ORIGINAL ARTICLE

Increased microRNA-221/222 and decreased estrogenreceptor α in the cervical portion of the uterosacralligaments from women with pelvic organ prolapse

Zhonghua Shi & Ting Zhang & Lei Zhang & Jing Zhao &

Jian Gong & Chun Zhao

Received: 22 August 2011 /Accepted: 9 February 2012 /Published online: 21 April 2012# The International Urogynecological Association 2012

AbstractIntroduction and hypothesis The aim of this study was tocompare the expression of microRNA (miR)-221, miR-222,and estrogen receptor α (ERα) in the uterosacral ligamentsof women with and without pelvic organ prolapse (POP).Methods Histologically confirmed full-thickness uterosacralligament biopsies were procured during hysterectomies from40 POP patients and 40 postmenopausal women withoutprolapse. Expression of miR-221/222 was determined byquantitative real-time polymerase chain reaction (PCR), andERα protein expression was analyzed byWestern blotting andimmunohistochemistry.Results The mean expression levels of miR-221/222 bothincreased by approximately twofold in women with POPrelative to controls, while ERα protein levels in POPpatients were significantly lower than controls. Negativecorrelations were observed between ERα protein expression

and both miR-221 (r0−0.8542) and miR-222 (r0−0.861) inPOP patients.Conclusions Elevated miR-221/222 expression levels areassociated with, and may be responsible for, reduced ERαexpression in the cervical portion of uterosacral ligaments ofpatients with POP. MiR-221/222 may serve as potentialtherapeutic targets for POP.

Keywords Pelvic organ prolapse .MicroRNA-221 .

MicroRNA-222 . Estrogen receptorα . Uterosacral ligament

AbbreviationsPOP Pelvic organ prolapseER Estrogen receptormiR MicroRNAERT Estrogen replacement therapyPOP-Q Pelvic Organ Prolapse Quantification system

Introduction

Pelvic organ prolapse (POP) is a common gynecologicaldisease with a multifactorial etiopathogenesis, which nega-tively affects quality of life in millions of women [1–3]. As lifeexpectancy increases, increased numbers of women will pres-ent with POP and require surgical intervention. The pelvicorgans are supported by pelvic floor muscles, ligaments, pel-vic bones, and fascial supports [4]. It is hypothesized thatchanges in the pelvic connective tissues, including the utero-sacral ligaments, may contribute to POP [5, 6]. Uterosacralligaments are the main supportive structures of the uterus andvagina and are often weakened in women with POP [7].

Zhonghua Shi, Ting Zhang, and Lei Zhang contributed equally to thiswork.

Z. ShiState Key Laboratory of Reproductive Medicine,Nanjing Maternity and Child Health Care Hospital Affiliatedto Nanjing Medical University,Nanjing 210004, China

L. Zhang : C. Zhao (*)Nanjing Maternity and Child Health Care Hospital Affiliatedto Nanjing Medical University,Nanjing 210004, Chinae-mail: zhaochun2008@yeah.net

T. Zhang : J. Zhao : J. Gong (*)Wuxi Maternity and Child Health Hospital Affiliatedto Nanjing Medical University,Wuxi 214002, Chinae-mail: gongjianwx@126.com

Int Urogynecol J (2012) 23:929–934DOI 10.1007/s00192-012-1703-5

The prevalence of uterine prolapse increases after meno-pause, which suggests that a hypoestrogenic state may con-tribute to the etiology of prolapse [8]. The biologicalactivities of estrogens are mediated via two geneticallydistinct receptors, estrogen receptor (ER) α and β, whichfunction as hormone-inducible transcription factors [9].Both ER genes are expressed in pelvic muscle and ligaments[10], prompting further research on the role of ER expres-sion in the urogenital support system of POP patients. ERβexpression in the cardinal ligaments does not differ betweenPOP and control postmenopausal women [11]; however,Lang et al. [12] and Bai et al. [13] reported that ERα proteinexpression is reduced in both premenopausal and postmen-opausal patients with POP. ERα protein deficiency maycontribute to the induction or progression of POP; however,the factors regulating ERα deficiency are not known.

MicroRNAs (miRNAs) are small noncoding RNAs thatbind to target mRNA regulatory sites to modify their ex-pression, either by translational repression or mRNA degra-dation [14]. MiRNA expression is deregulated in manydiseases, suggesting that miRNAs play an important rolein the pathogenesis and progression of different conditions.ERα is a target gene for both miR-221 and miR-222, andERα protein expression can be inhibited by miR-221/222[15]. This study aims to compare miR-221/222 and ERαprotein expression in the cervical portion of uterosacralligaments of women with and without POP to determinethe association between miR-221/222 and ERα in POPpathogenesis.

