|Home | About | Journals | Submit | Contact Us | Français|
The developmental origins of health and disease hypothesis holds that inappropriate environmental cues in utero, a period marked by tremendous developmental sensitivity, facilitate cellular reprogramming to ultimately predispose disease in adulthood. In this review, we analyze if stress during early stages of development can affect future health. This has wide clinical importance, given that 5 million children have been conceived with assisted reproductive technologies (ART). Because the primary outcome of assisted reproduction procedures is delivery at term of a live, healthy baby, the postnatal effects occurring outside of the neonatal period are often overlooked. To this end, the long-term outcome of ART is appropriately the most relevant concern of the field today. Evidence of adverse consequences is controversial. The majority of studies have concluded no obvious problems in IVF-conceived children, although a number of isolated cases of imprinted diseases, cancers, or malformations have been reported. Given that animal studies suggest alteration of metabolic pathways following preimplantation stress, it will be of great importance to follow-up ART individuals as they enter later stages of adult life.
Mammalian preimplantation development is a period of elegantly orchestrated molecular events, extending from fertilization of the oocyte to invasion of the uterine epithelium by the blastocyst [reviewed in (Cockburn and Rossant, 2010)]. These early stages are remarkably sensitive to environmental states, which stimulate signaling events and changes in gene regulatory networks (Fernandez-Gonzalez et al., 2004). As will be described in detail, gametes and embryos are exposed to a myriad of conditions throughout the fertilization and preimplantation periods. For example, the progression of the embryo from the fallopian tube to the uterus consists of successive changes in nutrient composition and growth factor availability, as well as serial decreases in pH from alkaline to mildly acidic, all of which have measurable influences on development and viability. Alterations to or disruptions within the circumstances of preimplantation development have profound implications for an organism’s health and predisposition to disease. Given that the number of individuals conceived by assisted reproductive technologies (ART), such as in vitro fertilization, has now reached 5 million (ICMART, 2012), the impact of unusual or suboptimal fertilization and preimplantation environments is particularly relevant. As will be described, many of the conditions associated with ART have the potential to be perceived as stressful.
This review evaluates our current understanding of preimplantation stress and its consequences for disease susceptibility. We analyze the physiologic events that occur before implantation in mammals, and describe how maternal disease or culture in vitro can have long-term health consequences. Finally we will focus on the possible different conditions in vitro that can modify metabolism and growth trajectories.
Importantly, we define stress broadly, as anything that jeopardizes homeostasis. In addition to conditions that threaten macromolecular damage or the integrity of various cell functions, stress may also occur as inappropriate physiological signals. These forms of stress do not necessarily evoke a stress response, but have a measurable effect on cell viability and/or alter typical development on the order of growth, gene expression, and overall competence. Stress therefore encompasses stimuli that perturb normal physiological functions with the potential to induce adverse pathological outcomes.
Preimplantation development encompasses fertilization through the invasion of the hatched blastocyst into the uterus (Fig. 1). Soon after the oocyte is fertilized by a spermatozoan to form a single-celled zygote, the cellular constituents are replicated and partitioned over several rounds of cleavage without increasing whole embryo volume (Cockburn and Rossant, 2010). At eight cells, embryo polarization allows compaction, providing a foundation for establishing distinct cell lineages: the trophectoderm, precursor to extraembryonic tissues like the placenta, and the inner cell mass, which will give rise to the embryo proper as well as part of its extraembryonic membrane. These lineages evolve over subsequent asymmetric cleavage divisions, continuing through the formation of the blastocoel cavity. Cavitation permits full expansion of the blastocyst, which hatches from the protective zona pellucida before implanting into the uterus. The first attachment of the human blastocyst to the uterus (apposition stage) occurs on day 8 postfertilization (Dey et al., 2004).
Analysis of metabolic physiology in preimplantation embryos indicates diverse nutritional requirements during development from zygote to blastocyst. Progression from the zygote to two-cell stage has an absolute requirement for pyruvate, and further cleavage stages are synergistically supported by lactate (Brinster, 1965; Biggers et al., 1967). Glucose cannot exclusively sustain development until the late four-cell/early eight-cell stage, but becomes the principal carbohydrate metabolized beginning around compaction (Brinster and Thomson, 1966). These requirements are relatively conserved across different mammalian species, with slight variations in time of increase in glucose oxidation. Effectively, aerobic and anaerobic respiration are not mutually exclusive but together function to meet the changing energy demands of early embryos (Wilding et al., 2009). This transition in energy substrate preference is influenced by the availability of carbohydrates, amino acids, and extracellular pH, which affect glycolytic activity, oxidative phosphorylation, and membrane transport, with potential consequences for proliferation and growth (Lane and Gardner, 2000, 2003). For example, the absense of glucose in culture prevents the characteristic decline of pyruvate uptake after compaction (Brinster, 1965). Consequently, preimplantation development occurs over a dynamic range of conditions, demonstrating strong metabolic plasticity and the ability of embryos to compensate for nutrient fluctuations at this time (Gardner and Leese, 1988; Edwards et al., 1998; Horsthemke and Ludwig, 2005).
