← Writing

Origins of Apoptosis: Mitigating Injury or Regulating Development?

evolutionary biologyapoptosiscell death

The nervous system deploys the same weapon to solve two problems: to prune neurons which are overproduced during development [Kerr et al., 1972], and to mitigate the effect of damaged neurons from injury, allowing them to die quietly instead of necrotically [Nagata, 2018; White et al., 2014]. Though these scenarios are vastly different, the brain employs the same core mechanism for both. Yet apoptosis is very costly metabolically — it is energy dependent, requires new protein synthesis, and demands extensive reshaping of the cell membrane so the cell can be engulfed by macrophages [Nagata, 2018; Lauber et al., 2003]. As a result, there is likely a strong selection pressure that drove the development of this mechanism. Whether apoptotic machinery was selected for development or for injury containment is an open question among researchers [Kaczanowski, 2020]. I will argue that apoptosis for injury likely evolved first, with developmental apoptosis representing a later co-option of that same machinery.

Apoptosis Mechanism

First, it is useful to clarify terms. Apoptosis, first defined by Kerr et al. [1972], is a form of programmed cell death that involves nucleus condensation, cell shrinkage, and membrane blebbing, where the resulting apoptotic bodies are phagocytosed [Kerr et al., 1972]. This is in contrast to necrosis. Regulated cell death (RCD) is a broader term for any kind of cell death that is controlled by genes; apoptosis is just one of several forms of RCD. Programmed cell death (PCD) refers to physiologically scheduled RCDs — parts of normal development and homeostasis. Apoptosis, therefore, is a specific program of cell death that can be triggered either by PCD or by injury.

Apoptosis is generally divided into two categories: intrinsic, activated in response to intracellular stressors like DNA damage, and extrinsic. The intrinsic pathway is controlled by proteins of the BCL-2 family [Cory and Adams, 2007; Tait and Green, 2010]. Pro-death proteins BAK and BAX are induced, while pro-survival proteins BCL-XL, BCL-W, and MCL-1 are neutralized [Cory and Adams, 2007]. This causes mitochondrial outer membrane permeabilization (MOMP), releasing cytochrome c into the cytoplasm [Tait and Green, 2010], which promotes formation of the apoptosome. The apoptosome recruits caspase-9 and promotes its activation [Tait and Green, 2010]. Caspase-9 is the initiator caspase that triggers a cascade activating effector caspases 3 and 7, where the intrinsic and extrinsic pathways converge [Tait and Green, 2010]. Caspase activation serves several purposes. First, activated caspases suppress mtDNA-triggered cGAS/STING signaling and thereby prevent IFN-β secretion, helping render mitochondrial apoptosis immunologically silent [White et al., 2014]. Second, caspase-3 cleaves ICAD, a protein that inhibits caspase-activated DNase (CAD), so CAD becomes active and performs DNA fragmentation. Third, caspase-3 cleaves and activates scramblase proteins like XKR8 to promote the exposure of phosphatidylserine on the cell surface, which signals phagocytes to engulf the corresponding cells [Nagata, 2018; Lauber et al., 2003].

The extrinsic pathway is initiated when ligands bind cell-surface death receptors, such as Fas/CD95 or DR4/TRAIL-R1 and DR5/TRAIL-R2 [Galluzzi and Kroemer, 2022; Seyrek et al., 2021]. Ligand-bound death receptors assemble the death-inducing signaling complex (DISC), containing Fas, the adaptor FADD, and caspase-8, which initiates extrinsic apoptosis [Seyrek et al., 2021; Sprick et al., 2002]. At the DISC, procaspase-8 is activated, leading to activation of caspase-8/-10 [Seyrek et al., 2021; Sprick et al., 2002], which triggers the same caspase cascade that activates caspase-3 and caspase-7, linking it to the intrinsic pathway [Galluzzi and Kroemer, 2022].

Injury-First Hypotheses

The first collection of reasons supporting the idea that apoptosis evolved for injury mitigation is molecular, with two major pieces of evidence. First, most of the pathways described above are, fundamentally, a damage sensor repurposed to achieve cell death. The cell monitors damage to DNA, mitochondria, or proteins through machinery that likely evolved before apoptosis itself. PARP likely originally evolved as a DNA repair enzyme; the ER unfolded protein response likely evolved to degrade misfolded proteins. Only at high stress levels does either trigger cell death. This suggests that the cell had already evolved some form of damage-response architecture before apoptosis existed, and that apoptosis is essentially a response triggered at the upper end of that damage spectrum — in mitochondria, only at high levels of BAK/BAX does the membrane become fully permeable to cytochrome c [Tait and Green, 2010]. Second, apoptosis is explicitly framed as a noninflammatory way to handle cell loss, as opposed to necrotic cell death [Nagata, 2018; White et al., 2014]. This is particularly relevant to injury, which otherwise causes inflammation through DAMP release, histamine, and cytokines. The way this problem is solved — through macrophage engulfment — closely resembles the body's standard way of dealing with injury, suggesting that apoptosis evolved to solve exactly this kind of problem [Nagata, 2018; Lauber et al., 2003]. If the root cause were developmental pruning, it is unclear why apoptotic cell death, which requires more energy than other forms of cell death, would be favored.

