inferring that infants in socially-deprived areas in the 1930s had experienced a suboptimal in utero or postnatal environment, particularly poor nutrition, a link was developed between an adverse intrauterine environment and the risk of CVD in adult life. Several subsequent large-scale epidemiological studies that used cohorts in Hertfordshire (UK), Preston (UK) and Helsinki (Finland) as well as analysis of historical data from the 1946 Dutch famine have confirmed that nutrient restriction during gestation increases the incidence of several disorders, including obesity(Reference Barker, Gluckman and Godfrey4–Reference Painter, Roseboom and Bleker6). Researchers investigating the developmental origins of health and disease (DOHaD) hypothesis concurrently began to study the potential associations between a suboptimal fetal and/or postnatal environment and have found that several pathologies such as obesity and diabetes are increased in the offspring(Reference Anguita, Sigulem and Sawaya7, Reference Tamashiro, Terrillion and Hyun8). The finding that suboptimal conditions in utero can lead to obesity will come as no surprise to farmers who have known for many years that ‘runt’ pigs grow more slowly and produce pork containing excess fat(Reference Powell and Aberle9). There are numerous experimental studies now investigating the impact of early-life events on later health.
As described in other presentations from the ‘Frontiers in adipose tissue biology’ symposium, over the past 15 years adipose tissue has emerged as a dynamic endocrine organ with an extent of anatomical and physiological plasticity(Reference Cinti10, Reference Cannon and Nedergaard11). Epidemiological and animal studies have determined that obesity and type 2 diabetes are pathologies programmed by adverse events in utero or in early postnatal life(Reference Barker, Gluckman and Godfrey4, Reference Cettour-Rose, Samec and Russel12–Reference Whorwood, Firth and Budge14); therefore, it is unsurprising that adipose tissue is a prime target. There are several features of adipose tissue that make it a highly-programmable tissue: plasticity and/or expansion capacity; cell number set in early life; insulin signalling; obesity; adipokine production; type 2 diabetes; energy storage and regulation; depot specificity.
Fat cell turnover
One of the most striking features of adipose tissue and its role in obesity is that cell number is set early in life. Recent work has demonstrated convincingly that fat cell number, but not volume, is set during childhood and adolescence(Reference Spalding, Arner and Westermark15). Earlier work that informed the latter research had indicated that the development of fat depots of obese children differ both quantitatively and qualitatively from that of non-obese children of the same age(Reference Knittle, Timmers and Ginsberg-Fellner16). Obese children display marked increases in fat cell size, up to non-obese adult values, early in life at a time when non-obese subjects display no change or reduced lipid content. These findings strongly suggest that the foundations for excessive adiposity are laid down during infancy. Furthermore, it is suggested that there are certain time intervals that have important consequences for ultimate adulthood cellularity and size of fat depots(Reference Knittle, Timmers and Ginsberg-Fellner16). Continuation of the time course of adipose tissue study into adulthood has confirmed that adipose tissue cellularity is set in childhood and adolescence, as adipocyte number remains stable through adulthood(Reference Spalding, Arner and Westermark15). Interestingly, cell number, but not volume, appears to be irreversible when weight is lost after bariatric surgery(Reference Spalding, Arner and Westermark15).
Animal models
In order to produce offspring in which to study adipose tissue, experimenters must produce a suboptimal in utero or postnatal environment. There are several different methodologies employed by researchers in the DOHaD field, including placental restriction (sheep, rodent), glucocorticoid infusion (sheep, rodent, avian species), nutritional manipulation (sheep, pig, rodent, non-human primate) and natural birth-weight variation (pig; for a review of these methodologies, see McMullen & Mostyn(Reference McMullen and Mostyn17)). Such in utero challenges can result in the long-term programming of adipose tissue abundance and function. A range of hormones, enzymes, transcription factors and other metabolic signalling molecules have been implicated in adverse adipose tissue development, including leptin, glucocorticoids, members of the PPAR family, fatty acid-binding proteins (FABP) and adipokines(Reference Budge, Gnanalingham and Gardner18, Reference Budge, Sebert and Sharkey19). For the purpose of the present review only evidence from two large-animal models, sheep and pig, will be considered.
