1330 Research Article

IntroductionOvarian Graafian follicles are comprised of somatic cells thatexhibit both endocrine and developmental functions. Theendocrine functions, including steroidogenesis, are carried outprimarily by the mural granulosa cells (Fig. 1). The cumuluscells, which are closely associated with the oocyte, promoteoocyte growth and developmental competence throughcomplex bi-directional interactions with the oocyte. In fact,oocyte stimulation of cumulus cell function is crucial for itsown subsequent development (Eppig, 2001; Hussein et al.,2006; Su et al., 2004). To fulfil both endocrine anddevelopmental functions, follicle growth must involve thecoordinated development of both cumulus and mural granulosacells, as a deficiency in either hormone production ordevelopment of oocyte competence would impair fertility.

Although granulosa cells express many mRNA transcriptsand proteins in common, the mural and cumulus cells eachexpress a subset of different transcripts, with markedlydifferent steady state levels (Fig. 1). Preantral granulosa cellsare the common precursors of both mural and cumulus cells.Large preantral follicles (secondary follicles) develop intoantral follicles (tertiary follicles) after forming a fluid-filledantrum under stimulation of follicle stimulating hormone(FSH). Antrum formation physically separates the preantralgranulosa cells into a mural cell compartment along the folliclewall and a cumulus cell compartment surrounding the oocyte(Fig. 1). The mural cells produce increasing amounts ofestrogen, which eventually initiates the surge of luteinizinghormone (LH) from the pituitary resulting in oocytematuration, expansion of the cumulus oophorus and ovulation.The cumulus cells promote oocyte growth and developmentalcompetence (Eppig, 2001; Hussein et al., 2006; Su et al.,2004). A defining feature of cumulus cells is their ability to

undergo expansion following the LH surge. Cumulusexpansion involves the production of a mucified matrix by thecumulus cells that is necessary for ovulation and thereforerequired for fertility (Chen et al., 1993).

The mechanisms responsible for differentiation of preantralgranulosa cells into mural and cumulus cells are not fullydefined, but involve factors secreted by oocytes and FSHproduced by the pituitary. FSH is essential, as follicles fail todevelop beyond the early antral stage in mice lacking FSH(Kumar et al., 1997) or FSH receptors (Dierich et al., 1998).FSH stimulates expression of Lhcgr (luteinizing hormonereceptor) mRNA in mural cells (Fig. 1). Lhcgr mRNA is a keymarker of mural cell differentiation (Lei et al., 2001). Bothcumulus and mural cells express FSH receptors (Richards andMidgley, 1976); however, oocyte suppression of Lhcgr incumulus cells restricts Lhcgr mRNA to mural cells (Eppig etal., 1997). In addition to suppressing Lhcgr mRNA, paracrinefactors secreted by oocytes have profound effects on folliclegrowth and development both before and after the LH surge.For example, absence of oocyte-derived growth anddifferentiation factor 9 (GDF9) results in failure to progressbeyond the early preantral stages of follicle development(Dong et al., 1996; Elvin et al., 1999). Lack of bonemorphogenetic protein 15 (BMP15), another oocyte-secretedprotein, leads to defective cumulus cell differentiation andcumulus expansion (Su et al., 2004; Yan et al., 2001; Yoshinoet al., 2006). Oocyte-derived factors are also required to enablecumulus expansion after the preovulatory LH-surge (Buccioneet al., 1990; Diaz et al., 2006). The major impact that oocyteshave in directing follicle development was demonstrated byrecombining mid-growth stage oocytes from large preantralfollicles with follicular somatic cells from newborn ovaries.The resulting re-aggregated ovary not only forms follicles

The two principal functions of ovarian follicles aredevelopmental and endocrine. The cumulus cellssurrounding the oocyte are specialized to serve thedevelopment of the oocyte and steroidogenesis is a principalrole of mural granulosa cells that line the follicle wall. Thefindings in this report demonstrate that oocytectomy ortreatment with an inhibitor of SMAD2/3 activation resultsin decreased cumulus marker mRNA transcript levels andallows FSH to induce mural marker transcripts in cumuluscells. In addition, SMAD2/3 signaling is involved inenabling cumulus expansion and EGF-induced increases in

Ptx3, Ptgs2 and Has2 mRNA levels. By contrast, follicle-stimulating hormone (FSH) stimulated expression of muraltranscripts, but suppressed levels of cumulus transcripts.Thus, FSH and oocyte-stimulated SMAD2/3 signalingestablish opposing gradients of influence in the follicle.These specify the mural and cumulus granulosa cellphenotypes that are pivotal for appropriate endocrinefunction and oocyte development.

Key words: Ovarian follicle, Oocytes, FSH, SMAD2/3

Summary

Oocytes determine cumulus cell lineage in mouseovarian folliclesFrancisco J. Diaz, Karen Wigglesworth and John J. Eppig*The Jackson Laboratory, 600 Main Street, Bar Harbor, ME 04609, USA*Author for correspondence (e-mail: john.eppig@jax.org)

Accepted 13 February 2007Journal of Cell Science 120, 1330-1340 Published by The Company of Biologists 2007doi:10.1242/jcs.000968

Jour

nal o

f Cel

l Sci

ence

1331Oocytes promote cumulus phenotype

when transplanted to recipient mice, but the chronologicallymore advanced oocytes actually accelerate the development ofthe newborn somatic cells (Eppig et al., 2002). Theseobservations provided strong evidence that oocyte-derivedfactors are important mediators of follicle growth, cumulus celldifferentiation and cumulus expansion.

Oocyte factors and FSH signal from opposite compartments.The oocyte signals emanate from a central follicular location,whereas FSH signals arrive from outside the follicle. This leadsto a gradient of expression of mural and cumulus transcripts.Levels of the mural marker transcript Lhcgr are highest nearthe basal lamina and lowest in the cumulus cells (Meduri et al.,1992) (Fig. 1). Conversely, expression of Amh mRNA, whichis stimulated by oocytes (Salmon et al., 2004) is highest in thecumulus cells (Baarends et al., 1995) (Fig. 1). Oocyte-secretedparacrine factors are also required to enable increases in Has2,Ptgs2, Ptx3 and Tnfaip6 transcripts during cumulus expansion(Diaz et al., 2006). Each of these transcripts is absolutelyrequired for cumulus expansion as the phenotype of nullmutations in the Ptgs2, Ptx3 or Tnfaip6 genes or inhibition ofHAS2 activity severely compromises cumulus expansion(Chen et al., 1993; Fulop et al., 2003; Ochsner et al., 2003a;Ochsner et al., 2003b; Varani et al., 2002).

The signaling pathways activated by oocytes and requiredfor cumulus cell function remain largely unclear. However,pathways activated by TGF�-related proteins, such as GDF9and BMP15, are probably crucial for mediating the effect ofoocytes on granulosa cells. BMP15 and GDF9 activateSMAD1/5/8 and SMAD2/3 signaling pathways, respectively,in granulosa cells (Moore et al., 2003; Roh et al., 2003).Signaling through either SMAD1/5/8 or SMAD2/3 pathways

requires that these receptor-regulated SMADs bind thecommon SMAD, SMAD4. Conditional deletion of the Smad4gene in granulosa cells of preantral follicles leads to severedefects in subsequent follicle development and differentiationowing to a lack of SMAD4-dependent signaling (Pangas et al.,2006). Moreover, mice deficient in Smad3 show increased ratesof cell death and abnormal cell differentiation (Tomic et al.,2004). Thus, SMAD signaling is crucial for proper follicledevelopment. More recently, the BMP receptor type II(BMPR2) extracellular domain and an inhibitor of SMAD2/3activation, SB431542, were used to block oocyte-stimulatedproliferation of mural and cumulus cells (Gilchrist et al., 2006).However, the exact TGF�-related signaling pathways activatedin cumulus cells by oocytes that lead to cumulus celldifferentiation and expansion remain undefined. The workreported here tests the hypothesis that oocyte-stimulatedSMAD2/3 signaling mediates the ability of oocytes tostimulate the cumulus cell phenotype during antral follicledevelopment.