Second, recent research suggests apoptosis arose along the same timeline as mitochondrial domestication, around 1.8 billion years ago [Kaczanowski et al., 2023]. In that time frame, injury was a far more common phenomenon than developmental overcrowding of neurons — neurons only evolved roughly 600–800 million years ago. Additionally, apoptosis-inducing factors (AIFs) are shared among modern eukaryotes and share a bacterial/mitochondrial origin [Kroemer et al., 1998; Kaczanowski et al., 2023]. There is evidence in single-celled organisms that apoptosis is overwhelmingly triggered by stress and damage. Kaczanowski et al. [2023] found that apoptotic proteins from different eukaryotic kingdoms — plants, animals, bacteria — could all substitute for one another functionally in inducing apoptosis. The problem of overcrowded neurons during development simply does not exist in bacteria, from which apoptosis likely evolved. A host of papers document how injury and stress trigger apoptotic cell death in early organisms: oxidative stress induces apoptosis in Entamoeba histolytica [Picazarri et al., 2010], Schizosaccharomyces pombe [Alseth et al., 2009], and Giardia lamblia [Ghosh et al., 2012; Chowdhury et al., 2009]. These stress factors have existed for as long as cells have.

Third, when directly compared with theories that developmental abundance caused apoptosis, injury-based hypotheses are stronger on three fronts. First, on scale: every cell, in our bodies and in all other organisms, faces injuries like protein misfolding, oxidative stress, and infection, whereas developmental pruning is only relevant in neural cells, which evolved comparatively recently. Second, on time: developmental pruning is only relevant during an organism's development, while injury is a constant threat throughout life. Since evolution is shaped by probability accumulated over an enormous sample size — it is more likely that organism A survives over organism B given certain traits, rather than certain — the fact that injury affects far more of an organism's life makes it more likely to have shaped the machinery for apoptosis [Ratcliff et al., 2012]. Third, the consequences of failure — the negative selection pressure in evolution — are more severe for injury. Failure of apoptosis to mitigate injury causes cells to rupture necrotically and initiate a cascade of tissue destruction [Nagata, 2018]. Failures of developmental apoptosis produce smaller problems that other mechanisms can compensate for; if a nervous system ends up over-wired, for instance, an organism might cope through reduced energy demands elsewhere, such as in the gut, or through increased astrocyte–neuron metabolic cooperation.

Development-First Hypothesis

There are, however, strong arguments that apoptosis evolved to prune development. Many of the reasons above conflate any form of RCD with apoptosis specifically — evidence that injury triggers some form of RCD is not sufficient to show that apoptosis is the specific culprit. On the other hand, there are clear empirical observations that a large number of neurons die by apoptosis as a normal part of development. In rats, neuron counts decline substantially over the first week after birth in the cervical sympathetic ganglion as a result of apoptosis. Second, an important mechanism of apoptosis involves competition for survival-promoting trophic factors — the neurotrophic theory of neuronal death. Target tissues secrete a limited amount of neurotrophic factor, and the neurons that fail to capture enough die through apoptosis: in neurotrophic-factor-deprived neurons, BAX is upregulated and cytochrome c is released. The very existence of this mechanism is fundamental to building a nervous system, which is itself a strong selection pressure for developmental apoptosis.

Evolutionarily, apoptosis is also observed in the worm C. elegans [Conradt and Xue, 2005], following an incredibly similar molecular pathway: a series of CED proteins are the direct equivalents of BCL-XL, the BH3-only proteins, and APAF-1 in mammalian apoptosis [Conradt and Xue, 2005]. The developmental apoptosis pathway has therefore been conserved between mammals and nematodes across 600 million years without a shared common ancestor. In both C. elegans and mammals, the pathway for developmental apoptosis is tightly integrated [Conradt and Xue, 2005] — surprising if apoptosis had evolved primarily for injury and only later been co-opted for development. Ameisen has argued that the C. elegans evidence is nonetheless inconclusive: the machinery for self-destruction, likely ancient and shaped by host–pathogen conflict and injury, was simply co-opted once multicellular organisms emerged, to help prune development.

Third, if apoptosis had evolved as a defense against injury, we might expect organisms to use dedicated, separate systems for pruning excess neurons. Instead, the same regulators that execute apoptosis are the exact molecules that shape the nervous system. p53, essential for killing damaged cells as a tumor suppressor, is also essential for pruning excess neurons — its knockout inhibits developmental neuron death. Knocking out caspase-3 causes brain overgrowth and is lethal, suggesting caspase-3 is also crucial for normal development. There is also redundancy, where different caspases compensate for one another, suggesting sustained selection pressure over millions of years. Overall, the machinery is deeply integrated into the development of any nervous system.

