Molecular and Cellular Endocrinology 294 (2008) 70–80

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology

journa l homepage: www.e lsev ier .com/ locate /mce

Transforming growth factor beta (TGF�1, TGF�2 and TGF�3)null-mutant phenotypes in embryonic gonadal development

Mushtaq A. Memona, Matthew D. Anwayb,1, Trevor R. Covertb,Mehmet Uzumcub,2, Michael K. Skinnerb,∗

a Center for Reproductive Biology, Department of Veterinary Clinical Sciences, Washington State University, Pullman, WA 99164-4234, United Statesb Center for Reproductive Biology, School of Molecular Biosciences, Washington State University, Pullman, WA 99164-4234, United States

a r t i c l e i n f o

Article history:Received 4 February 2008Received in revised form 9 August 2008Accepted 11 August 2008

Keywords:TestisOvary

a b s t r a c t

The role transforming growth factor beta (TGFb) isoforms TGFb1, TGFb2 and TGFb3 have in the regulationof embryonic gonadal development was investigated with the use of null-mutant (i.e. knockout) mice foreach of the TGFb isoforms. Late embryonic gonadal development was investigated because homozygoteTGFb null-mutant mice generally die around birth, with some embryonic loss as well. In the testis, theTGFb1 null-mutant mice had a decrease in the number of germ cells at birth, postnatal day 0 (P0). In thetestis, the TGFb2 null-mutant mice had a decrease in the number of seminiferous cords at embryonic day15 (E15). In the ovary, the TGFb2 null-mutant mice had an increase in the number of germ cells at P0.TGFb isoforms appear to have a role in gonadal development, but interactions between the isoforms is

Knockout

TGF-beta 1TGF-beta 2TE

speculated to compensate in the different TGFb isoform null-mutant mice.© 2008 Elsevier Ireland Ltd. All rights reserved.

1

ddepopdtceco(

l9

I

D

pagintfoacl1

0d

GF-beta 3mbryonic development

. Introduction

Gonadal development during sex determination requires a coor-inated somatic and germ cell differentiation and proliferationuring the embryonic period. In the mouse, crucial morphogenicvents are initiated in the testis at embryonic day 11.5 (E11.5;lug day considered to be E0) (Karl and Capel, 1998). At this stagef male gonadal development, SRY expression is initiated in therecursor Sertoli cells and male sex determination ensues. In theeveloping testis at E12.5 the pre-peritubular cells migrate fromhe adjacent mesonephrose to enclose precursor Sertoli and germell aggregates to promote development of cords. Testis seminif-

rous cords can be identified by E14 in rats. After seminiferousord formation, somatic and germ cells undergo the highest levelf cellular proliferation that occurs during testis developmentLevine et al., 2000). In the rodent ovary there is no major mor-

∗ Corresponding author at: Center for Reproductive Biology, School of Molecu-ar Biosciences, Washington State University, Abelson Hall Room 507, Pullman, WA9164-4234, United States. Tel.: +1 509 335 1524; fax: +1 509 335 2176.

E-mail address: skinner@mail.wsu.edu (M.K. Skinner).1 Current address: Department of Biological Sciences, University of Idaho, Moscow,D 83844, United States.2 Current address: Department of Animal Science, Rutgers University, 84 Lipmanrive, New Brunswick, NJ 08901, United States.

m

gtwSeToa1ao

303-7207/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.mce.2008.08.017

hological alteration until after birth when primordial folliclesssemble (Skinner, 2005). During ovarian sex determination theerm cells initially proliferate and then enter meiosis and arrestn prophase I to form clusters (i.e. nests) of germ cells. Theseests of germ cells after birth degenerate through random apop-osis of selected germ cells to form several primordial folliclesrom each nest. The somatic cells that surround the nests of mei-tically arrested germ cells form the precursor granulosa cellsssociated with individual oocytes to form the primordial folli-les after birth. The current study was designed to investigate theate embryonic testis and ovary morphology at embryonic day5 (E15) and at birth, postnatal day 0 (P0), in TGFb null-mutantice.Transforming growth factor beta’s (TGFbs) are critical for

rowth, differentiation and development of many different cellypes within an organism. Five TGFb isoforms have been identifiedith three (TGFb1, TGFb2, TGFb3) present in mammals (Roberts and

porn, 1988, 1990; Sporn and Roberts, 1987). The TGFb isoforms arencoded from individual genes located on different chromosomes.he primary functions of the TGFbs include enhancing formation

f the extracellular matrix, regulation of cellular differentiation,nd inhibition of proliferation of most cells (Sporn and Roberts,987; Lawrence, 1996). Inhibition of growth by TGFbs occur throughn arrest of the cell cycle in late G1 phase and requires actionsn the retinoblastoma protein and cyclin-dependent kinases. The

http://www.sciencedirect.com/science/journal/03037207mailto:skinner@mail.wsu.edudx.doi.org/10.1016/j.mce.2008.08.017

ellula

eri1

cOtWesoiAmsSaTtm1

wfsaidmRbtomdt2hTp(meskiaT22tg

inmscTwhcnm

2

2

eh(alaapC

2

vws193w9rtaCrc

2

zgaerTptCpPd

2

2SeR

2

wwawTmpawsiwt

M.A. Memon et al. / Molecular and C

ffects of TGFbs are mediated through activation of the membraneeceptors TGFb-receptor I (TGFbRI) and TGFb-receptor II contain-ng serine/threonine kinase activity (Lawrence, 1996; Howe et al.,991), with all TGFb isoforms binding to TGFbRII (Cupp et al., 1999).

TGFb isoforms display similar, although not identical, biologi-al activity and differential tissue expression (Graycar et al., 1989;hta et al., 1987; Rosa et al., 1988). For example, in the mature

estis TGFb3 is the major isoform expressed (Miller et al., 1989;atrin et al., 1991). TGFb1 is expressed by the Sertoli cell (Esposito

t al., 1991; Skinner and Moses, 1989) and may be important forpermatogenesis (Nargolwalla et al., 1990). Sertoli cell productionf TGFb1 may be targeted to the germinal cell population duen part to the blood–testis barrier (Mullaney and Skinner, 1993).ll 3 isoforms (TGFb1, TGFb2 and TGFb3) are expressed in theale gonad and their receptors are present in the testis in both

omatic cells and germ cells (Cupp et al., 1999; Mullaney andkinner, 1993; Caussanel et al., 1997; Olaso et al., 1998; Teerdsnd Dorrington, 1993). The embryonic (E14) testis has TGFb1 andGFb2 expression by Sertoli cells, germ cells and selected intersti-ial cells, while TGFb3 is expressed at high levels at the testis and

esonephrose interface by precursor peritubular cells (Cupp et al.,999).

Gene knockout null-mutant and over-expression experimentsith TGFb have demonstrated that precise regulation of each iso-

orm is essential for survival. The null-mutant TGFb1 embryoshow defects in vasculogenesis and hematopoiesis and often dieround birth. However, some TGFb1 knockouts can be phenotyp-cally normal until approximately 3 weeks after birth and thenevelop a severe wasting syndrome related to excessive inflam-atory responses (Shull and Doetschman, 1994; Shull et al., 1992).

eproductive tracks and organs within TGFb1 knockout mice haveeen studied (Ingman and Robertson, 2007). Significant devia-ions from normal Mendelian ratios are observed with decreasedffspring for both heterozygotes and homozygotes of TGFb1 null-utants. TGFb2 knockout mice show multiple developmental

efects including cardiopulmonary, skeletal, ocular, and urogeni-al systems defects (Bartram et al., 2001; Dunker and Krieglstein,003; Molin et al., 2002). Urogenital defects include ectopic andypoplastic testis. TGFb2 knockout mice generally die by birth.GFb3 knockout mice show delayed pulmonary and defectivealate development, and die prenatally due to failure to suckleKaartinen et al., 1995; Proetzel et al., 1995). Although knockout

ice for each TGFb isoform have been generated, all animals gen-rally die at the neonatal stage making it difficult to investigate theignificance of TGFbs in gametogenesis in the adult gonads. Doublenockout null mutants of combinations of TGFb isoforms die earlyn embryonic development prior to gonadal development (Dunkernd Krieglstein, 2002). Although functions have been found forGF� family members in the adult testis and ovary (Drummond,005; Findlay et al., 2002; Knight and Glister, 2003; Lin et al.,003), studies are needed with conditional knockouts to pinpointhe physiological significance of TGFbs null mutants in the postnatalonad (Lui et al., 2003).