ResultsOocyte regulation of SMAD2 activation and localizationduring follicle developmentA strong pSMAD2 signal was detected byimmunofluorescence staining of tissue sections in preantralgranulosa cells of small and large preantral follicles (Fig.2A,B). In antral follicles, pSMAD2 localized more strongly tocumulus cells than to mural cells (Fig. 2C). Surprisingly,treatment with hCG in vivo resulted in decreased pSMAD2levels by 8 hours (Fig. 2D). No pSMAD2 signal was detectedin ovary sections incubated with secondary antibody alone or

Preantral Follicle

Antral Follicle

Proliferation &Differentiation LH Surge

Cumulus Expansion

Slc38a3 Amh

Cumulus Marker Transcripts

Rel

ativ

e Le

vels

of L

hcgr

mR

NA

Rel

ativ

e Le

vels

of C

yp11

a1 m

RN

A

Rel

ativ

e Le

vels

of C

d34

mR

NA

Lhcgr Cyp11a1 Cd34

00.010.020.030.040.050.060.070.080.090.1

0

0.02

0.04

0.06

0.08

0.1

0.12

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0

0.1

0.2

0.3

0.4

0.5

0.6

0

0.1

0.2

0.3

0.4

0.5

0.6

0

0.005

0.01

0.015

0.02

0.025Ar

COC Mural COC Mural COC Mural COC Mural

* *

* * *

*

COC Mural COC Mural

Rel

ativ

e Le

vels

of S

lc38

a3 m

RN

A

Rel

ativ

e Le

vels

of A

mh

mR

NA

Rel

ativ

e Le

vels

of A

r m

RN

AMural Marker Transcripts

Post- LH surgePre-LH surge

MuralLhcgr

Cyp11a1Cd34

CumulusSlc38a3

AmhAr

Has2Ptgs2Ptx3

Tnfaip6

Fig. 1. Model depicting follicular growth and cumulus expansion. Preantral follicles under the stimulation of gonadotropins and oocyte factorsdevelop into antral follicles. The formation of the antrum separates the granulosa cells into mural and cumulus cells. Mural cells (green) expressLhcgr, Cyp11a1 and Cd34 transcripts more highly than cumulus cells (red), whereas Slc38a3, Amh and Ar transcripts are highly expressed incumulus cells. At ovulation, the LH surge induces cumulus expansion with associated increases in expansion-related transcripts (Has2, Ptgs2,Ptx3 and Tnfaip6). Real-time PCR analysis of mural and cumulus marker transcripts in freshly isolated COCs and mural cells is also shown.Transcripts were normalized to Rpl19 mRNA. *Indicates significant difference by Student’s t-test P<0.05.

Jour

nal o

f Cel

l Sci

ence

1332

when the pSMAD2 antibody was preincubated with a SMAD2phosphopeptide (pSer465/467, Abcam) (Fig. 2E and data notshown). To verify the results observed in tissue sections, wecompared pSMAD2 levels in freshly isolated cumulus-oocytecomplexes (COCs) and mural cells by western blot and foundthat levels of total SMAD2 were similar in mural and COCs,but levels of phosphorylated SMAD2 were higher in COCs

Journal of Cell Science 120 (8)

compared with mural cells (Fig. 2F). The levels of pSMAD2observed in cumulus cells were stimulated by the oocyte.Levels of pSMAD2 detected by western blot were high inuntreated COCs (Fig. 3). Removal of the oocyte greatlyreduced the pSMAD2 signal to an almost undetectable level,which was restored upon co-culture with oocytes. SB431542suppressed pSMAD2 in COCs and OOX cells co-cultured withoocytes (Fig. 3). Thus, the oocyte potently stimulates pSMAD2in cumulus cells.

Specificity of inhibitorsThe effect of SB431542 and SIS3 on activation of varioussignaling molecules in cultured COCs was determined toassess the specificity of the inhibitors. SB431542 blockedSMAD2 and SMAD3 activation in COCs, but had no effect ontotal SMAD2 levels or levels of pSMAD1/5/8 (Fig. 4A). Bycontrast, SIS3 blocked pSMAD3 activation without affectingtotal SMAD2, pSMAD2 or pSMAD1/5/8 levels (Fig. 4A). Theefficacy of SB431542 was determined by a dose-responseexperiment where COCs were treated with concentrations ofSB431542 ranging from 0.1 to 100 �M (Fig. 4B). Theminimum dose that effectively blocked pSMAD2 levels inCOCs was 3 �M (Fig. 4B). Although SB431542 and SIS3 arehighly specific for inhibiting the kinase activity of ALK-4,ALK-5 and ALK-7, the possibility exists that these moleculescould have effects on other untested cellular kinases or evenother non-kinase proteins.

Suppression of mural transcripts by oocytesTranscripts encoding the LH receptor, Lhcgr, the steroidogenic

Fig. 2. Immunolocalization of pSMAD2 in preantral (A), transitionalpreantral-antral (B) and antral follicles before (C) and after (D) 8hours hCG (5 IU) treatment in vivo. (E) No staining was observedwhen the primary antibody was pre-incubated with pSMAD2phosphopeptide. Arrows indicate cumulus-oocyte complexes. Bars,50 �m. (F) Western blot immunostained for pSMAD2, total SMAD2and ACTB in freshly isolated COCs and mural cells from antralfollicles.

Fig. 3. Oocytes stimulate pSMAD2 in cumulus cells. Western blotimmunostained for pSMAD2 and ACTB in COCs cultured in mediaalone or with SB431542 (10 �M) and in OOX cells cultured alone,OOX cells co-cultured with FGO and co-cultured OOX cells treatedwith SB431542 (10 �M) for 15 hours. Graphical representation ofrelative pixel intensity normalized to ACTB pixel intensity is alsoshown. a,bP<0.05, significant differences.

Jour

nal o

f Cel

l Sci

ence

1333Oocytes promote cumulus phenotype

enzyme P450 side chain cleavage, Cyp11a1, and the immunemolecule Cd34 were higher in mural than in cumulus cells asmeasured by real-time PCR (Fig. 1). We used these transcriptsas markers of the pre-LH-surge mural cell phenotype. Oocyte-derived factors suppress expression of Lhcgr mRNA incumulus cells (Eppig et al., 1998; Eppig et al., 1997). Wehypothesized that oocytes might suppress Lhcgr, Cyp11a1 andCd34 through activation of the TGF� signaling pathway. Totest this idea, the effects of oocyte extirpation (OOX), FSH(100 ng/ml), SB431542 (an inhibitor of SMAD2 and SMAD3activation, 10 �M) and SIS3 (a SMAD3-specific inhibitor, 20�M) on expression of mural transcripts in COC and mural cellswere measured. FSH was required to maintain expression ofmural cell markers in vitro (Fig. 5A,B). Fully-grown, germinalvesicle (GV) stage oocytes (FGO) potently suppressed thesteady-state levels of the three mural transcripts that had beenmaintained by FSH (Fig. 5A,B). The suppression of muraltranscripts by oocytes in mural cells was blocked by treatmentwith SB431542, but not SIS3 (SMAD3 inhibitor) suggestingthat SMAD3 signaling was not involved (Fig. 5B). In COCs,we hypothesized that oocytes would also suppress muraltranscripts through TGF� signaling pathways. Removal of theoocyte (OOX) allowed FSH to stimulate increased expressionof mural transcripts in cumulus cells (Fig. 6A). Co-culture withoocytes prevented FSH-induced increases of mural transcripts,but treatment with SB431542, but not SIS3, blocked the abilityof oocytes to suppress mural transcript induction by FSH (Fig.6B,C), indicating that oocyte suppression of mural transcriptsin cumulus cells is likely via pSMAD2.