Mammalian preimplantation development is additionally characterized by coordinated reprogramming of the genome and establishment of epigenetic marks that are maintained after birth (Bermejo-Alvarez et al., 2011). Epigenetic regulation occurs at the DNA level via methylation of cytosine bases residing in CpG dinucleotides (Ooi et al., 2009), or by modification of histone proteins through methylation, acetylation, ubiquitination, sumoylation, or phosphorylation (Zhang and Reinberg, 2001). These covalent moieties combine into cooperative epigenetic signatures that affect gene expression through a variety of mechanisms, including (but not limited to) control of gene or promoter accessibility by relaxation or condensation of chromatin, the recruitment of chromatin remodeling enzymes, and occlusion of transcriptional machinery (Fischer et al., 2008; Wang et al., 2008; Burdge and Lillycrop, 2010; Khorasanizadeh, 2011). Shortly after fertilization, the complementary parental genomes undergo dramatic changes in DNA methylation status via the TET family of dioxygenases, allowing for subsequent re-methylation to establish somatic differentiation patterns after implantation (Reik, 2007). Because preimplantation development is typified by epigenetic rearrangement, the embryo may be particularly vulnerable to disturbances. Environmental perturbations could cause errors or epimutations, or affect the programming of cell states and response pathways.
Emerging research over the past several decades has supported a clear biological basis for health and disease. More specifically, there is now appreciable evidence that exposure to stressful environments during critical periods of development can increase an organism’s susceptibility to a variety of diseases later in life. The administration of thalidomide to alleviate morning sickness in the late 1950s and early 1960s—resulting in an epidemic of phocomelia and other related birth defects—was paramount to demonstrating that events in utero could affect developmental outcomes (Brent and Holmes, 1988; McBride, 1961). Although previous researchers had suggested that developmental experiences could inform postnatal health outcomes (Kermack et al., 1934; McCance and Widdowson, 1974; Wadsworth et al., 1985), Barker et al. crystallized this theory in the 1980s by correlating nutritional stress in utero, manifested by low birth weights, with heightened risk of adult cardiovascular disease in particular regions of England and Wales (Barker and Osmond, 1986). Their fetal origins of adult disease hypothesis posited that fetal undernutrition could reprogram gene pathways to permanently affect body composition and metabolism. Today, the Developmental Origins of Health and Disease (DOHaD) hypothesis states that different environmental conditions occurring at critical points during development have the potential to shape developmental trajectories.
One of the most frequently cited examples in support of the DOHaD hypothesis is the Dutch Hunger Winter study, and deserves recognition here. In 1944, during the German occupation of specific regions of the Netherlands, rations were reduced to as few as 400 to 800 calories per day for a period of 5 months (Schulz, 2010). As with Barker’s study, the results are undeniable: nutrient deprivation in utero is associated with glucose intolerance, obesity, and cardiac dysfunction in adulthood (de Rooij et al., 2006a, b, c, d, 2010).
It is now apparent that fertilization and preimplantation development are additional periods of vulnerability to stressful environments. This becomes particularly relevant for individuals conceived by ART, given that many facets of fertilization and culture in vitro may be perceived as stressful (Fig. 2). Unfortunately, there is little data available for long-term epidemiological analysis due to the relatively brief history of ART. IVF children are at most 34 years old and too young to exhibit signs of chronic middle-aged syndromes, such that the impact of ART on adult health and age-related diseases remains obscure. Although the majority of children are apparently healthy, the increased prevalence of birth defects, epigenetic disorders, and metabolic discrepancies indicates a potential danger associated with assisted reproductive technologies.
With respect to human development, it is challenging to differentiate a plausible stress limited to the preimplantation period that does not persist for the remainder of pregnancy. However, there is evidence from animal studies suggesting that stress limited to this period can have both short and long-term consequences. In particular, low protein diet administered exclusively during the preimplantation period in rats is associated with cardiovascular pathologies, as well as perturbations to renin-angiotensin homeostasis in adulthood (Watkins et al., 2008, 2010). Specifically, the authors of this study observed significantly elevated systolic blood pressure, impaired arterial vasodilation, and increased expression of angiotensin-converting enzyme, an important mediator of vasoconstriction, in the lung. Within the oviduct and uterus, maternal diet affects nutrient availability and hormone composition (Leese et al., 2008). It is therefore possible to speculate that because embryo utilization of energy substrates is based primarily upon nutrient accessibility, altered concentrations of key metabolites and growth signals during the preimplantation period will affect embryo development and competence. Given that the worldwide prevalence of obesity and diabetes are at all-time highs, the effects of unhealthy maternal nutrition during preimplantation development are of serious concern.
An excellent study by Moley et al. evaluated the impact of high glucose during early gestation in mice (Wyman et al., 2008). Either zygotes or blastocysts obtained from streptozotocin-induced diabetic or control mothers were transferred into nondiabetic surrogate dams. Fetuses derived from zygotes and blastocysts of the diabetic mice displayed significantly higher neural tube closure problems and abdominal wall or limb deformities at midgestation. In addition, both groups of fetuses were significantly growth retarded, demonstrating a significant effect of maternal environment as early as the one-cell stage.
