DE-cadherin-mediated cell adhesion is essential for maintaining somatic stem cells in the Drosophila ovary.

Abstract

Evidence from many systems has shown that stem cells are maintained in "niches" or specific regulatory microenvironments formed by stromal cells. The question of how stem cells are maintained in their niches is important, and further studies will lead to a better understanding of stem cell regulation and enhance the future use of stem cells in regenerative medicine. Here we show that cadherin-mediated cell adhesion is required for anchoring somatic stem cells (SSCs) to their niches in the Drosophila ovary. DE-cadherin and Armadillo/beta-catenin accumulate in the junctions between SSCs and their neighboring cells, inner germarial sheath cells. Removal of DE-cadherin from SSCs results in stem cell loss in the adult ovary. Furthermore, the cadherin-mediated adhesion is also important for maintaining SSCs in their niches before adulthood. This study provides further support that SSCs are located in a niche formed by their neighboring cells. We have previously shown that DE-cadherin-mediated cell adhesion is essential for anchoring germ-line stem cells to their niches in the Drosophila ovary. This study further implicates cadherin-mediated cell adhesion as a general mechanism for anchoring stem cells to their niches in a variety of systems.

Full text

Page 1

DE-cadherin-mediated cell adhesion is essential
for maintaining somatic stem cells in the
Drosophila ovary
Xiaoqing Song* and Ting Xie*†‡
*Stowers Institute for Medical Research, 1000 East 50th Street, Kansas City, MO 64110; and†Department of Anatomy and Cell Biology, University of Kansas
School of Medicine, 3901 Rainbow Boulevard, Kansas City, KS 66160
Edited by Anthony P. Mahowald, University of Chicago, Chicago, IL, and approved September 18, 2002 (received for review July 5, 2002)
Evidence from many systems has shown that stem cells are main-
tained in ‘‘niches’’ or specific regulatory microenvironments
formed by stromal cells. The question of how stem cells are
maintained in their niches is important, and further studies will
lead to a better understanding of stem cell regulation and enhance
the future use of stem cells in regenerative medicine. Here we
show that cadherin-mediated cell adhesion is required for anchor-
ingsomaticstemcells(SSCs)totheirnichesintheDrosophilaovary.
DE-cadherin and Armadillo??-catenin accumulate in the junctions
between SSCs and their neighboring cells, inner germarial sheath
cells. Removal of DE-cadherin from SSCs results in stem cell loss in
the adult ovary. Furthermore, the cadherin-mediated adhesion is
also important for maintaining SSCs in their niches before adult-
hood. This study provides further support that SSCs are located in
a niche formed by their neighboring cells. We have previously
shown that DE-cadherin-mediated cell adhesion is essential for
anchoring germ-line stem cells to their niches in the Drosophila
ovary. This study further implicates cadherin-mediated cell adhe-
sionasageneralmechanismforanchoringstemcellstotheirniches
in a variety of systems.
S
directly or indirectly shown to exist in many adult tissues, and are
responsible for generating differentiated cells that replace lost
cells throughout an organism’s lifetime (1–3). The molecular
mechanisms controlling stem cell function in vivo are crucial to
the future use of stem cells in regenerative medicine and also in
understanding the aging process, tumor formation, and degen-
erative diseases (4–6). Likely, one of the important niche
functions is to control stem cell behavior (i.e., self-renewal,
proliferation, and differentiation) through secreted growth fac-
tors and cell–cell contacts. Among the top priorities in stem cell
research is to further define the structures and functions of
different stem cell niches and to reveal the molecular mecha-
nisms involved in communication between stem cells and their
niches.
Two stem cell types, somatic stem cells (SSCs) and germ-line
stem cells (GSCs) that exist in the Drosophila ovary represent an
excellent system in which to study adult stem cells at the cellular
and molecular levels in vivo in the adult Drosophila ovary (7, 8).