Oocytes promote cumulus transcriptsOocyte-derived factors promote expression of Slc38a3 (Eppig

et al., 2005) and Amh mRNA (Salmon et al., 2004) in cumuluscells. These transcripts along with Ar (androgen receptor)mRNA are more highly expressed in COCs than mural cells(Fig. 1). We hypothesized that oocytes may promote expressionof the cumulus transcripts Slc38a3, Amh and Ar though theTGF� signaling pathway. To test this idea, the effects of oocyteextirpation (OOX), FSH (100 ng/ml), SB431542 (an inhibitorof SMAD2 and SMAD3 activation, 10 �M) and SIS3 (aSMAD3-specific inhibitor, 20 �M) on expression of cumulustranscripts was tested. OOX resulted in a decrease in cumulusmarkers (Slc38a3, Amh and Ar mRNA), which were restoredby oocyte co-culture (Fig. 7A) indicating that oocyte factorsstimulate these transcripts. SB431542 but not SIS3 (Fig. 7B,C)caused a decrease in cumulus marker transcript levels.Surprisingly, FSH completely suppressed Slc38a3 anddecreased Ar transcript levels by ~50%, even in the presenceof oocytes (Fig. 7B). Mural cells express lower levels ofSlc38a3, Amh and Ar mRNAs (Fig. 1). Since oocytes canstimulate pSMAD2 levels in mural cells (Gilchrist et al., 2006)

Fig. 4. Specificity of inhibitors. (A) Western blot of lysates collectedfrom COCs treated with media only, SB431542 (10 �M) or SIS3 (20�M) for 4 hours and immunostained for pSMAD2, total SMAD2,pSMAD3, pSMAD1/5/8 and ACTB proteins. (B) Western blotimmunostained for pSMAD2 and ACTB of COCs treated withincreasing concentration of SB431542 (0.1 �M to 100 �M) for 4hours.

Rel

ativ

e Le

vels

of L

hcgr

mR

NA

Rel

ativ

e Le

vels

of C

yp11

a1 m

RN

AR

elat

ive

Leve

ls o

f Cd3

4 m

RN

A

00.020.040.060.080.1

0.120.140.160.180.2

00.10.20.30.40.50.60.70.80.91

0

0.01

0.02

0.03

0.04

0.05

0.06

Lhcgr

Cyp11a1

Cd34

FSH

FGO FGOSB431542

FSH

FSH

cc

c

c

bb

a

a

a

b

b

a

0

0.05

0.1

0.15

0.2

0.25

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0.0035

Rel

ativ

e Le

vels

of L

hcgr

mR

NA

Rel

ativ

e Le

vels

of C

yp11

a1 m

RN

AR

elat

ive

Leve

ls o

f Cd3

4 m

RN

A

b

aa

a

aa

aa

a

b

b

B. Mural-SIS3

a

A. Mural-SB431542

Control Control

FGO FGOSB431542

Control Control

FGO FGOSB431542

Control Control FGO FGOSIS3

Control Control

FSH

FGO FGOSIS3

Control Control

FSH

FGO FGOSIS3

Control Control

FSH

Lhcgr

Cyp11a1

Cd34

Fig. 5. Regulation of mural transcripts Lhcgr, Cyp11a1 and Cd34.Effect of SB431542 (10 �M) (A) and SIS3 (20 �M) (B) on levels ofmural transcripts in mural cells cultured alone or with FGOs (2oocytes/�l) for 12 hours. Transcripts were measured by real-timePCR and normalized to Rpl19 mRNA. abcP<0.05, significantdifferences.

Jour

nal o

f Cel

l Sci

ence

1334

and pSMAD2 appears important for expression of cumulustranscripts (Fig. 7B), we hypothesized that oocytes wouldstimulate cumulus transcripts in mural cells. Levels of Slc38a3and Ar were not induced by oocytes in mural cells. However,oocytes did stimulate Amh transcript levels in mural cells,which was blocked by SB431542, but not SIS3 (Fig. 8A,B).

Cumulus expansion requires SMAD signalingCumulus expansion requires one or more oocyte-secretedfactors (Vanderhyden et al., 1990). SB431542 and SIS3 wereused to test the effect of blocking both SMAD2 and SMAD3or SMAD3 only on cumulus expansion. As expected,expansion was stimulated by EGF treatment in vitro (Fig. 9A).Treatment with SB431542 completely blocked cumulusexpansion (Fig. 9A). By contrast, expansion was not severelyaffected by SIS3 treatment, but some cells did attach to theculture dish suggesting that expansion was not completelynormal (white arrows Fig. 9A). The minimum effective doseof SB431542 that blocks expansion was 1-3 �M (Fig. 9B) andis similar to the minimum dose that blocks pSMAD2 levels inCOC (Fig. 4B).

Effect of SB431542 and SIS3 on expansion-relatedtranscriptsTo better define the molecular effects of SB431542 and SIS3on cumulus expansion, an analysis of cumulus-expansion-related transcripts at several time points during cumulusexpansion was undertaken. EGF stimulated expression ofHas2, Ptgs2, Ptx3 and Tnfaip6 mRNA in COCs with the

Journal of Cell Science 120 (8)

highest level measured occurring 6 hours after EGF treatment(Fig. 9C). The effect of the inhibitors on EGF-inducedexpression of expansion-related transcripts was dependent onthe specific transcript and on the time after EGF treatment.Levels of Ptgs2 mRNA were not affected by SIS3, whereastreatment with SB431542 suppressed levels of Ptgs2 mRNA at8 and 10 hours (Fig. 9C). Both SIS3 and SB431542 reducedlevels of Ptx3 mRNA by 75-95% (Fig. 9C). Levels of Tnfaip6mRNA were not affected by either SB431542 or SIS3 at 6 or8 hours after EGF, but by 10 hours, both inhibitors suppressedTnfaip6 mRNA levels (Fig. 9C). Levels of Has2 mRNA werenot affected by SIS3, but were completely suppressed bySB431542 at all times examined (Fig. 9C). Thus, regulation ofexpansion transcripts is likely controlled through a complicatednetwork of pathways which includes pSMAD2 and pSMAD3signaling.

Interaction of MAPK and SMAD signaling pathwaysCumulus expansion requires both activation of MAPK3/1 andMAPK14 and oocyte-stimulated SMAD2/3 signaling.SB431542 can inhibit MAPK14 activation in a cell-free assaysystem (Inman et al., 2002). Therefore, we analyzed the effectof SB431542 on phosphorylation of MAPK3/1 and MAPK14by EGF and the effect of EGF on pSMAD2 levels in order tobegin to identify at what level the MAPK and SMAD signalingpathways interact. EGF stimulated MAPK3/1 and MAPK14phosphorylation in COCs (Fig. 10A). SB431542 treatmentcompletely blocked pSMAD2 levels in COCs, but lack ofSMAD2 phosphorylation did not prevent MAPK3/1 or

Rel

ativ

e Le

vels

of L

hcgr

mR

NA

Rel

ativ

e Le

vels

of C

yp11

a1 m

RN

AR

elat

ive

Leve

ls o

f C

d34

mR

NA

B. COC-SB431542

Lhcgr

Cyp11a1

Cd34

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0.0035

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

COC FSH FSHSB431542

SB4315420

0.005

0.01

0.015

0.02

0.025

0

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Rel

ativ

e Le

vels

of L

hcgr

mR

NA

Rel

ativ

e Le

vels

of C

yp11

a1 m

RN

AR

elat

ive

Leve

ls o

f C

d34

mR

NA

A. COC-OOX

Lhcgr

Cyp11a1

Cd34

COC FSH FSHSB431542

SB431542

COC FSH FSHSB431542

SB431542

FSH

FSH

FSH

Ctrl OOXCtrl OOXFGO

Ctrl OOXCtrl OOXFGO

Ctrl OOXCtrl OOXFGO

b

a

c

a a

b

b

a a

a a a

a a

a a

a a a

a

b

b

b

c

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0.0035

0.004

0.0045

0.005

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0.0035

0.004

0.0045

0.005

Rel

ativ

e Le

vels

of L

hcgr

mR

NA

Rela

tive L

eve

ls o

f C

yp11a1

mR

NA

Rel

ativ

e Le

vels

of C

d34

mR

NA

C. COC-SIS3

COC FSH FSHSIS3

SIS3

COC FSH FSHSIS3

SIS3

COC FSH FSHSIS3

SIS3

Lhcgr

Cyp11a1

Cd34

Fig. 6. (A) Effect of OOX and FSH (100ng/ml) on levels of mural transcripts incumulus cells treated for 12 hours.Effect of SB431542 (10 �M) (B) orSIS3 (20 �M) (C) alone or incombination with FSH (100 ng/ml) onlevels of mural transcripts in COCscultured for 15 hours. Transcripts weremeasured by real-time PCR andnormalized to Rpl19 mRNA. abcP<0.05,significant differences.

