Neurobiology of Aging 26 (2005) 1083–1091
Age-dependent changes of gene expression in the Drosophila head
Se Nyun Kimb , Ji-Hwan Rheea , Young-Hwa Songb , Dong Yoon Parkb , Mina Hwanga ,
Sung-ll Leea , Ja Eun Kimb , Byung Soo Gimb , Jeong Ho Yoonb ,
Young-Joon Kimc,∗∗ , Jeongsil Kim-Haa,∗
a
Department of Molecular Biology, School of Natural Sciences, Sejong University, Gunja-dong, Gwanjin-gu, Seoul 143-747, South Korea
b Digital Genomics Inc., Seoul 120-749, South Korea
c Department of Biochemistry, Yonsei University, 134 Shinchon-dong, Seodeamun-gu, Seoul 120-749, South Korea
Received 23 December 2003; received in revised form 8 June 2004; accepted 26 June 2004
Abstract
Previous gene expression profiling studies in Drosophila have provided clues for understanding the aging process at the gene expression level.
For a detailed understanding, studies of specific regions of the body are necessary. We therefore employed microarray analysis to examine
gene expression changes in the Drosophila head during aging. Six hundred and eighty-four of the 5405 genes present in the microarray
showed significant age-dependent changes as determined by significance analysis of microarray (SAM) (q < 0.05). The biological significance
of the changes was analyzed using the gene annotations provided by the Gene Ontology Consortium. Major changes involved genes affecting
energy metabolism (proton transport, energy pathways, oxidative phosphorylation) and neuronal function, especially responses to light. Genes
involved in protein catabolism and several other metabolic processes also showed age-dependent changes. Most of the changes were reductions
in gene expression and occurred before day 13 of adult life. After day 13, the age-dependent gene expression changes were relatively smaller
than earlier life. Interestingly, the two biological processes of major gene expression changes are related to the two known environmental
changes that increase life span in Drosophila: caloric restriction and light reduction. Our findings suggest that light signaling and energy
metabolism may be important biological processes affected by aging and be interesting targets for the further investigation related to the
longevity in Drosophila.
© 2004 Elsevier Inc. All rights reserved.
Keywords: Age; Gene expression profiling; Drosophila; Head; Microarray; Energy metabolism; Response to light; Neuronal function; Longevity
1. Introduction
Aging is the gradual change in the structure and function of
multicellular organisms that eventually leads to an increased
probability of death. The most prominent characteristics of
aging are increased mortality, progressive decrease in physiological capacity, reduced ability to respond adaptively to
environmental stimuli, and increased susceptibility to disease. One idea in relation to age-dependent gene expression
is that aging represents a decline of regulatory function in the
∗
Corresponding author. Tel.: +82 2 3408 3644; fax: +82 2 312 8834.
Co-corresponding author. Tel.: +82 2 2123 2628; fax: +82 2 312 8834.
E-mail addresses: [email protected] (Y.-J. Kim), [email protected]
(J. Kim-Ha).
∗∗
0197-4580/$ – see front matter © 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.neurobiolaging.2004.06.017
organism. According to this idea, the organized regulation of
gene expression is gradually disturbed with aging, resulting
in disruption of homeostasis, and physiological decline. Another idea is that aging is an active process that is genetically
controlled. The latter idea predicts that gene expression in
the aged organism is as tightly controlled as in earlier life.
Previous studies of gene expression changes with aging in
model organisms generally support the latter idea. Analysis
of a variety of mammalian tissues revealed age-dependent
increases in the expression of genes related to inflammation,
stress, and metabolism [7,4,8–10].
Drosophila has been a favored organism for aging
studies because of its short life span, ease of maintenance, and the availability of genetic tools, mutants, and
large homogeneous populations. There are three reports
1084
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
on age-dependent genome-wide gene expression profiles
in Drosophila [2,12,18]. However, these studies were performed on the whole animal, hence, could not reveal regionspecific gene expression profiles.
In the present study, we focused on gene expression
changes in the head region during Drosophila aging. Genes
with significant age-dependent expression were selected with
significance analysis of microarray (SAM) data [15]. The biological meaning of the gene expression changes was analyzed
using the gene annotation information provided by the Gene
Ontology (GO) Consortium (http://www.geneontology.org).
