Slowed aging during reproductive dormancy is reflected in genome-wide transcriptome changes in Drosophila melanogaster

Abstract

Background: In models extensively used in studies of aging and extended lifespan, such as C. elegans and Drosophila, adult senescence is regulated by gene networks that are likely to be similar to ones that underlie lifespan extension during dormancy. These include the evolutionarily conserved insulin/IGF, TOR and germ line-signaling pathways. Dormancy, also known as dauer stage in the larval worm or adult diapause in the fly, is triggered by adverse environmental conditions, and results in drastically extended lifespan with negligible senescence. It is furthermore characterized by increased stress resistance and somatic maintenance, developmental arrest and reallocated energy resources. In the fly Drosophila melanogaster adult reproductive diapause is additionally manifested in arrested ovary development, improved immune defense and altered metabolism. However, the molecular mechanisms behind this adaptive lifespan extension are not well understood.

Results: A genome wide analysis of transcript changes in diapausing D. melanogaster revealed a differential regulation of more than 4600 genes. Gene ontology (GO) and KEGG pathway analysis reveal that many of these genes are part of signaling pathways that regulate metabolism, stress responses, detoxification, immunity, protein synthesis and processes during aging. More specifically, gene readouts and detailed mapping of the pathways indicate downregulation of insulin-IGF (IIS), target of rapamycin (TOR) and MAP kinase signaling, whereas Toll-dependent immune signaling, Jun-N-terminal kinase (JNK) and Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathways are upregulated during diapause. Furthermore, we detected transcriptional regulation of a large number of genes specifically associated with aging and longevity.

Conclusions: We find that many affected genes and signal pathways are shared between dormancy, aging and lifespan extension, including IIS, TOR, JAK/STAT and JNK. A substantial fraction of the genes affected by diapause have also been found to alter their expression in response to starvation and cold exposure in D. melanogaster, and the pathways overlap those reported in GO analysis of other invertebrates in dormancy or even hibernating mammals. Our study, thus, shows that D. melanogaster is a genetically tractable model for dormancy in other organisms and effects of dormancy on aging and lifespan.

Fig. 1

Diagrams summarizing differentially expressed transcripts under diapause and control conditions, and their functional classification based on gene ontology (GO) terms. a Out of 4624 differentially expressed transcripts, 3387 are detected in flies diapausing for 3 weeks (3wD) compared to both 1 (1wN) and 3 (3wN) week controls. Very few transcripts (19) differ due to the age difference between control flies. The transcripts shown display at least an absolute two-fold change and q < 0.05. b The functional classification of GO terms reveals that the majority of the affected genes are involved in metabolic (44 %) and cellular processes (21 %)

Fig. 2

The 25 most up- and downregulated transcripts in diapausing flies. Flies diapausing for three weeks (3wD) were compared to one week old flies (1wN) kept at normal conditions. Upregulated transcripts are predominantly associated with the intestinal structures, fat body, spermatheca and heart, while the most downregulated ones are expressed in the ovaries intestinal structures and fat body (only Fad2). The changes are given as logarithmic fold change (logFC). For a complete list of altered gene expression see Additional file 3: Dataset S1

Fig. 3

Quantitative real-time PCR confirms transcript changes in diapausing flies. Estimate of relative gene expression, comparing (ΔΔCt) three week old diapausing flies (3wD) with three week old flies kept under normal conditions (3wN). Directionality and relative magnitude of change matches microarray data in all cases, except Vm32E. We used Welch’s unpaired t-test with corrections for non-equal variances, comparing expression at 3wD and 3wN for each gene, ns, not significant; *, p < 0.05; **, p < 0.01. Error bars show standard error of the mean, n = 4. Acronyms used: FAD2 - Fatty Acid Desaturase 2, HS67Bc - Heat Shock 67 Bc, LSPgamma - Larval Serum Protein gamma, NeurpepRecF - Neuropeptide receptor F, OBP19b - Odorant Binding Protein 19 b, Per – Period, Tim-1 - Timeless, (targets all Timeless isoforms except N and O), Tim-2 - Timeless, (targets N and O isoforms of Timeless), Tim-Total - Timeless, (targets all Timeless isoforms), VME32e - Vitelline Membrane 32e

Fig. 4

Altered gene expression in AKH, insulin/IGF- and TOR-signaling pathways in diapausing flies. The scheme displays a generalized assembly of relevant genes (regardless tissue-specificity) in the three signal pathways. Transcript levels (logarithmic fold change, LogFC) are given in red for upregulated, blue for downregulated and black for no significant change (ns; LogFC close to 0). The acronyms are listed in Additional file 6: Table S3 where also references and details of gene/protein functions are given. The AKH signaling is generally upregulated and the expression of the target gene tobi increased (X depicts an unidentified transcription factor). The Insulin/IGF signaling is generally downregulated and as a result several of the FOXO-regulated target genes are upregulated. The TOR-signaling is also downregulated, together with a number of its target genes. The AKH signaling scheme was compiled after [68, 171], the insulin/IGF signaling based on elements from [29, 36] and TOR signaling assembled after [35, 57]