The current study was designed to investigate the role of TGFbsn late embryonic testis and ovary development using knockoutull-mutant mice. To determine actions of TGFb on testis develop-ent, two time points were evaluated. The first was E14–E15 after

eminiferous cord formation occurs and the second was P0 whenells of the testis are actively proliferating. To evaluate the role ofGFb on ovarian development, ovaries from P0 mice were collected,

hen oocytes are starting to assemble into primordial follicles. The

ypothesis tested was that TGFb1, TGFb2, and TGFb3 have criti-al roles in embryonic gonadal development and are necessary forormal cell–cell interactions during the process of testis and ovaryorphogenesis.

2

da

r Endocrinology 294 (2008) 70–80 71

. Materials and methods

.1. Animals

TGFb1 mouse (strain 129) and TGFb2 and TGFb3 mouse (C57black6) het-rozygous (+/−) mice were bred to generate TGFb1, TGFb2 and TGFb3 knockoutomozygous (−/−) mice, respectively. The testes from TGFb1, TGFb2 and TGFb3−/−), (+/−) and wild type (+/+) embryos were collected from timed pregnant micet embryonic day 14–15 (E14–E15, E0 = plug date). The testes and ovaries were col-ected from TGFb1, TGFb2 and TGFb3 pups at P0–P1 (P0 = postnatal day 0). Whennimals survived the testes from TGFb1 (−/−), (+/−) and (+/+) pups were collectedt P10–P20. Tails from embryos and the pups were collected for genotyping. Allrotocols were approved by the Washington State University Animal Care and Useommittee.

.2. Genotyping and SRY analysis

Tail fragments were collected and genomic DNA was isolated by procedures pre-iously reported (Levine et al., 2000; Cupp et al., 1999). Briefly, the tails were digestedith proteinase K (0.15 mg/ml) overnight at 55 ◦C, followed by precipitation with

aturated NaCl and 100% ethanol, and re-suspended in sterile water. Approximately00 ng of genomic DNA was used for PCR analyses. For TGFb1 analyses, 30 cycles of4 ◦C/30 s, 58 ◦C/30 s and 72 ◦C/90 s using forward 5′-GAGAAGAACTGCTGTGTGCG-′ and reverse 5′-GTGTCCAGGCTCCAAATATAG-3′ primers containing 1.2% DMSOere used per reaction. For TGFb2 analyses, 30 cycles of PCR was performed at4 ◦C/30 s, 55 ◦C/30 s and 72 ◦C/90 s using forward 5′-CTCCATAGATATGGGCATGC andeverse 5′-AATGTGCAGGATAATTGCTGC primers containing 1.2% DMSO per reac-ion. For TGFb3 analyses, 30 cycles of PCR was performed at 94 ◦C/30 s, 58 ◦C/30 snd 72 ◦C/90 s using forward-1 5′-TGGGAGACATGGCTGTAACT and forward-2 5′-ACTCACACTGGAAGTAGT and reverse 5′-GATGGGATGTTTGGTTGGT primers pereaction. SRY analysis was utilized to confirm sex of the embryos using the PCRonditions previously reported (Levine et al., 2000).

.3. Semi-quantitative PCR analysis

RNA was extracted from testis sections on slides with the use of Tri-ol reagent (Sigma, St. Louis, MO). The TGFb1, TGFb2, and TGFb3 null mutantonadal RNA was reverse transcribed with 3′ TGFb isoform specific primersnd then the PCR reaction performed with similar primers as listed abovexcept for the TGFb2, which used forward 5′-CAGGAGTGGCTTCACCACAAAG andeverse 5′-TGGCATATGTAGAGGTGCCATCA, and the TGFb3, which used forward 5′-CGACATGATCCAGGGACTG and reverse 5′-CCACTGAGGACACATTGAAACG primerser reaction. A constitutively expressed S2 gene expression was used to normalizehe data with the use of forward 5′-TGCCAGTGCAGAAGCAGACT and reverse 5′-ACCAAGACCAACGTGACCA primers in the same RT-reaction. TGFb isoform specificroducts were extracted and sequenced to confirm the identify of the PCR product.CR electrophoretic band intensity was measured with a scanner to quantify theata.

.4. Histology

The testes and ovaries were fixed in Bouin’s fixative (Sigma, St. Louis, MO) forh, washed in 70% ethanol and embedded in paraffin using standard procedures.erial ribbon sections from each testis and ovary were stained with hematoxylin andosin (H&E) using standard procedures for morphological analyses. The Center foreproductive Biology Histology Core Laboratory assisted with this analysis.

.5. Morphological analysis

Testis and ovary sections from genotyped animals at each developmental ageere analyzed as described earlier (Cupp et al., 2002). Briefly, the testes sectionsere evaluated for the presence of seminiferous cords, area of seminiferous cords,

nd area of the interstitium. Using computer aided morphometry seminiferous cordsere circled within a section to calculate the total area of seminiferous cords (mm2).

he seminiferous cord area was then subtracted from the total testis area to deter-ine the interstitial area of each section. The data for each averaged area (for a

articular genotype) were depicted as the number of mm2 per designated testisrea (mm2). Because the sections were relatively small, serial sections on each slideere utilized to obtain these measurements. Therefore, for each genotype repre-

ented, at least six sections from each testis were utilized to obtain the cord area,nterstitial area, and number of seminiferous cords. All morphological observations

ere normalized per a defined testis section area (mm2) to correct for variation inestis and section size.

.6. Germ cell nuclear antigen (GCNA) staining

Germ Cell Nuclear Antigen (GCNA) staining was analyzed as previouslyescribed (Cupp et al., 2003). Briefly, the testis and ovary sections were deparafinizednd rehydrated through a series of alcohols. Sections were boiled in a microwave

7 ellula

octsowAmd

2

a(ctpoew

2

owciaitoga

pGcaTtPs

2

DTgtDdta

3

ttm

TM

(

T

O

(

T

O

Sb

2 M.A. Memon et al. / Molecular and C

ven for 15 min in sodium citrate buffer for antigen retrieval and blocked in 10%alf serum for 30 min at room temperature. The sections of testes and ovaries werehen incubated with GCNA1 primary antibody (1:10 dilution) at 4 ◦C for 16 h. Theections were then incubated at 20 ◦C for 2 h with biotinylated goat anti-mouse sec-ndary antibody (1:300; Vector Laboratories, Burlington, CA). Secondary antibodyas detected using the Histo stain-Sp Kit (Zymed Laboratories, San Francisco, CA).t least two experiments were conducted using GCNA1 antibodies at each develop-ental time. In each experiment, 3 serial sections of 4–5 testes and ovaries for each

evelopmental age were analyzed.