Of the three organizing principles of GO, we chose the principle of biological process ontology that refers to the broad
functional goal accomplished by a given gene product or
group of products. Most of the age-dependent gene expression changes that we detected were reductions of gene expression, and in many instances, they involved genes implicated
in energy metabolism and response to light. In addition to
genes involved in light signaling, other genes closely related
to neuronal function including synaptic transmission and protein catabolism also showed noticeable expression changes.
As the central nervous system is the locus where information
on physiological status is recorded, these gene expression signatures may provide valuable indicators of physiological age.
cDNA elements representing 5405 different genes (based
on the NCIB UniGene database on August 1, 2003).
Information concerning the cDNA elements is available as supplementary information (ftp://genome.yonsei.
ac.kr/kim et al supplementary information/). For hybridization experiment, 50 ␮g of total RNA was used for reversetranscription with aminoallyl-modified dUTP, and Cy3 or
Cy5 fluorescent dyes were chemically coupled after reversetranscription reaction. The detailed protocols for labeling
and hybridization are available as supplementary information. After hybridization, slides were scanned with ScanArray Lite (Perkin-Elmer) and tiff image files were quantified with GenePix 3.0 (Axon Instruments). Spots were excluded from analysis if any spot had less than 55% of pixels
brighter than the median background intensity at both wavelengths.
2.4. Data analysis
We normalized each array data set in an intensity- and
location-dependent manner using ‘lowess’ implemented in
the software package S-plus (InSightful). For normalization,
we transformed all the data to M and A values with the following equation:
2. Materials and methods
M = log2
2.1. Determination of survival curve of Drosophila
W1118 was used as wild type Drosophila. Flies were grown
on standard Drosophila medium at 25 ◦ C until eclosion. Agesynchronized flies were obtained by collecting newly emerging flies within 12 h time window. They were placed in fresh
vial at 25 ◦ C for 1 day. Subsequent cultures were maintained
at 29 ◦ C and flies were transferred to new vials with fresh
food every 3 days. Forty wild type flies were collected and
tested for the climbing ability and survival. Climbing ability was measured by counting the number of flies that climb
up to the top of the vial within 20 s after application of the
negative geotactic pressure.
R
,
G
√
A = log2 R × G
where R is the background-subtracted Cy5 signal, and G is
the background-subtracted Cy3 signal. We then performed
within-print tip group normalization using ‘lowess’ function.
After the within-print tip group normalization, the scale of
each print tip group was adjusted so that the variance with
in each print tip group is the same. All the normalization
procedures were performed as previously described [16] and
the user-defined parameter f for the ‘lowess’ function was
set at 0.2. Pearson’s correlation coefficient was calculated
with M values (RM ) and background-subtracted intensities
(RI ) from each microarray data using MicroSoft Excel 2000
(MicroSoft).
2.2. Sample preparation
To isolate adult fly heads from the rest of the body, whole
flies were chilled with liquid nitrogen and vortexed rapidly.
Heads were separated using sieves with 710- and 300-␮m
diameter openings. Total RNA was prepared from days 1, 13,
and 25 flies using TriZol reagent (InVitrogen). The integrity
of the RNA was assessed by electrophoresis with a 2100
Bioanalyzer (Agilent).
2.3. cDNA microarray hybridization
The cDNA microarrays (Digital Genomics, Seoul, Korea) used in this study were constructed with 5929
2.5. Gene ontology analysis
In this analysis, the probability (P) that a given number
of genes or more belonging to a certain GO category will be
observed by chance in a selected Gene list (from SAM) was
calculated. Thus, smaller P values mean higher significance
for the association of gene expression changes with a certain
class of functions. Genes not included in the GO annotation
were not considered in this analysis. To calculate the P value,
we employed the hyper-geometric distribution [14], where
the probability of observing at least k genes from a gene list
of size n by chance in a category containing C genes, from a
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
1085
total number of genes, G, is given by:
C
G−C
k−1
i
n−i
P =1−
G
i=0
n
The number of categories of biological process in GO on August 1, 2003 was 7309. Among them, 461 categories that contain at least four genes present in our microarray were tested
for significant association with gene expression changes. In
the categories with less than four genes, sufficient significance for single false positive could not be obtained even if
all the genes in such categories showed an age-dependent
expression change. The categories with P value less than
1.8 × 10−3 (one false positive in 461 tests) were considered
highly significant.
