A

i(iwh�efte©

K

1

oods2augc

Available online at www.sciencedirect.com

Theriogenology 74 (2010) 1707–1712

0d

Technical note

Immunohistochemical localization of aromatase during thedevelopment and atresia of ovarian follicles in prepubertal horses

Wieslawa Mlodawskaa,*, Maria Slomczynskab

a University of Agriculture, Department of Animal Reproduction and Anatomy, 30-059 Krakow, Al. Mickiewicza 24/28; Polandb Department of Endocrinology and Tissue Culture, Institute of Zoology, Jagiellonian University, 30-060 Krakow, Poland

Received 21 July 2009; received in revised form 15 April 2010; accepted 18 April 2010

bstract

Ovarian steroidogenesis from the neonatal to pubertal period in horses is poorly understood. This study was designed tommunolocalize cytochrome P450 aromatase in the ovarian follicles of slaughtered fillies ages approximately (I) 6–9 mo�10MF); (II) 1 y (1YF); and (III) 1.5 y (1.5YF). The ovaries of adult mares were used as controls. In each age group,mmunoreactivity for P450arom was observed in the mural granulosa of nonatretic follicles �5 mm in diameter. Staining intensityas dependent on the size and morphology of the follicle. In nonatretic follicles 5–10 mm in diameter, the reaction was weak andeterogeneous, while most intense staining was observed in preovulatory follicles. In follicles (diameter �20 mm) in the groups10MF and 1YF, the reaction was less intense than in adult mare follicles of similar size. In each age group, several follicles with

arly or advanced signs of atresia exhibited a heterogeneous staining pattern, which subsequently disappeared in late atreticollicles. No immunoreactivity was detected in the theca interna, preantral follicle, or stroma cells. Our observations reveal thathe mural granulosa of viable follicles in fillies about 6–18 mo old contains aromatase, indicating that the ovary is capable ofstrogen synthesis. Immunoreactivity for P450arom was dependent on follicle size and disappeared in atretic follicles.

2010 Elsevier Inc. All rights reserved.

eywords: Fillies; Follicles; Aromatase; Immunohistochemistry

www.theriojournal.com

atctcr

pSe[taw

. Introduction

In females, the development of ovarian follicles,vulation, and puberty are strictly related and dependn the hypothalamic-pituitary-ovarian axis. Studies toate have suggested that, unlike the situation in rats [1],heep [2], or cows [3,4], prepubertal horses (at least–10 months old) show no feedback between the ovarynd the hypothalamic-pituitary axis [5–7]. Prior to sex-al maturity in fillies, FSH secretion and follicularrowth seem to occur as independent processes. Folli-les grow in a wave-like fashion and then regress, while

* Corresponding author. Tel: �4812662-41-12.

iE-mail address: rzmlodaw@cyf-kr.edu.pl (W. Mlodawska).

093-691X/$ – see front matter © 2010 Elsevier Inc. All rights reserved.oi:10.1016/j.theriogenology.2010.04.019

low LH concentration prevents the development ofhe dominant follicle [6,7]. In prepubertal rats, in-reased estradiol secretion has been shown to stimulatehe first, preovulatory surge of LH [1,8]. In 6 mo oldows, exogenous estradiol was found to induce theelease of LH and ovulation [4].

Ovarian steroidogenesis from the neonatal to theubertal period in horses has been poorly investigated.o far, studies have focused on examining blood serumstradiol and/or progesterone concentrations in fillies9,10]. These studies have identified high concentra-ions of both hormones in newborn fillies, followed by

decrease to undetectable concentrations in the firsteeks of life. Progesterone concentrations seem only to