At the anterior end of each ovariole of an ovary, or germarium
(Fig. 1A), are two or three GSCs whose progeny eventually
develop into mature oocytes. Recently, GSCs have been shown
to be located in a niche and are directly regulated by their niche
cells (9–14). Two or three SSCs reside in the middle of the
germarium, likely representing a model system in which to study
epithelial stem cells in adult tissues (15, 16). They divide and
generate a population of mitotically active follicle progenitor
cells that reside in the middle of germarium. Like stem cells in
other systems, these progenitors continuously proliferate in the
egg chambers of stage 1 to stage 6, and generate several types of
differentiated cells that cover egg chambers. These differenti-
ated somatic follicle cell types have characteristics similar to
tem cells are defined by their ability to self-renew and to
continuously generate differentiated cells. They have been
epithelialcellsinmammals,suchasbasal-lateralandbasal-apical
polarities. Recent research indicates that hedgehog (hh), pro-
duced primarily by cap cells located a few cells away from SSCs,
directly regulates SSC maintenance and proliferation, suggesting
that SSCs are located in a niche (16, 17). Hyperactivation of hh
signaling by means of overexpressing hh or by removing negative
regulators causes the overproduction of somatic follicle cells and
increases the number of SSCs. The reduction of hh signaling by
removing the functions of hh or downstream positive regulators
results in rapid SSC loss.
In mice and humans, hyperactivation of the hh pathway, either
by mutations or overexpression, causes over-proliferation of skin
cells, resulting in the formation of skin tumors (18–20). These
studies, along with those from Drosophila, suggest that mecha-
nisms regulating epithelial cells may be conserved from Dro-
sophila to humans. Thus, SSCs in the Drosophila ovary could
represent an effective system in which to study epithelial stem
cell self-renewal, proliferation, and differentiation. Although
SSCs lack distinctive morphology and unique molecular mark-
ers, SSCs in the Drosophila ovary can still be studied in great
detail by using a cell-marking system to monitor their activities.
SSCs mutant for a given gene can be effectively marked genet-
ically,andtheirbehaviorcanbestudiedthroughoutalongperiod
so that the effect of any gene on stem cell regulation can be
assessed (16, 17).
Cadherin gene family members, encoding Ca2?-dependent
transmembrane adhesion molecules, mediate interactions be-
tween different cell types in a variety of organisms through
homophilic interactions (21). DE-cadherin, encoded by shotgun
(shg), is required for positioning oocytes in egg chambers and for
border cell migration in the Drosophila ovary (22–25). Our
recent research demonstrates that DE-cadherin is required for
anchoring GSCs in their niche (26). Here we report that
DE-cadherin is also required for maintaining SSCs in their niche
in the adult Drosophila ovary. DE-cadherin-mediated cell adhe-
sion also requires ?-catenin, encoded by armadillo (arm) in
Drosophila (27). We further show that DE-cadherin-mediated
cell adhesion plays an essential role in SSC maintenance before
adulthood. This work further suggests that cadherin-mediated
cell adhesion may represent a general mechanism by which adult
stem cells are anchored to their niches in different systems.
Materials and Methods
Drosophila Stocks and Genetics. The following fly stocks were used
in this study and are described either in Flybase or as specified:
X-15-29, X-15-33, MKRS hs-FLP (15, 28); FLP recombination
target (FRT)42DshgR69(24), FRT42Dshg10469(26), and FRT42D.
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations:IGS,innergermarialsheath;SSC,somaticstemcell;GSC,germ-linestemcell;
Fas3, Fasciclin III; FRT, FLP recombination target.
‡To whom correspondence should be addressed. E-mail: tgx@stowers-institute.org.
www.pnas.org?cgi?doi?10.1073?pnas.232389399 PNAS ?
November 12, 2002 ?
vol. 99 ?
no. 23 ?
14813–14818
CELL BIOLOGY

Page 2

All Drosophila stocks were maintained at room temperature on
standard cornmeal?molasses?agar media.
Generating lacZ-Positive Clones to Mark SSCs.We generated mitotic
clones to mark SSCs according to the published procedures (15).
One- or two-day-old females of the genotype X-15-29?X-15-33;
MKRS hs-FLP?? were heat-shocked twice in a 37°C water bath
for1hseparatedbyanintervalof8h.Fliesweretransferreddaily
to fresh, yeasted food to maintain optimal conditions for
oogenesis. The SSC clones were identified by the expression of
the lacZ gene as detected by antibody staining.
