Macrophages gain a partner at the table: Epidermal cells digestperipheral dendritic debris in Drosophila

In this issue of Neuron, Han et al.(2014) develop powerful methods to visualize phagocytosis of Drosophila peripheral sensory neuron dendrites. Remarkably, epidermal cells rather than professional phagocytes are the primary mediators of debris clearance, using both familiar and new molecular players.

In the Drosophila larval peripheral nervous system there are some gorgeous sensory neurons that tile along the barrier epidermis. The development and arborization of these cells has been intensively studied by a number of groups (Jan and Jan, 2010). During metamorphosis (Williams and Truman, 2005) or following injury (Sugimura et al., 2003), their peripheral processes degenerate and are rapidly cleared prior to remodeling or regeneration. Previous studies showed that Drosophila plasmatocytes, a macrophage-like circulating blood cell, can be found at sites of injury or programmed remodeling and can contain dendritic cell material during the clearance process (Williams and Truman, 2005). This observation was consistent with the view that in organisms that possess them, circulating macrophage-like cells are the undisputed champions at clearing cell debris. The question that opens the Han study is a simple sufficiency experiment- are plasmatocytes necessary for dendritic clearance during metamorphosis or following injury? The answer to this question was a surprising no. Although plasmatocyte ablation slightly delayed dendrite fragmentation following injury, ultimately degenerating cells were cleared. This begged the question, if not plasmatocytes, then who?

“Non-professional” phagocytosis, a job for neighboring cells?

There are at least two prominent examples where macrophages are not the primary phagocyte. The first is in organisms like C. elegans, which lack circulating macrophage-like cells entirely. Such organisms must rely on other cell types to clear both apoptotic and injury-induced cell debris. Indeed, in C. elegans apoptotic cells of the larval ventral nerve cord are engulfed by local hypodermal cells during development and it has also been speculated that dendritic debris after axonal injury may also be cleared by these same cells (Chisholm and Xu, 2012). The second example is in the vertebrate central nervous system (CNS) where the blood brain barrier prevents macrophages from gaining access. Here, microglia derived from the same monocyte lineage as macrophages act as primary phagocytes, displaying macrophage-like filopodia that draw in and engulf cellular debris (Neumann et al., 2009). Additionally, in certain areas of the brain where microglia do not populate, such as in the external granule layer of the rat, Bergmann glia have been found to act as the primary phagocytes that engulf apoptotic cells (Parnaik et al., 2000). These studies bolster the view that neighboring cell types can act as non-professional phagocytes in cases where “professional” phagocytes either do not exist or cannot reach the site of clearance.

In fly larvae, the epidermal basal lamina (BL) acts as a physical barrier between plasmatocytes circulating in the open body cavity and the neuronal dendrites intimately associated with epidermal cell bodies. Some sessile plasmatocytes co-localize with the cell bodies of peripheral sensory neurons in the space between the muscle and epidermal layer (Makhijani et al., 2011). However, the spatial relation of these sessile cells to the epidermal BL is not entirely clear. During metamorphosis, the integrity of the BL as a barrier to plasmatocyte traffic is not known, but the Han ablation experiment rules them out as the primary phagocytic tissue. Although the central projections of the peripheral sensory neurons are wrapped by glia inlarvae (Stork et al., 2008), the dendritic processes are not. Thus, in searching for an alternative phagocyte, epidermal cells have the dual advantages of being abundant and in closest proximity to the degenerating dendrites. Han et al blocked epidermal phagocytosis through inhibition of dynamin, a component necessary for pseudopod extension and subsequent engulfment. Importantly, they found significant dendrite clearance defects in both damage-induced and pruning models when dynamin activity was lacking. These experiments add to the growing evidence that epidermal cells can play a significant role in phagocytosis of cellular debris (Monks et al., 2005).

Upon finding epidermal cells as the primary phagocytes for degenerating neurons, the authors developed a novel tool (membrane-associated pHsensor: MApHS) that allowed them to visualize and dissect the process of internalization and breakdown of dendritic debris through the various phagolysosomal compartments of epidermal cells. Their probe contains apH-quenchable extracellular GFP as well as an internal proteolysis-resistant RFP to create a membrane-associated pH sensor that can report on the journey of internalized debris through the endocytic pathway. This probe is likely to be useful in a number of other Drosophila contexts where endocytosis or phagocytosis is prominent. One example among many would be glial-mediated pruning of mushroom body projections during metamorphosis (Awasaki et al., 2006). The probe also has the potential to be adapted for use in other model organisms such as zebrafish, where live visualization of peripheral remodeling following injury is also a strength (Sugimura et al., 2003). Lastly, this tool allowed the authors to identify a number of candidate genes required for debris clearance and pinpoint the stage of the phagocytic process affected by loss of each gene in epidermal cells.