In addition to dietary or metabolic conditions, toxicologic agents can impact early pregnancy. Developmental exposure to xenoestrogens, such as diethylstilbestrol or bisphenol A, have been documented activating estrogen-responsive genes, affecting proliferation, and increasing adult cancer susceptibility (Nadal et al., 2000; Markey et al., 2001). This could result in reprogramming of hormone-response pathways, disruption of normal endocrine function, and altered establishment of hypothalamic-pituitary homeostasis (Maffini et al., 2006). Exposure of mouse embryos to environmentally relevant quantities of bisphenol A advanced blastocyst development and resulted in higher weight at weaning (Takai et al., 2001), whereas incubation with higher doses impaired embryo development (Takai et al., 2000).
Several interesting studies investigating the effects of tobacco on preimplantation development have shown that smoking before pregnancy can affect early embryonic growth. Tobacco exposure is correlated with ovulation of fewer oocytes, altered tubal function, (Knoll et al., 1995; Magers et al., 1995; DiCarlantonio and Talbot, 1999; Gieseke and Talbot, 2005), and impaired embryo development in female mice (Hassa et al., 2007). In rats, smoke reduces both preimplantation embryo development and sperm count (Polyzos et al., 2009). Similarly, maternal exposure to tobacco smoke in the immediate postconception period is associated with a dose-related retardation in embryo growth, increased pregnancy wastage (Seller and Bnait, 1995; Bnait and Seller, 1995), and delayed migration of embryos from the fallopian tubes into the uterus (Tachi and Aoyama, 1989). Yoshinaga et al. (1979) obtained similar results after daily maternal nicotine injections, observing a 12-hr delay in cleavage at the two-cell stage, and subsequent retardations at each step thereafter, including implantation.
Human in vitro fertilization (IVF) as a treatment for infertility is regarded as one of the most outstanding accomplishments of the 20th century, leading to the awarding of a Nobel Prize to its visionary Sir Robert Edwards in 2010. However, IVF has been viewed with some skepticism since it deviates rather severely from natural conception. The dynamic conditions existent in the genital tract are not only lost with IVF, but the procedure introduces many additional stressors. It is therefore not entirely surprising that numerous health complications are present following manipulation of embryos in vitro (Table 1).
However, it is important to emphasize the limitations of ART epidemiological studies: because ART patients represent a rather exclusive population, reliable data is often confounded by variables such as increased maternal age, infertility, or lack of appropriate control groups. In fact, subfertility itself is a risk factor for several health complications, including cardiovascular disease, depression, and certain reproductive cancers (Brinton et al., 2005; Wilkins et al., 2010; Parikh et al., 2012). It is therefore difficult to parse an effect of preimplantation stress attributable to ART procedures from outcomes secondary to the infertility status of the patients themselves. Nevertheless, important conclusions about embryo stress, adaptation, and reprogramming may be drawn from human ART analyses.
The major risk associated with ART in humans is the problem of multiple gestations, which are linked with numerous health complications. In 2010, the American Society of Reproductive Technologies (SART) reported that the risk of a woman under the age of 35 bearing twins after ART was 32.9% (Center for Disease Control and Prevention ASfRM, 2012). This relative risk is increased 25- to 30fold in contrast to twinning in spontaneous conception, which occurs in 1/80 pregnancies (1.25%) (Jaddoe and Witteman, 2006). Monozygotic twinning is twice as common in ART pregnancies, and the incidence of higher order multiples (triplets or more) is 2 orders of magnitude greater (1.5%) than in natural pregnancy (1/6400 or 0.01%) (Vitthala et al., 2009).
There is an overall increased risk of preterm delivery (PTD; defined as <37 wks gestation), low birth weight (LBW < 2500 g), and very low birth weight (VLBW < 1500 g) for ART pregnancies, even in singleton gestations (Helmerhorst et al., 2004; Jackson et al., 2004; Ombelet et al., 2005). Both PTD and LBW are correlated with several unfavorable neonatal outcomes, including neuromotor, behavioral, and cognitive dysfunctions, necrotizing enterocolitis, and cerebral palsy (Saigal et al., 1991; Lems et al., 1993; Holman et al., 1997; O’Shea et al., 1998; Buck et al., 2000; Saigal, 2000; Hille et al., 2001; Jackson et al., 2004). Compared with infants weighing 2500 to 2999 g and 3000 to 3499 g at birth, low birth weight alone increases the risk of neonatal death by 4- and 10-fold, respectively (Ashworth, 1998). ART is also associated with placenta abnormalities, including placenta previa, placental abruption, and abnormal cord insertion (Jackson et al., 2004; Shevell et al., 2005; Romundstad et al., 2006), which may influence fetal nutrition and growth, gestational age, and size at birth. These conditions are exacerbated by twin or higher order multiple gestations (McDonald et al., 2010). Additional obstetric risks correlated with ART include pregnancy-induced hypertension, and labor complications such as neonatal asphyxia and increased delivery by Cesarean section, although the latter may reflect a more cautious management of ART pregnancies (Gillet et al., 2011).
