Abrams et al. concluded that fly legs, which normally do not regenerate after amputation, show some regeneration ability when the fly’s diet is supplemented with insulin, leucine, and glutamine. We carefully followed their methods to replicate the fly leg regeneration experiments in Figure 3 of their paper. We amputated legs of 1283 flies, one hind leg per fly, at the midpoint of the tibia (Figure 1; Table 1). The majority of these flies, 1083, were of the same wild-type fly strain (Canton-S) used in their study. After amputation, we raised 240 flies on control food and 843 on treated food. Three weeks later, we examined the legs at high magnification using bright-field microscopy, with the experimenter blind to experimental condition.

We did not observe any regrown tibias, either in the control group or the treated group. The outcome of the two groups was qualitatively similar (Figure 1). All tibia stumps had bristle deterioration and darkened cuticle, indicating necrosis. The site of the darkened cuticle varied. It was usually near the cut site, but sometimes farther up the leg, closer to the tibia-femur joint. In 4% of control cases and 3% of treated cases, we observed a white blob near the cut site (Table 1; Figure 1, bottom row). The white blob was also observed by Abrams et al. to occur at a similar frequency and was called “white tissues protruding from the end”. Overall, based on close inspection of the cuticle three weeks after amputation, we found no evidence of leg regeneration.

Most of the internal structures of the fly leg, including muscle and neurons, are partially transparent. Inspection of amputated limbs using bright-field microscopy alone might be insufficient to detect surviving or regenerated tissue. We therefore used fluorescent labels to test for the presence of muscles, neurons, and other cells in the tibia before and after amputation.

Each fly leg has ~500 tactile bristles, including ~120 distributed uniformly along the tibia (Schubiger and Hadorn, 1968). Each bristle is innervated by a single sensory neuron. To ask whether bristle sensory neurons can regenerate after injury, we amputated 200 legs in a fly strain with a fluorescent reporter that labels bristles and other sensory neurons (ChAT-Gal4 >UAS GFP). We raised half of the flies on control food and half on food with the supplemented diet. We then used confocal imaging to image GFP expression in the leg.

Although bristle neurons were present immediately after amputation, they were absent three weeks later, presumably due to deterioration following cell death. The external bristle hairs also deteriorated and did not reappear. The results were indistinguishable between the control and treated groups (Figure 2A; Table 1). In the femur and other leg segments proximal to the tibia, the bristle sensory neurons and hairs appeared normal. In summary, we failed to find any evidence that leg sensory neurons regenerate following tibia amputation.

Each fly leg has twelve muscles, including four in the tibia (Soler et al., 2004). We next tested whether these tibia muscles can regenerate after amputation. We fluorescently labeled wild-type (Canton-S) leg muscles with phalloidin, a stain that labels F-actin. Three weeks after tibia amputation, neither the control group nor the treated group had any muscle staining in the tibia stump (Figure 2B; Table 1). In the femur and other leg segments proximal to the amputation site, muscle staining was normal. In summary, we failed to find any evidence that leg muscles regenerate following tibia amputation.

Since we did not find any sensory neurons or muscles in the amputated tibia stumps, we asked whether any other tissues, possibly hemocytes, glia, or epithelial cells, survive and/or regenerate. We stained legs of wild-type (Canton-S) flies with DAPI to label nuclei. Three weeks after amputation, we did not observe any DAPI staining in the cuticle of the tibia stumps, either in the control or treated groups (n=137 control, n=544 treated; Figure 2C). In the femur and other leg segments proximal to the tibia, DAPI staining looked normal. This pattern is consistent with previous work showing that cell death was constrained to the injured leg segment in the adult cockroach (Bodenstein, 1955). In summary, our evidence supports the conclusion that all cells in the tibia stump die after amputation and fail to regenerate. The amputated stump appears to be an empty tube of cuticle, devoid of living cells.

Abrams et al. reported the occasional appearance of a white blob at the tip of the amputated tibia stump in flies fed the supplemented diet. They called this “white tissue” and interpreted it to be an intermediate regeneration morphology. We observed the white blob form with a similar probability to that reported by Abrams et al. (3–4% of amputations, Table 1); however, we found that the blob occurred in both the control and experimental groups. Nonetheless, we sought to determine the nature of the white blob, and if it was, in fact, a sign of regeneration.

In many regeneration model systems, tissue regrowth is mediated by a blastema: a concentrated group of undifferentiated cells near the amputation site that proliferates to grow and repattern the missing tissue (Kiehle and Schubiger, 1985; Lehrberg and Gardiner, 2015). One possibility we considered is that the white blob at the cut site in amputated fly legs is a blastema. Alternatively, we thought that the blob could be fly tissue (e.g., muscle) extruded from the leg, a phenomenon which we sometimes observed immediately following amputation. To distinguish between these possibilities, we performed 5-ethynyl-2′-deoxyuridine (EdU) labeling on amputated legs (Figure 3A–B). EdU is a thymidine analog that is incorporated into the DNA of proliferating cells during S-phase. It has been used extensively to label cell proliferation in regeneration blastemas of many animals, including Drosophila imaginal discs (Lehrberg and Gardiner, 2015; Worley et al., 2022). We amputated wild-type (Canton-S) fly legs, as above, and fed flies EdU continually for three weeks with either control food or the supplemented diet. When we harvested the legs for EdU staining, we also saved the fly gut as a positive control. Gut is one of the few adult fly tissue types that exhibits homeostatic cell proliferation during the adult stage (Xiang et al., 2017). All fly guts (n=5) had robust EdU staining (Figure 3B). However, none of the tibia stumps had EdU staining (Table 1), including legs with the white blob (Figure 3A). We conclude that the amputated stump lacks regenerating cells and that the white blob is not a regeneration blastema.