ncrease during the first estrous cycle, while estradiol

1708 W. Mlodawska, M. Slomczynska / Theriogenology 74 (2010) 1707–1712

rtbiTobpoiti

2

2

t(fi6afam(wbsemaaf

olmffosealmlts

2

nc(naaw

rhiGdntsl

Fhga

(

(

(m

(kn

(t

(i

1709W. Mlodawska, M. Slomczynska / Theriogenology 74 (2010) 1707–1712

emains unchanged when progesterone peaks or fluc-uates [10]. On the other hand, most fillies show estrousehavior prior to first ovulation [9], which suggestsncreased estradiol secretion at least during this period.o determine the mechanism that controls maturationf the hypothalamic-pituitary-ovarian axis in horses, aetter understanding of ovarian steroidogenesis in pre-ubertal and pubertal fillies is required. The objectivef the present study was to evaluate the cellular local-zation of cytochrome P450 aromatase (P450arom) inhe follicles of fillies about 6- to 18 mo old using anmmunohistochemical technique.

. Materials and methods

.1. Tissue collection and preparation

The study was performed on 54 ovaries of slaugh-ered fillies (n � 27) and 8 ovaries of adult maresAdM; n � 4, control group). The ovaries were classi-ed to one of the 3 groups according to filly age: (I)–9 mo (�10MF); (II) approximately 1 year (1YF);nd (III) approximately 1.5 years (1.5YF). Three to sixollicles were isolated from each ovary, and assignedccording to diameter to one of the 4 groups: 1) �10m, 2) 11–20 mm, 3) 21–30 mm or 4) �30 mm

preovulatory). Follicles or sections of follicle wallsere fixed in 4% paraformaldehyde, dehydrated, em-edded in paraplast, and cut into serial sections (6 �m),ome of which were stained with hematoxylin andosin for histomorphological analysis. According to theodified criteria of Kenney et al [11] and Driancourt et

l [12], follicles were classified as nonatretic (viable) ortretic: early (AI), advanced (AII) or late (AIII). Viableollicles showed a mural granulosa with several layers

ig. 1. Filly and mare nonatretic (A–I) and atretic (J–R) folliclistomorphology of the follicles visualized by hematoxylin and eosinranulosum (A, D and G); panel 2: immunohistochemical staining; pntral cavity (a); scale bar � 50 �m: (A–H and J–O); scale bar �

A–C): Preovulatory filly follicle �35 mm in diameter; note the int

D–F): Mare follicle 15–18 mm in diameter; note the lower amount

G–I): Filly follicle of about 15 mm in diameter; the reaction for P4are follicles of comparable size (E).

J–L): Early atretic filly follicle of about 25 mm in diameter showaryorrhexis (white arrows) and pyknosis of GC nuclei (black arrowonatretic follicles.

M–O): Advanced atretic mare follicle of about 18 mm in diameter she weak immunostaining for P450arom (N).

P–R): Late atretic filly follicle of diameter 8–10 mm showing no g

mmunopositive staining (arom-). No staining was observed in any of the c

f epithelial cells, columnar arrangement of the basalayer of granulosa cells (GCs) adjoining the basementembrane, and numerous blood vessels visible on the

ollicular wall (Figs. A, D and G). Early atretic (AI)ollicles had close to normal granulosa cells, foci of kary-rrexis and apoptotic GC nuclei were could be seen on theection (Fig. J). Advanced atretic (AII) follicles exhibitedxtensive pyknosis in mural granulosa, numerous GCsnd apoptotic bodies floating in the antrum, bi- or mono-ayer granulosa, and gradual thickening of basementembrane (Fig. M). Late atretic (AIII) follicles showed a

oss of the separate identity of granulosa and theca cells, ahickened basement membrane, and few or no blood ves-els (Fig. P).