It has become increasingly clear that different forms of stress correlate with unique results. One interesting assessment—albeit with limited evidence—explored whether certain infertility etiologies are associated with particular outcomes. In a Finnish study, the risk of PTD was highest in women diagnosed with anovulation or endometriosis, whereas admission into neonatal intensive care was higher in pregnancies associated with male factor infertility (Kuivasaari-Pirinen et al., 2012). More complete analyses are required, but may inform the use of different ART procedures or pregnancy management options within ART populations.
Several groups have reported an increased incidence of epigenetic diseases in association with ART, implying that preimplantation stress can alter the epigenome. This includes the imprinting disorders Beckwith-Wiedemann Syndrome (BWS) and Angelman Syndrome (AS), which are associated with impaired development, growth defects, cognitive dysfunction, and cancer (Williams et al., 1995; Maher et al., 2003). Imprinted genes are genes in which a particular allele is inactivated in a parent-of-origin-dependent manner, and imprinting disorders arise when a maternal or paternal allele is inappropriately expressed because of abnormal DNA methylation [reviewed in (Manipalviratn et al., 2009; Batcheller et al., 2011)]. ART has previously been linked with altered methylation at several imprinting control regions (Doherty et al., 2000; Tremblay et al., 2000; Mann et al., 2004; Li et al., 2005; Rivera et al., 2008; Giritharan et al., 2010, 2012; Bloise et al., 2012). Although imprinted genes encompass 0.1 to 0.5% of the genome, they play a pivotal role in early development by controlling processes such as nutrient consumption and the cell cycle (Fowden et al., 2006). Inappropriate regulation of imprinted genes—inherited, sporadic, or environmentally induced—has been associated by some with diseases of growth and development (Miozzo and Simoni, 2002), but not by others (Rancourt et al., 2012; Puumala et al., 2012). While rare, the increased observation of BWS and AS may reflect more widespread, as-of-yet unidentified epigenetic alterations occurring after preimplantation stress. For example, assisted conception has been connected to widespread methylation changes in several mammalian species, including humans (Shi and Haaf, 2002; Zaitseva et al. 2007; Katari et al., 2009; Deshmukh et al., 2011). Mouse embryos cultured in vitro exhibit global hypermethylation (Wright et al., 2011), which could influence higher-order organization of chromatin structure. Interestingly, in the same study, IVF embryos were overall hypomethylated at 1500 selected foci. In another report, extensive DNA methylation analyses in placenta and cord blood cells revealed no obvious changes in global methylation between IVF- and in vivo-conceived newborns, indicating that IVF individuals may not have a noticeable epigenetic fingerprint (Katari et al., 2009). Yet, there were locus-specific methylation and gene expression discrepancies, suggesting there may be epigenetic differences in children born after IVF. As many imprinted genes are involved in growth and development, stress-induced epigenetic aberrancies could have a significant effect on postnatal health (Miozzo and Simoni, 2002). Furthermore, epigenetic changes may also account for previously described transgenerational effects (Holemans et al., 1991; Harrison et al., 2009). These investigations have revealed that culture-derived mice can produce offspring exhibiting organomegaly of the brain, pituitary, and kidney, with lower body weights at weaning (Mahsoudi et al., 2007).
Because the primary objective of assisted reproduction procedures is delivery at term of a live, healthy baby, postnatal effects occurring outside of the neonatal period are often overlooked. To this end, the long-term outcome of ART is appropriately the most relevant concern of the field today. Evidence of adverse consequences is controversial. The majority of studies have concluded no obvious problems in IVF-conceived children, although a number of isolated cases of rare diseases, cancers, or malformations have been reported (Table 1; see these excellent reviews and meta-analyses: Finnstrom et al., 2011; Fortunato and Tosti, 2011; Kallen et al., 2005; Labouesse, 2011; Steel and Sutcliffe, 2009; Tosti et al., 2006; Wennerholm et al., 2009). However, the relatively short history of IVF has restricted long-term analyses, preventing a more complete understanding of middle-age health and susceptibility to age-related diseases following preimplantation stress.
Arguably the most compelling evaluation of growth and metabolic outcome of IVF treatments compared approximately 225 IVF-conceived children with 225 children conceived naturally by subfertile parents (Wagenaar, 2008; Ceelen et al., 2007, 2008a, b, 2009). The subfertility control is an ideal way to remove a potential infertility effect, hence it is more reasonable to conclude that observed differences were induced by IVF procedures. The authors discovered that IVF babies were significantly smaller regardless of gestational age, and more likely to be born prematurely. Comparisons of weight, height, and BMI between the two cohorts during the first four years of life demonstrated that the growth velocity of IVF infants was considerably faster, such that their physiological characteristics were statistically indistinguishable from controls by age 3 (Ceelen et al., 2009). Of note, rapid weight gain in early postnatal life is an independent risk factor for metabolic syndrome (Khuc et al., 2012), adult hypertension (Eriksson et al., 2000, 2007), and coronary heart disease (Eriksson et al., 1999; Forsen et al., 1999). In the same study, children (average age 12.3 yrs) born through IVF had significantly higher systolic and diastolic blood pressure, peripheral fat deposition, and fasting glucose levels, with trending increases in bone mineral density and total fat mass (Ceelen et al., 2007; Ceelen et al., 2008a). IVF did not appear to affect the onset of puberty, although pubertal girls had significantly increased circulating levels of LH and DHEAS (Ceelen et al., 2008b). These data indicate that ART might alter long-term metabolic homeostasis, and it will be of great importance to continue follow up analyses of IVF individuals as they enter later stages of adult life.
