Egg laying progressively decreases with age, and its cessation is predictive of death of an individual female within approximately 5 days. Rauser et al., 2005 reported egg laying declined rapidly in the 5 days prior to death, with most, but not all, flies declining to zero eggs within 1–3 days prior to death (Rauser et al., 2005). Rogina et al. (2007) also reported rapid decline in the 5 days prior to death, with an average of 0.2 eggs at 2 days prior to death, and zero eggs at 1 day prior to death. Curtsinger compared data from two inbred laboratory strains, four outbred laboratory strains, and three recently collected wild-type stocks (Curtsinger, 2020). The average number of days of survival for individual flies after the last egg ranged from 2.4 days for one inbred strain, to 9.5 days for one wild-type stock, with an average over all strains of 4.7 days.

Green autofluorescence in the eye and body of male and female Drosophila increases several hours prior to death, and continues to increase after death (Tower et al., 2019). Similarly, when GFP expression is targeted to the retinal tissue of adult male and female flies, GFP fluorescence increases several hours prior to death, and continues to increase after death ((Tower et al., 2019); see video Clip 1 and Clip 2). In C. elegans and mammals, cell death has been associated with increased green autofluorescence resulting from altered subcellular localization and redox-sensitive changes in the fluorescence of flavins (Levitt et al., 2006; Pincus et al., 2016). Retinal cell death is therefore one candidate for the origin of the increased green autofluorescence observed prior to death of the fly. Indeed, progressive retinal degeneration and photoreceptor cell death during aging are particularly apparent in the w[1118] reference strain, which lacks normal eye pigments (Ferreiro et al., 2017). The mechanism for the increased fluorescence of GFP targeted to the retina is currently unclear, but conceivably could be due to alterations in subcellular localization associated with retinal cell death. The increased green autofluorescence and transgenic GFP fluorescence in the Drosophila retina has not yet been specifically correlated with any independent markers of cell physiology or cell death, and this may be an useful area for future research.

Consistent with the ability of Drosophila eye fluorescence changes to predict imminent death, the eye has been reported to be a diet-sensitive regulator of life span. When w[1118] strain flies were maintained on a low-yeast diet, strong visible light shortened life span, and this effect was absent on a high-yeast diet (Shen et al., 2019). In turn, constant darkness extended life span on a high-yeast diet, and these effects were absent on a low-yeast diet (Hodge et al., 2022). Because the high yeast protein diet decreases life span, one possibility is that the negative effects of strong visible light and high yeast protein act through the same pathway (Shen et al., 2019).

Notably, changes in the optical and structural properties of the eye have emerged as promising predictive biomarkers of physiological age, neurodegenerative disease progression, and risk of death in male and female humans. Drusen is an autofluorescent pigment containing proteins derived from blood and retinal epithelial cell debris (Bergen et al., 2019). Fluorescence imaging of the human eye reveals increased drusen abundance in the retina of AD patients (Ashok et al., 2020; Singh and Verma, 2020). Optical coherence tomography reveals thinning of the retinal nerve fiber layer and macular ganglion cell layer in patients with reduced cognitive function and AD (Kesler et al., 2011; Ward et al., 2020). Finally, machine learning models based on images of the fundus and retina provided a predictive biomarker for mortality risk (Ahadi et al., 2023; Zhu et al., 2023). In summary, increased autofluorescence in the retina has emerged as a predictive biomarker of death in Drosophila, and as a marker of the #7 AD cause of death in humans, consistent with the idea that neurodegeneration may be a cause of death in Drosophila as well as in humans.