2.6. In situ hybridization of fly sections
Digoxigenin (DIG)-labeled antisense RNA probes were
made from clones GH10824 (ninaC), GH08311 (Gycalpha99B), GH19095 (Arr1), GH24781 (inaC), HL07966
(Gbeta76C), GM13193 (CG1746), GH01760 (Oscp),
HL08087 (AQP), CK02342 (AP-50), LD09042 (Rop), and
GH14812 (Ddc). To detect Rp49 transcripts, we used the
312 base pair genomic DNA region (nucleotide number,
394–705 according to the numbering of AE003772 in GenBank database). Fifteen-micrometer frozen sections were
prepared from days 1 and 25 fly heads, fixed with 4%
paraformaldehyde in 1 × PBS for 1 h, and washed in PBST
(0.1% Triton X-100 in PBS). After acetylation in acetylation
buffer (0.25% acetic anhydride in 0.1 M triethanolamine),
they were washed in PBST, and prehybridized in hybridization buffer (5× SSC, 50% formamide, 100 ␮g/ml salmon
sperm DNA, 50 ␮g/ml heparin, and 0.1% Tween-20) at room
temperature for 10 min. Hybridizations were performed at
55 ◦ C overnight. After washing with PBST, slides were incubated with alkaline phosphatase-conjugated anti-DIG antibody (Roche, Mannheim, Germany), and hybridization signals were visualized with BCIP and NBT. Reactions were
performed until the hybridization signals accumulated to a
sufficient level and were stopped by washing in 1× PBS.
3. Results
3.1. Aging of Drosophila
For the study to understand the age-related gene expression changes in the Drosophila head, we examined the survivorship of the Drosophila along time. We initially performed aging experiment at 25 ◦ C. Climbing ability dropped
to 20% at day 49, but most of the flies remain survived (Fig. 1,
blue open circles). After 50 days of culture at 25 ◦ C, survivorship of the flies decreased and all of the flies died at day 63
Fig. 1. Life span and climbing ability of flies. Percentages of flies survived at
25 ◦ C (blue closed circles) and at 29 ◦ C (red closed squares) are shown along
the days after eclosion. Percentages of flies with normal climbing abilities
among the survived flies at the indicated time point are shown (blue open
circles, 25 ◦ C; red open squares, 29 ◦ C). (For interpretation of the references
to color in this figure legend, the reader is referred to the web version of the
article.)
(blue filled circles). As increase of the culture temperature
generally fasten metabolic rate of the flies, we raised flies at
29 ◦ C and examined the climbing ability and survival (Fig. 1,
red squares). We found that flies age faster at 29 ◦ C than at
25 ◦ C. After 33 days, survivorship of the flies decreased fast
and all of the flies died at day 39 (red filled squares). At this
condition, climbing ability did not change dramatically until
day 17 (red open circles). Most of the flies remained alive
by day 25, but they showed reduction in climbing ability to
25%. Since growing condition at 29 ◦ C was advantageous for
experimental set up and showed no other observable defect,
we decided to collect the samples for gene expression study
from the flies grown at 29 ◦ C. Therefore, day 13 was selected
as the middle of adult life and day 25 was selected as the very
last phase of adult life.
3.2. Selection of genes whose expression changes with
age
We examined gene expression in the Drosophila head at
days 13 and 25 compared to day 1. The microarray experiments were performed separately in males and females by
dye-swapping. To estimate the similarities between microarray data for different experimental factors, we used Pearson’s
correlation between experiments. The average correlation coefficient between M values (RM ) of upper and lower arrays
was 0.80 (correlation coefficient of background-subtracted
intensities, RI , was 0.95) and between males and females
0.71 (RI = 0.88). For two time points (days 1/13 and days
1/25), the average correlation coefficient of M values was
0.56 (RI = 0.78), suggesting that the difference in gene expression between animals of different ages is much greater
than between the sexes. In addition, we also analyzed the
gene expression changes separately in each sex. The biological processes associated with the gene expression changes in
males and females were generally similar, but showed some
discrepancy (supplementary information). Although the sex
1086
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
difference in age-dependent gene expression change is an
important issue for the understanding of aging process, we
focused on the common factors of aging and ignored the sex
difference in this study. We therefore treated the data from
males and females as independent replications for selecting
genes whose expression displayed age-dependent changes.