.2. Immunohistochemistry

Since homologous antibody for equine aromatase isot commercially available, we used a mouse mono-lonal IgG against human cytochrome P450 aromataseSerotec, Milan, Italy) at a 1:50 dilution. A heteroge-eous antibody has been used by other authors to detectromatase in equine gonads [13,14]. Biotinylated horsenti-mouse IgG (1:400; Vector Laboratories, Inc; USA)as used as a secondary antibody.Before immunohistochemical staining, sections were

outinely deparaffinized and rehydrated. Sections wereeated in a microwave oven in 0.01M citrate buffer andncubated in 3% H2O2 (Sigma-Aldrich, Co. Schnelldorf,ermany) in TBS (Tris-Buffered Saline); as previouslyescribed [15]. Sections were incubated first with 3%ormal horse serum (Vector Laboratories INC; USA) andhen with primary antibody overnight at 4 °C. Finally, theections were incubated with secondary antibody fol-owed by avidin-biotinylated horseradish peroxidase com-

s correspond to the wall section of the same follicle; panel 1:g; note the columnar arrangement of the basal layer of the stratum

negative control sections; granulosa cells (g); theca interna cells (th);(I and P–R).

munopositive staining for P450arom in the cytoplasm of GCs (B).

ning (E) than that seen in the GCs of preovulatory follicles (B).

is heterogeneous, well visible (H) but less intense than in GCs of

e lost columnar orientation of the basal layer of the GCs; foci ofte the reduced staining for P450arom compared to that seen in the

a thin granulosa and thickened basement membrane (asterisk). Note

a or theca cells and a thickened basement membrane (asterisks); no

es; rowstainin

anel 3:100 �m

ense im

of stai

50arom

ing ths); no

howing

ranulos

ontrol sections (C, F, I, L, O and R).

pwmCCfs

2

tYwc

3

3a

sgtsp

dhisAii

vw1tldf(mtc

4

Pm

TI

G

�

1

1

A

Ne

1710 W. Mlodawska, M. Slomczynska / Theriogenology 74 (2010) 1707–1712

lex (Vector Laboratories, Inc; USA). The color reactionas visualized using 0.05% diaminobenzidine (MP Bio-edicals, Inc, USA), 0.07% imidazole (Sigma-Aldrich,o. Schnelldorf, Germany) and 0.01% H2O2 in TBS.ontrols were processed as described above, except that

or negative controls, the primary antibody was omitted orubstituted with irrelevant IgG.

.3. Statistical analysis

The effects of filly age on number of immunoposi-ive follicles were analyzed by the chi-squared test.ates’ correction for continuity was applied when dataere presented in a two by two table. Differences were

onsidered statistically significant when P � 0.05.

. Results

.1. Cellular localization of cytochrome P450romatase

Regardless of the age of the females, an immunopo-itive reaction for P450arom was observed in the muralranulosa cells cytoplasm in nonatretic follicles largerhan 5 mm in diameter (Table 1). In all the age groups,taining intensity was dependent on the size and mor-hology of the follicle. Thus, nonatretic follicles of

able 1mmunolocalization of aromatase in the mural granulosa of filly and

roup Follicle category Diameter of fol

� 10 mm*

10MF (N � 8) Viable 6/8AI 0/2AII 0/5AIII 0/31Total arom�/n (%) 6/46 (13.0)

YF (N � 10) Viable 6/9AI 1/5AII 0/9AIII 0/16Total arom�/n (%) 7/39 (17.9)

.5YF (N � 9) Viable 5/7†AI 2/5AII 0/4AIII 0/13Total arom�/n (%) 7/29 (24.1)

dM (N � 4) Viable 3/3AI 0/1AII 0/3AIII 0/8Total arom�/n (%) 3/15 (20.0)

: number of females (pairs of ovaries); arom�/n: number of folliclearly (AI), advanced (AII) and late (AIII).* an immunopositive reaction for P450arom was only found for follicle

† detected in the 1.5YF group. Values in the last column with different su

iameter 5–10 mm (in each group), showed a weak,eterogeneous reaction while greater amounts of stain-ng were seen in the larger follicles. The most intensetaining was observed in preovulatory follicles (Fig. 1.–B, D–E and G–H). In follicles (diameter � 20 mm)

n groups �10MF and 1YF, immunostaining was lessntense than in adult mare follicles of similar size.