One way that Drosophila activity rhythms can be assayed is by using activity monitors, where fly activity is measured by the frequency at which the movement of the fly disrupts an infrared light beam. Activity monitors revealed periods of increased arrhythmic behavior prior to death in Drosophila (Levine et al., 2002). Both male and female Drosophila exhibit reduced strength of sleep:wake cycles and more fragmented sleep with age, similar to changes observed in humans (Koh et al., 2006). Zhao et al. examined male flies containing transgenic luciferase reporters for the circadian regulatory genes timeless and period (Zhao et al., 2018). They found that whereas reporter expression declined with age, as expected, there was a surprising spike in expression that lasted approximately 3–4 days, and was centered approximately 3–4 days prior to death. These flies also became completely arrhythmic over a period of 3.7 ± 0.1 days before death, and therefore the spike in gene expression corresponded to a final loss of rhythmic movement. Null mutation of period caused reduced negative geotaxis behavior, neurodegeneration and shortened life span, consistent with a possible causative role for loss of rhythmic function in Drosophila death (Krishnan et al., 2009). Notably, disruptions in circadian rhythms are causally implicated in the #1 heart disease, #2 cancer, and #7 AD causes of death in humans (Chellappa et al., 2019; Olejniczak et al., 2023).

There have long been anecdotal reports in the Drosophila research community that flies exhibit periods of abnormal behavior prior to death, sometimes referred to as a “death dance”. Video analysis has proven to be a powerful method for assay of Drosophila behavior, aging and death (Zou et al., 2011). In an elegant study, Gaitanidis et al. performed longitudinal analysis of over 1,000 individual male and female flies using video recordings (Gaitanidis et al., 2019). They found two general trajectories for the deterioration of motor function with age. Some flies became progressively incapacitated over several days before death (termed “illderlies”), whereas others became incapacitated only several hours before death (termed “wellderlies”). Both trajectories were observed to converge on a terminal state with stereotypical signs of functional collapse followed by death within ∼3 h. Common features observed included decreased walking and climbing speed, leg joint defects (leg permanently extended or retracted; see Movie 2), climbing defects including falls to a supine position, and a terminal phase typically associated with permanent supine position (see Movie 3). Notably, many flies, in particular among the “wellderlies”, exhibited bouts of increased locomotor activity that lasted up to 2 h, and peaked about 5 h prior to death. These bouts involved increased walking behavior, increased climbing efforts often associated with falls, seizure-like events, and “paradoxical” behaviors including rearing up with front-leg boxing movements followed by a fall to supine position (see Movie 1). No significant sex differences in terminal behaviors were observed. Supine behavior is also predictive of impending death in male medflies (Ceratitis capitata), where temporary supine behavior began on average 16 days prior to death, and time spent in supine position increased exponentially until death (Papadopoulos et al., 2002).

Tower et al. used two synchronized video cameras to track male and female Drosophila movement through 3D space (Tower et al., 2019). Several flies were analyzed as they died from normal aging, and several young flies were placed in empty vials, and tracked as they died from dehydration/starvation. Tracking was conducted using either reflected visible light, reflected infrared light, or fly fluorescence using flies containing fluorescent transgenic reporter constructs. Old flies exhibited reduced total movement, negative geotaxis and centrophobism behaviors relative to young flies, as expected. The frequency of directional heading changes (FDHC) was calculated to measure erratic movement. Notably, the majority of the flies exhibited one or more bouts of erratic movement 0–8 h prior to death, as indicated by spikes in FDHC, and sometimes leading to falls to supine position (Tower et al., 2019).

Supine behavior is consistent with a deterioration in nervous system function. Supine behavior can be promoted in young Drosophila by mutations that disrupt brain function (Akitake et al., 2015; Sakakibara et al., 2018), as well as by traumatic brain injury (Katzenberger et al., 2013; Lee et al., 2019; Saikumar et al., 2020). Several mutations in mitochondrial genes and ion channel genes result in the bang-sensitive phenotype, where the impact of the fly against a surface causes seizure-like activity, paralysis and supine position (Fergestad et al., 2006). Notably, old wild-type flies often exhibit a bang-sensitive phenotype, and this might be induced by a fall. A fly in supine position cannot access water or food, and even in young flies a lack of water and food results in death within 24–48 h (Schriner et al., 2013; Parkash et al., 2014; Tower et al., 2019). As discussed above, supine position is commonly observed at death for Drosophila (see Figure 1D). Here, ∼200 flies were scored for position at death, and this analysis yielded 23% upright, 42% supine, and 35% on side.