.7. Testicular cell apoptosis

The Fluorescein In Situ Cell Death Detection Kit (Roche Applied Science, Indi-napolis, IN) was utilized to detect apoptosis of testicular cells as described earlierUzumcu et al., 2004). The assay measures fragmented DNA from apoptotic cells byatalytically incorporating fluorescein-12-dUTP at the 3′DNA ends using the enzymeerminal deoxynucleotidyl transferase (TdT), which forms a polymeric tail using therincipal of the TdT-mediated dUTP Nick-End Labeling (TUNEL) assay. All the flu-rescent cells in each testis section were counted at 100× magnification. In eachxperiment, 3 serial sections of 4–5 testis and ovaries for each developmental ageere analyzed. The average number of fluorescent cells/whole testis was calculated.

.8. Microarray analysis

RNA was collected from E13, E14, E16, P0 and P6 testes and E13, E14 and E16varies from Sprague–Dawley rats as previously described (Small et al., 2005). RNAas hybridized to the Affymetrix (Affymetrix, Santa Clara, California) rat 230A gene

hip. The Genomics Core Laboratory of the Center for Reproductive Biology at Wash-ngton State University performed the analysis as previously described (McLean et

l., 2002; Shima et al., 2004). Briefly, RNA from the cells was reverse transcribednto cDNA then was transcribed into biotin labeled RNA. Biotin labeled RNA washen hybridized to the Affymetrix rat 230A gene chips. Each gene set is composedf 11 pairs of 24-mer oligonucleotides, with one anti-sense strand specific for theene and one anti-sense strand with single point mutations for use as compar-tive negative control. Biotin labeled RNA was then visualized by labeling with

lett(

able 1icroarray expression levels (mean raw signal) of Transforming Growth Factor (TGF) �1,

A) Rat gonadal development

E13 E14

estisTGF�1 A ATGF�2 70.2 69.1TGF�3 36.9 59.4�-receptor I 53.6 59.6�-receptor II 203.9 213.5�-receptor III 86.6 97.9

varyTGF�1 A ATGF�2 76.1 77.0TGF�3 41.2 59.4�-receptor I 34.0 43.2�-receptor II 190.3 193.7�-receptor III 95.9 90.8

B) Mouse gonadal development

E11.5 E12.5

estisTGF�1 A ATGF�2 42.4 64.1TGF�3 83.7 124.3�-receptor I 32.7 40.3�-receptor II A A�-receptor III A A

varyTGF�1 A ATGF�2 51.4 48.4TGF�3 92.4 105.3�-receptor I A 48.4�-receptor II A A�-receptor III 41.2 61.7

tatistically significant present call microarray signals are presented with E = embryonic day the array analyses. “ND” denotes samples not analyzed.

r Endocrinology 294 (2008) 70–80

hycoerythrin-coupled avidin. The microarray was scanned on a Hewlett-Packardene Array Scanner (Hewlett-Packard Co., Palo Alto, CA). Two different microarrayhips from two different RNA samples involving different groups of animals werenalyzed for each E13, E14, E16, P0 and P6 testis and E13, E14 and E16 ovary samples.he microarray data set can be observed at www.skinner.wsu.edu and was submit-ed to GEO. Transcript levels from mouse microarray analyses of E11.5, E12.5, E16.5,0 and P6 testis and ovary samples were extracted from previously reported dataets (Small et al., 2005).

.9. Statistical analysis

Data were analyzed with GraphPad Prism (GraphPad Prism Software, Inc., Saniego, CA). All values are expressed as the mean ± S.E.M. of the parameter measured.he n value was based on different testis and ovary and not replicates from the sameonad. Each tissue section analysis involved a minimum of 6 different areas. Statis-ical analysis was performed using one-way ANOVA followed by post-hoc test usingunnett test for comparison to controls and the Tukey–Kramer honestly significantifference test for multiple comparisons between different treatment groups. Statis-ical difference was confirmed at p < 0.05. Specific comparisons, analyses, replicatesnd results are presented in the different figure legends.

. Results

The gene expression of TGFb1, TGFb2 and TGFb3 isoforms andheir receptors were analyzed by microarray analyses for E13–P6 ratestes, E11.5–P6 mouse testes, E13–E16 rat ovaries and E11.5–E16.5

ouse ovaries (Table 1). TGFb1 mRNA was below the detection

imits in both rat and mouse samples studied. TGFb2 mRNA lev-ls increased from E14–P0 and then decreased at P6 in the ratestis (Table 1A). TGFb2 mRNA levels were constant in the mouseestes development with the highest level of expression at E16.5Table 1B). TGFb3 mRNA levels gradually increased from E13–E16

�2, �3 and receptors I, II and III during (A) rat and (B) mouse gonadal development

E16 P0 P6

A A A101.2 119.5 65.0120.2 85.0 95.450.6 A 46.2240.2 233.2 726.8102.3 369.1 92.9

A ND ND80.0 ND ND70.5 ND NDA ND ND215.0 ND ND77.9 ND ND

E16.5 P0 P6

A A A69.6 59.7 36.5158.3 184.6 196.3A 15.9 26.3A A A51.1 68.5 90.1

A ND ND50.1 ND ND175.8 ND NDA ND NDA ND ND50.5 ND ND

y (E0 = post-mating plug), P = postnatal day. “A” denotes transcript was not detected

http://www.skinner.wsu.edu/

M.A. Memon et al. / Molecular and Cellular Endocrinology 294 (2008) 70–80 73

Fig. 1. Sex ratios (A) and percentages of genotypes (B) from TGFb1, TGFb2 and TGFb3litters. Total litters in analysis for TGFb1 n = 40, TGFb2 n = 38, and TGFb3 n = 26. In(B) open bar represents wild type (+/+), black bar represents heterozygous (+/−) andhashed bar represents homozygote knockouts (−/−). Data represents mean ± S.E.M.

Fig. 3. Number of seminiferous cords per mm2 testis area for E15 embryos (A) and P0postnatal pups (B) from TGFb1, TGFb2 and TGFb3 animals. Total number of embryosfor E15 testis in the analysis are TGFb1 (+/+) n = 3, (+/−) n = 4, (−/−) n = 5; TGFb2 (+/+)n = 3, (+/−) n = 6, (−/−) n = 3; TGFb3 (+/+) n = 3, (+/−) n = 5, (−/−) n = 6. Total numberof P0 animals in the analysis are TGFb1 (+/+) n = 5, (+/−) n = 4, (−/−) n = 8; TGFb2(+/+) n = 3, (+/−) n = 5, (−/−) n = 2; TGFb3 (+/+) n = 4, (+/−) n = 4, (−/−) n = 3. Open barrepresents wild type (+/+), black bar represents heterozygous (+/−) and hashed barrepresents homozygote knockouts (−/−). An (a) represents statistically significantvalue from wildtype (+/+). Data presented as mean ± S.E.M.

Fig. 2. Morphology of TGFb2 E15 (A–C) and TGFb1 P0 (D–F) testes cross-sections from (+/+) (A, D), (+/−) (B, E) and (−/−) (C, F) animals. An H&E stained micrograph at 400×magnification is presented and representative of a minimum of three different animals examined.

7 ellula

dpmcTwTPTgsTr

dmsTamtwesEtbTt

ifRgpiaoct

Twe((ep3lw(lMt

3T0r(psaM

ef

tTfT(timPiTd

pdrmasbnificant difference was observed between TGFb2 (−/−) and (+/−)testes. No significant difference in the number of seminiferouscords per mm2 testis area for E15 embryos was present for eitherTGFb1 or TGFb3 animals (Fig. 3A). The number of seminiferous

Fig. 4. Average cross-sectional area of seminiferous cords per mm2 testis area fromE15 embryos (A) and P0 postnatal pups (B) from (+/+), (+/−) and (−/−) TGFb1, TGFb2,and TGFb3 animals. Total number of embryos for E15 testis in the analysis are TGFb1

4 M.A. Memon et al. / Molecular and C

uring the rat testis development, then slowly decreased in theostnatal samples. TGFb3 had a consistent increase until P6 duringouse testis development (Table 1B). In the rat testis, TGFbRI had

onsistently low expression during testis development (Table 1).GFbRII expression levels were high during testes developmentith the highest expression detected in the P6 sample (Table 1A).