Jour

nal o

f Cel

l Sci

ence

1335Oocytes promote cumulus phenotype

MAK14 activation 4 hours after EGF treatment (Fig. 10A).Similar to in vivo hCG-treated follicles (Fig. 2E), levels ofpSMAD2 began to decrease significantly (P>0.05) by 8 hourswith a further decrease by 12 hours after EGF treatment in vitro(Fig. 10B). The levels of total SMAD2 did not change duringEGF stimulation indicating that the decrease in SMAD2activation observed at 8 and 12 hours was due to decreasedpSMAD2 activation and not to loss of SMAD2 protein. Thus,activation of MAPK is independent of SMAD2/3 activationduring cumulus expansion. However, activation of MAPKpathways led to decreased pSMAD2 signaling during cumulusexpansion.

DiscussionEvidence is presented here that oocyte-stimulated SMADsignaling is crucial for defining the differentiation and functionof cumulus cells. Moreover, oocyte-secreted factors oppose theaction of FSH, which would, in the absence of the oocyte,promote the expression in cumulus cells of characteristics moretypical of mural granulosa cells. Soluble oocyte-derived factors

strongly promoted phosphorylation of SMAD2 in cumuluscells. Activated SMAD2 promoted expression of markercumulus transcripts and suppressed expression of muraltranscripts in cumulus cells, and together with pSMAD3,enabled cumulus expansion. The SMAD3-specific inhibitor,SIS3, had no effect on expression of mural or cumulus markertranscripts, but did block increases in Ptx3 and Tnfaip6 mRNA,two transcripts induced by EGF during cumulus expansion,suggesting that SMAD2 and SMAD3 have partially divergenteffects on cumulus cell function. By contrast, FSH stimulatedexpression of mural marker transcripts and suppressed theexpression of some cumulus marker transcripts. In the presenceof oocytes, FSH was unable to induce mural transcripts, butdid suppress cumulus transcripts in vitro. Thus, we present amodel (Fig. 11) where oocyte-stimulated pSMADs and FSHestablish opposing gradients of influence that define thecumulus and mural granulosa cell phenotypes, respectively.These mechanisms are important for promoting optimalendocrine and developmental functions in the ovarian follicle.

Mural and cumulus cells both originate from preantral

Slc38a3

0

0.01

0.02

0.03

0.04

0.05

0.06

Amh

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

B. Mural-SIS3

Control FGO FGOSiS3

Control FGO FGOSiS3

Rel

ativ

e Le

vels

of S

lc38

a3 m

RN

AR

elat

ive

Leve

ls o

f Ar

mR

NA

a

bb

Rel

ativ

e Le

vels

of S

lc38

a3 m

RN

AR

elat

ive

Leve

ls o

f Am

h m

RN

A

Slc38a3

Amh

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

FGO FGOSB431542

A. Mural-SB431542

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Rel

ativ

e Le

vels

of A

r m

RN

A

Ar

FGO FGOSB431542

FGO FGOSB431542

ba

c

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Rel

ativ

e Le

vels

of A

r m

RN

A

Ar

Ctrl FGO FGOSIS3

Control

Control

Control

Rel

ativ

e Le

vels

of S

lc38

a3 m

RN

AR

elat

ive

Leve

ls o

f Am

h m

RN

A

Slc38a3

Amh

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Rel

ativ

e Le

vels

of A

r m

RN

A

Ar

COC FSHSB431542

0

0.01

0.02

0.03

0.04

0.05

0.06

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0

0.005

0.01

0.015

0.02

0.025

0.03

B. COC-SB431542

COC OOX OOXFGO

A. COC-OOX

Rel

ativ

e Le

vels

of S

lc38

a3 m

RN

AR

elat

ive

Leve

ls o

f Am

h m

RN

AR

elat

ive

Leve

ls o

f Ar

mR

NA

Slc38a3

Amh

Ar

COC OOX OOXFGO

COC OOX OOXFGO

COC FSHSB431542

COC FSHSB431542

aa

aa

aa

b

b

b

b b

a

a a

a

b

c

b

Slc38a3

0

0.01

0.02

0.03

0.04

0.05

0.06

Amh

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

COC SIS3

C. COC-SIS3

COC SIS3

Rel

ativ

e Le

vels

of S

lc38

a3 m

RN

AR

elat

ive

Leve

ls o

f Am

h m

RN

A

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Rel

ativ

e Le

vels

of A

r m

RN

A

Ar

COC SIS3

Fig. 7. Regulation of cumulus transcripts (Slc38a3, Amh, and Ar) by oocytes.(A) Effect of OOX and oocyte co-culture on cumulus transcript levels in COCsand OOX cumulus cells cultured in vitro for 12 hours. Effect of SB431542 (10�M) and FSH (100 ng/ml) (B) or SIS3 (20 �M) (C) on cumulus transcript levelsin COCs cultured for 15 hours. Transcripts were measured by real-time PCR andnormalized to Rpl19 mRNA. abcP<0.05, significant differences.

Fig. 8. Effect of SB431542 (10 �M) (A) and SIS3 (20�M) (B) on expression of cumulus transcripts in muralgranulosa cells cultured alone or with FGOs (2oocytes/�l) for 15 hours. Transcripts were measuredby real-time PCR and normalized to Rpl19 mRNA.abcP<0.05, significant differences.

Jour

nal o

f Cel

l Sci

ence

1336

granulosa cells. The divergent functions(endocrine vs developmental) of these twolineages develop within the context of afollicle and thus the specification of thecumulus vs mural cell fate must depend inpart on the local microenvironment. In theantral follicle, the oocyte is emerging as acentral regulator of cumulus cell function.We hypothesized that SMAD2/3 signalingpathways might be involved in specifyingthe cumulus cell phenotype because of tworecent reports showing: (1) an inhibitor ofSMAD2/3 activation, SB431542, blocks theability of the oocyte to stimulate cumuluscell proliferation (Gilchrist et al., 2006) and(2) deletion of the Smad4 gene in granulosacells results in defective cumulus celldifferentiation (Pangas et al., 2006). InSMAD4-deficient granulosa cells, activatedforms of SMAD2/3 and SMAD1/5/8 wouldbe unable to alter target gene transcriptionbecause of the absence of the co-SMAD,SMAD4, effectively ablating signalingthrough these pathways. As a first step inanalyzing SMAD2/3 signaling in the follicle,activated SMAD2 was localized in tissuesections by immunofluorescence andimmunoblot analyses. Strong pSMAD2immunostaining was evident in granulosacells of preantral follicles. However, in largeantral follicles, 48 hours after eGCtreatment, pSMAD2 levels were muchhigher in cumulus cells compared withmural cells, similar results were obtained byimmunoblotting. The pattern of pSMAD2staining suggested that the oocyte isresponsible for activating this pathway incumulus cells. Oocytectomy completelyablated pSMAD2 levels in cumulus cells,which were restored upon oocyte co-cultureshowing that oocytes stimulate pSMAD2signaling in cumulus cells. Theseobservations are consistent with a role ofSMAD2/3 signaling in cumulus cellfunction.

