Autophagy's Expanding Role in Development: Implantation Is Next

Autophagy is an intracellular process distinct from endocytotic lysosomal protein catabolism that sequesters cytoplasmic components in autophagosomes for subsequent degradation when these double-membrane vesicles fuse with lysosomes (1). Consumed products are returned to cellular metabolism to generate energy and provide building blocks for new macromolecules and organelles (2). The field has advanced significantly with the identification of autophagy-related (Atg) genes that mediate the initiation and assembly of autophagosomes, as well as their fusion with lysosomes to form autolysosomes (1). Autophagy is activated as an adaptation to cellular stress, starvation, growth factor deprivation, and infection, providing pathways for removing and recycling damaged cellular elements and to consolidate the remaining molecular resources for survival.

Although much emphasis has been placed on understanding the physiological role of autophagy in homeostasis and survival, recent contributions to the literature suggest that it also comes into play during development. Autophagy is induced by starvation as yeast and cellular slime molds transition from a vegetative state to dormancy and is essential for spore differentiation, based on genetic ablation of Atg genes (3, 4). In worms, larval development and dauer formation rely on autophagy (5). Loss of a gene required for the protein conjugation step during autophagy in Drosophila results in larvae that are unable to induce heterophagy in fat body cells before pupariation, causing death during metamorphosis (6). These are generally circumstances where developmental events are triggered by a scarcity of nourishment, forcing the organism to metamorphose into a dormant form until conditions improve. Autophagy provides the means to convert nonessential resources into molecules that can be mobilized for energy and macromolecule synthesis during emergence from dormancy as well as a rapid mechanism for cellular remodeling in response to environmental and hormonal signals.

The programmed attrition of oocytes near the end of fetal development in mice, which results in a sudden 40% decrease in germ cells at birth, is surprisingly not associated with high levels of germ cell apoptosis. Instead, markers of autophagy are abundant in oocytes around birth and an inhibitor of autophagy, 3-methyl adenine, blocks the accumulation of lysosomes in serum-deprived ovarian cultures (7). Germ cell attrition is probably regulated by multiple mechanisms that are linked to parturition-induced starvation, where autophagy contributes to the programmed demise of many primordial follicles while maintaining an adequate number for reproductive function.

Murine Atg5 knockout embryos survive to birth but soon die when a wave of postnatal autophagy fails to materialize to maintain the amino acid pool during parturition-induced starvation (8). An early embryonic phenotype in Atg5 knockouts was initially missed because the ATG5 protein is stockpiled during oogenesis, eliminating the need for its transcription during preimplantation development. Oocyte-specific deletion of Atg5 to remove maternal stores of the protein produces oocytes that fail to develop past the eight-cell stage after fertilization (9), demonstrating the requirement for autophagy during preimplantation development. Indeed, these authors find a dramatic increase in autophagosomes immediately after fertilization, perhaps to down-regulate existing proteins and provide amino acids for subsequent development. Aggressive recycling through autophagy could be important at this stage of development because mammalian ova do not contain large stores of nutrients. As more Atg knockouts have been studied in mice, other prenatal and postnatal deficiencies in cell differentiation have come to light (10).

An article in this issue by Lee et al. (11) suggests a new developmental function for autophagy in mice that possibly influences blastocyst implantation. The authors use an experimental model (diapause) that delays implantation in vivo by withdrawal of estrogen just before the zona-free blastocyst attaches to the uterine epithelium. Diapause is characterized by a dormancy period in which blastocyst growth and development cease until estrogen is reintroduced and implantation proceeds forth (12, 13). Normally in mice, the ovaries produce a surge of estrogen on the morning of gestation d 4, leading to attachment of the blastocyst to the uterus by 2300 h that evening (14). Evidence of uterine responsiveness to the presence of a blastocyst is first apparent by 1600 h (15), about 6 h after the nidatory estrogen surge. Lee et al. (11) found that the lack of estrogen in ovariectomized mice was recognized as early as 1800 h based on molecular and morphological evidence of autophagy well beyond the basal response of embryos in nonovariectomized females. As diapause was extended over several days, LC3/ATG8 expression and LysoTracker Red staining increased, indicating an accumulation of autophagosomes and autolysosomes, respectively. Along with these changes, the length of dormancy was inversely correlated with developmental competence, assessed by transferring the embryos to surrogate dams and examining d-14 conceptuses.

Dormant embryos reactivated by estrogen injection go on to implant. Lee et al. (11) report that LC3 expression persisted in the trophectoderm of activated blastocysts, but was reduced dramatically in the metabolically quiescent inner cell mass (ICM). The ICM contains embryonic stem cells that are set aside until the beginning of fetal development, whereas trophectoderm cells provide a transporting epithelium and will soon invade the uterus. Notably, molecular and morphological observations indicated that multivesicular bodies began to appear in the trophectoderm of reactivated embryos (11). The multivesicular bodies, derived from autophagosomes, degrade their cargo (16), suggesting that blastocyst activation switches the trophectoderm from sequestering cytoplasmic components to actively recycling them into ATP and molecular building blocks to be used for differentiation and invasion during implantation.

The ability of blastocysts to engage autophagy during delay provides an important means of survival. Inhibition of autophagy by injection of 3-methyl adenine significantly increased terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling in dormant blastocysts (11), suggesting that the blastocyst juggles a delicate balance between autophagy and apoptosis. An autophagic response to the absence of nidatory estrogen spares embryos from apoptosis and provides a pathway for adaptation to an environment that is not yet compatible with implantation. Dysregulation of the signals that link these pathways could shift the balance away from autophagy and trigger apoptosis as a pathological outcome. Diapause has never been observed in humans, although it is a normal physiological response in other mammals (12). However, it is conceivable that suboptimal conditions in the uterine environment that defer blastocyst implantation in humans could induce autophagic activity, similarly to that described in delayed mouse embryos, with the potential to veer off toward apoptosis. Indeed, there is variability in the timing of implantation in humans and, while extended diapause in mice reduces developmental competency, late implantation drastically increases the probability that a woman will have a miscarriage (17).

It is now clear that autophagy could contribute significantly in developmental programming. The responsiveness of autophagy to environmental stress and hormones positions it to execute developmental decisions that involve cellular remodeling or changes in metabolism. This concept is illustrated in the findings of Lee et al. (11) where it was observed that trophectoderm and ICM use autophagy to very different degrees. With more exposure to external stimuli and stress, the trophectoderm may be primed for autophagy and use it to facilitate the transition to invasive differentiation during implantation. Autophagosomes were observed accumulating in the trophectoderm of normally implanting blastocysts at 2200 h on gestation d 4, indicating that autophagy could play an important role during on-time implantation. With the availability of an expanded set of molecular and genetic tools for probing autophagy, it seems likely that investigators will find more examples of autophagy driving developmental events.