Thus, for each point, the relative gene expression was measured four times (twice for each sex). To identify significant
changes, SAM was performed in a one-class response format
for a single comparison (day 1 versus day 13 or day 1 versus
day 25) with the quadruple data as described.
Among 5929 probes (representing 5405 genes defined by
NCBI UniGene clusters on August 1, 2003) on the cDNA
microarray, probes with at least two data points at a single
comparison were analyzed with SAM. With a cut-off of 5%
of q value, we identified 489 probes (468 genes) differentially expressed between days 1 and 13. Among them, 422
probes (404 genes; 146 genes annotated in the Gene Ontology biological process category) showed decreased expression and 67 probes (64 genes; 20 genes annotated in the biological process category) showed increased expression at
day 13. We also identified 471 probes (455 genes) in a comparison between days 1 and 25. Among them, 332 probes
(319 genes; 103 annotated in the GO biological process category) decreased and 139 probes (136 genes; 36 genes annotated in the GO biological process category) increased at
day 25 compared to day 1. In summary, 712 probes (684
genes) showed significantly altered expression at least once,
and 249 probes (241 genes) showed altered expression both
at days 13 and 25. Most of the significant gene expression
changes observed with aging were down-regulations. Gene
lists obtained with SAM are available in the supplementary
information.
Fig. 2. Heat map for the categories significantly associated with agedependent gene expression changes. Gene expression changes in the categories most significantly associated with age-dependent gene expression
changes are represented by color-coding. The order of samples is female/day
13, male/day 13, female/day 25, and male/day 25 (from left to right). Green
color indicates down-regulation of gene expression with aging, red color
up-regulation. Gray color indicates missing data. (For interpretation of the
references to color in this figure legend, the reader is referred to the web
version of the article.)
tal functions are also discussed, although they are of only
moderate significance (P < 0.1).
3.3. Identification of biological process GO functions
associated with the age-dependent gene expression
changes
3.4. Functions significantly associated with gene
expression changes
One of the useful ways of analyzing genome-wide gene
expression data is to see whether the genes that show significant expression changes are associated with particular functions. For example, if the number of genes with significant
expression changes in a given functional group is larger than
expected by chance, one may assume that the changes are
related to their particular function. In this study, we used the
annotation information from the Gene Ontology Consortium
to examine the functional implications of the age-dependent
gene expression changes. Among the three organizing principles used by the Gene Ontology Consortium, biological process was the most useful for our purpose. The significance for
the association of the category with a gene expression change
was tested for 684 selected genes and the results are provided
as supplementary information. We will mainly consider biological process categories significantly associated with gene
expression changes (P < 1.8 × 10−3 ). Processes represented
by a number of differentially expressed genes related to vi-
3.4.1. Proton transport
The proton transport (GO:0015992) category showed
the most significant association, and all the gene expression
changes were down-regulations (8/13, P = 0.000086; Fig. 2).
Genes in the proton transport category can be divided into
two groups. One group of genes encodes components of
the proton-transporting ATP synthase complex (CG4692,
CG1746, AQP, Oscp, and CG3321). This complex synthesizes ATP using the proton gradient generated by the
electron transport system in mitochondria. Thus, these genes
are critically involved in aerobic ATP generation. Another
group of genes encode subunits of the hydrogen-transporting
ATPase V1 domain (Vha36, Vha55, and Vha68). The
hydrogen-transporting ATPase is responsible for acidification of intracellular compartments such as lysosomes
and endosomes. Since the major role of lysosomes is the
degradation of components no longer needed by the cell,
the down-regulation of hydrogen-transporting ATPase genes
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
1087
may result in impairment of cellular homeostasis dependent
on lysosome function.
3.4.2. Energy pathways
Ten of the 24 genes in the energy pathway (GO:0006091)
category showed reduced expression (P = 0.00074; Fig. 2).
Genes in the energy pathway category are involved in aspects of energy metabolism such as glycolysis (Pfk, Hex-t1,
and Gapdh1), TCA cycle (CG3861, CG6439, CG14740, and
SdhB), glycine catabolism (CG7430), glutamate catabolism
(Gdh), and glycogen biosynthesis (CG6904).
3.4.3. Oxidative phosphorylation
Oxidative phosphorylation (GO:0006119) is another important functional category related to ATP generation (Fig. 2).