The TGF� signaling pathways areemerging as a key regulator of ovarianfunction. In this report, two inhibitors ofTGF�-activin-GDF9 signaling, SB431542and SIS3 were used. The activin-like kinase(ALK) inhibitor, SB431542, specificallyinhibits the activity of ALK4, ALK5 andALK7 without effect on other cellularkinases in living cells (Inman et al., 2002; Laping et al., 2002)(Figs 4, 10). ALK4 and ALK5 are the type 1 receptors involvedin GDF9 and activin signaling. When activated by a ligand,ALK4 and ALK5 pair with a specific type II receptor, resultingin phosphorylation of SMAD2 and SMAD3 signalingmolecules (Moore et al., 2003; Roh et al., 2003). SIS3specifically blocks SMAD3 activity by inhibiting ALK5-mediated phosphorylation of SMAD3 and association with

Journal of Cell Science 120 (8)

SMAD4 (Jinnin et al., 2005). In COCs, SB431542 is a specificinhibitor of SMAD2 and SMAD3 activation, whereas SIS3specifically blocks SMAD3 activation. Neither inhibitor hadany effect on total SMAD2 levels or on phosphorylation of therelated SMAD1/5/8 pathway.

Increasing evidence supports a direct role of the oocyte inpromoting cumulus cell function before the LH surge. Oocytesare known to promote glycolysis (Sugiura et al., 2005),

Fig. 9. Involvement of SMAD signaling in cumulus expansion. (A) Effect of mediumonly, SB431542 (10 �M) or SIS3 (20 �M) on cumulus expansion in COCs treated withEGF (10 ng/ml) for 15 hours. White arrows indicate cumulus cells attached to the culturedish in COC treated with SIS3. (B) Expansion of COCs treated with EGF (10 ng/ml)alone or with 1 �M SB431542 or 0.1 �M SB431542 for 15 hours. (C) Effect ofSB431542 (10 �M) or SIS3 (20 �M) on expression of cumulus expansion transcripts inCOCs treated with EGF (10 ng/ml) for 6, 8 and 10 hours. Transcripts were measured byreal-time PCR and normalized to Rpl19 mRNA. abcP<0.05, significant differences.

Jour

nal o

f Cel

l Sci

ence

1337Oocytes promote cumulus phenotype

proliferation (Gilchrist et al., 2006; Vanderhyden et al., 1992),amino acid transport (Eppig et al., 2005) and cell survival(Hussein et al., 2005), and to suppress inappropriate expressionof mural transcripts such as Lhcgr in cumulus cells (Eppig etal., 1997). Results presented in this study confirm and extendthese observations to include oocyte regulation of othercumulus and mural transcripts before the LH surge. Forexample, we show that oocytes promote expression of threecumulus transcripts, Slc38a3, Amh and Ar. Each of thesetranscripts has important functions. Expression of SLC38A3 incumulus cells probably mediates, at least in part, the transportof specific amino acids from the cumulus cells to the oocyte insupport of oocyte growth (Eppig et al., 2005). Expression ofAR in the follicle is required for full fertility because Ar nullmice are severely subfertile and show defects in cumulus cellmorphology and differentiation (Hu et al., 2006). In addition,androgens augment the ability of oocytes to stimulate cellularproliferation (Hickey et al., 2005). AMH in early antralfollicles antagonizes FSH-induced follicle growth (Durlingeret al., 2001; Durlinger et al., 1999). Moreover, we show thattreatment with SB431542, but not SIS3, resulted in reducedcumulus marker transcript levels. These observations suggestthat SMAD2 and SMAD3 have divergent functions in theregulation of the steady-state levels of these transcripts.Surprisingly, FSH suppressed Slc38a3 and Amh mRNA levelsin cultured COCs and is likely to be one reason why Slc38a3and Amh are not expressed in mural cells.

The mural cell phenotype is produced to a great extent byFSH stimulation of transcripts involved in steroidogenesis(Cyp11a1), ovulation (Lhcgr) and immune function (Cd34).However, oocytes potently suppressed mural transcripts in

mural cells even in the presence of FSH. Because oocytefactors and FSH form opposing concentration gradients in thefollicle, a gradient of Lhcgr transcript levels is observed inantral follicles, where granulosa cells farther away from theoocyte express higher levels of mural transcripts (Eppig et al.,2002; Lei et al., 2001). Cumulus cells are closely associatedwith the oocyte and therefore it is not surprising that expressionof mural transcripts is low in cumulus cells because of thesuppressive action of oocyte-derived factors. This helps toexplain why FSH did not promote expression of the muralmarker transcripts in cultured COCs even though cumulus cellsrobustly express FSH receptors (Richards and Midgley, 1976).However, FSH is able to stimulate mural transcript expressionin cumulus cells in the absence of oocytes (OOX) or in COCstreated with SB431542, but not SIS3 indicating that SMAD2activity is required for suppression of mural transcripts byoocyte factors. That FSH suppressed cumulus markertranscripts and stimulates mural transcripts in vitro, suggeststhat oocytes must continuously overcome these two potentialeffects of FSH on cumulus cells before the LH surge in vivo.It is remarkable that oocyte-stimulated SMAD2 signalingpromotes elevated levels of cumulus marker transcripts andsuppresses mural transcripts in cumulus cells, while onlysuppressing mural transcripts in mural cells. Theseobservations might reflect recruitment of different co-activators or co-repressors to the promoter regions of thesegenes. However, further studies are needed to define how acommon signal can generate both positive and negativeresponses in the same cell type.

In addition to regulating the levels of cumulus transcriptsbefore the LH surge, oocyte-secreted factors also modify

Fig. 10. (A) Western blot immunostained for pSMAD2, pMAPK14, pMAPK3/1 and ACTB in COCs treated with EGF (10 ng/ml) alone ortogether with SB431542 (10 �M) for 4 hours. Graphical representations of pSMAD2, MAPK14 and MAPK3/1 pixel intensity normalized toACTB are shown below a representative western blot (n=3). (B) Western blot immunostained for pSMAD2, total SMAD2 and ACTB in COCstreated with EGF (10 ng/ml) for 0, 4, 8 and 12 hours. Graphical representations of pSMAD2, total SMAD2, MAPK14 and MAPK3/1 pixelintensity, normalized to ACTB, are shown below a representative western blot (n=4). Note that each image shown is of the same blotsequentially stained for the proteins indicated. abcP<0.05, significant differences.

Jour

nal o

f Cel

l Sci

ence

1338

cumulus cell function after the LH surge by enabling cumulusexpansion. The process of cumulus expansion requires thepresence of cumulus expansion enabling factors (CEEFs)secreted by the oocyte. The nature of the CEEFs has remainedelusive, but could include GDF9 (Elvin et al., 1999; Gui andJoyce, 2005) and/or BMP15 (Su et al., 2004; Yoshino et al.,2006) since both recombinant proteins stimulate expansion.Here, we demonstrated that SB431542 completely blocksEGF-induced expansion, whereas the effect of SIS3 is moresubtle, expansion still occurs in SIS3 treated COCs, but manycells from attachments to the culture dish, indicating anabnormal expansion. Cumulus expansion is dependent onincreases in transcript levels of Has2, Ptx3, Ptgs2 and Tnfaip6(Chen et al., 1993; Fulop et al., 2003; Ochsner et al., 2003a;Ochsner et al., 2003b; Varani et al., 2002). The potentialrequirement of pSMAD2 and pSMAD3 in driving theincreased levels of these transcripts in cumulus cells wasunknown. Here we demonstrate a complex pattern of transcriptregulation during cumulus expansion. One transcript, Tnfaip6,was largely unaffected by either inhibitor during EGF-inducedexpansion. At the other extreme, EGF-induced increase in Ptx3mRNA was suppressed by both inhibitors at all time pointsexamined. Ptgs2 mRNA levels were not suppressed at 6 hours,but by 8-10 hours, SB431542, but not SIS3, suppressed levelsof this transcript. Finally, Has2 was suppressed by SB431542,but not by SIS3. These results indicate that the oocyte-enabledincrease in Ptx3 requires activation of both SMAD2 andSMAD3, whereas SMAD2 activation is sufficient for enablingincreases in Ptgs2 and Has2 mRNA. However, the CEEFactivity responsible for enabling increases in Tnfaip6 mRNAsignal through pathways that are not acutely dependent oneither pSMAD2 or pSMAD3. This suggests the possibleexistence of one or more CEEF factors secreted by the oocyte.