Generating Mutant shg SSC Clones. Clones of mutant cells were
generated by FLP-mediated mitotic recombination, as described
(11).Togeneratethestocksforstemcellclonalanalysis,FRT42D,
FRT42D shg10469?CyO and FRT42D shgR69?CyO males were
mated with virgin females yellow (yw) hs-FLP; FRT42Darma-
dillo–lacZ, respectively. One- or two-day-old adult non-CyO
females carrying an armadillo–lacZ transgene in trans to the
mutant-bearing chromosome were heat-shocked six times at
37°C for 1 h separated by intervals of 8–12 h. The females were
transferred to fresh food every day at room temperature, and
ovaries were removed 1, 2, or 3 weeks after the last heat-shock
treatment and then further processed for antibody staining. To
determine stem cell maintenance, the percentages of the ovari-
oles carrying a marked SSC clone at different time points were
calculated by dividing the number of germaria carrying marked
follicles by the number of total germaria examined.
Generating Mutant shg Pre-Somatic Stem Cells to Test SSC Establish-
ment Before Adulthood. To generate the stocks for marked
preSSCs, FRT42D?, FRT42Dshg10469?CyO, and FRT42DshgR69?
CyO males were mated with virgin females yw hs-FLP; FRT42D
armadillo–lacZ, respectively. The culture tubes containing the
progeny from these crosses that had not reached the late
third-instar stage were heat shocked twice for 1 h at an interval
of 6 h. One- or two-day-old adult non-CyO females carrying an
armadillo–lacZ transgene in trans to the mutant-bearing chro-
mosome were dissected and processed for immunostaining. To
determine the effect of shg mutations on SSC establishment
before adulthood, the percentages of the ovarioles carrying a
marked SSC clone for a given genotype were calculated by
dividing the number of germaria carrying marked SSCs by the
number of total germaria examined.
Immunohistochemistry. The following antisera were used: mono-
clonal anti-Hts antibody 1B1 (1:4), monoclonal anti-Fasciclin III
(Fas3) antibody 7G10 (1:4), monoclonal antibody anti-
Armadillo N7A1 (1:4), Rat anti-DE-cadherin (1:100) (23) and
rabbit polyclonal anti-?-galactosidase (1:300; Molecular Probes)
antibodies.
Ovaries were dissected in Grace’s media and fixed in PBS with
4% formaldehyde for 12–15 min, then washed with PBT (PBS
and 0.2% Triton X-100) 5 times for 15 min each. The ovaries
were incubated in 0.5% goat serum diluted with PBT for 1 h.
Appropriate primary antibodies were added to PBS and incu-
bated at 4°C overnight, then washed with PBT 5 times for 15 min
each wash. Lastly, appropriate secondary antibodies were added
and incubated overnight, then washed with PBT five times for 15
min each. After the last wash, the stained ovaries were mounted
in Vectashield mounting media (Vector Laboratories). All mi-
crographs were taken with a Leica SPII confocal microscope.
Results
DE-Cadherin and Arm Are Expressed in the Interfaces Between SSCs
and Their Neighboring Cells. To directly determine whether DE-
cadherin-mediated cell adhesion is important for anchoring
SSCs to niche cells, we examined the expression of DE-cadherin
and Arm proteins in the SSCs of the germarium. Because no
reliable markers for SSCs are available, we used the FLP-
mediated FRT recombination to positively mark SSCs. This
technique takes advantage of a heat-shock-inducible FLP re-
combinase (29) to produce an active tubulin–lacZ reporter gene
by driving recombination between two complementary inactive
alleles (28). In this way, mitotically active cells in the germarium,
including SSCs, are labeled by lacZ expression and visualized by
staining for ?-galactosidase proteins. This technique also labels
early follicle progenitor cells that are mitotically active in the
germarium. However, because it takes 3–4 days for marked
follicle progenitor cells to migrate from the germarium, all
Fig. 1.