Eight relevant genes among the 16 genes present in the
microarray showed reduced expression (P = 0.000591). The
products of these genes play roles in electron transport
and oxidative phosphorylation in mitochondria. The downregulation of genes in this category, together with the previously described categories (proton transport and energy pathways), points to impairment of the ATP generating system in
aged organism.
3.4.4. Response to light
The response to light (GO:0009408) category also showed
highly significant association (Fig. 2). Eight genes among
16 genes present in the microarray showed reduced expression with aging (P = 0.000591). Among them, three genes
(ninaC, Arr1, and Gβ76C) are involved in the termination
of rhodopsin-mediated signaling that is necessary for adaptation to light. Others encode guanylate cyclase (Gycα99B),
protein serine/threonine kinase (ninaC and inaC), a G protein coupled photoreceptor (cry), neurotransmitter secretion
related protein (Rop), and a G protein (Gγ30A). This result
suggests that light signal transduction and light adaptation
processes degenerate with age because of a reduction in the
relevant gene expression.
3.5. Categories with moderately significant association
with gene expression changes
3.5.1. Synaptic transmission
Of the 51 genes whose functions are related to synaptic
transmission (GO:0007268), 11 genes showed reduced expression with aging (P = 0.086413; Fig. 3). In this category,
genes related to the synaptic vesicle functions (sec5, endoA,
Csp, Rop, n-syb, AP-50, and γ-SNAP) were noticeable. All
of the genes with significant expression change were downregulated with age. Though the significance of the synaptic transmission category is rather low, the number of the
down-regulated genes is not negligible. In addition, the downregulation of the genes in this category may account, at least
in part, for the degeneration of neuronal function in the aging
organism.
Fig. 3. Heat map for the categories with moderately significant association
with age-dependent gene expression changes. Data are represented as in
Fig. 2. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of the article.)
3.5.2. Protein catabolism
Genes in the protein catabolism (GO:0030163) category
are also moderately over-represented among the differentially
expressed genes (18/88, P = 0.053405; Fig. 3). Six genes
belonging to this category encode subunits of the proteasome complex (Pros54, Rpn7, ProsMA5, Rpn9, Rpn5, and
Pros26), and all showed reduced expression with aging. Five
genes involved in ubiquitin-dependent proteolysis were also
in this category. However, the direction of the gene expression
changes was not consistent: two (CG9934 and DG: 25E8.2)
were down-regulated and three (ago, CG5384, and faf) upregulated. Seven genes encoding proteolytic activities were
also included in this category, the expression of six being reduced. These results suggest that the activity of the ubiquitindependent proteolytic system and of proteasomes declines in
the aged organism. Since the proteasome complex is responsible for the removal of damaged proteins and plays a vital
role in cell survival and proper cellular functioning especially
in the neuron, the decline of its activity could contribute to
reduced neuronal function with age.
3.5.3. Metabolism
Several metabolism-related categories were also moderately over-represented (Fig. 3). Among them were alcohol
metabolism (GO:0046164), aromatic compound biosynthesis (GO:0019438), carbohydrate metabolism (GO:0005975),
and organic acid metabolism (GO:0006082). Genes in these
categories were consistently down-regulated with age, suggesting that metabolic processes generally decline with age.
1088
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
3.5.4. Miscellaneous categories
Previous studies in mammals showed increased expression of inflammation-related genes in aged organism [7,4,10].
Three inflammation-related categories are present in biological process GO (anti-inflammatory response, inflammatory
response, and killing of inflammatory cells), but our microarray did not contain genes in those categories. Stress responserelated genes have also been reported to increase in aged organism [7,4,9,10,12]. Sixty-two genes belonging to this category are present in our microarray and eight genes showed
differential expression with aging. Among them, two heat
shock genes (Hsc70-5 and Hsp26) and two DNA repair related genes (Ercc1 and rad50) were up-regulated with aging.
Four genes related to diverse stress responses (CycT, bsk,
GstE1, and CG12013) were down-regulated.
3.6. Differences in gene expression between days 13 and
25
Among the 684 genes selected by SAM, only 241 genes
showed a significant change in expression on both days 13
and 25. To examine the temporal differences in the gene expression changes, we examined the number of significantly
changed genes in the selected categories at each time point
(Table 1). Gene expression changes were not substantially
different between two time points in the categories of proton
transport, responses to light, and certain metabolic processes.