We found that the white blob stained robustly with the nuclear label DAPI (Figure 3A, C and D). However, the nuclei in the white blobs were about one tenth the size of other cells in the fly leg, including leg hemocytes (Figure 3E), leg sensory neurons (Figure 3F), and leg muscle (Figure 3G). This discrepancy made us doubt that the white blob consisted of Drosophila cells. Abrams et al. also performed DAPI staining on this type of structure, but the image in their paper (Figure 4f) lacked a scale bar, making it difficult to determine the source of the nuclei.

In some animals that do regenerate amputated limbs, such as newts and axolotls, blastema cells in an amputated leg de-differentiate, undergoing morphological and transcriptional changes to become more like younger cells (Leigh et al., 2018; Tanaka et al., 2016). To address this possibility in the fly leg, we compared DAPI staining in the white blob to DAPI staining in leg precursor cells of the larval leg imaginal disc (Figure 3H). The nuclei in the white blob were again approximately one tenth the size of those of leg imaginal disc cells. This size difference is inconsistent with the idea that the white blob consists of de-differentiating fly tissue.

We compared DAPI staining of the white blob to other cells present in a Drosophila food vial: baker’s yeast and bacteria. The size of baker’s yeast nuclei was comparable to that of fly and other eukaryotic cells — significantly larger than the nuclei in the white blob (Figure 3I). However, the size and density of DAPI labeling of the white blob did resemble DAPI staining of bacteria from fly excrement (Figure 3J).

Finally, to further investigate the possibility that the white blob is a growth of bacteria, we stained the white blob with phalloidin (Figure 3D). If the cells were an outgrowth of fly tissue, we would expect them to have an f-actin-cytoskeleton that stains robustly with phalloidin (Figure 3H). If, on the other hand, the cells are bacteria, we would expect no phalloidin staining because the bacterial cytoskeleton is built from actin homologs, primarily MreB, that do not bind phalloidin (Shih and Rothfield, 2006; Dempwolff et al., 2011). Consistent with the second scenario, we observed that the cells in the white blob did not stain with phalloidin (n=12; Figure 3D). Furthermore, we speculate that the white blob did not incorporate EdU in the earlier experiment (Figure 3A–B) because the bacteria are not feeding on EdU-fly food, but on decomposing fly tissue in the leg stump that is non-proliferating and thus EdU-free. We conclude that the white blob is most likely a colony of bacteria and not an intermediate fly tissue related to regeneration morphology.

Drosophila were raised on a standard cornmeal-molasses-yeast food fly food at 25 °C with a 14 hr dark/10 hr light cycle. We used male and female adults, 1–2 days post-eclosion, reasoning that young flies would be more likely to regenerate and more likely to survive the three-week recovery period than old flies. For leg amputation, flies were anesthetized in groups of 20 on CO₂ plates for 5 minutes or less. One hind-leg per fly was amputated at the mid-point of the tibia with a fine double-edge super-stainless razor blade (ASR 72–003). Amputated flies were included in our analysis only if the amputation site was within ~50 μm of the tibia midpoint, using leg bristles as fiducial markers. Regeneration of cuticle was assessed according to whether the tibia length three weeks later fell outside of that range. After amputation, flies were immediately returned to a vial with either standard lab food or treated food, with random assignment.

To make treated food, vials of standard fly food were microwaved to liquefy the food. Before adding supplements, we let it cool to lukewarm to prevent the insulin from denaturing (Kaufmann et al., 2021). We added supplements in an aqueous stock solution and mixed the food for a final homogeneous concentration of 5 mM L-Leucine, 5 mM L-Glutamine, and 0.1 mg/ml insulin (Abrams et al., 2021). Food was mixed and allowed to set at room temperature for one hour. Flies were moved onto freshly prepared food every 2–3 days.

Legs or imaginal discs were fixed in 4% formaldehyde (PFA) PBS solution for 20 min followed by rinsing in PBS with 0.2% Triton X-100 (PBT) three times. To label muscle, legs were incubated in 1:50 phalloidin in a PBS solution with the following reagents to improve tissue penetrance: 1% triton X-100, 0.5% DMSO, 0.05 mg/ml Escin (Sigma-Aldrich, E1378), and 3% normal goat serum. Legs were allowed to incubate for one week at 4 °C with occasional rocking. After staining, legs were rinsed 3 x with PBS-Tx, 1 x with PBS, and mounted onto slides in Vectashield with or without DAPI.

Each slide was labeled according to experimental condition. Prior to analysis, we taped-over the labels. Categorizations in Table 1 were performed with the experimenter blinded to experimental condition.

Cells were transferred to a slide, diluted in water, flame-fixed, then mounted in Vectashield with DAPI.

We supplemented the food recipe above with 2 mg/mL EdU. Flies were moved to freshly prepared food every 2–3 days. After three weeks, legs and guts (positive control) were dissected and fixed in 4% formaldehyde (PFA) PBS solution for 20 min and processed according to Click-&-Go kit instructions. After staining, legs were rinsed 3 x with PBS-Tx, 1 x with PBS, and mounted in Vectashield with DAPI.

Mounted legs were imaged on a Confocal Olympus FV1000 (phalloidin, ChAT, EdU, and cuticle autofluorescence images) Leica DMI6000 Widefield (brightfield images), and Leica SP8X (DAPI images). Image stacks were processed in FIJI (Schindelin et al., 2012). Bright-field images were processed in Photoshop with the color channel mixer to correct a bluish background to truer white background.