Introduction

Research in the field of biogerontology is traditionally focused on the later life stages. There are numerous lines of evidence, however, that support the idea that rate of age-related functional declines and life expectancy may be “programmed” by nutrition and other environmental factors through development (de Magalhães 2012; Walker 2011; Monaghan and Haussmann 2015; Vaiserman 2014, 2015; Vaiserman et al. 2018). Accumulating data indicate that organism is exceedingly sensitive to environmental cues during early developmental periods (critical windows), with outcomes impacting later stages of life cycle (Wells 2017). This type of plasticity is a highly valuable characteristic necessary to generate phenotypes suited to various environmental conditions on the basis of the same genotype (Bateson 2015; Projecto-Garcia et al. 2017). Early-life tuning of epigenetic profiles (alterations in gene expression that do not involve changes to the underlying DNA sequence) during critical developmental periods characterized by high rate of cell proliferation in developing tissues, is considered to play a crucial role in this process (Hochberg et al. 2011; Bianco-Miotto et al. 2017).

The fruit fly, Drosophila melanogaster, seems particularly suitable model for investigating the developmental programming phenomenon, even despite large differences in ontogeny of flies and humans. This organism undergoes significant changes, including metamorphosis, at the stage of development. These morphological changes are accompanied by active cell proliferation and differentiation and, obviously, global epigenetic reprogramming. The adult fruit fly, by contrast, is almost completely postmitotic organism (except for gonads and gut), thus allowing studying simultaneously aging cells without any influence of newly dividing cells. Therefore, it is considered to be an excellent model system in aging research (Rogina 2011). In addition, due to the above features of ontogeny of fruit fly, it might be a useful model in studying epigenetic mechanisms underlying developmental programming.

Different traits in adult flies were repeatedly shown to be dependent on conditions of larval development. Manipulations with dietary macronutrients in larval diet were shown to induce obesity and oxidative stress in adult flies (Lushchak et al. 2011; Rovenko et al. 2015a, b). Moreover, significant changes in flies can be induced with larval treatments with toxic compounds (Lozinsky et al. 2012, 2013; Perkhulyn et al. 2017).

According to the developmental theory of aging proposed by Lints (1978), aging is considered as a part of development following growth and differentiation. Thereby, it is assumed that the rate of pre-adult development and life span (LS) are causally correlated. Most experiments aimed at testing the developmental theory of aging were carried out in the 70 s–80 s using Drosophila as a model system. Currently, after a long break, this topic has returned to center stage. The interest in this theory has been stimulated by modern discussions, in particular, on the potential role of neoteny (the persistence of larval or fetal features in the adult form of an animal) in exceptional longevity of naked mole-rats and some other species (Skulachev et al. 2017). Being strongly affected by developmental conditions, fruit flies are extensively used to test the predictions of the developmental theory of aging. In these studies, the growth rate of the flies was manipulated by varying the larval rearing temperature or the amount of yeast provided to the developing larvae, e.g., in overcrowded conditions in which emerging flies tend to have decreased body size and increased longevity compared to those raised at a normal larval density (Economos and Lints 1984a; Shenoi et al. 2016). The first experimental evidence for this theory was obtained in the study by Lints and Lints (1969), where increasing of both pre-adult developmental time and adult longevity was recorded in flies reared at obviously unfavorable condition of high larval density. From these findings, authors concluded that developmental time is an important causing factor in determining the aging rate. A strong negative association between the growth rate and longevity in Drosophila was confirmed in subsequent research of the same authors (Lints and Lints 1971). Along with this research, the extended LS in flies raised in conditions of high larval density was observed in some other studies (Klepsatel et al. 2018; Lints and Lints 1971; Miller and Thomas 1958; Zwaan et al. 1991), though it seems counterintuitive to many researchers (Martin et al. 1996), and several studies, e.g., by Zwaan et al. (1991), have failed to confirm the link between the developmental time and longevity.