GFbRIII expression increased from E13–P0 then decreased in the6 samples (Table 1). In the mouse testis, the expression of TGFbRI,GFbRII and TGFbRIII were low during all time points (Table 1B). Ineneral the TGFb’s pattern of gene expression were similar or con-erved during mouse and rat testis development. In contrast, theGFb receptors pattern of gene expression were distinct betweenat and mouse.

In the rat ovary samples, the TGFb1 mRNA levels were belowetection in the E13–E16 samples (Table 1A). TGFb2 and TGFb3RNA levels were consistent in the E13–E16 samples. The expres-

ion of TGFbRI and TGFbRIII were low in the E13–E16 ovary samples.GFbRII was the most abundant of the three receptors in the ovarynd it remained consistent between E13–E16 (Table 1A). In theouse ovary samples, the TGFb1 mRNA levels were below detec-

ion in the E11.5–E16.5 samples (Table 1B). TGFb2 mRNA levelsere consistent in the E11.5–E16.6 samples. The TGFb3 mRNA lev-

ls gradually increased from E11.5–E16.5. Similar to the rat ovaryamples, the expression of TGFbRI and TGFbRII were low in the11.5–E16.5 samples and the TGFbRIII was the most abundant ofhe three receptors in the mouse ovary and remained consistentetween E11.5–E16.5 (Table 1B). As seen in the testis, the ovarianGFb expression patterns were similar between rat and mouse withhe TGFb receptor expression being distinct.

The gene expression of the TGFb isoforms was assessed in RNAsolated from P0 testis sections to confirm that alternate TGFb iso-orms were expressed in the individual isoform null mutant mice.T-PCR data was normalized with the constitutively expressed S2ene to provide a semi-quantitative analysis. The identity of PCRroducts was confirmed with DNA sequence analysis. The specific

soform null mutants had no expression of that TGFb isoform. Thelternate TGFb isoforms were expressed in each of the TGFb1, TGFb2r TGFb3 null mutant P0 testis samples (data not shown). No majorhange in expression levels was observed when compared to con-rol wild-type P0 testis samples (data not shown).

To generate TGFb null-mutant homozygote (−/−) animals theGFb1, TGFb2 and TGFb3 heterozygous (+/−) females were housedith age matched (+/−) males and examined every day for the pres-

nce of vaginal plugs. The plug day was considered embryonic day 0E0). Forty litters from TGFb1 (+/−) breedings, 38 litters from TGFb2+/−) breedings and 26 litters from TGFb3 (+/−) breedings werexamined. The average TGFb1 litter size was 2.95 ± 0.29 femaleups and 3.35 ± 0.29 male pups. The average TGFb2 litter size was.35 ± 0.35 females and 3.19 ± 0.25 males pups. The average TGFb3

itter size was 3.4 + ± 0.43 females and 3.8 ± 0.41 males pups. Thereas no difference in pup sex ratio between TGFbs (+/−) breedings

Fig. 1A). The control wild type 129 and C57BL/6 mice strains haveitter sizes of 6.0 and 6.6, respectively (Jackson Labs, Bar Harbor,

aine), so the TGF/b (+/−) breeding litters were smaller in size tohe control wild type.

The compiled genotype ratios for TGFb1 were 2.18 ± 0.29,.10 ± 0.32, and 1.03 ± 0.17 for (+/+), (+/−) and (−/−), respectively.he compiled TGFb2 genotype ratios were 3.0 ± 0.37, 3.3 ± 0.33, and.32 ± 0.11 for (+/+), (+/−) and (−/−), respectively. The genotypeatios for TGFb3 were 2.30 ± 0.44, 4.08 ± 0.51, and 1.12 ± 0.22 for

+/+), (+/−) and (−/−), respectively. Fig. 1B presents the data asercentages of the total litter. The average genotype ratios wereignificantly different for TGFb1 (p < 0.0001), TGFb2 (p < 0.0001)nd TGFb3 (p < 0.0025), based upon the observed and expectedendelian ratio (Fig. 1B). Observations demonstrate significant

((anbk

r Endocrinology 294 (2008) 70–80

mbryonic loss for homozygote (−/−) mutants for all TGFb iso-orms.

The testes morphology was examined at E15 and P0 for wild-ype (+/+), heterozygote (+/−) and homozygote (−/−) samples fromGFb1, TGFb2 and TGFb3 mice. The gross morphology of the testesrom the (−/−) mice was similar to the (+/−) and (+/+) for TGFb1,GFb2 and TGFb3 mice. Representative micrographs for TGFb2+/+), (+/−) and (−/−) E15 testes are shown in Fig. 2A–C. Similaro the E15 testis, the gross testis morphology at P0 was also sim-lar between the (+/+), (+/−) and (−/−) TGFb1, TGFb2 and TGFb3

ice. Representative micrographs for TGFb1 (+/+), (+/−) and (−/−)0 testes are shown in Fig. 2D–F. The ovary morphology was exam-ned at P0 for (+/+), (+/−) and (−/−) samples from TGFb1, TGFb2 andGFb3 mice. As observed with the testis, no gross morphologicalifferences were found in the ovaries (data not shown).

Testis morphology was analyzed in more detail in various sam-les. Testis and testis section size was variable so all morphologicalata was normalized per defined testis section area (mm2) to cor-ect for variation in size. The number of seminiferous cords perm2 testis area for E15 embryos were determined for (+/+), (+/−)

nd (−/−) TGFb1, TGFb2 and TGFb3 mice (Fig. 3A). There was aignificant (p < 0.01) decrease in the number of seminiferous cordsetween TGFb2 (−/−) and (+/−) testes and the (+/+) testes. No sig-

+/+) n = 3, (+/−) n = 4, (−/−) n = 5; TGFb2 (+/+) n = 3, (+/−) n = 6, (−/−) n = 2; TGFb3+/+) n = 3, (+/−) n = 5, (−/−) n = 6. Total number of pups in the P0 testis analysisre TGFb1 (+/+) n = 5, (+/−) n = 3, (−/−) n = 8; TGFb2 (+/+) n = 3, (+/−) n = 5, (−/−)= 2; TGFb3 (+/+) n = 4, (+/−) n = 4, (−/−) n = 3. Open bar represents wild type (+/+),lack bar represents heterozygous (+/−) and hashed bar represents homozygotenockouts (−/−). Data represents mean ± S.E.M.

ellular Endocrinology 294 (2008) 70–80 75

cmnT

eadt(ntT

ETtfn(gTc(acwb(o(

eao(wb

Fig. 6. Number of GCNA stained cells per mm2 testis area for E15 (A) and P0 (B) testisfrom TGFb1, TGFb2 and TGFb3 animals. Total number of embryos for E15 testis inthe analysis are TGFb1 (+/+) n = 4, (+/−) n = 4, (−/−) n = 3; TGFb2 (+/+) n = 3, (+/−)n = 5, (−/−) n = 3; TGFb3 (+/+) n = 4, (+/−) n = 4, (−/−) n = 4. Total number of P0 ani-mals in the analysis are TGFb1 (+/+) n = 3, (+/−) n = 4, (−/−) n = 4; TGFb2 (+/+) n = 4,(+/−) n = 3, (−/−) n = 2; TGFb3 (+/+) n = 5, (+/−) n = 4, (−/−) n = 3. Open bar representswild type (+/+), black bar represents heterozygous (+/−) and hashed bar represents

Fm

M.A. Memon et al. / Molecular and C

ords per mm2 testis area for P0 pups from TGFb1, TGFb2 and TGFb3ice was determined. No significant difference was observed in the

umber of seminiferous cords per mm2 testis area for P0 pups fromGFb1, TGFb2 nor TGFb3 animals (Fig. 3B).