However, larger numbers of gene expression changes were
observed on day 25 in the energy pathways and carbohydrate
metabolism categories. On the other hand, the categories of
oxidative phosphorylation and synaptic transmission showed
fewer gene expression changes on day 25. These observations
led us to analyze the gene expression change from days 13 to
25 in more detail.
To examine the gene expression change occurring from
days 13 to 25 in detail, we examined the gene expression
profiles in Drosophila head samples only from male. Four
hybridizations with dye-swapping were performed and the
changes in gene expressions were analyzed with SAM.
With a cut-off of 5% q value, we identified 178 genes (181
probes) increased and 53 genes (53 probes) decreased (list
of the genes were provided in supplementary information).
In this case, we could not find significant association of
certain biological process category with gene expression
change at P < 1.8 × 10−3 . In addition, we could not find
remarkable gene expression changes even in the categories
that showed different numbers of significantly changed
genes at days 13 and 25 compared to day 1 (such as synaptic
transmission and carbohydrate metabolism), suggesting that
the difference of the numbers of changed genes might be
caused by artifacts from experiments or statistical selection
of the genes. These results suggest that the gene expression
changes we described occur mostly in earlier stage of life
and the age-dependent gene expression changes are greater
in earlier life than in later life.
The most noteworthy gene expression changes were observed in the category of response to stress (GO:0006950).
Among 61 genes, 6 genes showed increased expression
(P = 0.036412) and 3 heat shock protein genes (HSP26,
HSP27, and HSP23), one oxidative stress-related gene
(GSTE1), and two DNA repair related genes (rad50 and
EG:33C11). Since the role of the proteins encoded by these
genes are adaptation to environmental or internal stresses, it
is likely that the increase of gene expression in this category
may represent adaptation of the organism toward accumulating stresses. The biological process categories associated
with the gene expression changes from days 13 to 25 are
summarized in Table 2.
3.7. Histological analysis of gene expression changes in
the aged Drosophila head
To verify the gene expression change in the histological
context, we examined the expression pattern of some of the
genes identified from the microarray analysis by in situ hybridization. We selected genes involved in the response to
light, since this group of genes was not reported to show significant changes when the whole body was examined. Expression of Arr1, inaC, ninaC, and Gβ76C in eye region
(Fig. 4C–F) and Gycα99B in the cortex region of the cen-
Table 1
Biological process ontology categories associated with gene expression changes from days 1 to 13 and 25
GO number
Ontology name
Ratio
P value
Day 13
Day 25
Up
Down
GO:0015992
GO:0006091
GO:0006119
GO:0009416
GO:0007268
GO:0030163
GO:0006066
GO:0019438
GO:0005975
GO:0006082
Proton transport
Energy pathways
Oxidative phosphorylation
Response to light
Synaptic transmission
Protein catabolism
Alcohol metabolism
Aromatic compound biosynthesis
Carbohydrate metabolism
Organic acid metabolism
8/13
10/24
8/16
8/16
11/51
18/88
7/19
4/7
12/37
14/50
0.000086
0.00074
0.000591
0.000591
0.086413
0.053405
0.010552
0.009101
0.002945
0.006128
8
5
8
6
11
14
5
4
6
11
7
9
3
5
3
10
7
3
11
9
–
–
–
–
–
3
–
–
–
1
8
10
8
8
11
14
7
4
12
13
Categories with lower P value than 1.8 × 10−3 (one false positive from 461 tests) were indicated by boldface. Ratio indicates the number of genes with
expression changes in a given category divided by the number of genes present in the microarray in a given category. The number of genes selected by SAM
at each time point (days 13 and 25) and those of up-regulated (up) or down-regulated (down) were indicated.
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
1089
Table 2
Biological process ontology categories associated with gene expression changes from days 13 to 25
GO number
Ontology name
Ratio
P value
Up
Down
GO:0008595
GO:0006360
GO:0009168
GO:0008293
GO:0006950
GO:0006364
GO:0006731
Determination of anterior–posterior axis, embryo
Transcription from Pol I promoter
Purine ribonucleoside monophosphate biosynthesis
Torso receptor signaling pathway
Response to stress
rRNA processing
Coenzymes and prosthetic group metabolism
3/11
3/13
2/5
2/7
6/61
2/8
3/21
0.008725
0.014254
0.015468
0.030791
0.036412
0.039975
0.052423
3
3
2
2
6
2
3
–
–
tral nervous system (Fig. 4B) were strong at day 1, and the
level of gene expression dramatically decreased at day 25 in
all cases. Expression of the Rp49 gene that did not change
in the microarray also did not show any reduction in RNA
expression in day 25 flies (Fig. 4A).