Among the plausible causal mechanisms underlying the longevity-promoting effects of larval crowding, there are increased expression of antioxidant enzymes (Martin et al. 1996) and heat-shock proteins (Buck et al. 1993; Sørensen and Loeschcke 2001), alterations in body size (Economos and Lints 1984a) and subcellular organelle numbers, size and/or functional properties (Economos and Lints 1985), modulating the composition of phospholipids and shaping cell membrane bilayers (Moghadam et al. 2015), as well as increased starvation resistance caused by increased relative fat content in adult flies (Zwaan et al. 1991). Collectively, these changes induced during the larval development can result in a long-term post-eclosion hardening/acclimation (Sørensen and Loeschcke 2001), thereby affecting the longevity. In recent years, the long-lasting epigenetic changes of gene expression, triggered by environmental cues during the pre-adult development, have been suggested to be key players in mediating these associations. Most evidence for the role of epigenetic pathways in shaping the developmental outcomes in later life has been obtained in rodent models (Langley-Evans 2015; Tarry-Adkins and Ozanne 2014) and in epidemiological studies (Vaiserman 2015). In Drosophila, such findings are rather scarce. The larval crowding-mediated life extension was found to be accompanied by an elevated level of expression of Hsp70 gene and an increased heat-stress resistance in adult flies (Sørensen and Loeschcke 2001, 2004). In the study by Vaiserman et al. (2013), the larval dietary restriction resulted in increased LS and enhanced expression of gene encoding insulin-like receptor (InR), known to be regulated by insulin signalling to peripheral tissues and associated with longevity, in adult male flies.

To study the possible effect on adult longevity of developmental programming, we used the eclosion order (i.e., day of emergence) as an index of developmental rate. The emergence of flies developed in non-crowded conditions from synchronously laid eggs lasts for about a day. Under the conditions of larval crowding, flies eclosed for a week or even more, even if they are genetically homogeneous and emerge from synchronously oviposited eggs, and their body weight is significantly reduced as compared to normal level. In such conditions, those insects emerging in the last days are much smaller (sometimes 2 times or even more) than those that emerge first. This can be likely caused by nutritional limitation and/or by accumulation of the toxic waste products such as urea and uric acid through high larval density (Joshi et al. 1996; Scheiring et al. 1984). It can be assumed that flies emerged first have particular competitive advantages that allow them avoid, at least partly, the influence of these, or other, unfavorable factors. In the present study, we examined LS, fly weight, fecundity and expression levels of longevity-associated genes in adult flies reared in normal or in crowded conditions by using the order (day) of emergence as an index of the developmental duration.

Materials and methods

Origin and maintenance of flies

A wild type Oregon-R strain of Drosophila melanogaster was used in the study. Throughout the whole study period, flies were reared and maintained at 25 °C, 70% relative humidity and a 12-h light/dark cycle on a standard cornmeal-sugar-yeast-agar medium. To avoid confounding effects related to genetic variation, flies from a single isofemale line, i.e. individuals with a relatively homogeneous genetic background, were used as a parental population. The parental flies were placed in one-liter glass jars containing fresh medium where they could lay eggs. In low-density (LD) jars, egg-laying was performed by 20 pairs of 10–12-days flies per jar for ~ 24 h until reaching sufficient egg numbers (~ 300–400 eggs per jar), and in high-density (HD) jars, egg-laying was performed by 200 pairs of flies per jar for ~ 24 h until reaching egg density more than 3000 eggs per jar. After oviposition, parental flies were removed from the jars.

Measurement of the egg-to-adult developmental time and body weight

In each group, the median egg-to-adult developmental time was measured from the midpoint of an egg-laying period to the midpoint of adults’ eclosion. The eclosion continued for about a day in LD jars, while in overcrowded HD jars it continued for 5 days (groups HD1-HD5, respectively). Virgin flies were collected within 6–8 h after eclosion. They were lightly etherized, separated according to sex, and six replicate sets of 25 flies per sex per group were weighed to the nearest 0.001 mg.