The area of the seminiferous cords per mm2 testis area for E15mbryos and P0 postnatal pups were determined for (+/+), (+/−)nd (−/−) TGFb1, TGFb2 and TGFb3 animals (Fig. 4). No significantifference was found in area of seminiferous cords per mm2 for E15estes between (−/−), (+/−) and (+/+) for TGFb1, TGFb2 or TGFb3Fig. 4A). Similar to the E15 seminiferous cord analyses, there waso significant difference in the area of seminiferous cords per mm2

estis area for P0 pups from (+/+), (+/−) and (−/−) TGFb1, TGFb2 orGFb3 animals (Fig. 4B).

The number of developing germ cells per mm2 testis area for15 and P0 samples were determined for (+/+), (+/−) and (−/−)GFb1, TGFb2 and TGFb3 mice (Fig. 5). Fig. 5A–C are representa-ive micrographs of the GCNA staining of E15 testis cross-sectionsrom (+/+), (+/−) and (−/−) TGFb1 embryos. There was a sig-ificant (p < 0.05) increase in germ cell number between TGFb1+/+) and (+/−) E15 testes (Fig. 6A). No significant differences inerm cell numbers were observed for the (+/+), (+/−) and (−/−)GFb2 and TGFb3 mice (Fig. 6A). The number of developing germells per mm2 testis area for P0 samples were determined for+/+), (+/−) and (−/−) TGFb1, TGFb2 and TGFb3 mice. Fig. 5D–Fre representative micrographs of the GCNA staining of P0 testisross-sections from (+/+), (+/−) and (−/−) TGFb1 animals. Thereas a significant (p < 0.05) decrease in developing germ cell num-ers between TGFb1 (−/−) testes and (+/−) and (+/+) testesFig. 6B). No significant difference in germ cell numbers wasbserved for the (+/+), (+/−) and (−/−) TGFb2 and TGFb3 animalsFig. 6B).

The number of apoptotic cells per mm2 testis area for E15mbryos were determined for (+/+), (+/−) and (−/−) TGFb1, TGFb2nd TGFb3 mice. Fig. 7A and B are representative micrographs

f the TUNEL staining of E15 testis cross-sections from (+/+) and−/−) TGFb2 embryos. The majority of the positively labeled cellsere outside the seminiferous cords indicating germ cells may note affected at this time (Fig. 7). There was an apparent increase

homozygote knockouts (−/−). An (a) represents statistically significant value fromwildtype (+/+). Data represents mean ± S.E.M.

ig. 5. GCNA immunohistochemistry for TGFb1 E15 (A–C) and TGFb1 P0 (D–F) testis cross-section from (+/+) (A, D), (+/−) (B, E) and (−/−) (C, F) animals. Representativeicrographs are presented from three animals different animals examined.

76 M.A. Memon et al. / Molecular and Cellular Endocrinology 294 (2008) 70–80

F s-sectim

icsdfndmsAlsTN(

aaAm((pl(

ympPmaapoT

4

haeaggTTaltrfbcpTibecetdGipm

ig. 7. Apoptosis TUNEL stained TGFb2 E15 (A, B) and TGFb2 P0 (C, D) testis crosicrograph with apoptotic cells (yellow) at 400× magnification.

n apoptotic cell number in the TGFb2 (+/−) and (−/−) samplesompared to the (+/+) samples, however, this increase was onlytatistically significant for (−/+) (Fig. 8A). There was no significantifference in apoptotic cell numbers in the testis of E15 embryosrom (+/+), (+/−) and (−/−) TGFb1 and TGFb3 mice (Fig. 8A). Theumber of apoptotic cells per mm2 testis area for P0 samples wereetermined for (+/+), (+/−) and (−/−) TGFb1, TGFb2 and TGFb3ice. Fig. 7C and D are representative micrographs of the TUNEL

taining of P0 testis cross-sections from (+/+) and (−/−) TGFb2 mice.s with the labeling in the E15 testis sections, the majority of the

abeling was located outside the seminiferous cords. There was aignificant (p < 0.05) increase in apoptotic cell number between theGFb3 (+/−) testis and the (+/+) and (−/−) testis samples (Fig. 8B).o significant difference in number of apoptotic cells was found for

+/+), (+/−) and (−/−) TGFb1 or TGFb2 P0 testis (Fig. 8B).The homozygote (−/−) TGFb2 and TGFb3 animals all died shortly

fter birth or during embryogenesis. However, a small percent-ge of the TGFb1 (−/−) pups survived after birth up to 20 days.t postnatal day 20 (P20) the testicular size of TGFb1 (−/−) ani-als appeared 30–40% smaller than their litter mate (+/+) and

+/−) animals. Analyses of apoptotic cells in these (+/+), (+/−) and−/−) TGFb1 P10 to P20 testes demonstrated an increase in TUNELositive labeling inside the seminiferous tubules (i.e. germ cell

abeling) in the (−/−) mice compared to the (+/+) and (+/−) animalsdata not shown).

Ovary morphology was also investigated in more detail. Anal-ses of the ovaries of P0 (+/+) and (−/−) TGFb1, TGFb2 and TGFb3ice are shown in Fig. 9. Representative micrographs of the mor-

hology of the ovaries and GCNA staining for (+/+) and (−/−) TGFb20 animals are presented. The number of developing germ cells perm2 of ovarian area for P0 samples was determined for (+/+), (+/−)

nd (−/−) TGFb1, TGFb2 and TGFb3 animals (Fig. 9E). There wasn increase in germ cells in the homozygote (−/−) samples com-ared to the (+/+) samples for TGFb2. No significant difference wasbserved in the number of germ cells for (+/+) and (−/−) TGFb1 orGFb3 samples (Fig. 9E).

etpf

on from wildtype (+/+) (A, C) and homozygote (−/−) (B, D) mice. Representative

. Discussion

The transforming growth factor-beta family of growth factorsas been shown to be important for testis and ovary function atnumber of stages of development (Drummond, 2005; Findlay

t al., 2002; Knight and Glister, 2003; Lin et al., 2003; Ingmannd Robertson, 2002; Itman et al., 2006). This large family ofrowth factors involves the bone morphogenic proteins (BMP’s),rowth and differentiation factors (GDF’s), activins, inhibins and theGF isoforms (Drummond and Findlay, 2006). The TGFb1, TGFb2,GFb3 isoforms are critical to regulate cell growth, differentiationnd development. This involves growth inhibition and extracel-ular matrix production (Roberts and Sporn, 1988). In the testis,he TGFbs are localized to most cell types, as well as the TGFbeceptor isoforms (Konrad et al., 2006a). The expression of TGFb iso-orms in the testis alters during development and can be influencedy other hormones and growth factors, as well as physiologicalonditions (Muller et al., 2005; Wagener et al., 2005). An exam-le of the actions of TGFbs in the testis is the ability to integrateGFb signal transduction events with other signaling pathways tonfluence extracellular matrix, junction complex proteins and thelood–testis barrier (Xia and Cheng, 2005; Xia et al., 2006). Anotherxample of TGFbs actions in the testis is the ability to regulate germell apoptosis and spermatogenesis (Konrad et al., 2006b; Mairet al., 2005). In the ovary, TGFbs have a role in cellular prolifera-ion and differentiation associated with preantral and antral follicleevelopment (Drummond, 2005; Findlay et al., 2002; Knight andlister, 2003; Lin et al., 2003). The current study was designed to

nvestigate the role of TGFb isoforms in the embryonic and earlyostnatal testis and ovary using null-mutant (i.e. knockout) mouseodels.