Evidence from porcine studies
Low-, normal- and high-birth-weight pigs
There is overwhelming evidence from agriculture to demonstrate that small (or ‘runt’) piglets grow more slowly and become ‘fat’ adults(Reference Powell and Aberle9, Reference Rehfeldt, Tuchscherer and Hartung20). In an investigation of a cohort of low-, middle- and heavy-birth-weight piglets up to 180 d postnatal age it was found that the low-birth-weight piglets grow more slowly but contain more perirenal adipose tissue and intramuscular fat than the middle-birth-weight piglets(Reference Poore and Fowden21).
Several experimental studies have now compared the effects of physiological and biochemical outcomes in low- and normal-birth-weight piglets on adipose tissue in later life. Low-birth-weight piglets have been demonstrated to have significantly higher back fat at 12 months (P<0·05)(Reference Poore and Fowden22). However, despite the increase in fat depth, male pigs at 12 months of age have reduced plasma leptin and there are no associations between leptin and back fat(Reference Poore and Fowden22). Gender-specific effects of birth weight and other in utero and postnatal challenges are a common feature in DOHaD studies, with many experimental findings being gender, and also age, specific. The same cohort of low-birth-weight pigs have impaired glucose tolerance at 12 months of age(Reference Poore and Fowden21).
Although the early epidemiological studies of DOHaD focused on nutrient restriction and/or socio-economic background linked to a low birth weight, offspring with high birth weight may also be at risk from obesity and other diseases in later life; the ‘U’-shaped curve(Reference Baker, Olsen and Sørensen23). On this basis, adipose tissue from piglets with low, normal or high birth weights were investigated on days 7 and 14 of age (neonatal period). It was found that uncoupling proteins (UCP) 2 and 3, which have a number of cellular roles including energy regulation and inflammation, are reduced in adipose tissue from low-birth-weight piglets on day 7, but not day 14(Reference Mostyn, Litten and Perkins24). In addition, UCP3 is positively related to plasma leptin in normal-sized pigs but negatively correlated in low- and high-birth-weight piglets on day 14, suggesting reduced activation(Reference Scarpace, Nicolson and Matheny25). Although the study only investigated the neonatal effects of low birth weight, a loss of the relationship between leptin and UCP3 may promote adipose tissue deposition. However, at this early stage of development there is no difference in adipose tissue TAG content. Low-birth-weight piglets do possess markedly more adipocytes(Reference Williams, Marten and Wilson26), in keeping with the earlier findings(Reference Spalding, Arner and Westermark15) and suggesting an increased capacity for adipose tissue expansion and lipid accumulation. FABP3 and 4 (Fig. 1) and PPARγ are down regulated in low- and high-birth-weight piglets on day 7, but not day 14 in adipose tissue. Conversely, FABP3 and 4 are down-regulated in low- and high-birth-weight piglets in skeletal muscle on day 14. The reduced expression of PPARγ2 in adipose tissue of the low- and high-birth-weight piglets suggests delayed adipocyte development. This finding is partly confirmed by histological data, as the low-birth-weight piglets, but not the high-birth-weight piglets, have more adipocytes per visual field on day 7(Reference Williams, Marten and Wilson26). The reduction in FABP4, the ‘adipose tissue’-specific FABP, may be protective against the development of metabolic syndrome, as a number of recent studies suggest FABP4 is central to its onset through interaction with the insulin signalling and inflammatory pathways(Reference Boord, Maeda and Makowski27–Reference Maeda, Cao and Kono29). A protective effect in this case seems unlikely, given the well-described ‘U’- or ‘J’-shaped relationship between birth weight and obesity and the metabolic syndrome in human studies(Reference Baker, Olsen and Sørensen23, Reference Martorell, Stein and Schroeder30), but cannot be excluded because these animals have not been studied into adulthood.
Fig. 1. Adipose tissue gene expression on day 7 of postnatal age in low-body-weight (□), normal-body-weight (▪) and high-body-weight (
) piglets. Values are means with their standard errors represented by vertical bars for five animals per group. a,bMeans with unlike superscript letters were significantly different for each gene (P<0·05). BW, birth weight; FABP, fatty acid-binding protein. (Adapted from Williams et al.(Reference Williams, Marten and Wilson26).)
