PerspectiveCell Biology

Eat Me or Die

See allHide authors and affiliations

Science  28 Nov 2003:
Vol. 302, Issue 5650, pp. 1516-1517
DOI: 10.1126/science.1092533

Apoptosis is a physiological program of cell suicide that directs engulfment and safe destruction of cell corpses by healthy neighboring cells or professional phagocytic scavengers such as macrophages (1). Cells dying by apoptosis provide molecular instructions for their own funeral, sometimes releasing “come hither” signals that summon scavenger cells (2). These phagocytes then recognize “eat me” flags, such as phosphatidylserine (PS)—a phospholipid normally limited to the inner leaflet of the plasma membrane bilayer—that is displayed prominently on the surface of dying cells. A receptor (PSR) on macrophages that recognizes PS (3) then orchestrates compliance with the last line of the cellular suicide note: “Don't get angry when you dispose of me.” This triggers suppressive pathways—such as the release of transforming growth factor-β1 (TGF-β1)—that prevent phagocytes from mounting a proinflammatory response to the dying cells (1). Two new studies in this issue using mice (4) and worms (5) lacking PSR provide important insights into how PSR governs the clearance of apoptotic cells. Furthermore, one of these studies (4) reveals a sinister postscript to the funeral invitation—“Eat me or die.”

The PSR is one of a number of receptors on phagocytic cells that have been implicated in the uptake of dying cells. In addition, there are a variety of “bridging” molecules and “eat me” signals other than PS that are also important players in this process (6). The molecular structure of the PSR differs from that of other scavenger receptors recognizing “altered self,” hinting that this receptor has additional functions. These may include oxoglutarate-dependent dioxygenase activity that is reminiscent of the peptidyl hydroxylases involved in oxygen sensing (7, 8). Even though it is expressed by many cell types, the PSR has proved elusive principally because the few anti-PSR antibodies available are not ideal for tracking its subcellular distribution. Indeed, some features of PSR's structure suggest that it may be a nuclear protein rather than a transmembrane receptor (7, 8). Reassuringly, a “knockout and rescue” experiment in the developing worm by Wang and colleagues on page 1563 of this issue (5) provides strong evidence that PSR is indeed the phagocytic receptor responsible for engulfment of dying cells.

These investigators exploited a powerful and predictable model of embryonic cell deletion and clearance: that of the developing worm Caenorhabditis elegans. The worm has the added advantage of simplicity: Only a few proteins are involved in the uptake of dying cells, and most of those identified from mutant screens are intracellular molecules that induce cytoskeletal rearrangements in the cells engulfing corpses (9). The worm proteins CED-2, CED-5, CED-10, and CED-12 are structurally and functionally homologous to the mammalian signaling molecules CrkII, DOCK180, Rac guanosine triphosphatase, and ELMO, respectively. Worm embryos deficient in psr-1 (a PSR homolog) exhibit a clear defect in the clearance of cell corpses. This defect could be rescued by overexpression of ced-2, -5, -10, or -12 but not by overexpression of the ced-6 or -7 genes in a parallel pathway that is downstream of the scavenger receptor encoded by ced-1. Indeed, the clearance defect in the psr-1 mutant was rescued not only by overexpression of wild-type psr-1 but also by overexpression of human PSR. This finding confirms that PSR does participate in corpse engulfment, engaging a signaling pathway that is crucial for rearrangement of the cytoskeleton of phagocytic cells so that they can surround and “swallow” doomed corpses. It will be interesting to see whether PSR engages CrkII, DOCK180, and ELMO in mammalian phagocytes. The signaling pathway in which these proteins reside is also engaged by αv integrins expressed by phagocytes. These surface receptors have been implicated in phagocytosis of apoptotic cells and TGF-β1-mediated suppression of macrophage responses.

Worms suffer few ill effects if their cell corpses are not engulfed. In contrast, the nonengulfment of cell corpses profoundly upsets higher organisms with their sophisticated innate immune systems, which present autoantigens derived from noningested apoptotic cells to the adaptive immune system. For example, mice lacking functional C1q—a component of the complement cascade that forms a bridge between macrophages and apoptotic cells (10)—develop multisystem autoimmune disease in response to demonstrably defective clearance of dying cells.

In contrast to the worm subjects of Wang et al., Li and colleagues (4) in their study of PSR-deficient mice on page 1560 of this issue find a dramatically different phenotype during fetal development. These animals exhibit fatal neonatal respiratory failure associated with a reduction in the number of airways formed and accumulation of noningested dying cells, cellular debris, and recruited inflammatory leukocytes in the developing lungs (see the figure). Perhaps the PSR is particularly important for clearing apoptotic cells from the developing lungs. This would be consistent with the lack of evidence of pathology in most other PSR-deficient tissues (although the authors did not undertake detailed searches for diminished apoptotic cell clearance). Indeed, different clearance mechanisms may predominate depending on the tissue—for example, C1q-deficient mice exhibit impaired clearance of apoptotic cells from the peritoneum and kidney but not from the skin. Whatever mechanisms underlie the lung phenotype of PSR-deficient mice, the new data establish PSR at the epicenter of developmental and inflammatory lung diseases.

Neonatal lung failure in PSR-deficient mice.

During lung development in the mouse embryo, certain populations of epithelial and mesenchymal cells undergo apoptosis as part of the sculpting process that forms the airways. In mice lacking PSR, the corpses of apoptotic cells are not cleared efficiently and ultimately disintegrate. Such “secondary necrosis” releases cell contents that may directly or indirectly incite inflammation (via macrophage activation), resulting in the recruitment of neutrophils. In addition, binding of proteins called collectins by apoptotic cells may tether dying cells to the proinflammatory macrophage receptor CD91, further promoting the inflammatory cascade.


Unexpectedly, 15% of PSR-deficient mice also exhibit hyperplastic brain malformations due to overproliferation of brain cells. This aberrant phenotype is also seen in mice lacking the Apaf-1-caspase-9-caspase-3 apoptosome pathway that normally mediates apoptotic deletion of excess cells from the developing brain. However, unlike these animals, affected brains in PSR-deficient mice show increased numbers of apoptotic cells and the recruitment of macrophages. These observations imply that in wild-type mice binding of dying cells to macrophages may trigger macrophage-mediated killing of apoptotic cells as reported in human (11) and worm, a phenomenon termed “grab and stab” (9). Alternatively, the brain malformations seen in PSR-deficient animals might reflect loss of a nonphagocytic PSR function, such as oxygen sensing.

PSR is now established as a key player in the clearance of dying cells during development (4, 5). Animals engineered to have only certain tissues deficient in PSR (conditional knockouts) are needed to generate viable adult mice in which the functions of PSR during inflammation and immunity can be tested. PSR and the downstream signaling pathways that it shares with other suppressive phagocytic receptors, such as the αv integrins, may represent new therapeutic targets for treating common inflammatory diseases of the lung where leukocyte clearance by apoptosis has gone awry.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.

Navigate This Article