(A) A cross-section diagram of a Drosophila germarium. (B) A confocal section
of a germarium containing a marked lacZ?SSC labeled for LacZ (red), Fas3
(green), and nuclei (blue). One marked lacZ?SSC (arrow) is evident in the
middle of germarium by its low Fas3 expression and the production of LacZ-
positive follicle cells (one indicated by arrowhead). (C) A germarium contain-
ing two marked lacZ?IGS cells labeled for LacZ (red), Fas3 (green), and nuclei
(blue). Two lacZ?IGS cells (arrows) with low levels of Fas3 expression are
apparent,butLacZ-positivefolliclecellsarenotpresentintheovariole.(Dand
E) A germarium carrying a marked lacZ?SSC labeled for DE-cadherin (green),
LacZ (red), and nuclei (blue). (D) The SSC is identified by its location in the
middle of the germarium and capacity to generate LacZ-positive follicle cells
(oneindicatedbyarrowhead).(E)Theenlargedarea(dottedrectangleshown
in D) showing the existence of DE-cadherin-positive foci (arrowheads) be-
tween a marked lacZ?SSC and neighboring IGS cells. (F and G) A germarium
carrying a marked lacZ?SSC labeled for Arm (green), LacZ (red), and nuclei
(blue). (F) The SSC is identified by its location in the middle of the germarium
and capacity to generate LacZ-positive follicle cells (one indicated by arrow-
head). (G) The enlarged area (dotted rectangle shown in F) showing the
existence of Arm-positive foci (arrowheads) between a marked SSC and
neighboring IGS cells. CB, cystoblast; CPC, cap cell; DCs, developing cysts; FC,
follicle cell; FS, fusome; SS ? spectrosomes; TF ? terminal filament cell. (Bars
in B and C represent 10 ?m.)
DE-cadherin and Arm in the interface between SSCs and IGS cells.
14814 ?
www.pnas.org?cgi?doi?10.1073?pnas.232389399 Song and Xie

Page 3

marked, transient follicle cells will completely exit the germa-
riumwithina1-weekperiod.Therefore,theonlymarkedsomatic
follicle cells in the germarium 1 week after clone induction must
be produced by marked SSCs. The ovaries from females 1 week
after the last heat-shock treatment were harvested and immu-
nolabeled for LacZ and Fasciclin III (Fas3). In some germaria,
one or two marked somatic cells were observed in the middle of
the germarium with marked follicle cells present in the same
germarium (Fig. 1B). Discrimination between these two cell
types is made possible by the expression of Fas3, which is
expressed at high levels in somatic follicle cells but at much lower
levels in SSCs (16). We concluded that the marked somatic cells
in the middle of the germarium with low Fas3 expression were
SSCs. While in other germaria, we observed lacZ-positive so-
matic cells located in the middle of the germarium that had low
levels of Fas3 expression; however, there were no lacZ-positive
follicle cells in the germarium or early egg chambers of the same
ovariole (Fig. 1C). We concluded that these labeled cells are
most likely inner germarial sheath (IGS) cells in the same
locationasSSCs.Therefore,weidentifiedmarkedSSCsbasedon
their level of Fas3 expression, location, and production of
marked follicle cells.
To study the expression of DE-cadherin and Arm in SSCs, the
ovaries from the treated females were immunolabeled for LacZ
and DE-cadherin or for LacZ and Arm. DE-cadherin proteins
accumulated at high levels on several contact sites between inner
sheath cells and SSCs, and among early follicle cells at their
contact sites (Fig. 1 D and E). DE-cadherin proteins also
accumulated at the contact sites between germ-line cysts and
IGS cells, and between germ-line cells and follicle cells. DE-
cadherin-mediated cell adhesion requires Arm, the Drosophila
homolog of ?-catenin, which was consistently present at the
contact sites between SSCs and IGS cells, between germ-line
cells and follicle cells, and among follicle cells themselves (Fig.
1 F and G). The presence of DE-cadherin-mediated cell adhe-
sion molecules between IGS cells and SSCs suggests a potential
role in SSC anchorage.