Materials and methods

Tissue collection

Women undergoing hysterectomy at the Department ofGynecology of Wuxi and Nanjing Maternity and ChildHealth Hospital, affiliated to Nanjing Medical University,between 2007 and 2009 were asked to participate in thestudy and provided informed consent. The protocol wasapproved by the Institutional Review Board of NanjingMedical University. The exclusion criteria were gynecologicalmalignancy or any history of pelvic radiation therapy, estro-gen replacement therapy (ERT), previous pelvic surgery, col-lagen deficiency syndromes, paralysis of the pelvic floormuscles, or chronic lung disease manifested by chronic cough.

The grade of POP was estimated according to the Interna-tional Continence Society Pelvic Organ Prolapse Quantifica-tion system (POP-Q). Forty women undergoing hysterectomyas part of reconstructive pelvic surgery with POP stage II orabove were enrolled in the POP group. In the POP group, 22women had stage II POP and 18 women had stage III POP.The womenwith POP stage II underwent hysterectomy as part

of the reconstructive pelvic surgery. These women wereelderly and their quality of life was obviously decreased astheir condition was also complicated by cystocele and/orrectocele. We performed the hysterectomy with repair of theanterior vaginal wall and/or posterior vaginal wall. Fortywomen undergoing hysterectomy due to uterine leiomyomaswith POP stage I or less served as controls. These postmeno-pausal women underwent hysterectomy as their leiomyomaswere too large and did not reduce in size after menopause.Three 5-mm thick biopsies were obtained from the medialends of the bilateral uterosacral ligaments with the use of ascalpel held between two clamps during abdominal or vaginalhysterectomy. Biopsies were made at the level of the cervicalinsertions to the uterus.

The specimens were divided into two, which were eitherimmediately frozen in liquid nitrogen or fixed in formalinand embedded in paraffin (FFPE). All of the tissue speci-mens and slides were examined by an experienced patholo-gist to ensure that they were indeed uterosacral ligamenttissue.

RNA extraction and miRNA quantification

Total RNA, including mRNA, was extracted from the ute-rosacral ligament samples using TRIzol reagent. MiR-221,miR-222, and RNU6B-specific cDNA were synthesizedusing gene-specific primer sets according to the TaqManmRNA assay protocol (Applied Biosystems, Foster City,CA, USA). RNU6B was used as internal control. Reversetranscription was performed in 7.5 μl reactions containing10 ng total RNA, 50 nmol stem loop RT primer, 1× RTbuffer, 0.25 mmol dNTPs, and 3.33 U/μl RNase inhibitor inthe GeneAmp PCR System 7500 (Applied Biosystems) for30 min at 16 °C, 30 min at 42 °C, and 5 min at 85 °C.Quantitative real-time polymerase chain reaction (PCR) wasperformed using the ABI StepOne Real-Time PCR System(Applied Biosystems) in 10 μl reactions containing 0.67 μlRT product, 1× Taqman Universal PCR master mix, and1 μl primer and probe mix in 48-well optical plates, at 95 °Cfor 10 min, followed by 40 cycles at 95 °C for 15 s and 60 °Cfor 30 s. Each sample was analyzed in duplicate. Relativequantification of mRNA expression was calculated using the2-ΔΔCT method, where ΔΔCT0ΔCT (CTmiR-221/222-CTRNU6B) POP-ΔCT (CTmiR-221/222-CTRNU6B) normal. Data arepresented as the relative quantity of target mRNA normalizedto RNU6B, which was expressed relative to a calibratorsample.