In each age group, several follicles with early or ad-anced signs of atresia showed heterogeneous staining,hich completely disappeared in late atretic follicles (Fig.. J–K, M–N and P–Q). The most atretic and at the sameime immunonegative follicles were observed among fol-icles � 10 mm in diameter. No immunoreactivity wasetected in theca interna cells, preantral follicles, or antralollicles � 5 mm. An exception was a small folliclediameter 2–3 mm) in the 1.5YF group, in which theural granulosa yielded a color reaction. No color reac-

ion was detected in ovarian stroma cells or in negativeontrol sections (Fig. 1. C, F, I, L, O and R).

. Discussion

The present study reports the immunolocalization of450arom in the ovarian follicles of fillies approxi-ately 6–18 mo old. As in the adult mares, in all the

ollicles.

rom�/n) Total arom�/n

20 mm 21–30 mm � 30 mm

/2 1/1 — 9/11/4 1/1 — 2/7/6 — — 0/11/3 — — 0/34(20.0) 2/2 (100) — 11/63a (17.5)/7 1/1 2/2 16/19/2 2/2 1/1 4/10/15 0/7 — 2/31/12 0/2 — 0/30(25.0) 3/12 (25.0) 3/3 (100) 22/90a,b (24.4)/5 2/2 2/2 14/16/3 1/1 — 6/9/9 2/7 — 3/20/8 — — 0/21(36.0) 5/10 (50.0) 2/2 (100) 23/66b,c (34.8)/6 2/2 1/1 12/12/2 1/1 — 2/4/6 1/2 — 2/11/1 0/1 — 0/10(53.3) 4/6 (66.7) 1/1 (100) 16/37c (43.2)

nopositive for P450arom/total number of follicles; Atretic follicles:

m in diameter, except for one positive follicle about 2-3 mm in diameter

mare f

licles (a

11–

2100

3/157020

9/365310

9/256110

8/15

s immu

s �5 m

perscripts differ significantly: a,b,c � P � 0.05.

amwfapndGWlirmssfwtit3tme

gasmslatcseocgllt

cIfersg

aIhttmrdcsfiscb2aoftocdatdlofigrlhsfomslPpnpba

A

1711W. Mlodawska, M. Slomczynska / Theriogenology 74 (2010) 1707–1712

ge groups of fillies, aromatase was detected in theural granulosa of viable follicles � 5 mm in diameter,ith the single exception of a small 2–3 mm diameter

ollicle also showing immunostaining for aromatase. Indult mare ovaries, Watson and Thompson [13] re-orted immunoreactivity for P450arom in the GCs ofonatretic follicles, which were larger than 5 mm iniameter. Similar findings have been reported byoudet et al [16] and Belin et al [17], who used theestern blot method to detect this enzyme. In these

atter studies, aromatase contents were only determinedn GCs collected from follicles � 5 mm. Thus, itemains unclear whether equine follicles smaller than 5m contain aromatase and are capable of estradiol

ecretion. Other studies have demonstrated that theynthesis of P450arom mRNA is strictly related toollicular diameter and begins in early antral follicleshose size varies according to species. Thus, in cows,

he presence of mRNA for P450arom was first detectedn the GCs of follicles 4–6 mm diameter and attributedo follicle recruitment [18]. In pig follicles of diameter–4 mm, not all granulosa cells showed immunoposi-ive staining [15]. In prepubertal goats P450aromRNA was detected in follicles 0.2–0.8 mm in diam-

ter [19].In the present study, we show that in the mural

ranulosa of the follicles of fillies and adult mares, themount of aromatase present varies according to theize and morphology of the follicle. Most intense im-unostaining was observed in preovulatory follicles. A

imilar relationship between reaction intensity and fol-icular size in adult mares was reported by Almadhidi etl [20]. In cycling mares, the amount of aromatase inhe follicles was dependent on the stage of the estrousycle and was higher in GCs from dominant thanubordinate follicles [16,17]. In our study, somearly and advanced atretic follicles exhibited heter-geneous staining for P450arom but this stainingompletely disappeared with ongoing atresia. Inoats, follicular atresia has been associated with theoss of mRNA for P450arom in GCs [19]. In cattle,ow amounts of P450arom mRNA have been de-ected in early and advanced atretic follicles [18].