DE-cadherin Is Essential for Anchoring SSCs to Their Niches. To
investigate whether DE-cadherin-mediated cell adhesion is es-
sential for maintaining stem cells in their niche, we used
FLP-mediated FRT recombination to genetically mark SSCs in
the adult Drosophila ovary and to disrupt the function of shg in
the marked SSC clones. Marked SSC clones were identified by
thelossofexpressionofanarm–lacZtransgene(thefusionofthe
arm promoter and the bacterial lacZ gene) from follicle cells in
the germarium (Fig. 2A). When this technique is used, the only
lacZ?folliclecellclonespresent1weekaftercloneinductionwill
consist of mutant SSCs and their progeny from the last 3–4 days
(8, 11). With this marking system, we can deduce the loss of
marked shg SSCs due to differentiation by the disappearance of
mutant lacZ?follicle cells in the germaria. Furthermore, the
position in which the most recent lacZ?mutant follicle cells are
found on egg chambers in the ovariole also indicates when the
marked SSCs were lost.
To determine the importance of DE-cadherin-mediated ad-
hesion in maintaining SSCs in their niches, we used two shg
alleles, the deletion allele (shgR69) and a weak allele (shg10469). As
a control, 46% of marked lacZ?wild-type SSCs observed during
the first week after clone induction were maintained for 2 more
weeks based on the presence of lacZ?follicle cell clones in the
germarium (Fig. 2 A and B, and Table 1), suggesting that the
half-life of SSC is about 3 weeks. This finding is consistent with
previous results (15). Three weeks after clone induction, marked
lacZ?follicle cell patches derived from marked lacZ?wild-type
SSCs were still frequently observed. In some germaria, only
marked lacZ?follicle cells were present, suggesting that non-
clone (lacZ?) SSCs were lost and?or replaced by marked lacZ?
SSCs (Fig. 2B). In contrast, 8.9% of marked lacZ?shg10469SSCs
observedduringthefirstweekweremaintainedfor2moreweeks
(Fig. 2C and Table 1), suggesting that shg10469mutant SSCs were
severely destabilized. Most of the mutant SSC clones were lost
during the initial 2-week period (Fig. 2 D and E). Mutant shg10469
oocytes were always localized to the posterior end of the egg
chambers, indicating that homozygous shg10469germ-line clones
have no obvious defects in the DE-cadherin-based cell adhesion
between oocytes and follicle cells (data not shown). Possibly, the
shg10469mutation induced by a P element primarily affects the
Fig.2.
(red), Hts (green), and nuclei (blue). Marked lacZ?SSC clones or follicle cell
clones are outlined. (A) A germarium carrying a wild-type SSC clone that is
evident by the presence of LacZ-negative follicle cells. (B) A 3-week-old
marked lacZ?wild-type SSC clone showing that all follicle cells are LacZ-
negative. In this germarium, the lacZ?SSC was lost and?or replaced by a
marked lacZ?SSC. (C) A 3-week-old germarium carrying only marked lacZ?
follicle cells from a shg10469SSC clone. This has also resulted from the loss
and?or replacement of a lacZ?wild-type SSC by a marked lacZ?SSC. (D) A
germarium carrying no marked lacZ?shg10469SSCs with marked lacZ?follicle
cell patches on egg chambers (E) 3 weeks after clone induction. (F) A 1-week-
old shgR69SSC clone. (G) A germarium with no shgR69SSCs but having marked
lacZ?follicle cell patches on egg chambers (H) 2 weeks after clone induction.
The marked mutant follicle cells in E and H were identified by loss of expres-
sion of cytoplasmic LacZ and increased expression of nuclear LacZ because
shg10469and shgR69still have a nuclear LacZ gene inserted in the locus capable
of expressing it in late follicle cells. All of the germaria are shown at the same
scale. (Bar in A represents 10 ?m.)
DE-cadherinanchoringSSCsintheirniches.GermarialabeledforLacZ
Table 1. DE-cadherin is required for keeping SSCs in their niches
Genotypes
1 week,
%
2 weeks,
%
3 weeks,
%
WT
shg10469
shgR69
54.8 (104)
21.3 (141)
7.2 (250)
42.2 (225)
1.9 (216)
0.3 (286)
25.3 (217)
1.9 (263)
0.0 (360)
Numbers indicate the percentage of germaria carrying a marked SSC at a
given time point ? the number of germaria carrying a marked SSC?total
germaria examined ?100%. Numbers in parentheses indicate total germaria
examined for each genotype at that time point.
Song and XiePNAS ?
November 12, 2002 ?
vol. 99 ?
no. 23 ?