The Meishan breed
Early programming of adipose tissue in the Meishan breed of pig may provide protection against hypothermia and hypoglycaemia in this ancient oriental breed. The comparatively low birth weight of Meishan piglets compared with commercial lean breeds, associated with a high litter number, is not detrimental to offspring survival as the mortality rate is lower than that of commercial breeds(Reference Le Dividich, Mormede and Catheline31). Adipose tissue PPARγ expression during the first week of life is reduced in Meishan piglets compared with that of piglets of a commercial breed(Reference Mostyn, Litten and Perkins32) and the expression of adiponectin and its receptor (adiponectin receptor 1) are below the level of detection at ⩽21 d of age(Reference Treece, Williams and McGivern33). A greater percentage of plurilocular adipocytes has been reported in Meishan piglets at ⩽30 d of age compared with a lean breed(Reference Hauser, Mourot and De Clercq34). Taken together these findings suggest that adipose tissue development is delayed in Meishan piglets, which may have a beneficial impact on nutrient partitioning and energy regulation during the neonatal period. Although older but prepubertal Meishan pigs (i.e. aged 80 d) have increased lipogenic potential compared with Large White pigs, the opposite has been found after puberty at 100 d, when several fat depots exhibit less lipogenic potential(Reference Mourot, Kouba and Bonneau35). These findings suggest that the mechanisms responsible for high fat accretion occur after weaning but before puberty(Reference Hauser, Mourot and De Clercq34, Reference Treece, Williams and McGivern36).
Interestingly, milk leptin of Meishan sows is lower than that of commercial sows, suggesting that in the pig milk leptin does not reflect maternal fat stores or serum leptin(Reference Guay, Palin and Jacques Matte37, Reference Mostyn, Sebert and Litten38). Milk leptin is, however, positively related to piglet growth rate, girth and body, gut, heart and spleen weight in Meishan piglets, suggesting a growth-promoting effect(Reference Mostyn, Sebert and Litten38). Despite milk leptin not reflecting adiposity, serum leptin is associated with backfat thickness in early pregnancy in Meishan–Landrace sows and increases significantly (P<0·05) up to day 25 of gestation (Fig. 2), when leptin receptor (total and long form) mRNA expression in homogenates prepared from whole fetuses is higher in Meishan–Landrace pigs than in Yorkshire–Landrace pigs(Reference Guay, Palin and Jacques Matte37).
Fig. 2. Serum leptin concentrations in two breeds of pregnant sow: Meishan–Landrace (□); Yorkshire–Landrace ().Values are means with their standard errors represented by vertical bars. a,bMeans with unlike superscript letters were significantly different between breeds (P<0·05). Mean value was significantly different from that at 25 d of gestation: *P<0·05. (Adapted from Guay et al.(Reference Guay, Palin and Jacques Matte37).)
The in utero factors that regulate Meishan adipose tissue are as yet unknown, but may reflect differences in glucocorticoid signalling(Reference Klemcke and Christenson39, Reference Desautes, Sarrieau and Caritez40). Plasma cortisol concentrations of Meishan fetuses are 30% greater than those of age-matched white cross-bred (Yorkshire×Landrace×Large White×Chester White) fetuses, despite similar plasma adrenocorticotropin and cortisone(Reference Klemcke and Christenson39). At 6 weeks old Meishan pigs exhibit higher circulating cortisol concentrations under basal conditions yet demonstrate a reduced response to a stressful experience(Reference Desautes, Sarrieau and Caritez40). Furthermore, 7-week old Meishan piglets exhibit an enhanced response to adrenocorticotropin infusion compared with Large White piglets(Reference Hazard, Liaubet and SanCristobal41). These increased basal and stimulated cortisol concentrations are likely to be a result of differential regulation in the Meishan adrenal gland, particularly via regulators of steroidogenesis(Reference Hazard, Liaubet and SanCristobal41). The reported increased plasma cortisol levels, along with greater glucocorticoid receptor gene expression observed in adipose tissue from 4-d-old Meishan piglets, may represent a potential for greater lipid uptake(Reference Mostyn, Sebert and Litten38). Fig. 3 summarises the early-life programming of adipose tissue in the Meishan breed.