Gene knockout null-mutant mice have been generated for

ach TGFb isoform. Due to the importance of TGFb to all tissueshese null-mutants will generally die late embryonically or earlyostnatally. Therefore, these TGFb knockout models are not use-ul to investigate pubertal or adult testis function, but can be

M.A. Memon et al. / Molecular and Cellula

Fig. 8. Number of TUNEL labeled cells per mm2 testis area for E15 (A) and P0 (B)TGFb1, TGFb2 and TGFb3 animals. Total number of embryos in the analysis are TGFb1(+/+) n = 5, (+/−) n = 3, (−/−) n = 6; TGFb2 (+/+) n = 3, (+/−) n = 10, (−/−) n = 3; TGFb3(+/+) n = 5, (+/−) n = 5, (−/−) n = 5. Total number of P0 animals in the analysis areTGFb1 (+/+) n = 3, (+/−) n = 3, (−/−) n = 5; TGFb2 (+/+) n = 6, (+/−) n = 5, (−/−) n = 2;Tbor

uUbgtbiTigabt

ao(al(bl

th

esocsenrnct(cpbTduag

shmId(mtstkdinksdbi

htnhTTodutisiTonoco

GFb3 (+/+) n = 3, (+/−) n = 3, (−/−) n = 3. Open bar represents wild type (+/+), blackar represents heterozygous (+/−) and hashed bar represents homozygote knock-uts (−/−). An (a) represents statistically significant value from wildtype (+/+). Dataepresents mean ± S.E.M.

sed to investigate embryonic gonadal development and function.nfortunately, double knockout TGFb null-mutants of various com-ination of isoforms die early in embryonic development prior toonadal development (Dunker and Krieglstein, 2002). Therefore,he combined functions of the different TGFb isoforms could note analyzed with double knockout mice. The current study exam-

nes the gonadal phenotype of the individual TGFb1, TGFb2 andGFb3 knockouts at embryonic day 15 (E15) after cord formationn the testis and at the day of birth, postnatal day 0 (P0). The back-round mouse strain was 129 for TGFb1 and C57BL/6 for TGFb2nd TGFb3. The phenotypes observed can be influenced by theackground strain, so should be considered in any data interpre-ations.

The sex ratio of the different TGFb isoform knockouts was notffected demonstrating no preferential sex specific effect on embry-nic survival or death. In contrast, the genotype of homozygotes−/−) for all the TGFb1, TGFb2 and TGFb3 knockouts demonstratednon-Mendelian distribution and suggests an apparent embryonic

oss of some homozygote (−/−) embryos. The reduced number of−/−) null-mutant animals required a large number of litters to

e obtained such that sufficient numbers of animals could be ana-

yzed.Previous analysis has localized TGFb isoform expression in

he developing embryonic testis (Cupp et al., 1999). Immuno-istochemical localization of TGFb1 and TGFb2 demonstrated

dCema

r Endocrinology 294 (2008) 70–80 77

xpression in Sertoli cells at E14 and P0, along with selected expres-ion in interstitial cells. TGFb3 was highly expressed at the junctionf the mesonephros and gonad, in peritubular cells around theords and in the gonocytes after birth. As shown in the currenttudy, TGFb2 and TGFb3 expression is more predominant in thembryonic testis, while TGFb1 is more highly expressed post-atally (Cupp et al., 1999), Table 1. All cells are anticipated toespond to TGFbs and the TGFb receptor type II was predomi-ant in the whole gonad. The administration of TGFb1 to organultures of embryonic testis was found to reduce cell growth andissue size, correlating to the growth inhibitory actions of TGFbsCupp et al., 1999). The current study used microarray analysis toonfirm the previous expression analysis. TGFb2 and TGFb3 wereredominant in both the embryonic testis and ovary, with negligi-le levels of TGFb1. Previous observations support the presence ofGFb1 expression (Cupp et al., 1999), but levels are at the limit ofetection for the microarray analysis. Therefore, as previously doc-mented TGFb isoforms are expressed in the developing gonadsnd the sites of action are likely at most cell types within theonad.

Analysis of the null-mutant (−/−) TGFb knockouts demon-trated no major effect on testis histology (Fig. 2) or ovarianistology (Fig. 9) and (data not shown). Therefore, no grossorphological alterations on gonad development were observed.

nvestigation of the number of cords that develop in the E15 testisemonstrated a reduced number of cords in the TGFb2 (+/−) and−/−) knockout animals, but no effect in the TGFb1 or TGFb3 ani-

als (Fig. 3). No effect on cord formation was observed in the P0estis suggesting the effect in TGFb2 knockout at E15 was tran-ient and rescued by P0 of development. Analysis of the area ofhe testis cords demonstrated no effect in any of the TGFb isoformnockouts. The primary morphological effect on embryonic testisevelopment observed was a transient decrease in cord numbers

n the TGFb2 knockout at E15. The TGFb1 and TGFb3 knockouts didot appear to effect testis morphogenesis. None of the TGFb isoformnockouts had an ovarian phenotype at either E15 or P0 (data nothown). The alterations in the TGFb2 knockout observed could beue to changes in cell populations, particularly the germ cell num-ers. Therefore, the germ cell numbers and cellular apoptosis were

nvestigated.The germ cell number was examined with the GCNA immuno-

istochemical stain in the TGFb null-mutant animals. Neitherhe TGFb2 nor TGFb3 knockouts showed any effect on germ cellumbers in the testis. The TGFb1 E15 heterozygote (+/−) animalsad an increase in germ cell number in the testis, while the P0GFb1 null-mutant (−/−) had a decrease in germ cell numbers.his inverse relationship between E15 and P0 in the TGFb1 knock-uts likely reflects different functions of TGFb1 at these differentevelopmental periods. Analysis of apoptosis with TUNEL stain wassed to complement the germ cell number analysis. Interestingly,he cellular apoptosis observed in the testis was primarily in thenterstitial space around and between cords (Fig. 7) with no majortaining in germ cells. The number of these apoptotic cells in thenterstitial space was increased at E15 in the TGFb2 (+/−) and P0GFb3 (+/−) heterozygote animals, but no effects were observedn apoptosis in the TGFb (−/−) null-mutant animals. Therefore,o major effect on cellular apoptosis or germ cell numbers wasbserved. The reduction in cord formation in the TGFb2 knockoutould be related to the increase in apoptosis in the interstitiumbserved. At E15 if the migrating mesonephros cells were reduced

ue to increased apoptosis, then cord formation may be reduced.ord formation is dependent on mesonephros cell migration. Givennough time, the migrating cells may accumulate to eventually pro-ote a similar number of cords reflected by a lack of effect observed

t P0. As shown in Fig. 7, the apoptotic cells observed in the TGFb2

78 M.A. Memon et al. / Molecular and Cellular Endocrinology 294 (2008) 70–80

F m TGFo ber of( combk

kcaif

cTgkotao

cAT

rrftl

ig. 9. Ovary micrograph of H&E (A, B) and GCNA (C, D) ovarian cross-sections frovarian cross-section for P0 pups from TGFb1, TGFb2, and TGFb3 animals. Total num+/−) n = 5, (−/−) n = 2; TGFb3 (+/+) and (+/−) n = 5, (−/−) n = 3. Open bar representsnockouts (−/−). Data represents mean ± S.E.M.

nockout testis often associated with the cells that surround theords. These are the cells that migrate from the mesonephrosnd become peritubular cells. Therefore, the observed reductionn these cells in the E15 testis could relate to the reduced cordormation observed in the TGFb2 null-mutant animals.