Fig. 5

Altered gene expression in the JAK/STAT signaling pathway during diapause. This scheme displays a generalized assembly (regardless tissue specificity) of relevant genes in the JAK/STAT signal pathway. Transcript levels (logarithmic fold change, LogFC) are given in red for upregulated, blue for downregulated and black for no significant change (ns; LogFC close to 0). The acronyms are listed in Additional file 6: Table S3 where also references and details of gene/protein functions are given. At the whole organism level we observed a decreased expression of the core components of the JAK/STAT signaling and some read-out genes in diapausing flies. Some exceptions are seen in certain genes, involved in humoral immune response and stress resistance, which were upregulated (probably tissue-specific). The scheme is based partly on [172] and [80]. Further details are given in Additional file 6: Table S3 and text

Fig. 6

Tissue specific JAK/STAT signaling in diapausing flies. The Stat92E reporter 10xStat92E-GFP expression in the whole body is compared between control flies (1wN) (a) and flies in diapause conditions (3wD) (b) and shows an increased fluorescence in 3wD, as quantified in (c) with corrected total fly fluorescence (CTFF). The statistical significance was verified by Welch’s unpaired t-test, * p < 0.05, n = 6 (error bars show standard error of the mean). In (a) and (b) arrows point to hemocytes that accumulated in the legs of the fly, which largely contributes to the fluorescence signal, and the asterisk in B indicates GFP signal in the ovaries. d and e Hemocytes exhibit a strong accumulation of the 10xStat92E-GFP signal upon diapause, 3wD (e) compared to controls, 1wN (d). In these figures green label is 10xStat92E-GFP, and blue label hemocyte nuclei stained with DAPI. E´ and D´ The corresponding hemocytes displayed in bright field. The scale bar corresponds to 20 μm. f Corrected total cell fluorescence (CTCF) for diapause and control hemocytes is calculated as an average from n = 15, statistical significance verified by Welch’s unpaired t-test, ***, p < 0.001. Error bars show standard error of the mean. g and h Portions of ovaries from control (1wN) and diapausing (3wD) flies. In diapausing flies there is an increased expression of the 10xStat92E-GFP reporter (green) in polar cells of ovaries (arrow in h), which is not detectable in directly developing ovaries of control flies (g). Blue label, nuclei stained with DAPI; red label phalloidin-TRITC. Scale bars: D, E 50 μm, G, H 25 μm

Fig. 7

Altered gene expression in Toll (Tl) signaling pathway during diapause. The scheme displays a generalized assembly (regardless tissue specificity) of relevant genes in the Toll signal pathway. Transcript levels (logarithmic fold change, LogFC) are given in red for upregulated, blue for downregulated and black for no significant change (ns; LogFC close to 0). The acronyms are listed in Additional file 6: Table S3 where also references and details of gene/protein functions are given. The Toll pathway is activated in diapausing flies as seen from increases of target genes such as Drosomycin (Drs), Drosomycin-like 5 (Drsl5), IM1 and IM2. Also several extracellular members of the Toll-activating proteolytic cascade were upregulated. The signaling scheme is based on [97]

Fig. 8

Drosomycin (Drs) and Drosomycin-like 5 (Drsl5) transcription during diapause is regulated by Spätzle and Toll signaling. Upregulation of Drosomycin (a) and Drosomycin-like 5 (b) is significant in the control genotype (Canton S), as well as in a mutant in a component of the Imd pathway, the transcription factor Relish (RelE20). Conversely transcription of Drs and Drsl5 was blocked in a mutant in the Toll ligand Spätzle (spzRM7). Significance compared to one-week-old Canton S flies (1wN) was verified by ANOVA, followed by Tukey test (ns, not significant; * p < 0.05; ** p < 0.01; *** p < 0.001). Error bars show standard error of the mean, n = 3

Fig. 9

Altered gene expression in the JNK signaling pathway during diapause. This scheme displays a generalized assembly (regardless tissue specificity) of relevant genes in the JNK signal pathway. Transcript levels (logarithmic fold change, LogFC) are given in red for upregulated, blue for downregulated and black for no significant change (ns; LogFC close to 0). The acronyms are listed in Additional file 6: Table S3 where also references and details of gene/protein functions are given. At the whole organism level diapausing flies demonstrate a moderate activation of JNK signaling as judged from an elevated expression of some target genes. There is also enhanced expression of JNK-responsive genes, known to be under FOXO (IIS) transcriptional control (Sesn and Fer1HCH)

Fig. 10

Venn diagrams comparing differentially regulated genes in genome-wide transcriptome studies of D. melanogaster. a Venn diagram comparing differentially regulated genes after 3 weeks of diapause (3wD, blue area; our study) with genes responding to aging [136] and a list of candidate genes affecting lifespan [135]. In Additional file 10: Figure S5 we show the results of Overrepresentation test of GO terms, performed in PANTHER for genes altered in dormancy in comparisons in groups A-D. b Comparisons with starvation, one-week diapause (abdomen) and cold tolerance. The blue area displays the number of differentially regulated genes after 3 weeks of diapause (3wD; our study). The grey area shows number of genes regulated in abdomens of flies after one week of diapause (1wD Abdomen) [140]. The other two sets contain genes regulated due to cold intolerance [141] and starvation [143]. In Additional file 11: Figure S6 we present a gene set enrichment analysis (GSEA) of Gene Ontology (GO) terms for comparisons among the four genome-wide transcription studies for each of the fields A-H in 10B