Longevity and fecundity testing

After the weighing, flies were placed in groups of 25 of the same sex to glass vials (14 cm length and 1.5 cm diameter), containing 3 ml of a culture medium. Both control and experimental vials were placed into the incubator with the temperature of 25 °C. The flies were transferred into vials containing fresh medium thrice a week. Dead flies were removed during every transferring, and their numbers were recorded until the last death. Three males and five females were lost during the transfers. About 145–150 flies per group were used for the mean LS testing in each group. Maximum LS was determined as the mean LS of the longest living 10% in each group. For evaluating the fecundity, flies (males and females mixed) were kept up to 10–14 days of age in one-liter glass jars (100–150 flies per jar per group). At the age of 10–14 days, inseminated females were individually placed in glass vials (14 cm length and 1.5 cm diameter), containing 3 ml of a culture medium. Fecundity was determined for 1 day by counting total number of eggs laid per female per day in 17–20 females per group.

mRNA isolation and cDNA synthesis

The mRNA levels of selected genes were measured in 10–14 days old flies. Five genes such as HSP70, dTOR, dFOXO, InR and dSir2 were selected. All these genes were previously shown to be substantially contributed to regulate longevity in fruit fly. A detailed description of these genes and their functions is provided in Table 1. Total RNA was isolated from homogenized flies (four samples per group; five flies per sample) with RNA extraction kit—RIBO-Sorb (InterLabService, Russia). RNA concentration was determined using a NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies Inc., USA). The A260/A280 ratio of the RNA samples was 1.8–2.0. Single-strand cDNA was synthesized using 1 μg of total RNA pretreated with REVERTA-L reagents kit (InterLabService, Russia) at 37 °C during 30 min on Veriti Thermal Cycler (Applied Biosystems, USA).

Table 1 List of genes selected for expression analysis
Full size table

Quantitative real time PCR

Real-time PCR was carried out on the 7500 Real-Time PCR System (Applied Biosystems, USA) using MasterMix with intercalating fluorescent dye SYBR Green and reference ROX dye (Ukrainian Genetic Technologies, Ukraine). Each reaction was run in triplicates three times in final volume of 20 μL PCR mix containing SybrGreen MasterMix, 500 nM of reverse and forward primers and 10–50 μg cDNA. The threshold cycle Ct was determined (7500 Software v2.0.5, Applied Biosystems, USA). The amplification efficiency values were calculated using ΔCt method. GapDH2, Hsp70, InR and dSir2 primers sequences were designed using PrimerExpress Software v3.0 (AppliedBiosystems, USA) and dFoxO, dTOR primers sequences were synthesized by Ukrainian Genetic Technologies, Ukraine, as described in Chattopadhyay et al. (2017). Primer sequences are listed in Table 2. GapDH2 was used as reference gene to calculate relative expression level.

Table 2 Primer sequences used for real-time PCR
Full size table

Statistical analysis

All the physiological parameters studied as well as gene expression data were analyzed by one- or two-way ANOVA followed by Dunn’s multiple comparison test to detect differences between groups. The longevity data were also analyzed using ANOVA, which is preferable for analyzing LS data, particularly when sample sizes are equal or nearly equal (Le Bourg 2014). All analyses were done using GraphPad Prism 7 (GraphPad Software, USA). Differences were considered to be significant at p < 0.01.

Results

The mean developmental times in both LD group and in flies emerged during the first day in HD jars (HD1 group) were roughly the same (~ 202 h). The developmental durations of about 226, 250, 274, and 298 h were observed in groups HD2, HD3, HD4 and HD5, respectively. The mean weights showed a significant trend to decrease with increasing the timing of emergence (Fig. 1). According to the two-way ANOVA, the treatment (LD, HD1–5) effect, as well as the sex and interaction effects were statistically significant [treatment: F (5, 60) = 384.9, p < 0.0001; sex: F (1, 60) = 1687, p < 0.0001; interaction: F (5, 60) = 15.4, p < 0.0001].

Fig. 1
figure 1

Mean weights of flies reared in conditions of low larval density (LD group) and high larval density (HD1-HD5 groups). Each mean weight was calculated for six replicate sets of 25 flies per set. Data are given as means and S.E.M, n = 6 for each group. Significant difference from LD group at p < 0.01 according to the Dunn’s multiple comparison test is marked with a; significant difference from HD1 group is marked with b

Full size image

Development at high larval density significantly affected life span of both sexes (Figs. 2 and 3).