The factor(s) responsible for pSMAD2/3 activation incumulus cells are not known. Candidate molecules includeoocyte-derived GDF9 or TGFB1 or the granulosa-cell-derivedactivin (Knight and Glister, 2006). GDF9 is an attractivecandidate because mice lacking this protein do not form antralfollicles (Dong et al., 1996), rGDF9 activates SMAD2 ingranulosa cells in vitro (Roh et al., 2003) and RNAi inhibitionof GDF9 production in oocytes abrogates cumulus expansion(Gui and Joyce, 2005), which we now show is dependent onpSMAD2/3. Thus, GDF9 might have roles during bothcumulus cell differentiation and cumulus expansion. However,not all the data support a role of GDF9 during expansion, sincea GDF9-neutralizing antibody did not block cumulusexpansion enabled by oocytes (Dragovic et al., 2005). TGFB1is another possible candidate since it is expressed by the oocyteand activates the SMAD2/3-signaling pathway. Mice with nullmutations in the Tgfb1 gene have reduced fertility (Ingman etal., 2006) suggesting a possible involvement in follicledevelopment. However, a TGFB1-neutralizing antibody doesnot block cumulus expansion enabled by oocytes (Salustri etal., 1990). The possible involvement of activin in cumulus celldifferentiation or cumulus expansion has not been investigated.Regardless of which ligand is responsible for SMAD2/3activation in cumulus cells, oocyte-stimulated pSMAD2/3 incumulus cells is part of a differentiation mechanism requiredto specify the cumulus cell phenotype.

Communication between the cumulus cells and oocyte iscrucial for subsequent fertility. Disruption of these signaling

Journal of Cell Science 120 (8)

pathways between oocytes and cumulus cells leads to poordevelopmental potential or even infertility. Thus, resolvinghow cumulus cells and oocytes signal each other is key tounderstanding the mechanisms responsible for a fundamentalprocess in biology, the production of a mature female gamete.Fig. 11 summarizes our working model and highlights thecrucial role of oocyte-stimulated SMAD2/3 signaling indetermining aspects of the cumulus cell phenotype, includingexpression of cumulus transcripts, suppression of muraltranscripts before the LH surge and promotion of cumulusexpansion after the LH surge. Two opposing radial gradientswithin the follicle are created by FSH signaling from outsidethe follicle and oocyte-stimulated SMAD2 signaling fromwithin the follicle. Together, they specify characteristics andfunctions of the mural and cumulus cell compartments withinthe follicle to foster an optimal microenvironment for properendocrine (mural) and developmental (cumulus-oocytecomplex) functions.

Materials and MethodsAnimalsFemale B6SJLF1 mice (Mus musculus) were produced and raised in the researchcolony of the investigators. Ovaries were collected from mice on day 22, 48 hoursafter I.P. injection of 5 IU eCG (National Hormone and Peptide Program,NIDDK). For some experiments, ovaries from 12-day-old animals were used foranalysis of preantral granulosa cells. Animals were maintained according to theGuide for the Care and Use of Laboratory Animals (Institute for Learning andAnimal Research).

Isolation and culture of oocytes and granulosa cellsCOCs and fully grown denuded oocytes at the GV (germinal vesicle) stage werecollected from antral follicles. Following COC collection, clumps of muralgranulosa cells from approximately one primed ovary were equally divided into fourgroups. Mural granulosa cells, COCs (25 for mRNA or 50 for protein), OOXcomplexes (25 or 50) and co-cultures of mural or OOX complexes with fully grownoocytes (FGOs 2 oocytes/�l) were cultured in bicarbonate-buffered MEM-� (Life

Opposing effects of FSH and SMAD2/3 signaling

FSH

LhcgrCyp11a1

Cd34

Mural Cells

Slc38a3Ar

pSMAD2Slc38a3

AmhAr

LhcgrCyp11a1

Cd34

Cumulus Cells

OSFs

Fig. 11. Summary model depicting the opposing gradients of oocyte-stimulated SMAD2 and FSH signaling in the antral follicle thatresult in divergent expression of mural and cumulus transcripts.Cumulus cells (red) form under the influence of the oocyte secretedfactors (OSPs), whereas mural cells (green) form under the influenceof FSH. Granulosa cells located in between oocyte and FSH signalsexhibit intermediate levels of mural and cumulus transcripts (lightgreen or pink) depending on their proximity to the basal lamina orthe oocyte.

Jour

nal o

f Cel

l Sci

ence

1339Oocytes promote cumulus phenotype

Technologies, Grand Island, NY) with Earles salts, supplemented with 75 mg/lpenicillin G, 50 mg/l streptomycin sulfate, 10 �M milrinone (to prevent GVB inFGOs), 0.23 mM pyruvate, and 3 mg/ml crystallized lyophilized bovine serumalbumen. COC, mural cells or OOX complexes were treated for one or more of thefollowing time points 0, 4, 6, 8, 12 or 15 hours with control medium, FSH (100ng/ml) or EGF (10 ng/ml). Some cultures were pre-incubated with SB431542 (0.1to 100 �M, Calbiochem) or SIS3 (20 �M, Calbiochem) to block both SMAD2 andSMAD3 (SB431542) or SMAD3 (SIS3) activation for 1 hour before beginningtreatment. Expansion was stimulated in media (MEM-�) containing EGF 10 ng/ml)and 5% FBS. Samples were immediately frozen in liquid nitrogen and stored at–70°C until analyzed for protein or mRNA levels. All reagents were purchased fromSigma Chemical Company (St Louis, MO) unless otherwise noted. All experimentswere repeated at least three times.

Quantification of RNA transcriptsTotal RNA was isolated from frozen samples and reverse transcribed into cDNA asdescribed previously (Diaz et al., 2006). Quantification of mural and cumulus andexpansion-related transcripts was conducted using gene-specific primers asdescribed previously using Rpl19 as an internal control (Diaz et al., 2006). Onlyone product of the appropriate size was identified for each set of primers and allamplification products were sequenced to confirm specificity. All experiments wererepeated three to four times and values shown are the mean ± s.e.m.

ImmunoblottingSamples were prepared from 50 COCs or OOX complexes cultured alone or withFGOs (2 oocytes/�l). Some samples were pretreated with SB431542 (10 �M) orSIS3 (20 �M) for 1 hour. Groups were treated with medium only (control), FSH(100 ng/ml) or EGF (10 ng/ml) for 1, 4 or 15 hour(s). Samples were simultaneouslydenatured by boiling in 1� loading buffer for 5 minutes followed by quenching onice for 5 minutes. Proteins were separated on a 10% SDS PAGE gel and transferredto PVDF membrane. Membranes were blocked in 1� blocking buffer (OdysseyBlocking buffer, Licor Bioscience, Lincoln, NE) for 1 hour with shaking at roomtemperature followed by incubation with specific anti-pMAPK3/1 antibody (1:2000,Sigma), anti-pMAPK14 antibody (1:1000, Cell Signaling Technology, Danvers,MA), anti-pSMAD2 antibody (1:1000, Invitrogen), anti-total SMAD2 antibody(1:2000, Cell Signaling Technology, Danvers, MA), anti-pSMAD1/5/8 antibody(1:1000, Cell Signaling Technology, Danvers, MA), anti-pSMAD3 antibody(1:1000, Invitrogen) or �-actin (ACTB) antibody (1:6000, Sigma) diluted inblocking buffer with 0.1% Tween-20 for 2-12 hours at room temperature. Followingincubation, blots were washed three times for 10 minutes each with wash buffer(PBS, 0.1% Tween-20). Fluorescently labeled secondary antibodies (IRDyeTM 800anti-mouse or anti-rabbit, Rockland Immunochemicals, Gilbertsville, PA) werediluted at 1:5000 and incubated with the blots for 1 hour at room temperature. Blotswere washed as above with an additional final wash in PBS without Tween-20.Detection was accomplished with an infrared scanner (Licor Bioscience, Lincoln,NE). Representative blots of three to four independent experiments are shown inFigs 2-4 and 10.