14815
CELL BIOLOGY

Page 4

expression of DE-cadherin in somatic cells but not in germ-line
cellsoftheDrosophilaovary.Asshown(24,25),weobservedthat
shgR69mutant oocytes frequently failed to localize to the pos-
terior end of the egg chambers (data not shown). Interestingly,
we were able to produce marked, mutant shgR69SSC clones only
at a much lower frequency, 13% of the control under the same
conditions (Fig. 2F, Table 1). Two weeks after clone induction,
there were no germaria that carried mutant shgR69SSC clones,
but some ovarioles had mutant follicle cell patches on egg
chambers (Fig. 2 G and H, Table 1). Results from our mutant
clonal analysis clearly demonstrate that shg is essential for
maintaining SSCs in their niche.
DE-Cadherin-Mediated Signaling Does Not Play an Important Role in
the Proliferation of Somatic Follicle Cells. DE-cadherin-mediated
cell adhesion is often involved in regulating other intercellular
signaling events in many different systems (21). It has been
shown that signaling from anterior somatic cells is important for
SSC division (16, 17). It is very difficult to measure the division
rates of SSCs because their immediate progeny cannot be
effectivelyquantified.Instead,westudiedwhetherDE-cadherin-
mediated adhesion is important for the proliferation of somatic
follicle cells by comparing the sizes of marked lacZ?wild-type
and shg mutant follicle cell clones in the germaria and egg
chambers. To ensure that the removal of DE-cadherin disrupts
its accumulation between follicle cells and germ-line cells, we
immunolabeled the ovaries for LacZ and DE-cadherin or for
LacZ and Arm one week after heat-shock treatment. As ex-
pected, the removal of DE-cadherin from follicle cells suffi-
ciently disrupted its accumulation between germ cells and
follicle cells (Fig. 3 A and B). The sizes of shgR69mutant follicle
cellclonesweresimilartothatofmarkedwild-typefollicleclones
(compare Fig. 3 A and B with Fig. 2A). To further confirm
whether the accumulation of Arm in the junctions is also
affected, we examined its expression in shgR69mutant follicle
clones. As shown (24, 25), the accumulation of Arm in the
junctions between mutant follicle cells and germ cells in the
germarium was disrupted (Fig. 3C). Sometimes, these shgR69
mutant follicle cells failed to integrate into the follicle cell
epithelium of egg chambers and accumulated between egg
chambers (Fig. 3D). Although shg mutant follicles have obvious
defects in interacting with other cells, there were no obvious
defects in follicle cell proliferation.
DE-Cadherin-Mediated Cell Adhesion Is Also Required for Anchoring
SSCs During Development. We have previously demonstrated that
DE-cadherin-mediated cell adhesion is required for recruiting
GSCs and keeping GSCs to their niches before adulthood (26).
It is possible that this adhesion process is also required for SSC
recruitmentandanchoragetotheirnichesbeforeadulthood.The
establishment of SSC identity occurs during pupation, later than
the late-third-instar stage in which GSCs start to establish (26).
However, no molecular markers or reliable positional informa-
tion are available for identifying SSC precursors and SSCs
themselves during this early developmental process. Therefore,
we could not effectively determine the role of DE-cadherin in
SSC recruitment, but rather focus on investigating whether
DE-cadherin plays any role in maintaining SSCs before adult-
hood. To determine the role of DE-cadherin in maintaining
SSCs before adulthood, we randomly removed DE-cadherin
from the mitotically active preSSC population by using the
FLP-mediated FRT recombination technique. Randomly
markedlacZ?wild-type(asacontrol),mutantshg10469andshgR69
preSSC populations generated during the late-third-instar stage
were allowed to go through normal ovarian developmental
processes and to be further integrated into individual ovarioles.
To assay the ability of these labeled preSSCs to be maintained
in their niches, we examined the presence of marked lacZ?
follicle cells in 1- to 2-day-old germaria. Even in the control, we
expectedthatonlyafractionofthemarkedlacZ?preSSCswould
be incorporated into the germarium and become SSCs; however,
if DE-cadherin-mediated cell adhesion was important for inter-
actions between SSCs and their niche cells before adulthood, we
expected that mutant shg preSSCs would have a lower efficiency
of maintenance in their niches as compared with wild-type
clones.