The use of animal models to investigate the impact of preimplantation stress on subsequent development and health has been vital to our understanding of the mechanisms underlying the embryo stress response (Table 2) and potential consequences for adult fitness (Table 3). Animal models are valuable in corroborating human ART findings because they remove infertility as a potential confounding variable, permit investigation of a particular controlled stress, and extend analyses with available molecular tools (Jansson and Lambert, 1999; Walters and Edwards, 2003; Neitzke et al., 2008). However, studies in animals are not immune to bias and errors (Jansson and Lambert 1999; Walters and Edwards, 2003; Neitzke et al., 2008). An important caveat is that the conditions required for successful fertilization and preimplantation development can vary between different mammalian species, such that normal physiological signals for one species may be perceived as stressful by another. As a result, animal models may imprecisely or incompletely depict the needs of the human embryo (He et al., 2010).
It is well described that culture in vitro delays embryo development by 18 to 24 hr (Bowman and McLaren, 1970; Harlow and Quinn, 1982). Our laboratory has explored the impact of various forms of pre-implantation stress in mice, and has reported that factors, such as culture media composition, oxygen tension, method of fertilization, and culture dish rigidity will impact the percentage of embryos developing to blastocyst, ICM and TE lineage ratio and cell number, embryo morphology, and gene expression (Rinaudo and Schultz, 2004; Rinaudo et al., 2006; Giritharan et al., 2007, 2010, 2012; Kolahi et al., 2012). A recent study by Schwarzer et al. (2012) compared 13 ART culture protocols for mouse embryo development and observed distinct effects on developmental competence, lineage composition, and global gene expression with different conditions (Schwarzer et al., 2012). Each of these disparities could have a significant impact on fetal development after implantation, and thus influence organismal growth and health. In fact, there is evidence that placentation may be compromised following in vitro stress. TE cell number is reduced in blastocysts after IVF, ICSI, or various forms of culture (Rinaudo et al., 2006; Giritharan et al., 2007, 2010), with downregulated expression of genes involved in solute transport and placentation, as well as altered expression of several imprinted genes (Giritharan et al., 2012). This is particularly relevant since placental growth, efficiency, and nutrient transport capacity are frequently regulated by imprinted genes (Angiolini et al., 2006).
Mouse embryos fertilized and cultured under either optimal or substandard conditions in vitro exhibit similar implantation rates to in vivo embryos or to blastocysts flushed from oviducts and transferred into pseudo pregnant mothers (Delle et al., 2010). This suggests that despite the severity and type of stress, as well as variation in TE cell number, initial invasion of the uterine epithelium is a stable process. However, in these same embryos, abortion rates were significantly increased in the IVF groups and highest after suboptimal culture, possibly reflecting defective placentation or an inability to satisfy fetal needs. Impaired embryo viability following IVF or embryo culture has been confirmed by other groups (Holm et al., 1996; Khosla et al., 2001; Fernandez-Gonzalez et al., 2004; Block et al., 2010; Block et al., 2011; Bermejo-Alvarez et al., 2012).
Our laboratory has shown that IVF fetal and placental growth trajectories are significantly different throughout mouse fetal development. IVF placentae originate from fewer TE cells, are of normal weight by midgestation, and are significantly larger toward the end of pregnancy. Comparably, IVF fetuses are smaller throughout the majority of gestation, but display rapid catch-up growth to controls during the second half of gestation, possibly due to enlarged placentae (Bloise et al., 2012). At birth, weights are statistically indistinguishable between IVF and control animals, demonstrating that birth weight may be a weak indicator of fetal growth and nutrition.
The first record that preimplantation-specific stress can produce clear definable postnatal phenotypes came from Ecker et al. in 2004, who observed decreased anxiety and impaired spatial memory in adult mice after in vitro culture from the two-cell to blastocyst stage, without a significant effect of culture or embryo transfer on development to term (Ecker et al., 2004). Similar behavioral changes have been observed in mice transferred at the two-cell stage following ICSI (Fernandez-Gonzalez et al., 2008).
Animal models have additionally demonstrated that preimplantation conditions may affect long-term growth and metabolic health. Several groups have described increased body weights in mice after IVF, ICSI, or preimplantation culture (Fernandez-Gonzalez et al., 2004, 2008; Scott et al., 2010). Importantly, these differences may be present throughout postnatal growth, or appear later in adulthood, demonstrating latent effects of stress on physiology. Forms of in vitro stress reportedly affect organ size and integrity: Fernandez-Gonzalez et al. (2004) observed cases of pneumonia, kidney inflammation, and testicular atrophy in adult mice after embryo culture, as well as liver steatosis, possibly reflecting defective glycogen metabolism and resulting lipid mobilization as a compensatory energy source. Mice derived by IVF or ICSI exhibit signs of glucose intolerance and insulin resistance (Scott et al., 2010), and our group has noted similar detriment to long-term glucose homeostasis after IVF (Rinaudo et al., 2009). Embryo culture stress is also associated with increased systolic blood pressure in 21-wk mice, and alters the expression of several metabolic and cardiovascular genes (Watkins et al., 2007). These findings not only substantiate similar observations in humans, but suggest that metabolic reprogramming after assisted conception may not be an infertility effect or an outcome exclusive to ART populations.