Fig. 3. Summary of early-life differences in adipose tissue of Meishan pigs in comparison with a typical ‘commercial’ lean porcine breed(Reference Hauser, Mourot and De Clercq34, Reference Guay, Palin and Jacques Matte37, Reference Mostyn, Sebert and Litten38, Reference Mourot, Kouba and Bonneau62). ↑, Increased; ↓, decreased; GR, glucocorticoid receptor; adiporR1, adiponectin receptor 1.
Evidence from ovine studies
A number of in utero and postnatal challenges have been utilised to investigate the early-life programming of adipose tissue, including glucocorticoid infusion, suboptimal in utero nutrition (either through dietary or placental manipulation) and a postnatal obesogenic environment.
Glucocorticoid administration
Glucocorticoid infusion has been utilised in many studies of the DOHaD hypothesis, in particular in small animals in which a suboptimal maternal diet is linked to overexposure of the fetus or neonate to glucocorticoids(Reference Budge, Stephenson and Symonds42–Reference Langley-Evans44). However, the effect of a suboptimal maternal diet in a large-animal species has divergent responses in relation to maternal and offspring glucocorticoid concentrations, which are dependent on the type and timing of the nutritional challenge(Reference Budge, Stephenson and Symonds42). Studies of the effects of late-gestational glucocorticoid infusion on the growth of the fetus have demonstrated that the prepartum cortisol surge may be responsible for the normal decline in fetal growth rate observed towards term in the sheep(Reference Fowden, Szemere and Hughes45). Furthermore, glucocorticoid infusion has a pronounced impact on adipose tissue in the sheep when delivered near term. Maternal administration of the synthetic glucocorticoid dexamethasone on day 138 of gestation (term is approximately 147 d) enhances UCP1 abundance in perirenal adipose tissue (the most abundant adipose tissue depot in the fetal and newborn lamb) as well as thermogenic potential in prematurely-delivered sheep (day 140 of gestation)(Reference Clarke, Heasman and Symonds46). Infusing cortisol for 5 d to near-term (day 129 of gestation) fetal sheep produces an increase in UCP1 protein and UCP2 gene expression in perirenal adipose tissue. These early-life programming effects increase protein and mRNA for UCP1 to values comparable with those of a fetus at day 144 of gestation (Fig. 4), confirming another role for cortisol in the preparation of fetal adipose tissue for life after birth(Reference Mostyn, Pearce and Budge47).
Fig. 4. Uncoupling protein (UCP) 1 protein abundance in perirenal adipose tissue from fetal sheep that were infused for 5 d with cortisol (▪) or saline (9 g NaCl/l; □), adrenalectomised (Ax; ) or sham-operated (sham; 
). Values are means with their standard errors represented by vertical bars. a,bMeans with unlike subscript letters were significantly different between groups (P<0·05). (Adapted from Mostyn et al.(Reference Mostyn, Pearce and Budge47)).
In utero dietary challenges
Suboptimal nutrition during gestation has been demonstrated in several species to impact on offspring adiposity and can range from over- or undernutrition to placental restriction.