The presence of marked lacZ?SSCs was determined by the
presence of marked lacZ?follicle cells in the germarium because
Fig. 3.
genefromSSCs.(AandB)One-week-oldgermarialabeledforLacZ(red),DE-cadherin(green),andnuclei(blue).Bothpanelsillustratetheinterface(dottedline)
between lacZ?mutant shgR69follicle cells and germ cells, and the interface (solid line) between lacZ?wild-type follicle cells and germ cells. The removal of
DE-cadherin from follicle cells can disrupt its accumulation in the interface between follicle cells and germ cells. (C) One-week-old germarium labeled for LacZ
(red), Arm (green), and nuclei (blue). The dotted line indicates the interface between lacZ?mutant shgR69follicle cells and germ cells, and the solid line indicates
the interface between lacZ?wild-type follicle cells and germ cells. The removal of DE-cadherin from follicle cells disrupts Arm accumulation in the interface
betweenfolliclecellsandgermcells.(D)PartofanovariolelabeledforLacZ(red),Arm(green),andnuclei(blue)showingapatchoflacZ?shgR69mutantfollicles
(outlined) in the region between a germarium and a stage-2 egg chamber. All of the micrographs are shown at the same scale. (Bar in A represents 6.7 ?m.)
Diminished accumulation of DE-cadherin and Arm in the junctions between germ cells and mutant shg follicle cells after removal of a functional shg
14816 ?
www.pnas.org?cgi?doi?10.1073?pnas.232389399 Song and Xie

Page 5

labeledSSCprecursorsthatfailedtobemaintainedinnicheshad
already moved out of the germarium during pupation. As a
control, labeled, wild-type follicle patches were present in 34.3%
of germaria (n ? 300, Fig. 4A). The labeled shg10469mutant
follicle patches were present in 18.7% of germaria (n ? 246, Fig.
4B); however, labeled follicle cell patches homozygous for the
deletion allele, shgR69, were detected in only 0.8% of germaria
(n ? 249). We induced wild-type and mutant shgR69clones under
the same conditions (in the same experiment, 39% and 41% of
ovarioles carried marked lacZ?wild-type and mutant shgR69
cysts, respectively); therefore, the difference in the percentages
of germaria carrying mutant SSC clones reflects the role of
DE-cadherin in the interactions between SSCs and their niche
cells before adulthood. The presence of labeled follicle patches
in the germarium indicates the presence or recent presence of
labeled SSCs in the same ovariole. Ovarioles were observed in
which mutant follicle cell clones were present in egg chambers;
however, labeled follicle patches were absent in the germarium
(Fig. 4 C and D), indicating that labeled SSC precursors fail to
be integrated into niches as SSCs or that labeled SSCs are
integrated into niches and are lost during development. Cur-
rently, we cannot effectively separate the establishment and
maintenance role of DE-cadherin in SSCs; however, our results
can be accounted for at least by its maintenance role before
adulthood. We conclude, therefore, that disrupting the expres-
sion of DE-cadherin in preSSCs reduces their ability to be
maintained in their niches during development.
Discussion
In this study, we show that DE-cadherin accumulates at high
levels at the contact sites between SSCs and their neighboring
IGS cells. We demonstrate that DE-cadherin is essential for
holding SSCs against IGS cells and preventing SSCs from
moving away from their niche and, therefore, from differenti-
ating. Terminal filament cells?cap cells express the gene hh,
known for its role in maintaining SSC identity (16, 17). There-
fore, we propose that one function mediated by the DE-
cadherin-based cell adhesion is to ensure that SSCs stay close to
importantSSCgrowthfactors,suchasHh,tomaintaintheirstem
cell characteristics.
Different Stem Cell Types May Have Different Life Spans. Because
stem cells have the capacity to self-renew, it has been thought
that individual stem cells can live as long as host organisms.
Because of the difficulties in quantifying the self-renewal prop-
ertiesofstemcells,thisimportantquestionhasnotbeencarefully
tested in many stem cell systems. Fortunately, the life spans of
GSCs and SSCs, can be analyzed quantitatively in the Drosophila
ovary because we can mark both stem cell types and follow their
activities for a long period. The differences in the limited life
spans of GSCs (half-life of 4–5 weeks) and SSCs (half-life of 2–3
weeks) (refs. 11, 15, and 16, and this study) may reflect the
various capacities of diverse stem cells to self-renew. It currently
remains unclear what determines these self-renewal capacities
for different stem cells. One possibility is that the lost stem cells
are defective in their ability to receive critical niche cell signals.