In summary, erratic movement, falls and supine behavior are indicative of nervous system deterioration and are predictive of death in Drosophila, similar to the #4 unintentional injuries and #7 AD causes of death in humans. These observations are again consistent with the idea that neurodegeneration may be a cause of death in Drosophila as well as in humans.

Aging-associated deterioration in human IBI, sometimes referred to as “leaky gut”, is implicated in intestinal disease, chronic systemic inflammation, and neurodegeneration (Shemtov et al., 2022). Loss of IBI can also be detected in Drosophila, and is associated with increased risk of death. In 2011, Rera, Walker and coworkers introduced the “Smurf” assay, in which Drosophila flies are fed a high concentration (2.5% wt/vol) of the non-absorbable blue dye #1 (Rera et al., 2011). In healthy flies, the blue dye is limited to the lumen of the gut, which can be readily observed through the translucent cuticle. However, in a subset of flies, the blue dye is observed throughout all the tissues of the head (including eyes, antennal lobes and proboscis), thorax (including flight muscle and legs), and throughout the abdomen, indicating a loss of IBI and leakage of the dye into the circulatory system (for examples of Smurf, see Figure 1A; Figure 2A). The incidence of the Smurf phenotype was assayed in females, and found to increase with age (Rera et al., 2011), and a follow-up study reported that female Smurf flies had a median life span of 5.5 days or less (Rera et al., 2012). Clark et al. described criteria for scoring a “partial” Smurf, characterized by less intense blue staining of tissues, but still with detectable blue dye in tissues of both the head and the thorax (Clark et al., 2015).

Certain early studies suggested that every female Drosophila fly undergoes the transition to Smurf phenotype prior to death (Clark et al., 2015; Tricoire and Rera, 2015; Martins et al., 2018). However, more recent studies reveal that the majority of female and male flies die without ever undergoing Smurf, and that in some strains the Smurf phenotype is completely absent in males. Landis et al. assayed 284 individually-housed female flies continuously throughout their life span for Smurf phenotype, including partial Smurf, with every-other-day scoring of live flies, as well as scoring of dead flies (Landis et al., 2021a). Flies scored as Smurf when alive were dead at the next scoring in every case, and therefore the data indicated that Smurf phenotype robustly predicts death within 48 h or less. Notably, they found that only 40/284 female flies (14%) exhibited Smurf phenotype prior to or at death. In a large-scale and comprehensive study, Bitner et al. assayed 1982 individually-housed male and female flies continuously throughout their life span for Smurf phenotype, including partial Smurf, with every-day scoring of live and dead flies (Bitner et al., 2020). The majority of flies scored as Smurf were dead within 24 h or less. They found that only 22% of males and 34% of females exhibited Smurf phenotype prior to or at death. Therefore, whereas Rera and coworkers recently stated that their data “unequivocally show that every female Drosophila dies as a Smurf” (Zane et al., 2023), that conclusion is not consistent with the results from multiple other groups. Finally, Regan et al. reported that Smurf phenotype is completely absent throughout the life span of males of the wDah strain, and is present only at very low frequencies in males of the w[1118] strain (Regan et al., 2016). In summary, taken together, the data indicate that the Smurf phenotype is a robust biomarker of impending death, but only for a subset of flies.

Taken together, the studies discussed above suggest there are at least two mechanisms for death of Drosophila, and a possible third. The first mechanism is loss of IBI, as revealed by Smurf assay, which robustly predicts death within 0–∼48 h, but is only detected in a subset of flies. The second mechanism is nervous system malfunction, leading to locomotor malfunction, bouts of erratic behavior, and falls. The aged fly often cannot right itself after a fall to a side-ways or supine position, leading to inability to access the food and subsequent dehydration/starvation. The nervous system malfunction may be sufficient to cause death of the fly, but in the case of a fall to supine or side position, this likely synergizes with the lethal effects of dehydration/starvation. Consistent with the potential importance of nervous system malfunction, it is notable that the age-related increase in innate immune reporter expression is especially abundant in the brain (Figure 2). Finally, some flies die upright without Smurf phenotype (Figures 1B, C), suggesting a possible third mechanism. One possibility is that these flies die from nervous system malfunction without a fall. Alternatively, death might be due to failure of some other essential function or tissue, such as muscle as discussed above, or the Malphigian (renal) tubules (Lang et al., 2019), similar to the #10 kidney disease cause of death in humans. The relative frequency of these death mechanisms appears to vary between strains and culture conditions, which may affect the relative efficacy of various life span interventions.