Expression of the genes involved in proton transport
(Oscp, AQP, and CG1746) and synaptic transmission (Rop,
AP-50, and Ddc) was also examined by in situ hybridization.
All the tested genes showed reduced RNA levels in day 25
flies compared to those in day 1 flies (Fig. 5), suggesting that
the gene expression changes detected by microarray can be
verified by other detection method.
4. Discussion
To gain more detailed insight into the molecular basis of
the aging process in Drosophila, we examined gene expres-
–
–
–
–
sion difference between the heads of young and aged organisms. The gene expression profiles were systematically analyzed taking advantage of annotation information from the
GO Consortium. The result of the analysis showed that reductions in gene expression related to energy metabolism and
light signaling are the most significant changes with aging.
Genes involved in synaptic transmission, protein catabolism,
and several metabolic processes were also affected. These
gene expression changes mostly occur during the early phase
of aging (before the middle of adult life) and remained until
the last phase of life.
Based on the result of GO analysis, the age-dependent
changes controlled by gene expression could be divided into
two groups. One is a decline in cellular housekeeping activities, the other a decline in organism-level regulatory activities. The decline of cellular housekeeping activities could
be observed in the processes of energy metabolism (proton
transport, energy pathways, and oxidative phosphorylation),
Fig. 4. In situ RNA hybridization to days 1- or 25-old fly heads for genes involved in the response to light. In each panel, upper and lower lane is days 1 and 25
fly heads, respectively. Left side shows in situ hybridization results and right DAPI staining of the same section. (A) Rp49, (B) Gyc␣99B, (C) Arr1, (D) inaC,
(E) ninaC, and (F) Gβ76C.
1090
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
Fig. 5. In situ RNA hybridization to days 1 and 25 fly heads for genes involved in proton transporting or synaptic transmission. In each panel, upper and lower
lane is days 1 and 25 fly heads, respectively. (A) CG1746, (B) Oscp, (C) AQP, (D) AP-50, (E) Rop, and (F) Ddc. Arrows indicate Ddc expressing cells in day
1 fly heads. No Ddc transcript-specific hybridization signal was detected in the corresponding area of day 25 fly heads.
protein turnover (protein catabolism and proton transport),
and other metabolisms. Decline of regulatory activity could
be found in the categories of response to light and synaptic
transmission. These results suggest that control of the aging
process via regulation of gene expression is executed at both
the individual cell and whole organism level.
An interesting tendency evident in our data is that most of
the gene expression changes involve down-regulation. Especially, down-regulation of some group of genes, such as genes
involved in the energy metabolism and light signaling, might
result in the degeneration of particular function rather than the
activation. Up-regulation of the genes related to inflammation
and stress response was reported previously [7,4,9,10]. These
gene expression changes can be interpreted as adaptations to
environmental stress rather than consequences of active control of the aging process. Therefore, our observation suggests
that the degeneration of functions at the cellular and organism level is, at least in part, a way of genetic control of the
aging process.
Our results concerning age-dependent gene expression
changes in the Drosophila head region differ from those obtained in the studies of the whole body. In the latter, the biological process categories showing significant changes included protein transport, stress responses, responses to biotic
stimuli, and eggshell formation [12]. The categories related
to ATP generation and light signaling that showed the most
significant changes in the present study were not identified.
Protein metabolism and modification (currently referred to as
protein metabolism) were reported to be under-represented
among the gene expression changes. In contrast, the protein
catabolism category (the sub-category of protein metabolism)
was over-represented in our gene expression changes. These
discrepancies suggest that analysis of individual regions or
tissues, rather than the whole body, is essential for a detailed
understanding of the regulation of age-dependent gene expression. In addition, the gene expression signatures identified in this study may serve as indicators of aging in the
central nervous system.
Caloric restriction is known to increase life span in a variety of organisms in addition to Drosophila [6,3,5,17]. There is
a Drosophila gene, Indy, whose mutation leads to a doubling
of life span [13]. The life-extending effect of this mutation
is also thought to be due to an alteration of energy balance
that mimics caloric restriction. Another life-extending stimulus in Drosophila is reduction of light [1,11]. The effect
of light reduction on Drosophila aging is reminiscent of the
life-extending effect of caloric restriction. It is of interest
that energy metabolism and light signaling were identified as
the processes displaying the most significant gene expression
changes in the present study. We hypothesize therefore that
light may be an important signal affecting the aging process
and longevity. The effect of light on the aging process awaits
detailed investigation.