ImmunofluorescenceOvaries from 12-day-old mice, 22-day-old mice primed with eCG (44 hours) and22-day-old mice primed with eCG plus hCG (5 IU, 8 hours) were fixed overnightin 4% paraformaldehyde and embedded in paraffin wax. Ovarian sections weredewaxed in xylene (2� 5 minutes) followed by incubation for 5 minutes in each ofthe following: xylene:ethanol (1:1), 100% ethanol (2�), 95% ethanol, 85% ethanol,75% ethanol and ddH2O. Slides were then incubated in 1� antigen retrieval solution(DakoCytomation Retrieval solution) at 95°C for 25 minutes followed by washingwith ddH2O (2�) for 10 minutes and PBS (3�) for 5 minutes. Slides were thenincubated in blocking buffer (PBS, pH 7.4, 3% BSA, 10% goat serum and 0.05%Triton X-100) for 1 hour at room temperature followed by incubation with anti-pSMAD2 (1:400, Invitrogen) diluted in blocking buffer for 12 hours at 4 C. Slideswere then washed three times for 15 minutes with wash buffer (PBS 0.1% Triton-X 100), followed by incubation with anti-rabbit IgG Alexa Fluor 594 conjugate(1:1500, Molecular Probes) for 1 hour at room temperature, counterstained withDAPI and mounted with anti-fade solution (Slow-fade, Molecular Probes). Slideswere imaged using in an Olympus BX60 upright fluorescent microscope connectedto a 3CCD camera and computer.

Statistical analysesResults from real-time PCR and immunoblot experiments were analyzed by two-way ANOVA followed by Tukeys HSD post-hoc test or Dunnetts HSD (Fig. 10B)if a positive F-test was detected. Results in Fig. 1 were analyzed by Student’s t-test.The JMP 6.0 statistical analysis software was used for all analysis (SAS, Cary, NC).P<0.05 was considered statistically significant.

We thank Susan Ackerman, Robert Burgess and Mary Ann Handelfor their helpful suggestions in preparing this manuscript. This workwas supported by grant HD23839 from the NICHD (J.J.E.).

ReferencesBaarends, W. M., Uilenbroek, J. T. J., Kramer, P., Hoogerbrugge, J. W., Van

Leeuwen, E. C. M., Themmen, A. P. N. and Grootegoed, J. A. (1995). Anti-Mullerian hormone and anti-Mullerian hormone type II receptor messenger ribonucleicacid expression in rat ovaries during postnatal development, the estrous cycle, andgonadotropin-induced follicle growth. Endocrinology 136, 4951-4962.

Buccione, R., Vanderhyden, B. C., Caron, P. J. and Eppig, J. J. (1990). FSH-inducedexpansion of the mouse cumulus oophorus in vitro is dependent upon a specificfactor(s) secreted by the oocyte. Dev. Biol. 138, 16-25.

Chen, L., Russell, P. T. and Larsen, W. J. (1993). Functional significance of cumulusexpansion in the mouse: roles for the preovulatory synthesis of hyaluronic acid withinthe cumulus mass. Mol. Reprod. Dev. 34, 87-93.

Diaz, F. J., O’Brien, M. J., Wigglesworth, K. and Eppig, J. J. (2006). The preantralgranulosa cell to cumulus cell transition in the mouse ovary: development ofcompetence to undergo expansion. Dev. Biol. 299, 91-104.

Dierich, A., Sairam, M. R., Monaco, L., Fimia, G. M., Gansmuller, A., LeMeur, M.and SassoneCorsi, P. (1998). Impairing follicle-stimulating hormone (FSH) signalingin vivo: Targeted disruption of the FSH receptor leads to aberrant gametogenesis andhormonal imbalance. Proc. Natl. Acad. Sci. USA 95, 13612-13617.

Dong, J. W., Albertini, D. F., Nishimori, K., Kumar, T. R., Lu, N. F. and Matzuk, M.M. (1996). Growth differentiation factor-9 is required during early ovarianfolliculogenesis. Nature 383, 531-535.

Dragovic, R. A., Ritter, L. J., Schulz, S. J., Amato, F., Armstrong, D. T. and Gilchrist,R. B. (2005). Role of oocyte-secreted growth differentiation factor 9 in the regulationof mouse cumulus expansion. Endocrinology 146, 2798-2806.

Durlinger, A. L. L., Kramer, P., Karels, B., de Jong, F. H., Uilenbroek, J. T. J.,Grootegoed, J. A. and Themmen, A. P. N. (1999). Control of primordial folliclerecruitment by anti-Mullerian hormone in the mouse ovary. Endocrinology 140, 5789-5796.

Durlinger, A. L. L., Gruijters, M. J. G., Kramer, P., Karels, B., Kumar, T. R., Matzuk,M. M., Rose, U. M., de Jong, F. H., Uilenbroek, J. T. J., Grootegoed, J. A. et al.(2001). Anti-Mullerian hormone attenuates the effects of FSH on follicle developmentin the mouse ovary. Endocrinology 142, 4891-4899.

Elvin, J. A., Yan, C. N., Wang, P., Nishimori, K. and Matzuk, M. M. (1999). Molecularcharacterization of the follicle defects in the growth differentiation factor 9-deficientovary. Mol. Endocrinol. 13, 1018-1034.

Eppig, J. J. (2001). Oocyte control of ovarian follicular development and function inmammals. Reproduction 122, 829-838.

Eppig, J. J., Wigglesworth, K., Pendola, F. L. and Hirao, Y. (1997). Murine oocytessuppress expression of luteinizing hormone receptor messenger ribonucleic acid bygranulosa cells. Biol. Reprod. 56, 976-984.

Eppig, J. J., Pendola, F. L. and Wigglesworth, K. (1998). Mouse oocytes suppresscAMP-induced expression of LH receptor messenger RNA by granulosa cells in vitro.Mol. Reprod. Dev. 49, 327-332.

Eppig, J. J., Wigglesworth, K. and Pendola, F. L. (2002). The mammalian oocyteorchestrates the rate of ovarian follicular development. Proc. Natl. Acad. Sci. USA 99,2890-2894.

Eppig, J. J., Pendola, F. L., Wigglesworth, K. and Pendola, J. K. (2005). Mouseoocytes regulate metabolic cooperativity between granulosa cells and oocytes: aminoacid transport. Biol. Reprod. 73, 351-357.

Fulop, C., Szanto, S., Mukhopadhyay, D., Bardos, T., Kamath, R. V., Rugg, M. S.,Day, A. J., Salustri, A., Hascall, V. C., Glant, T. T. et al. (2003). Impaired cumulusmucification and female sterility in tumor necrosis factor-induced protein-6 deficientmice. Development 130, 2253-2261.

Gilchrist, R. B., Ritter, L. J., Myllymaa, S., Kaivo-Oja, N., Dragovic, R. A., Hickey,T. E., Ritvos, O. and Mottershead, D. G. (2006). Molecular basis of oocyte-paracrinesignalling that promotes granulosa cell proliferation. J. Cell Sci. 119, 3811-3821.

Gui, L.-M. and Joyce, I. M. (2005). RNA interference evidence that growthdifferentiation factor-9 mediates oocyte regulation of cumulus expansion in mice. Biol.Reprod. 72, 195-199.