In the future it may be helpful to further analyze the potential correlations between the different proposed death mechanisms. Death from Smurf and death from loss of nervous system function and falls might be caused by relatively independent mechanisms that proceed at different rates in different flies. Alternatively, there might be a common underlying mechanism, and what cause of death occurs first for a particular individual might be due to specific genetic vulnerabilities and/or be partly stochastic. The complete absence of Smurf in males of the wDah strain is consistent with significant sex differences in the relative frequencies of causes of death (Regan et al., 2016). It may be of interest to simultaneously assay Smurf and locomotor behaviors and falls to ask if Smurf occurs at similar frequency in male and female “illderlies” and “wellderlies”. Similarly, it may be of interest to quantify whether Smurf occurs at similar frequency in male and female flies found in upright, side-ways, or supine positions at death.

Whether aging results from a single underlying cause, or whether multiple (semi)-independent mechanisms are operating, such as the proposed “Hallmarks” of aging (Lopez-Otin et al., 2023), is an area of significant current interest and research. One appealing model for a single underlying cause that might explain multiple aging phenotypes and “Hallmarks” is the failure in mitochondrial maintenance with age (Tower, 2015; Ventura-Clapier et al., 2017). Mitochondrial maintenance failure is causally implicated in aging-associated proteostasis disruption (Moehle et al., 2019; Ruan et al., 2020). Moreover, mitochondrial malfunction with age is causally implicated in Drosophila nervous system malfunction and muscle malfunction, as well as in the induction of the hsp and innate immune response biomarkers of life span discussed above. Mitochondrial malfunction has also been reported to be involved in aging phenotypes in the Drosophila midgut (Boumard and Bardin, 2021). For example, disruption of mitochondrial function in the midgut can cause Smurf phenotype and shortened life span (Dai et al., 2020).

PGC1α is a conserved regulator of mitochondrial biogenesis, and conditional expression of PGC1α in adult Drosophila midgut using the Gene-Switch system was reported to reduce Smurf incidence and increase life span (Rera et al., 2011). However, another study on life span failed to reproduce the life span extension using the same strains (Landis et al., 2015). This does not rule out a central role for mitochondrial maintenance failure in midgut aging, but does indicate that transgenic upregulation of PGC1α in midgut is not a consistently effective life span intervention. Conceivably, differences in the efficacy of life span interventions, such as PGC1α over-expression in midgut, might be related to differences in the relative frequency of death mechanisms in different studies, as discussed above. Alternatively, or in addition, differences in results between certain studies may be related to the use of the Gene-Switch system. Using Gene-Switch to study aging phenotypes and life span in mated female Drosophila is complicated by the fact that the mifepristone drug used to trigger transgene overexpression can itself reduce midgut hypertrophy and extend life span (Landis et al., 2021b).

As discussed above, the markers and proposed mechanisms for death in Drosophila show many similarities to the leading causes of death in humans. Neurodegenerative disease and sarcopenia are both causative in the increased frequency of falls during aging (Yeung et al., 2019; Joza et al., 2020). In United States of America adults age 65 and older, 36 million falls occur each year, and 20% result in serious injury, including hip fracture and TBI (Moreland et al., 2020; Xu et al., 2022). Indeed, falls are the most common cause of TBI-related hospital admissions and deaths. Aging-associated deterioration in IBI, sometimes referred to as “leaky gut”, is implicated in intestinal disease, chronic systemic inflammation, and neurodegeneration (Shemtov et al., 2022). These similarities are consistent with the possibility that conserved mechanisms might underlie aging, and support the further use of Drosophila as a model for human aging and aging interventions.