For Drosophila ovarian GSCs, any defect in receiving the dpp
signal shortens their life span (11). For Drosophila ovarian SSCs,
the failure to receive the hh signal also results in a shortened life
span (16). Another hypothesis is that over time, stem cells
undergo intrinsic changes that shorten their life span. For
example,O-2oligodentricyteprecursorcellsundergosenescence
after a certain number of cell divisions due to continuous
accumulation of cell-cycle inhibitors, such as p21 (30). Damage
in DNA and protein machinery represent other possible mech-
anisms that could contribute to this stem cell instability. It will
be interesting to investigate what intrinsically determines GSC
and SSC life span in the Drosophila ovary.
Although stem cells in the Drosophila ovary have a limited life
span, the ovary can function well beyond the life span of
individual stem cells. One of the solutions is to replace lost stem
cells with daughter cells of other stem cells that reside in the
same or nearby niche. To repopulate an empty GSC niche in the
Drosophila ovary,adaughtercellofanotherstemcellinthesame
niche occupies the same position where a lost GSC normally
resides (12). Two pieces of independent evidence suggest that
lost SSCs can be replaced (15, 16). Our study also supports these
published results. The replaced SSCs must be derived from the
progeny of other SSCs within the same germarium. It needs to
be determined what mechanisms are used to replace lost SSCs
in the Drosophila ovary.
ExistenceofanSSCNiche.IntheDrosophilaovary,GSCshavebeen
shown unequivocally to be located in a niche (9–14). The
neighboring somatic cells, namely terminal filament, cap cells,
andIGScells,formanasymmetricalnichestructureinwhichtwo
daughter cells of a GSC interact with different cell types and,
therefore, adopt two completely different cell fates. The number
of cap cells directly correlates with the number of GSCs,
suggesting that cap cells are a major component of the niche.
Consistent with this idea, the daughter cell that stays in contact
with cap cells remains a stem cell, whereas the daughter cell that
fails to contact cap cells and is located closer to IGS cells
differentiates. Because they are located in the middle of the
germarium where IGS cells also reside, SSCs are in close contact
with a posterior group of IGS cells. After SSC division, one of
the two daughter cells remains in its original position as a stem
cell, whereas the other daughter cell differentiates and moves
away posteriorly. This study provides evidence supporting the
existence of SSC niches. The removal of DE-cadherin from SSCs
prevents them from anchoring to IGS cells and, thus, results in
SSC loss. It appears that close contact with IGS cells is essential
for maintaining SSC identity. These stem cells directly contact
underlying germ-line cysts and posterior IGS cells. Germ-line
cysts are transiently present in the germarium, whereas IGS cells
are permanent residents. Therefore, it is most likely that IGS
cells function as a niche for SSCs. However, it remains unclear
how the SSC niche is organized within one germarium.
Cadherin-Mediated Cell Adhesion Is a Common Strategy to Anchor
Stem Cells to Their Niches. Our previous research demonstrated
that DE-cadherin-mediated cell adhesion is required for anchor-
ing GSCs in the Drosophila ovary (26). Interestingly, DE-
Fig. 4.
maintenance before adulthood. Ovaries of the females that developed from
heat-shock-treated late third-instar larvae labeled for LacZ (red), Hts (green),
and nuclei (blue). (A) A wild-type germarium showing a marked lacZ?SSC
clone(outlined).(B)AgermariumshowingamarkedlacZ?mutantshg10469SSC
clone (outlined). (C) A germarium showing no mutant shgR69clones with a
lacZ?mutant follicle clone on late egg chambers (D). All of the micrographs
are shown at the same scale. (Bar in A represents 10 ?m.)
The requirement of DE-cadherin-mediated cell adhesion for SSC
Song and XiePNAS ?
November 12, 2002 ?
vol. 99 ?
no. 23 ?
14817
CELL BIOLOGY