Studies of embryo development under controlled conditions have identified a number of different sources of stress influencing preimplantation development. Because the environment in vitro can be exactly manipulated, we have included a review of specific conditions that can play an important role in affecting embryo growth.
Since the introduction of intracytoplasmic sperm injection (ICSI) as a method of fertilization in 1992, ICSI has become the most extensively used fertilization technique, accounting for approximately two-thirds of all fertilization treatments worldwide (Palermo et al., 1993; Ferraretti et al., 2012). Briefly, a micromanipulation device is used to pierce the zona pellucida of a denuded oocyte and inject a single spermatozoon into the ooplasm. ICSI has been confronted with certain suspicion, however, as the circumstances of the procedure itself employ many stressful techniques. Both sets of gametes must be handled to assess maturity, and in oocytes this is accomplished after enzymatic and mechanical removal of the protective surrounding cumulous and corona cells (Kimura and Yanagimachi, 1995). Moreover, direct penetration of the zona pellucida is physically destructive. Breach of the intra/extracellular barrier may result in cytoplasmic leakage or introduce culture media into the oocyte, thereby contaminating the cytoplasm. Animal studies show that ICSI-generated zygotes cleave at a slower rate, have a reduced hatching rate and reduced cell number in both trophectoderm and inner cell mass (Giritharan et al., 2010). ICSI zygotes also exhibit shorter calcium oscillations with an altered pattern (Kurokawa and Fissore, 2003). Blastocysts resulting from ICSI fertilization show approximately 1000 genes differentially expressed, compared with in vivo blastocysts (Giritharan et al., 2010). Importantly, as described earlier, ICSI-generated mice using fresh spermatozoa develop long-term consequences, such as obesity and organomegaly (Fernandez-Gonzalez et al., 2008).
The energy demands of the mammalian preimplantation embryo are perhaps the most vital aspects to consider when examining the various facets of preimplantation stress. Embryos exhibit diverse nutritional requirements, and the oviduct functions to fulfill those needs by providing energy substrates suitable for optimal development and metabolism [reviewed in (Leese et al., 2008; Leese, 2012)]. Tubal fluid is composed of nutrients, macromolecules, and electrolytes, the concentrations of which are remarkably conserved throughout different mammalian species (Gardner and Leese, 1990). Oviductal secretions are receptive to embryos and their evolving needs, providing different hormones, growth factors, and antioxidant defenses as embryos mature from zygote to blastocyst (Aviles et al., 2010). Surprisingly, the concentrations of many components found in culture media differ markedly from their physiological quantities (Gardner and Leese, 1990). This is important because the metabolism of preimplantation embryos relies heavily on feedback mechanisms (Summers and Biggers, 2003). For example, the concentration of lactate in culture media influences the degree of pyruvate metabolism (Lane and Gardner, 2000). In vitro, the build up of metabolic derivatives, including carbohydrate substrates or toxic byproducts, may contaminate culture media and affect energy usage by embryos (Lane and Gardner, 2003; Gardner and Lane, 2003). Although the use of sequential media is combative against this, in vitro culture is a closed system compared with the metabolic waste dissipation capacities provided by the oviduct (which to an embryo are essentially infinite) (Gardner and Lane, 1998). Variations in nutrient availability alter developmental competence (Edwards et al., 1998a, b), implying that culture in vitro is a source of great stress. Yet embryos can adapt to poor conditions, although embryo recovery rates after nutritional disturbances correlate with the severity of stress (Zhao and Baltz, 1996).
Independently, embryos culture is conducted under paraffin oil. Oil prevents evaporation of the medium to preserve a constant osmolarity, as well as minimizes fluctuations of pH and temperature (Tae et al., 2006). Oil also, however, can be toxic to gametes and embryos (Van Langendonckt et al., 1996).
pH is a powerful modulator of metabolic activity, and manipulations of intracellular pH are associated with changes to cellular proliferation, transcriptional activity, protein localization and synthesis, as well as glycolytic function [reviewed in (Shrode et al., 1997; Webb et al., 2011)]. In the mammalian reproductive tract, gametes and cleavage stage embryos are exposed to markedly alkaline conditions in the oviduct (pH ≥7.5), before implanting into a more acidic uterine environment (pH 6.96) (Zhao and Baltz, 1996). The stratified reproductive environment is suggestive of stage-specific pH requirements during fertilization and preimplantation development, a theory validated by a recent study demonstrating enhanced development of cultured embryos allowed to progress through cleavage stages at higher pH before transition into a more acidic pH after compaction (Hentemann et al., 2011).