The embryonic ovary was found to have no major morphologi-al alterations or effects on germ cell numbers or apoptosis in theGFb1, TGFb2 or TGFb3 null-mutant animals at E15. At birth, the P0erm cell numbers were found to be increased in the TGFb2 (−/−)

nockout ovaries, but no effect was observed in either the TGFb1r TGFb3 knockouts. Morphologically no major effects were seen inhe ovaries (Fig. 9). Since TGFb2 can influence apoptosis, the alter-tion in germ cell numbers could be due to an effect on the high ratef germ cell apoptosis observed in the oocyte at birth. This is asso-

tAdct

b2 P0 (+/+) (A, C) and (−/−) (B, D) animals. (E) Number of germ cells per mm2 ofanimals in the analysis are TGFb1 (+/+) and (+/−) n = 7, (−/−) n = 4; TGFb2 (+/+) andined wild type (+/+) and heterozygous (+/−) and black bar represents homozygote

iated with the assembly of the primordial follicle (Skinner, 2005).s seen in the testis, the primary phenotype observed was with theGFb2 null-mutant gonadal development.

Although different TGFb isoforms have different localization,egulation and developmental control, all the TGFb isoforms cross-eact with the different TGFb receptor types and can have similarunctions in regards to cell growth and differentiation. Therefore,he TGFb1, TGFb2 and TGFb3 can compensate for each other. Theack of major phenotypes observed is speculated to in part be due to

his compensation in the individual isoform null-mutant animals.nalysis of TGFb isoform expression in the null mutant P0 testisemonstrated the alternate isoforms are expressed, with no majorhange in expression. Therefore, the alternate isoforms are presento compensate in the various knockouts, but direct compensation

ellula

ekddptt

A

AepN

R

K

L

S

R

R

S

L

H

C

G

O

R

M

W

E

S

N

M

C

O

T

S

S

I

B

D

M

K

P

D

D

F

K

L

L

C

C

U

S

M

S

I

I

D

K

M

W

M.A. Memon et al. / Molecular and C

xperiments remain to be performed. Unfortunately, the doublenockouts die early in embryonic development prior to gonadalevelopment. The effects observed suggest unique roles for theifferent TGFb isoforms during gonadal development. The TGFb2henotype supports a potential role in mesonephros cell migra-ion and cord formation in the testis, but requires further investiga-ion.

cknowledgement

We acknowledge the technical assistance and discussions of Drsndrea Cupp, Vinayak Doraiswamy and Eric Nilsson. We acknowl-dge the assistance of Ms. Rochelle Pedersen and Ms. Jill Griffin inreparation of the manuscript. This study was supported by an NIHICHD grant to Michael K. Skinner.

eferences

arl, J., Capel, B., 1998. Sertoli cells of the mouse testis originate from the coelomicepithelium. Dev. Biol. 203, 323–333.

evine, E., Cupp, A.S., Miyashiro, L., Skinner, M.K., 2000. Role of transforming growthfactor-alpha and the epidermal growth factor receptor in embryonic rat testisdevelopment. Biol. Reprod. 62, 477–490.

kinner, M.K., 2005. Regulation of primordial follicle assembly and development.Hum. Reprod. Update 11, 461–471.

oberts, A.B., Sporn, M.B., 1988. Transforming growth factor beta. Adv. Cancer Res.51, 107–145.

oberts, A.B., Sporn, M.B., 1990. The transforming growth factor-betas. In: Sporn,M.B., Roberts, A.B. (Eds.), Handbook of Experimental Pharmacology. Springer,Heidelberg, pp. 419–472.

porn, M.B., Roberts, A.B., 1987. Peptide growth factors: current status and thera-peutic opportunities. Important Adv. Oncol., 75–86.

awrence, D.A., 1996. Transforming growth factor-beta: a general review. Eur.Cytokine Netw. 7, 363–374.

owe, P.H., Draetta, G., Leof, E.B., 1991. Transforming growth factor beta 1 inhibitionof p34cdc2 phosphorylation and histone H1 kinase activity is associated withG1/S-phase growth arrest. Mol. Cell Biol. 11, 1185–1194.

upp, A.S., Kim, G., Skinner, M.K., 1999. Expression and action of transforming growthfactor beta (TGFbeta1, TGFbeta2, and TGFbeta3) during embryonic rat testisdevelopment. Biol. Reprod. 60, 1304–1313.

raycar, J.L., Miller, D.A., Arrick, B.A., Lyons, R.M., Moses, H.L., Derynck, R., 1989.Human transforming growth factor-beta 3: recombinant expression, purifi-cation, and biological activities in comparison with transforming growthfactors-beta 1 and -beta 2. Mol. Endocrinol. 3, 1977–1986.

hta, M., Greenberger, J.S., Anklesaria, P., Bassols, A., Massague, J., 1987. Two forms oftransforming growth factor-beta distinguished by multipotential haematopoi-etic progenitor cells. Nature 329, 539–541.

osa, F., Roberts, A.B., Danielpour, D., Dart, L.L., Sporn, M.B., Dawid, I.B., 1988. Meso-derm induction in amphibians: the role of TGF-beta 2-like factors. Science 239,783–785.

iller, D.A., Lee, A., Matsui, Y., Chen, E.Y., Moses, H.L., Derynck, R., 1989. Comple-mentary DNA cloning of the murine transforming growth factor-beta 3 (TGFbeta 3) precursor and the comparative expression of TGF beta 3 and TGF beta 1messenger RNA in murine embryos and adult tissues. Mol. Endocrinol. 3, 1926–1934.

atrin, F., Scotto, L., Assoian, R.K., Wolgemuth, D.J., 1991. Cell lineage specificity ofexpression of the murine transforming growth factor beta 3 and transforminggrowth factor beta 1 genes. Cell Growth Differ. 2, 77–83.

sposito, G., Keramidas, M., Mauduit, C., Feige, J.J., Morera, A.M., Benahmed, M.,1991. Direct regulating effects of transforming growth factor-beta 1 on lactateproduction in cultured porcine Sertoli cells. Endocrinology 128, 1441–1449.

kinner, M.K., Moses, H.L., 1989. Transforming growth factor beta gene expressionand action in the seminiferous tubule: peritubular cell–Sertoli cell interactions.Mol. Endocrinol. 3, 625–634.

argolwalla, C., McCabe, D., Fritz, I.B., 1990. Modulation of levels of messenger RNAfor tissue-type plasminogen activator in rat Sertoli cells, and levels of messen-ger RNA for plasminogen activator inhibitor in testis peritubular cells. Mol. CellEndocrinol. 70, 73–80.

ullaney, B.P., Skinner, M.K., 1993. Transforming growth factor-beta (beta 1, beta 2,and beta 3) gene expression and action during pubertal development of the sem-iniferous tubule: potential role at the onset of spermatogenesis. Mol. Endocrinol.7, 67–76.

aussanel, V., Tabone, E., Hendrick, J.C., Dacheux, F., Benahmed, M., 1997. Cellular

distribution of transforming growth factor betas 1, 2, and 3 and their types Iand II receptors during postnatal development and spermatogenesis in the boartestis. Biol. Reprod. 56, 357–367.

laso, R., Pairault, C., Habert, R., 1998. Expression of type I and II receptors for trans-forming growth factor beta in the adult rat testis. Histochem. Cell Biol. 110,613–618.