Conclusion

How apoptosis evolved is necessarily ambiguous and hard to settle definitively. Still, I believe the evidence pushes toward the interpretation that injury prevention motivated the evolution of this mechanism. Injury is older and more common, and therefore provides a much stronger selection pressure to evolve the machinery in the first place [Kaczanowski et al., 2023; Kaczanowski, 2020], later "borrowed" by developing neurons that face a similar need for programmed cell death. The developmental selection pressure for apoptosis is clearly also ancient and deeply integrated [Conradt and Xue, 2005]. Nevertheless, the evidence suggests, to me, that apoptosis was originally used to handle injury.

References

  • Alseth, I., Tranbærg, K., Kvaløy, K., & Bjorås, M. (2009). Tdp1 protects against oxidative DNA damage in non-dividing fission yeast. The EMBO Journal, 28(21), 3302–3312.
  • Chowdhury, S. K., Ghosh, S., Das, S., & Ganguly, S. (2009). Oxidative stress-induced programmed cell death in Giardia lamblia. Drug Design, Development and Therapy, 3, 103–110.
  • Conradt, B., & Xue, D. (2005). Programmed cell death. WormBook, 1–13.
  • Cory, S., & Adams, J. M. (2007). The BCL-2 protein family: opposing activities that mediate cell death. Nature Reviews Molecular Cell Biology, 8(1), 47–59.
  • Galluzzi, L., & Kroemer, G. (2022). The concept of intrinsic versus extrinsic apoptosis. Biochemical Journal.
  • Ghosh, E., Ghosh, A., Ghosh, A. N., Nozaki, T., & Ganguly, S. (2012). Programmed cell death in Giardia. Parasitology, 139(7), 894–903.
  • Kaczanowski, S. (2020). Symbiotic origin of apoptosis. In Symbiosis: Cellular, Molecular, Medical and Evolutionary Aspects (Vol. 69, pp. 253–280). Springer.
  • Kaczanowski, S., et al. (2023). Apoptotic factors are evolutionarily conserved since mitochondrial domestication. Genome Biology and Evolution.
  • Kerr, J. F. R., Wyllie, A. H., & Currie, A. R. (1972). Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. British Journal of Cancer, 26(4), 239–257.
  • Kroemer, G., Dallaporta, B., & Resche-Rigon, M. (1998). Mitochondrial control of apoptosis. European Journal of Biochemistry, 60, 619–642.
  • Lauber, K., Bohn, E., Kröber, S. M., Xiao, Y.-J., Blumenthal, S. G., Lindemann, R. K., Marini, P., Wiedig, C., Zobywalski, A., Baksh, S., Xu, Y., Autenrieth, I. B., Schulze-Osthoff, K., Belka, C., Stuhler, G., & Wesselborg, S. (2003). Apoptotic cells induce migration of phagocytes via caspase-3-mediated release of a lipid attraction signal. Cell, 113(6), 717–730.
  • Nagata, S. (2018). Apoptosis and clearance of apoptotic cells. Annual Review of Immunology, 36, 489–517.
  • Picazarri, K., Nakada-Tsukui, K., Tsuboi, K., Miyamoto, E., Watanabe, N., Kawakami, E., & Nozaki, T. (2010). Hydrogen peroxide induces apoptosis-like death in Entamoeba histolytica trophozoites. Microbiology, 156(5), 1517–1529.
  • Ratcliff, W. C., Denison, R. F., Borrello, M., & Travisano, M. (2012). Experimental evolution of multicellularity. Proceedings of the National Academy of Sciences, 109(5), 1595–1600.
  • Seyrek, K., Richter, M., & Lavrik, I. N. (2021). Measuring CD95/Fas DISC formation. Journal of Visualized Experiments.
  • Sprick, M. R., Rieser, E., Stahl, H., Grosse-Wilde, A., Weigand, M. A., & Walczak, H. (2002). Caspase-10 is recruited to and activated at the native TRAIL and CD95 death-inducing signalling complexes in a FADD-dependent manner but cannot functionally substitute caspase-8. The EMBO Journal, 21(17), 4520–4530.
  • Tait, S. W. G., & Green, D. R. (2010). Mitochondria and cell death: outer membrane permeabilization and beyond. Nature Reviews Molecular Cell Biology, 11(9), 621–632.
  • White, M. J., McArthur, K., Metcalf, D., Lane, R. M., Cambier, J. C., Herold, M. J., van Delft, M. F., Bedoui, S., Lessene, G., Ritchie, M. E., Huang, D. C. S., & Kile, B. T. (2014). Apoptotic caspases suppress mtDNA-induced STING-mediated type I IFN production. Cell, 159(7), 1549–1562.