In a prior study, we noted that the ovaries of filliesontained mostly follicles � 10 mm in diameter [21].n vivo studies have revealed that in prepubertal fillies,ollicles grow to reach a certain size (few folliclesxceed a diameter of 24 mm), after which they undergoegression without differentiation into dominant andubordinate follicles [6,7]. Accordingly, among

rowing follicles, none are available for ovulation S

nd all of these will sooner or later undergo atresia.n our study, most filly follicles showed advanced oreavy atresia accompanied by a lack of immunoreac-ivity for P450arom. Simultaneously, the immunoposi-ive reaction in filly follicles was less intense than inare follicles of similar size, which may suggest the

educed ability of the filly follicles to synthesize estra-iol. However, it should be stressed that immunohisto-hemical analysis was only performed on randomlyelected (3–6) follicles from each ovary and thus de-nitive conclusions on the ability of the whole ovary toynthesize estrogens cannot be drawn. Analysis of LHoncentrations in filly plasma has revealed that a smallut appreciable increase in LH concentrations begins4–32 d before first ovulation [9]. At the same time, indult mares LH has been identified as the main gonad-tropin controlling the development of the dominantollicle after deviation [22]. These observations suggesthat the development of the first follicle destined forvulation takes a relatively long time in the horse. Itan therefore be assumed that the growing follicle pro-uces increasing amounts of estradiol, which addition-lly stimulates LH secretion. The increase in LH secre-ion would in turn promote further development of theominant follicle and ovulation. Consequently, it isikely that a positive feedback mechanism between thevary and the hypothalamic-pituitary axis develops inllies during this period. These observations also sug-est that the ovary’s ability to synthesize estradiol di-ectly increases during the period preceding first ovu-ation. Further research is needed to verify thisypothesis. In conclusion, our study for the first timehows the immunolocalization of aromatase in ovarianollicles in fillies around 6–18 mo old, indicating thevary’s ability to synthesize estrogens. As in adultares, the presence of aromatase in fillies was re-

tricted to the mural granulosa of growing folliclesarger than 5 mm in diameter. Immunoreactivity for450arom was dependent on follicular size and disap-eared in atretic follicles. However, further research iseeded to explain whether estrogens produced in therepubertal period are responsible for establishing feed-ack between the ovary and the hypothalamic-pituitaryxis in the horse.

cknowledgements

This work was supported by the Polish Ministry of

cience and Higher Education; grant No 2 P06D 042 30.

R

[

[

[

[

[

[

[

[

[

[

[

[

[

1712 W. Mlodawska, M. Slomczynska / Theriogenology 74 (2010) 1707–1712

eferences

[1] Ojeda SR, Urbanski HF. Puberty in the rat. In: E. Knobil, J.D.Neil. The Physiology of Reproduction, Second Edition. RavenPress, Ltd, New York, 1994. p. 363–409.

[2] Sonjaya H, Driancourt MA. FSH concentrations and sensitivityto feedback in infant lambs from breeds differing in prolificacy.J Reprod Fertil 1989;85:461–9.

[3] Day ML, Imakawa K, Garcia-Winder M, Zalesky DD, Schan-bacher BD, Kittok RJ, Kinder JE. Endocrine mechanisms ofpuberty in heifers: estradiol negative feedback regulation ofluteinizing hormone secretion. Biol Reprod 1984;31:332–41.

[4] Schoppee PD, Armstrong JD, Harvey RW, Washburn SP, FelixA, Campbell RM. Endocrine and ovarian responses to exoge-nous oestradiol-17� in 6-month-old heifers previously immu-nized against growth hormone-releasing factor. J Anim Sci1995;73:2071–8.