Fluctuations in pH are common in reproductive biology. pH oscillations are related to menstrual periodicity: parallel increases and decreases of pH in oviductal and uterine fluid (respectively) at the time of ovulation signify a role for pH regulation in optimizing fertilization and implantation events (Pommerenke and Breckenridge, 1952; Macdonald and Lumley, 1970; Maas et al., 1976). In humans, intercourse increases the pH of uterine fluid by 0.2 to 0.95 units for approximately 30 min, presumably to enhance capacitation and protect sperm from harsh acidic vaginal secretions, cervical mucus, and the uterine environment (Fox et al., 1973).
Importantly, the evolving pH environment from oviduct to uterus is lost with in vitro fertilization and culture systems. Culture media are routinely provided at pH 7.2 to 7.4, and most embryo culture protocols utilize a bicarbonate and CO2 buffering system. Incubation chambers maintaining CO2 concentrations of 5 to 7% under pressure alter the pH of culture media. In addition, the pH of culture conditions is unstable. Accumulation of pH-altering byproducts of embryo metabolism, such as lactate or ammonium, will pollute culture media over time (Edwards et al., 1998b; Lane and Gardner, 2003). Changes to pH (or lack thereof) could perturb metabolic homeostasis, stimulate or prevent induction of specific molecular pathways, and facilitate potentially irreversible metabolic changes.
Several groups have reported that extracellular pH manipulation during preimplantation development can initiate metabolic changes. For example, acid and alkaline loads acutely alter embryo intracellular pH, and are correlated with adjusted glycolytic activity to restore homeostasis (Gardner and Leese, 1988; Edwards et al., 1998a, b; Lane and Gardner, 2000, 2003). This implies the induction of an adaptive response to ensure optimal physiological activity. Separately, PFK1, a rate-limiting glycolytic enzyme, is released from allosteric regulation and increases its activity >100-fold between pH 7.0 and 7.5 (Bock and Frieden 1976a, b; Frieden et al., 1976). In embryo culture, the loss of normal pH shift from greater than 7.6 to below 7.0 during the progression from zygote to blastocyst would undoubtedly have an effect on metabolic activity, possibly through a PFK1-related mechanism. We have observed a 1.63-fold increase in PFK1 transcription in suboptimally cultured blastocysts after IVF, which due to the sensitivity of the enzyme, may have an enormous impact on glycolytic activity (Giritharan et al., 2007). Additionally, preimplantation stress modifies the expression of several genes encoding membrane-bound transporters that buffer pH within a narrow range (Zhao et al., 1995; Phillips et al., 2000), leading to changes in intracellular pH and affected transcriptional activity (Putney et al., 2004; Chen et al., 2008).
A cell’s location within its physical surroundings is tremendously important for different cellular behaviors. Mechanotransduction is the process by which cells interpret mechanical information obtained from the environment, providing a means of transducing biophysical stimuli into an appropriate biochemical response [reviewed in (Ingber, 2006)]. Variations in tensile force have been linked to several developmental processes, including proliferative capacity, morphogenesis, and feedback mechanisms of growth control (Ingber, 2006; Labouesse, 2011). Not only can the relative elasticity of an extracellular environment affect cell differentiation (Wang et al., 2001), but mechanical force is a key factor in the ability of tumor cells to invade and metastasize (Butcher et al., 2009). This depends in part upon the distribution of intercellular junctions, as well as extracellular matrix composition (Levental et al., 2009). Still, the relevance of tension and stiffness during pre-implantation development has been largely overlooked. Embryo blastomeres, oviducts, and uterine epithelia are relatively soft, with elasticities ranging from 100 to 1000 Pascals (Pa; the unit of pressure or tensile strength incorporating forces of stress and strain, used to describe force per unit area) (Kolahi et al., 2012). A polystyerene petri dish, routinely used to culture preimplantation embryos, is roughly six orders of magnitude stiffer at 1 GPa (Discher et al., 2005). As a result, development in an in vitro culture system, in conjunction with the physical manipulation of embryos or gametes, may be a source of mechanical stress influencing cell fate specification and developmental potential. Our laboratory has confirmed this hypothesis in a recent study demonstrating that fertilization and blastulation rates, as well as TE cell numbers, are higher for embryos developing on a collagen matrix versus a standard polystyrene petri dish (Kolahi et al., 2012).
Even small fluctuations in temperature are relevant to gamete and embryo viability. Sperm development occurs 1 to 2°C below core body temperature, and is impaired by heat stress. Ovarian follicles also maintain a temperature approximately 1.5 to 2.5°C lower than deep body temperature [reviewed in (Leese et al., 2008)], which has implications for in vitro maturation technology. Temperature gradients have been described along the Fallopian tube (David et al., 1971; Hunter and Nichol, 1986), with correlation to menstrual periodicity (Bahat et al., 2005). In human oocytes, microtubule depolymerization initiates after temperature decreases as small as 5°C, and is irreversible after a 10°C decrease to 27°C (Almeida and Bolton, 1995). Ten minutes at 33°C causes absolute microtubule disassembly, leaving spindles permanently disorganized (Wang et al., 2001). Unfortunately, exposure to room temperature conditions is unavoidable during in vitro culture of embryos. While this is minimized by culture under a protective coat of paraffin oil, longer intervals outside of 37°C incubators will result in a temperature change in culture.
