Symbiosis is a key source of ecological and evolutionary diversification of eukaryotic organisms throughout the animal and plant kingdoms. In insects that are obligatorily dependent on intracellular bacterial symbionts, novel host cells, the bacteriocytes, have evolved for harboring beneficial microbial partners. These cells constitute a fascinating riddle in developmental and evolutionary cell biology, as the molecular mechanisms governing their homeostasis in response to host physiology remain largely unsolved. We recently demonstrated that in the pea aphid (Acyrthosiphon pisum) model system, bacteriocytes degenerate through a hitherto unknown process distinct from evolutionarily conserved pathways, including apoptosis- or autophagy-dependent cell deaths (1,2). By combining electron microscopy, immunohistochemistry, and molecular analyses, we showed that the initial event of bacteriocyte cell death is the cytoplasmic accumulation of non-autophagic vacuoles derived from the endoplasmic reticulum. This early process is followed by a sequence of cellular stress responses, spanning aphid adulthood, that include the formation of autophagosomes in intervacuolar spaces and Buchnera aphidicola endosymbiont degradation by the lysosomal system. Interestingly, this mechanism of bacteriocyte degeneration is distinct from the process previously described in other insects. In the recently established symbiotic relationship [dating to less than 0.03 million years (My)] between Sitophilus weevils and the Sodalis pierantonius bacterium, bacteriocytes are totally and rapidly eliminated in young adults by a combination of apoptosis and autophagy (3). Given that bacteriocytes evolved independently in many different insect orders, it remains to be seen whether the bacteriocyte cell death identified in the pea aphid is a conserved mechanisms across insects or is restricted to evolutionary ancient symbioses like aphids and other hemipteran species. In order to understand the mechanisms underlying aphid bacteriocyte cell death, we carefully annotated the apoptosis pathway of A. pisum and studied the expression of apoptosis- and apoptosis inhibitors-related genes. We found that, while the executive pathway of apoptosis is incomplete in the pea aphid, with homologs for only four out of the eight proteins present in Drosophila melanogaster, apoptosis inhibitors underwent a huge gene expansion, with more than 80 inhibitor of apoptosis protein (IAP) homologs annotated. Indeed, in D. melanogaster, only four IAPs have been identified. We were able to classify the A. pisum IAPs in five different classes, based on the presence of different combinations of BIR and RING domains. While only a few apoptotic genes are induced in senescent bacteriocytes, apoptosis inhibitors from three of these classes are concomitantly over-expressed with time, possibly preventing the apoptosis in these cells and thereby delaying their elimination. Phylogenetic analyses showed that this expansion of IAP-encoding genes is widespread in the aphid lineage, even in the ancient, non-symbiotic, phylloxera species that contains 31 IAPs encoding genes. IAP gene expansion has not been found in other insect genomes. Complementary phylogenetic analyses and functional studies are ongoing to understand the role of this gene expansion and of the different IAPs in aphid physiology. (1) Simonet et al (2016) Direct flow cytometry measurements reveal a fine-tuning of symbiotic cell dynamics according to the host developmental needs in aphid symbiosis. Sci Rep. doi: 10.1038/srep19967. (2) Simonet et al (2018) Bacteriocyte cell death in the aphid/Buchnera symbiotic system. PNAS doi: 10.1073/pnas.1720237115. (3) Vigneron et al (2014) Insects recycle endosymbionts when the benefit is over. Curr Biol. doi: 10.1016/j.cub.2014.07.065.