[5] Wesson JA, Ginther OJ. Plasma gonadotropin levels in intactand ovariectomized prepubertal ponies. Biol Reprod 1979;20:1099–104.

[6] Nogueira GP, Ginther OJ. Dynamics of follicle populations andgonadotropin concentrations in fillies age two to ten months.Equine Vet J 2000;32:482–8.

[7] Nogueira GP. Follicle profile and plasma gonadotropin concen-tration in pubertal female ponies. Braz J Med Biol Res 2004;37:913–22.

[8] Parker CR Jr, Mahesh VB. Hormonal events surrounding thenatural onset of puberty in female rats. Biol Reprod 1976;14:347–53.

[9] Wesson JA, Ginther OJ. Puberty in the female pony: reproduc-tive behavior, ovulation, and plasma gonadotropin concentra-tions. Biol Reprod 1981;24:977–86.

10] Nogueira GP, Barnabe RC, Verreschi ITN. Puberty and growthrate in thoroughbred fillies. Theriogenology 1997;48:581–8.

11] Kenney RM, Condon W, Ganjam VK, Channing C. Morpho-logical and biochemical correlates of equine ovarian follicles asa function of their state of viability or atresia. J Reprod FertSuppl 1979;27:163–71.

12] Driancourt MA, Paris A, Roux C, Mariana JC, Palmer E. Ovar-ian follicular populations in pony and saddle-type mares. Re-

prod Nutr Develop 1982;22:1035–47.

13] Watson ED, Thomson SRM. Immunolocalization of aromataseP-450 in ovarian tissue from pregnant and nonpregnant maresand in ovarian tumours. J Reprod Fertil 1996;108:239–44.

14] Hess MF, Roser JF. Immunocytochemical localization of cyto-chrome P450 aromatase in the testis of prepubertal, pubertal,and postpubertal horses.Theriogenology 2004;61:293–9.

15] Slomczynska M, Tabarowski Z. Localization of androgen re-ceptor and cytochrome P450 aromatase in the follicle and cor-pus luteum of the porcine ovary. Anim Reprod Sci 2001;65:127–34.

16] Goudet G, Belin F, Bézard J, Gérard N. Intrafollicular contentof luteinizing hormone receptor, �-inhibin, and aromatase inrelation to follicular growth, estrous cycle stage, and oocytecompetence for in vitro maturation in the mare. Biol Reprod1999;60:1120–7.

17] Belin F, Goudet G, Duchamp G, Gérard N. Intrafollicular con-centrations of steroids and steroidogenic enzymes in relationto follicular development in the mare. Biol Reprod 2000;62:1335–43.

18] Bao B, Garverick HA, Smith GW, Smith MF, Salfen BE,Youngquist RS. Changes in messenger ribonucleic acid encod-ing luteinizing hormone receptor, cytochrome P450-side chaincleavage, and aromatase are associated with recruitment andselection of bovine ovarian follicles. Biol Reprod 1997;56:1158–68.

19] Yuan JH, Wang JZ, Lan GC, Sui HS, Yu JN, Tan JH. Expres-sion of steroidogenic enzymes and synthesis of steroid hor-mones during development of ovarian follicles in prepubertalgoats. Domest Anim Endocrinol 2008;34:451–60.

20] Almadhidi J, Seralini GE, Fresnel J, Silberzahn P, Gaillard JL.Immunohistochemical localization of cytochrome P450 aromatasein equine gonads. J Histochem Cytochem 1995;43:571–7.

21] Mlodawska W, Okolski A. Morphological characterization andmeiotic competence of oocytes collected from filly ovaries.Theriogenology 2009;71:1046–53.

22] Gastal EL, Bergfelt DR, Nogueira GP, Gastal MO, Ginther OJ.Role of luteinizing hormone in follicle deviation based on ma-nipulating progesterone concentrations in mares. Biol Reprod

1999;61:1492–8.