Hickey, T. E., Marrocco, D. L., Amato, F., Ritter, L. J., Norman, R. J., Gilchrist, R.B. and Armstrong, T. D. (2005). Androgens augment the mitogenic effects of oocyte-secreted factors and growth differentiation factor 9 on porcine granulosa cells. Biol.Reprod. 73, 825-832.

Hu, Y. C., Wang, P. H., Yeh, S., Wang, R. S., Xie, C., Xu, Q., Zhou, X., Chao, H.T., Tsai, M. Y. and Chang, C. (2006). Subfertility and defective folliculogenesis infemale mice lacking androgen receptor. Proc. Natl. Acad. Sci. USA 101, 11209-11214.

Hussein, T. S., Froiland, D. A., Amato, F., Thompson, J. G. and Gilchrist, R. B.(2005). Oocytes prevent cumulus cell apoptosis by maintaining a morphogenicparacrine gradient of bone morphogenetic proteins. J. Cell Sci. 118, 5257-5268.

Hussein, T. S., Thompson, J. G. and Gilchrist, R. B. (2006). Oocyte-secreted factorsenhance oocyte developmental competence. Dev. Biol. 296, 514-521.

Ingman, W. V., Robker, R. L., Woittiez, K. and Robertson, S. A. (2006). Null mutationin transforming growth factor �1 disrupts ovarian function and causes oocyteincompetence and early embryo arrest. Endocrinology 147, 835-845.

Inman, G. J., Nicolás, F. J., Callahan, J. F., Harling, J. D., Gaster, L. M., Reith, A.D., Laping, N. J. and Hill, C. S. (2002). SB-431542 is a potent and specific inhibitorof transforming growth factor-superfamily type I activin receptor-like kinase (ALK)receptors ALK4, ALK5, and ALK7. Mol. Pharmacol. 62, 65-74.

Jinnin, M., Ihn, H. and Tamaki, K. (2005). Characterization of SIS3, a novel specificinhibitor of smad3, and its effect on transforming growth factor-1-induced extracellularmatrix expression. Mol. Pharmacol. 69, 597-607.

Jour

nal o

f Cel

l Sci

ence

Knight, P. G. and Glister, C. (2006). TGF-� superfamily members and ovarian follicledevelopment. Reproduction 132, 191-206.

Kumar, T. R., Wang, Y., Lu, N. F. and Matzuk, M. M. (1997). Follicle stimulatinghormone is required for ovarian follicle maturation but not male fertility. Nat. Genet.15, 201-204.

Laping, N. J., Grygielko, E., Mathur, A., Butter, S., Bomberger, J., Tweed, C., Martin,W., Fornwald, J., Lehr, R., Harling, J. et al. (2002). Inhibition of transforminggrowth factor (TGF)-beta1-induced extracellular matrix with a novel inhibitor of theTGF-beta type I receptor kinase activity: SB-431542. Mol. Pharmacol. 62, 58-64.

Lei, Z. M., Mishra, S., Zou, W., Xu, B., Foltz, M., Li, X. and Rao, C. V. (2001).Targeted disruption of luteinizing hormone/human chorionic gonadotropin receptorgene. Mol. Endocrinol. 15, 184-200.

Meduri, G., Vuhai-Luuthi, M. T., Jolivet, A. and Milgrom, E. (1992). New functionalzonation in the ovary as shown by immunohistochemistry of luteinizing hormonereceptor. Endocrinology 131, 366-373.

Moore, R. K., Otsuka, F. and Shimasaki, S. (2003). Molecular basis of bonemorphogenetic protein-15 signaling in granulosa cells. J. Biol. Chem. 278, 304-310.

Ochsner, S. A., Day, A. J., Rugg, M. S., Breyer, R. M., Gomer, R. H. and Richards,J. S. (2003a). Disrupted function of tumor necrosis factor-alpha-stimulated gene 6blocks cumulus cell-oocyte complex expansion. Endocrinology 144, 4376-4384.

Ochsner, S. A., Russell, D. L., Day, A. J., Breyer, R. M. and Richards, J. S. (2003b).Decreased expression of tumor necrosis factor-alpha-stimulated gene 6 in cumulus cellsof the cyclooxygenase-2 and EP2 null mice. Endocrinology 144, 1008-1019.

Pangas, S. A., Li, X., Robertson, E. J. and Matzuk, M. M. (2006). Prematureluteinization and cumulus cell defects in ovarian-specific smad4 knockout mice. Mol.Endocrinol. 20, 1406-1422.

Richards, J. S. and Midgley, J. A. R. (1976). Protein hormone action: a key tounderstanding ovarian follicular and luteal cell development. Biol. Reprod. 14, 82-94.

Roh, J. S., Bondestam, J., Mazerbourg, S., Kaivo-Oja, N., Groome, N., Ritvos, O.and Hsueh, A. J. W. (2003). Growth differentiation factor-9 stimulates inhibinproduction and activates Smad2 in cultured rat granulosa cells. Endocrinology 144,172-178.

Salmon, N. A., Handyside, A. H. and Joyce, I. M. (2004). Oocyte regulation of anti-

Mullerian hormone expression in granulosa cells during ovarian follicle developmentin mice. Dev. Biol. 266, 201-208.

Salustri, A., Ulisse, S., Yanagishita, M. and Hascall, V. C. (1990). Hyaluronic acidsynthesis by mural granulosa cells and cumulus cells in vitro is selectively stimulatedby a factor produced by oocytes and by transforming growth factor �. J. Biol. Chem.265, 19517-19523.

Su, Y. Q., Wu, X., O’Brien, M. J., Pendola, F. L., Denegre, J. A., Matzuk, M. M. andEppig, J. J. (2004). Synergistic roles of BMP15 and GDF9 in the development andfunction of the oocyte-cumulus cell complex in mice: genetic evidence for an oocyte-granulosa cell regulatory loop. Dev. Biol. 276, 64-73.

Sugiura, K., Pendola, F. L. and Eppig, J. J. (2005). Oocyte control of metaboliccooperativity between oocytes and companion granulosa cells: energy metabolism.Dev. Biol. 279, 20-30.

Tomic, D., Miller, K. P., Kenny, H. A., Woodruff, T. K., Hoyer, P. and Flaws, J. A.(2004). Ovarian follicle development requires Smad3. Mol. Endocrinol. 18, 2224-2240.

Vanderhyden, B. C., Caron, P. J., Buccione, R. and Eppig, J. J. (1990). Developmentalpattern of the secretion of cumulus-expansion enabling factor by mouse oocytes andthe role of oocytes in promoting granulosa cell differentiation. Dev. Biol. 140, 307-317.

Vanderhyden, B. C., Telfer, E. E. and Eppig, J. J. (1992). Mouse oocytes promoteproliferation of granulosa cells from preantral and antral follicles in vitro. Biol. Reprod.46, 1196-1204.

Varani, S., Elvin, J. A., Yan, C., DeMayo, J., DeMayo, F. J., Horton, H. F., Byrne, M.C. and Matzuk, M. M. (2002). Knockout of pentraxin 3, a downstream target ofgrowth differentiation factor-9, causes female subfertility. Mol. Endocrinol. 16, 1154-1167.

Yan, C., Wang, P., DeMayo, J., Elvin, J. A., Carino, C., Prasad, S. V., Skinner, S. S.,Dunbar, B. S., Dube, J. L., Celeste, A. J. et al. (2001). Synergistic roles of bonemorphogenetic protein 15 and growth differentiation factor 9 in ovarian function. Mol.Endocrinol. 15, 854-866.

Yoshino, O., McMahon, H. E., Sharma, S. and Shimasaki, S. (2006). A uniquepreovulatory expression pattern plays a key role in the physiological functions of BMP-15 in the mouse. Proc. Natl. Acad. Sci. USA 103, 10678-10683.

Journal of Cell Science 120 (8)1340

Jour

nal o

f Cel

l Sci

ence