Fig. 2
figure 2

Survival curves of male a and female b flies developed in conditions of low larval density (LD group) and high larval density (HD1–HD5 groups)

Full size image
Fig. 3
figure 3

Mean and maximum LSs of flies developed in conditions of low larval density (LD group) and high larval density (HD1–HD5 groups). The symbols of significant difference are the same as on Fig. 1

Full size image

Both mean and maximum LSs in HD1 and HD2 groups were significantly increased in males and females compared to LD group (Fig. 3). In males, mean LS was increased by 24.0% and 23.5% in HD1 and HD2 groups, respectively. In females, corresponding increments in mean LS were 23.8% (HD1 group) and 29.3% (HD2 group). Significant effects were observed for treatment: F (5, 1779) = 24.44, p < 0.0001 and interaction between treatment and sex: F (5, 1779) = 3.3, p = 0.006, while the effect of sex was of borderline significance: F (1, 1779) = 4.4, p = 0.036. Male maximum LS was found to be significantly increased in HD1-3 groups, while female maximum LS was increased in all HD groups compared to respective LD groups (Fig. 3). ANOVA results for maximum LS were: treatment: F (5, 168) = 31.25, p < 0.0001, sex: F (1, 168) = 6.65, p = 0.01 and interaction: F (5, 168) = 11.07, p < 0.0001.

The female reproductive activity, assessed by mean fecundity levels, tended to decrease with subsequent days of eclosion [one-way ANOVA: F (5, 134) = 6.34, p < 0.0001; see also Fig. 4 for details]. The mean amount of laid eggs was about 20 in HD1 flies. Flies of LD, HD2 and HD3 groups laid virtually the same amount of eggs (9–12) whether substantial reduction was observed for flies of HD4 and especially HD5 groups.

Fig. 4
figure 4

Fecundity rate (mean number of eggs per female per day) in females reared in conditions of low larval density (LD group) and high larval density (HD1–HD5 groups). The symbols of significant difference are the same as in Fig. 1

Full size image

In HD groups, the mRNA levels for genes InR, dSir2, dTOR and dFOXO were revealed to be significantly dependent on the treatment condition in males, whereas this impact on expression of gene Hsp70 was of borderline significance only (Table 3, see also Fig. 5 for details). Interestingly, no gene expression differences were observed in females.

Table 3 Results of one-way ANOVA for the effects of treatment condition (LD, HD1-5) on the expression levels of longevity-associated genes in males and females
Full size table
Fig. 5
figure 5

Gene expression levels of flies reared in conditions of low larval density (LD group) and high larval density (HD1-HD5 groups). All levels are depicted as means normalized to those in reference (LD) group and expressed as fold over LD group, which was set as 1. Genes: InR (a–b), Hsp70 (c–d), dSir2 (e–f), dTOR (g–h), dFOXO (i–j). Left panels: males; right panels: females. The symbols of significant difference are the same as in Fig. 1

Full size image

Discussion

The main finding of our study was the life extension in both male and female flies reared in conditions of high larval density and emerged throughout the first and second eclosion days (HD1 and HD2 groups). On the whole, in HD flies, extension of developmental period was associated with reduced body weight and female fecundity, and also with life shortening. One plausible explanation for our findings could be the selective pre-adult mortality of flies raised in overcrowded conditions. This, however, seems unlikely because not only mean but also maximum LSs were found to be increased in HD groups compared to LD group. Indeed, it is obvious that even if the weakest flies died due to adverse events related to the larval crowding, that would only affect mean but not maximum LS. Gender differences in association between developmental rate and maximum LS in HD flies, namely, negative association in males and no association in females were revealed in our study. These differences can likely be explained by decreased reproductive activity in females reared in conditions of HD and re-allocation of resources from reproductive processes to maintenance of the soma (Flatt 2011). Indeed, the female maximum LS was significantly higher in all HD groups than that in LD group, while the female reproductive activity (fecundity) was significantly decreased in HD5 group only compared to LD group. Thus, lengthened maximum LS of females reared in HD conditions could be likely explained by their decreased reproductive activity. Remarkably, the larval crowding-induced changes in the expression of longevity-associated genes studied were less pronounced in females than in males. A limitation of this study, however, was that female fecundity was determined only once at the age of 10–14 days posteclosion, while repeated measurements would have yielded more information regarding the impact of the reproductive activity on the fly survival.