X

X

r Endocrinology 294 (2008) 70–80 79

eerds, K.J., Dorrington, J.H., 1993. Localization of transforming growth factor beta 1and beta 2 during testicular development in the rat. Biol. Reprod. 48, 40–45.

hull, M.M., Doetschman, T., 1994. Transforming growth factor-beta 1 in reproduc-tion and development. Mol. Reprod. Dev. 39, 239–246.

hull, M.M., Ormsby, I., Kier, A.B., Pawlowski, S., Diebold, R.J., Yin, M., Allen, R., Sid-man, C., Proetzel, G., Calvin, D., et al., 1992. Targeted disruption of the mousetransforming growth factor-beta 1 gene results in multifocal inflammatory dis-ease. Nature 359, 693–699.

ngman, W.V., Robertson, S.A., 2007. Transforming growth factor-beta1 null muta-tion causes infertility in male mice associated with testosterone deficiency andsexual dysfunction. Endocrinology 148, 4032–4043.

artram, U., Molin, D.G., Wisse, L.J., Mohamad, A., Sanford, L.P., Doetschman, T.,Speer, C.P., Poelmann, R.E., Gittenberger-de Groot, A.C., 2001. Double-outletright ventricle and overriding tricuspid valve reflect disturbances of looping,myocardialization, endocardial cushion differentiation, and apoptosis in TGF-beta(2)-knockout mice. Circulation 103, 2745–2752.

unker, N., Krieglstein, K., 2003. Reduced programmed cell death in the retinaand defects in lens and cornea of Tgfbeta2(−/−) Tgfbeta3(−/−) double-deficientmice. Cell Tissue Res. 313, 1–10.

olin, D.G., DeRuiter, M.C., Wisse, L.J., Azhar, M., Doetschman, T., Poelmann, R.E.,Gittenberger-de Groot, A.C., 2002. Altered apoptosis pattern during pharyngealarch artery remodelling is associated with aortic arch malformations in Tgfbeta2knock-out mice. Cardiovasc. Res. 56, 312–322.

aartinen, V., Voncken, J.W., Shuler, C., Warburton, D., Bu, D., Heisterkamp, N.,Groffen, J., 1995. Abnormal lung development and cleft palate in mice lackingTGF-beta 3 indicates defects of epithelial–mesenchymal interaction. Nat. Genet.11, 415–421.

roetzel, G., Pawlowski, S.A., Wiles, M.V., Yin, M., Boivin, G.P., Howles, P.N., Ding,J., Ferguson, M.W., Doetschman, T., 1995. Transforming growth factor-beta 3 isrequired for secondary palate fusion. Nat. Genet. 11, 409–414.

unker, N., Krieglstein, K., 2002. Tgfbeta2 −/− Tgfbeta3 −/− double knockoutmice display severe midline fusion defects and early embryonic lethality. Anat.Embryol. (Berl.) 206, 73–83.

rummond, A.E., 2005. TGFbeta signalling in the development of ovarian function.Cell Tissue Res. 322, 107–115.

indlay, J.K., Drummond, A.E., Dyson, M.L., Baillie, A.J., Robertson, D.M., Ethier, J.F.,2002. Recruitment and development of the follicle; the roles of the transforminggrowth factor-beta superfamily. Mol. Cell Endocrinol. 191, 35–43.

night, P.G., Glister, C., 2003. Local roles of TGF-beta superfamily members in thecontrol of ovarian follicle development. Anim. Reprod. Sci. 78, 165–183.

in, S.Y., Morrison, J.R., Phillips, D.J., de Kretser, D.M., 2003. Regulation of ovar-ian function by the TGF-beta superfamily and follistatin. Reproduction 126,133–148.

ui, W.Y., Lee, W.M., Cheng, C.Y., 2003. TGF-betas: their role in testicular functionand Sertoli cell tight junction dynamics. Int. J. Androl. 26, 147–160.

upp, A.S., Tessarollo, L., Skinner, M.K., 2002. Testis developmental phenotypes inneurotropin receptor trkA and trkC null mutations: role in formation of semi-niferous cords and germ cell survival. Biol. Reprod. 66, 1838–1845.

upp, A.S., Uzumcu, M., Suzuki, H., Dirks, K., Phillips, B., Skinner, M.K., 2003. Effect oftransient embryonic in vivo exposure to the endocrine disruptor methoxychloron embryonic and postnatal testis development. J. Androl. 24, 736–745.

zumcu, M., Suzuki, H., Skinner, M.K., 2004. Effect of the anti-androgenic endocrinedisruptor vinclozolin on embryonic testis cord formation and postnatal testisdevelopment and function. Reprod. Toxicol. 18, 765–774.

mall, C.L., Shima, J.E., Uzumcu, M., Skinner, M.K., Griswold, M.D., 2005. Profil-ing gene expression during the differentiation and development of the murineembryonic gonad. Biol. Reprod. 72, 492–501.

cLean, D.J., Friel, P.J., Pouchnik, D., Griswold, M.D., 2002. Oligonucleotide microar-ray analysis of gene expression in follicle-stimulating hormone-treated ratSertoli cells. Mol. Endocrinol. 16, 2780–2792.

hima, J.E., McLean, D.J., McCarrey, J.R., Griswold, M.D., 2004. The murine testiculartranscriptome: characterizing gene expression in the testis during the progres-sion of spermatogenesis. Biol. Reprod. 71, 319–330.

ngman, W.V., Robertson, S.A., 2002. Defining the actions of transforming growthfactor beta in reproduction. Bioessays 24, 904–914.

tman, C., Mendis, S., Barakat, B., Loveland, K.L., 2006. All in the family: TGF-betafamily action in testis development. Reproduction 132, 233–246.

rummond, A., Findlay, J., 2006. Focus on TGF-beta signalling. Reproduction 132,177–178.

onrad, L., Luers, G.H., Volck-Badouin, E., Keilani, M.M., Laible, L., Aumuller, G., Hof-mann, R., 2006a. Analysis of the mRNA expression of the TGF-Beta family intesticular cells and localization of the splice variant TGF-beta2B in testis. Mol.Reprod. Dev. 73, 1211–1220.

uller, R., Klug, J., Rodewald, M., Meinhardt, A., 2005. Macrophage migrationinhibitory factor suppresses transforming growth factor-beta2 secretion in cul-tured rat testicular peritubular cells. Reprod. Fertil. Dev. 17, 435–438.

agener, A., Fickel, J., Schon, J., Fritzenkotter, A., Goritz, F., Blottner, S., 2005. Seasonalvariation in expression and localization of testicular transforming growth factorsTGF-{beta}1 and TGF-{beta}3 corresponds with spermatogenic activity in roe

deer. J. Endocrinol. 187, 205–215.

ia, W., Cheng, C.Y., 2005. TGF-beta3 regulates anchoring junction dynamics in theseminiferous epithelium of the rat testis via the Ras/ERK signaling pathway: anin vivo study. Dev. Biol. 280, 321–343.

ia, W., Mruk, D.D., Lee, W.M., Cheng, C.Y., 2006. Differential interactionsbetween transforming growth factor-beta3/TbetaR1, TAB1, and CD2AP dis-

8 ellula

K

0 M.A. Memon et al. / Molecular and C

rupt blood–testis barrier and Sertoli-germ cell adhesion. J. Biol. Chem. 281,16799–16813.

onrad, L., Keilani, M.M., Laible, L., Nottelmann, U., Hofmann, R., 2006b. Effects ofTGF-betas and a specific antagonist on apoptosis of immature rat male germcells in vitro. Apoptosis 11, 739–748.

M

r Endocrinology 294 (2008) 70–80

aire, M., Florin, A., Kaszas, K., Regnier, D., Contard, P., Tabone, E., Mauduit, C.,Bars, R., Benahmed, M., 2005. Alteration of transforming growth factor-betasignaling system expression in adult rat germ cells with a chronic apop-totic cell death process after fetal androgen disruption. Endocrinology 146,5135–5143.

Transforming growth factor beta (TGFbeta1, TGFbeta2 and TGFbeta3) null-mutant phenotypes in embryonic gonadal developmentIntroductionMaterials and methodsAnimalsGenotyping and SRY analysisSemi-quantitative PCR analysisHistologyMorphological analysisGerm cell nuclear antigen (GCNA) stainingTesticular cell apoptosisMicroarray analysisStatistical analysis

ResultsDiscussionAcknowledgementReferences