New insight into draper and croquemort and a new player: debris buster

With their new tool for phagocytic pathway visualization in hand, Han et al. began examining a few of the usual suspects, uncovering novel functions of known phagocytosis genes as well as discovering a new player, debris buster (Figure 1). The first of the logical candidates examined was Draper; a well-known engulfment receptor required in Drosophila larval glial cells for phagocytosis of cell corpses and pruned axons during metamorphosis and after axonal injury in the CNS (Awasaki et al., 2006; MacDonald et al., 2006). Both analysis of mutants and epidermally-expressed RNAi transgenes targeting Draper indicated a role in the initial engulfment of fragmented dendrites. These results expand the repertoire of cell types that employ Draper in such a manner.

Figure 1. Cartoon of phagocytosis of dendrite debris by a Drosophila epidermal cell.

An epidermal cell (blue) between the apical cuticle (purple) and the basal lamina (black) is segregated from circulating plasmatocytes (red) in the body cavity. Following injury or during metamorphosis, dendritic debris (green) is engulfed in a draper and dynamin dependent manner (1). The early phagosome(2) then matures into a late phagosome (3) and a phagolysosome (4) in a manner that depends upon the functions of croquemort, debris buster, rab5, rab7, and LAMP1. See Han et al for further details.

Croquemort (crq), a member of the scavenger receptor Cluster Differentiation 36 (CD36) family, isvital for the phagocytosis of apoptotic cells in the embryo (Franc et al., 1996). Interestingly, in this study crq was not necessary for the engulfment of neurite debris by the epidermis, but rather impacted maturation of the phagosome past the initial acidification step. Confidence in these results is enhanced by the authors’ generation of a new genetic null allele of croquemort. Further investigation of the morphology of the stalled phagosomes via transmission electron microscopy and time lapse imaging found crq mutant phagosomes to be larger, more homogenous, and more intensely labeled by GFP, strongly suggesting homotypic fusion between phagosomes. The implication of crq in regulating later stages of epidermal phagocytosis was somewhat surprising given its previous description as a cell-surface receptor for apoptotic cells (Franc et al., 1996), and suggests that there may be some differences in how the phagocytic machinery is deployed between macrophages and epithelial cells.

Lastly, the authors identified a new CD36 family member they dubbed debris buster (dsb). An RNAi transgene targeting dsb displayed a very similar phenotype to crq, disrupting late phase phagosome maturation. Although these studies reveal interesting insights into gene function in previously undocumented phagocytic cells in Drosophila, the real power of the system will be in future gene discovery approaches. The powerful new tool unveiled here, combined with the possibility of conditional tissue-specific RNAi in Drosophila, makes possible identification of novel conserved genes that are essential for various stages of phagocytosis in “non-professional” epidermal cells.

Beyond Drosophila- who is at the table in the vertebrate PNS?

The results from the Jan group beg the question: Who clears neuronal debris in cases of peripheral degeneration in vertebrates? A prior study suggested that macrophages are recruited to sites of degeneration after peripheral nerve injury in the mouse and may be a source of signals for the regenerating neurons (Brown et al., 1991). Whether macrophages are sufficient to clear neurite debris however, has not been tested. There are a number of possible cellular candidates for clearing degenerating peripheral neuronal debris, including professional macrophages, Schwann cells, and local non-professional cells surrounding the axonal process. In some vertebrates, the length of degenerating neurite that needs to be cleared can be impressive, which raises the question- can circulating macrophages released from the vasculature at the site of injury work their way down the entire extent of the projection to engulf debris? Or cana combination of monocytes from the site of injury and tissue-resident macrophages dispersed along the rest of the axonal projection do the job? Alternatively, some of these processes are enwrapped by Schwann cells, which may be better positioned to mediate clearance, but what about those processes not wrapped in glia? Finally, as was found in this study, could other local cells along the projection route serve this function- for instance the epidermal cells surrounding the sensory nerve endings? What is now needed is a combination of ablation and visualization experiments with newly developed tools to attempt to resolve these questions- questions given a new urgency by the study here. Encouragingly, one group has already started and the results are not encouraging for professional phagocytes. Warchol et al, recently found that cochlear tissue-resident macrophages are not required for debris clearance or regeneration following ototoxic injury in the chick (Warchol et al., 2012).