The results obtained in our study are similar to those observed in a series of studies by Economos and Lints (1984b, 1985, 1986). By varying amount of yeast added on the medium, they revealed that LS depends on the developmental rate in a biphasic way. Across the entire growth rate range covered in this study, the relationship obtained had a parabolic form with a maximum at about 55–60 μg/day. Similar relationship between the growth rate and longevity was observed by varying larval density at a constant amount of added yeast. With both approaches used in this study, the variations in growth rate were due to opposite variations of both components of growth rate, i,e, developmental time and body size. The development in a medium with a varying amount of added yeast at constant larval density resulted in a similar biphasic relationship between growth rate and longevity although developmental time did not vary. The authors concluded that their findings did not confirm a simple causal relationship between the rate of development and longevity in Drosophila (Economos and Lints 1984b, 1985). Similar biphasic relationship was also found between the developmental rate and adult LS when variation in developmental temperature was used as a modulating factor (Economos and Lints 1986). Longevity was roughly independent of developmental temperature in the ‘physiological’ temperature range (16–29 °C), but both below and above this range, dramatically decrease the LS. Since growth rate varied twofold within this range, while longevity remained largely unchanged, the authors concluded that the flies’ LS is not determined by their growth rate per se, and negative relationship between growth rate and LS likely occurs only in a narrow range of environmental conditions.

The type of dose–response relationship obtained in our study can likely be explained by superposition of two processes—one promoting and the other inhibiting longevity. Such kind of relationship is characteristic of hormesis-like dose responses, in which high doses of exposure are deleterious but low doses are beneficial (Calabrese et al. 2015). Induction of hormesis-like response was proposed as a plausible explanation for advantageous effects of mild larval crowding in Drosophila (Henry et al. 2018). In this respect, larval crowding could be regarded as a condition inducing hardening effect and generating cross-tolerance to different stresses in adult life (Henry et al. 2018; Klepsatel et al. 2018; Sørensen and Loeschcke 2001). The types of effects observed in our study are similar to those reported in a Caenorhabditis elegans study, where clear hormetic effects were found after short-term exposures to heat, while debilitating effects were evident after long-term exposures, and intermediate durations caused a mixture of these effects (Yashin et al. 2002). According to the model of discrete heterogeneity proposed by the authors, each experimental worm population is a mixture of subcohorts of frail, normal, and robust animals; exposure to heat can change the proportion of worms in these subcohorts depending on the duration of exposure. It can be assumed that in our study, crowding-related exposure to stressors during larval development might, due to the induction of hormetic response, resulted in beneficial effects on viability in later life. In those flies eclosed in the last days, however, the impacts of detrimental factors including severe nutritional limitations and exposure to toxic waste products could accumulate, thereby resulting in leveling the favorable (hormetic) effects. On the basis of our findings, it should be assumed that the development in conditions of larval overpopulation (if not too extended) could trigger hormetic response thereby extending the longevity. Further studies are, however, needed to confirm this assumption.

Our data on gene expression may be also supportive for the “hormetic” explanation for the effects observed. Development in crowded conditions caused enhanced expression of all studied genes in males, suggesting that adaptive epigenetic response (Vaiserman 2008, 2010) could contribute to the effects observed. Indeed, “response to stimulus” is indicated in FlyBase (flybase.org, a primary online database of genetic and functional information about Drosophila species) as a major biological process for all genes studied. It might be assumed that the primary mode of expression changes of all these genes was a non-specific, generalized response to the overcrowding-mediated stress rather than a specific response. The gender differences in the effects on the gene expression levels (namely, lack of effects in females) require further study.

In conclusion, obtained data contradict predictions of the developmental theory of aging. Indeed, if causal relationship between the developmental time and LS would exist, one could expect that later-eclosed flies will have longer LS compared to early-eclosed ones. We, however, observed the opposite pattern. This contradiction can likely be explained either by incorrectness of the developmental theory of aging or by inadequacy of the “classical” design used to examine this theory, namely, modulating Drosophila growth rate by environmental stress (temperature, diet, larval crowding, etc.). Indeed, two potential outcomes could result from such an experimental setup – one related to modulating the duration of development per se, and other related to stress exposure. Therefore, precise cause-effect relationships cannot be inferred from the data obtained in such studies. This will be the subject of future research.