Thermal forces as a source of stress have become increasingly significant to preimplantation development after the introduction of cryopreservation techniques as a method of storing and preserving gametes and embryos. Human IVF laboratories conduct embryo freezing from 37°C to −196°C in two fashions, either by a controlled, slow-rate cooling, or by ultra-rapid (<1 s) vitrification; warming occurs over 1 to 2 min at room temperature. Freezing and thawing procedures can be damaging to intracellular structures and membrane integrity, to which delicate oocytes and embryos are particularly sensitive (Mandelbaum et al., 2004). Cryoinjury occurs in many forms, including ice crystal formation, structural damage to water-bound enzymes, separation of membrane proteins from lipids, altered membrane permeability resulting from phase transitions of phospholipids, and osmotic stress due to changes in cell volume (Carrell and Peterson, 2010). Indeed, cryodamage frequently results in embryo loss, although there are reports of healthy pregnancy in cases with up to 50% cell injury after freezing (Hartshorne et al., 1990). Freezing may additionally induce hardening of the zona pellucida (Matson et al., 1997), necessitating assisted blastocyst hatching (Tucker et al., 1991). Cryopreservation has been associated with zona cracking and resulting cytoplasmic leakage (Fabbri et al., 2001), microtubule disarray leading to meiotic spindle disruption, and chromosome dispersal (Boiso et al., 2002), as well as other cytoskeletal rearrangements and cytoplasmic disorder (Saunders and Parks, 1999).
The role of oxygen concentration in early development has become a subject of great interest. Culture of early embryos has historically been conducted similarly to somatic cells, under atmospheric conditions (~20% oxygen) (Gomes Sobrinho et al., 2011). However, several studies of oxygen concentration in the oviduct and uterus have demonstrated that oxygen tension fluctuates within 2–8% across different mammalian species (Mastroianni and Jones 1965; Mitchell and Yochim, 1968; Maas et al., 1976; Fischer and Bavister, 1993; Kaufman and Mitchell, 1994; Bavister, 2004). Despite the variation, this indicates that the female reproductive tract is absolutely capable of supporting aerobic metabolism in gametes and embryos (Leese, 1988). The generation of harmful reactive oxygen species (ROS) such as superoxide, hydrogen peroxide, and hydroxyl radicals is one major feature of aerobic respiration, although these may also originate from embryo surroundings (Guerin et al., 2001). Moreover, rises in oxygen concentration during culture are directly related to the increased production of oxygen free radicals in embryos (Goto et al., 1993). The accumulation of ROS threatens cellular viability, as they are damaging to all major cellular macromolecules, membrane and organelle integrity, as well as transcription. In fact, oxidative stress is considered to be a principal cause of the two-cell developmental block (Johnson and Nasr-Esfahani, 1994; Favetta et al., 2007). Although embryos are equipped with several endogenous means of countering oxidative stress, many antioxidant genes are not expressed until later stages of preimplantation development (Harvey et al., 1995). Production and accumulation of ROS from exposure to atmospheric oxygen concentrations during in vitro culture may therefore overwhelm embryo defense mechanisms.
In contrast, ROS at low concentrations function as critical physiological signals in many important cellular pathways, such as signal transduction, apoptosis, and transcriptional activity (Hancock et al., 2001). As a result, high oxygen tensions may induce unconventional downstream pathways. Oxidative stress can modulate gene expression, cause aberrant intracellular signaling, or affect the redox ratios of several key metabolites, and these changes are associated with supplementary rippling effects throughout metabolism (Guerin et al., 2001).
The embryo response to preimplantation stress, manifested by changes in cell number, lineage ratio, growth velocity, or gene expression patterns, highlights its remarkable plasticity and capacity for survival. Analysis of the literature reveals that many different sources of stress during the preimplantation period will have widespread effects on growth and metabolism. Stress can divert resources away from networks involved in coordinating optimal development, down alternative metabolic branches. The quiet embryo hypothesis postulates that stressed embryos will require more resources and energy in an attempt to alleviate molecular damage, leading to greater metabolic activity (Leese et al., 2008). An increase in mitochondrial oxidative phosphorylation to fuel cell repair machinery would be associated with an increase in free oxygen radicals, leading to aberrant signaling or cell damage. If this occurs during epigenetic reorganization, cellular defense strategies against transient stress exposure may be inappropriately programmed as homeostasis, leading to atypical responses to normal physiological signals, exhaustion of cell machinery, and long-term phenotypes. Alternatively, subtle metabolite fluctuations or atypical nutrient availability might not redirect resources towards damage responses, but rather be sensed as a ‘normal’ environment, eliciting an ‘appropriate’ programming reaction. However, these reprogramming events might be strategic for early survival, but inadequate for optimal postnatal life. This proves that long-term follow-up studies on the health of ART children are imperative, because although the great majority of children conceived by ART appear healthy, the long-term health impacts of these procedures are unknown.
Grant support: NICHD grant R01 HD062803-02 to PR.
Sky Feuer, Department of Obstetrics, Gynecology and Reproductive Sciences, Division of Reproductive Endocrinology and Infertility, University of California, San Francisco, California 94115.
Paolo Rinaudo, Eli and Edythe Broad Center for Regeneration Medicine and Stem Cell Research, University of California, San Francisco, California 94143.