Acknowledgements
This work was supported by a KISTEP grant (M1-010800-0070) to J. K-H and a Creative Research Initiatives Pro-
S.N. Kim et al. / Neurobiology of Aging 26 (2005) 1083–1091
gram from the Korean Ministry of Science and Technology
to Y.-J.K.
[8]
Appendix A. Supplementary data
[9]
Supplementary data associated with this article
can be found, in the online version, at doi:10.1016/
j.neurobiolaging.2004.06.017.
[10]
[11]
References
[1] Allemand R, Cohet Y, David J. Increase in the longevity of adult
Drosophila melanogaster kept in permanent darkness. Exp Gerontol
1973;8:279–83.
[2] Arbeitman MN, Furlong EEM, Imam F, Johnson E, Null BH, Baker
BS, et al. Gene expression during the life cycle of Drosophila
melanogaster. Science 2002;297:2270–5.
[3] Braeckman BP, Houthoofd K, Vanfleteren JR. Insulin-like signaling,
metabolism, stress resistance and aging in Caenorhabditis elegans.
Mech Ageing Dev 2001;122:673–93.
[4] Cao SX, Dhahbi JM, Mote PL, Spindler SR. Genomic profiling of
short- and long-term caloric restriction effects in the liver of aging
mice. Proc Natl Acad Sci USA 2001;98:10630–5.
[5] Chapman T, Partridge L. Female fitness in Drosophila melanogaster:
an interaction between the effect of nutrition and of encounter
rates with males. Proc R Soc Lond B Biol Sci 1996;263:
755–9.
[6] Jiang JC, Jaruga E, Repnevskaya MV, Jazwinski SM. An intervention
resembling caloric restriction prolongs life span and retards aging in
yeast. FASEB J 2000;14:2135–7.
[7] Kayo T, Allison DB, Weindruch R, Prolla TA. Influences of aging and caloric restriction on the transcriptional profile of skeletal
[12]
[13]
[14]
[15]
[16]
[17]
[18]
1091
muscle from rhesus monkeys. Proc Natl Acad Sci USA 2001;98:
5093–8.
Lee C-K, Allison DB, Brand J, Weindruch R, Prolla TA. Transcriptional profiles associated with aging and middle age-onset
caloric restriction in mouse hearts. Proc Natl Acad Sci USA
2002;99:14988–93.
Lee C-K, Klopp RG, Weindrugh R, Prolla TA. Gene expression
profile of aging and its retardation by caloric restriction. Science
1999;285:1390–3.
Lee C-K, Weindruch R, Prolla TA. Gene-expression profile of the
ageing brain in mice. Nature Genet 2000;25:294–7.
Massie HR, Whitney SJP. Preliminary evidence for photochemical
ageing in Drosophila. Mech Aging Dev 1991;58:37–48.
Pletcher SD, Macdonald SJ, Marguerie R, Certa U, Stearns SC,
Goldstein DB, et al. Genome-wide transcript profiles in aging and calorically restricted Drosophila melanogaster. Curr Biol
2002;12:712–23.
Rogina B, Reenan RA, Nilsen SP, Helfand SL. Extended life-span
conferred by cotransporter gene mutations in Drosophila. Science
2000;290:2137–40.
Tavazoie S, Hughes JD, Campbell MJ, Cho RJ, Church GM. Systematic determination of genetic network architecture. Nature Genet
1999;22:281–5.
Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci
USA 2001;98:5116–21.
Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, et al. Normalization for cDNA microarray data: a robust composite method
addressing single and multiple slide systematic variation. Nucleic
Acids Res 2002;30:4–15.
Yu BP, Masaro E, McMahan CA. Nutritional influences on aging of
Fisher 344 rats: physical, metabolic, and longevity characteristics. J
Gerontol 1985;40:657–70.
Zou S, Meadows S, Sharp L, Jan LY, Jan YN. Genome-wide study
of aging and oxidative stress response in Drosophila melanogaster.
Proc Natl Acad Sci USA 2000;97:13726–31.