Studies of maternal nutritional restriction have played a substantial role in DOHaD research; many have demonstrated effects on adipose tissue(Reference Budge, Sebert and Sharkey19). For example, offspring sampled at day 140 of gestation from ewes that have experienced nutrient restriction during early–mid gestation (day 28–day 80 of gestation), which is coincident with the period of maximal placental growth, exhibit increased perirenal adipose tissue at birth(Reference Bispham, Gopalakrishnan and Dandrea48). This adipose tissue also displays alterations in glucocorticoid signalling at day 140 of gestation and at 6 months postnatally(Reference Gnanalingham, Mostyn and Symonds49). Gene expression of the glucocorticoid receptor and 11β-hydroxysteroid dehydrogenase 1 are up regulated in perirenal adipose tissue from offspring of sheep that have experienced suboptimal nutrition during early–mid gestation; however, 11β-hydroxysteroid dehydrogenase 2 (which inactivates cortisol to cortisone) is reduced(Reference Gnanalingham, Mostyn and Symonds49). Cortisol has consistently been implicated in both obesity and the metabolic syndrome, which is predictable given the physiological traits observed in Cushing's syndrome(Reference Newell-Price, Bertagna and Grossman50). However, there is little evidence for raised plasma cortisol in obesity(Reference Rask, Olsson and Soderberg51). Although plasma cortisol per se is not consistently elevated with increased fat mass in human subjects, tissue sensitivity to glucocorticoids is increased via alterations in the enzymes and receptors that regulate local levels(Reference Gnanalingham, Mostyn and Symonds49). Tissue sensitivity to cortisol is regulated by the enzyme 11β-hydroxysteroid dehydrogenase and the glucocorticoid receptor. Taken together these results suggest that these offspring, who have experienced suboptimal nutrition in utero early in gestation, may be at a higher risk of developing obesity and thus metabolic syndrome, which is in keeping with human findings from epidemiological studies(Reference Painter, Roseboom and Bleker6). The findings of increased glucocorticoid sensitivity in these ovine offspring are similar to those observed in Meishan piglets, which are known to develop into obese adults.
In contrast, maternal nutrient restriction during late gestation (day 110–day147 of gestation), which is coincident with the period of maximal fetal growth and a parallel rise in fetal fat depots, results in reduced adiposity in the offspring. At days 1 and 30 postnatally offspring adipose tissue mRNA expression of glucocorticoid receptor and 11β-hydroxysteroid dehydrogenase 1 is lower and that of 11β-hydroxysteroid dehydrogenase 2 is higher in the nutrient-restricted group(Reference Gnanalingham, Mostyn and Symonds49). This adaptation may initially be protective against the development of visceral obesity or may simply represent a response to the reduced nutrient supply. However, the offspring have greater relative fat mass and impaired glucose tolerance at 1 year, which is associated with a reduction in GLUT4 protein abundance in adipose tissue(Reference Gardner, Tingey and Van Bon52). Again, this animal work complements the findings of human epidemiological studies. The Dutch Famine data highlighted that nutrient restriction during discrete periods of gestation leads to differential effects on body systems, with late-gestational nutrient restriction affecting intermediary metabolism (in particular glucose–insulin homeostasis) and increasing the risk of the offspring developing type 2 diabetes(Reference Painter, Roseboom and Bleker6).
Postnatal dietary challenges
The influence of a secondary ‘insult’ on ovine offspring who have experienced maternal nutrient restriction during early–mid gestation has been investigated by restricting offspring physical activity and providing ad libitum feed. From weaning to 1 year of age offspring from control and nutrient-restricted sheep were raised in an obesogenic environment that produces a 60% reduction in physical activity compared with pasture-reared sheep(Reference Sebert, Hyatt and Chan53, Reference Sharkey, Gardner and Fainberg54). At 1 year of age offspring that have experienced in utero nutrient restriction have a total body, subcutaneous and visceral fat weight similar to that of the animals born of control-fed mothers and reared in the obesogenic environment. The nutrient-restricted offspring do, however, exhibit a reduction in daily food intake when individually housed, an adaptation that is accompanied by increased plasma insulin, but not NEFA (Fig. 5). Indeed, raised plasma insulin may represent the primary adaptation promoting excess storage of nutrients in these offspring.
Fig. 5. Plasma insulin (ng/ml; □) and NEFA (mmol/l; ) from offspring of nutrient-restricted (NR) or control ewes that were reared in an obesogenic postnatal environment. Values are means with their standard errors represented by vertical bars. a,bMeans with unlike subscript letters were significantly different between groups (P<0·05). (Adapted from Sebert et al.(Reference Sebert, Hyatt and Chan53).)
There is evidence to suggest that metabolic syndrome and obesity disrupt endoplasmic reticulum function and cause protein misfolding, which can trigger the misfolded protein response, as documented by several reviews linking tissue stress responses to the metabolic syndrome(Reference Hotamisligil and Erbay55, Reference Ozcan, Cao and Yilmaz56). In an attempt to alleviate stress within the endoplasmic reticulum, insulin signalling pathways are inhibited through activation of c-Jun N-terminal kinase 1(Reference Muoio and Newgard57), which impairs systemic glucose regulation(Reference Ozcan, Cao and Yilmaz56). Emerging evidence from the obese offspring born to nutrient-restricted mothers demonstrates an increase in phosphorylated c-Jun N-terminal kinase protein in perirenal adipose tissue, suggesting a down-regulation of insulin signalling(Reference Sharkey, Gardner and Fainberg54). Several other aspects of the unfolded protein response, inflammation and infiltration of pro-inflammatory macrophages are enhanced in obese offspring born to nutrient-restricted mothers(Reference Sharkey, Gardner and Fainberg54). This outcome may represent an in utero programming effect whereby total adipocyte number is reduced, thus increasing its susceptibility for cellular hypertrophy(Reference Heilbronn, Smith and Ravussin58).
Experimental placental restriction
Experimental restriction of placental growth results in reduced birth weight, increased early postnatal growth and increased adiposity in the offspring at 6 weeks of age(Reference De Blasio, Gatford and Robinson59) and impairs insulin sensitivity. Feeding behaviour is also reset at 2 weeks of age, at least in the 90 min after 1 h of fasting, when placental-restriction offspring exhibit a longer suckling time, which is predictive of catch-up growth and increased central adiposity, at least up to 6 weeks of age. The molecular adaptations that produce these physiological responses in response to placental restriction may be set in fetal life. A reduction in insulin-like growth factor 1 and leptin mRNA expression has been demonstrated in perirenal adipose tissue from placental-restriction fetuses at day 140–day 145 of gestation(Reference Duffield, Vuocolo and Tellam60), which could impair adipocyte proliferation and differentiation, thereby potentially increasing their susceptibility for hypertrophy in later life(Reference Heilbronn, Smith and Ravussin58). A reduction in leptin may simply reflect a reduction in lipid stores of the fetal perirenal adipose tissue(Reference Symonds, Phillips and Anthony61). Together these findings suggest that offspring from an ovine model of intrauterine growth restriction display features that may predispose to obesity in later life, but to date have not demonstrated any long-term complications.
The various models of early-life programming of adipose tissue in the sheep are summarised in Fig. 6.
Fig. 6. Summary of adipose tissue programming effects in sheep subject to suboptimal in utero nutrition (either through dietary or placental manipulation) or a postnatal obesogenic environment in comparison with control animals(Reference Gnanalingham, Mostyn and Symonds49, Reference Gardner, Tingey and Van Bon52, Reference Sharkey, Gardner and Fainberg54, Reference De Blasio, Gatford and Robinson59, Reference Duffield, Vuocolo and Tellam63). ↑, Increased; ↓, decreased;=, similar; NR, nutrient-restricted; ER, endoplasmic reticulum; IGF1, insulin-like growth factor 1.
Conclusion
In summary, suboptimal conditions in utero have the potential to alter adipose tissue development and impact on long-term function and health of the offspring. The evidence from large-animal studies can inform both animal and clinical science; for example, runt piglets that have been cross-fostered to obtain greater milk intake become fatter than runts who remain with their litter(Reference Powell and Aberle9), highlighting the problems of rapid postnatal growth in low-birth-weight offspring. The molecular differences in Meishan piglets that provide improved neonatal mortality may provide an insight into potential strategies for improving survival in other breeds. Given that discrete periods of gestational nutrient restriction have been shown to have negative impacts on future adipose tissue function, maternal diet during pregnancy should be optimised to meet fetal requirements and avoid adverse effects in later life. Research into the emerging role of adipose tissue as a dynamic and endocrine organ is likely to yield further evidence for early programming of this tissue.
Acknowledgements
The authors declare no conflict of interest. Funding was received from the EU Sixth Framework Programme for Research and Technical Development of the European Community – The Early Nutrition Programming Project (FOOD-CT-2005-007036). A. M. developed the article outline, conducted the literature search, devised the Figures and wrote the article. M. E. S. assisted in the development of the body of the article and edited all drafts.