Western blotting

Frozen tissues were homogenized with an ULTRATURRAX homogenizer (Ika, Petaling Jaya, Malaysia) inlysis buffer containing 7 M urea, 2 M thiourea, 4 % (w/v)

930 Int Urogynecol J (2012) 23:929–934

CHAPS, 2 % (w/v) DTT, 1 % (v/w) Protease InhibitorCocktail Kit (Pierce Biotechnology, Rockford, IL, USA) at11,000 IU/min on ice with 10 bursts of 10 s. The suspen-sions were shaken at 4 °C for 1 h, after which the insolublefraction was removed by centrifugation at 40,000 g for 1 h at4 °C. Protein concentrations were determined by theBradford method using bovine serum albumin (BSA) asthe standard. Then, 50–100 μg protein was electrophoresedon 12 % sodium dodecyl sulfate (SDS) polyacrylamide gelsaccording to our previous study [16], transferred to nitrocel-lulose membranes (GE Healthcare, San Francisco, CA,USA), blocked in Tris-buffered saline (TBS) containing5 % nonfat milk powder for 1 h, and incubated with anti-ERα polyclonal antibody (1:500, Abcam, Cambridge, MA,USA) and anti-β-actin polyclonal antibody (1:1,000;Abcam) in TBS/5 % nonfat milk powder overnight.β-Actin was used as the loading control. Membranes werewashed in TBS, incubated with horseradish peroxidase(HRP)-conjugated goat anti-rabbit IgG (1:1,000, BeijingZhongShan Biotechnology, Beijing, China) for 1 h, andvisualized using ECL and the AlphaImager (FluorChem5500, Alpha Innotech, San Leandro, CA, USA). Proteinexpression was analyzed using AlphaEaseFC software(Alpha Innotech). The relative expression of ERα equaledthe intensity of the ERα band divided by the intensity of theβ-actin band.

Immunohistochemistry

FFPE biopsy specimens were sectioned into 5-μm thicksections, mounted on 3-aminopropyltriethoxysilane(APES)-coated slides, dewaxed, and rehydrated throughdescending grades of alcohol to distilled water. Endogenousperoxidase activity was blocked in 3 % (v/v) hydrogenperoxidase in phosphate-buffered saline (PBS) and antigenretrieval was performed in the microwave using 0.02 Methylenediaminetetraacetic acid (EDTA). Sections werewashed in PBS, blocked with goat or rabbit serum (BeijingZhongShan Biotechnology) for 2 h, and incubated overnightat 4 °C with polyclonal ERα antibody (1:50). Negativecontrols were incubated without primary antibody. Afterprimary antibody incubation, sections were washed in PBSand incubated with HRP-conjugated secondary antibody(1:1,000, Beijing ZhongShan Biotechnology) for 1 h atroom temperature. Immunoreactivity was revealed usingdiaminobenzidine peroxidase substrate (DAB, Sigma, St.Louis, MO, USA) and sections were counterstained withhematoxylin and mounted.

From each slide, 15 randomly selected × 40 fields werecaptured and the number of positive and negative nucleicounted using the image analysis software by a blindedexaminer [17].

Statistical analysis

SPSS (version 16, SPSS, Chicago, IL, USA) was used toperform t tests and Mann-Whitney tests. Pearson correlationcoefficients were used to determine the correlation betweenmiR-221/222 and ERα expression, and p values <0.05 wereconsidered to be statistically significant.

Results

Expression of miR-221/222 in the cervical portionof uterosacral ligaments of POP patients and controls

Clinical samples were obtained from 80 menopausal women,40 with POP and 40 controls. There was no difference in age,menopausal status, and body mass index (BMI) between thetwo groups (Table 1), and no patients were takinghormone replacement therapy at the time of enrolmentor surgery.

The normalized quantitative real-time PCR results showedthat expression of both miR-221 and miR-222 was upregu-lated in POP patients compared to women without prolapse(Fig. 1, p<0.01 for both). Additionally, the expression levelsof miR-221 and miR-222 were both significantly higher inpatients with stage III POP than in stage II POP (p00.037 andp00.009).

Expression of ERα protein in the cervical portionof uterosacral ligaments of POP patients and controls

Expression of ERα was evaluated using Western blottingand immunohistochemistry. Western blot analysis demon-strated that expression of ERα protein was 2.2-fold lower inthe cervical portion of uterosacral ligaments of POP patients(1.37±0.46) compared with the control group (3.02±0.44),despite the similarities in age, parity, menopausal status, andBMI between the two groups (Table 1). Furthermore, the

Table 1 Demographic and anthropometric characteristics of studyparticipants

Control(n040)

Pelvic organprolapse(n040)

p value

Median age (range) 58 (52–68) 61 (52–72) 0.052a

Mean body mass index±SD(kg/m2)

25.9±3.2 27.1±2.8 0.086b

Median years sincemenopause (range)

6 (1–11) 8 (1–15) 0.062a

Median parity (range) 2 (1–4) 2 (1–5) 0.105a

a t testbMann-Whitney test

Int Urogynecol J (2012) 23:929–934 931

expression of ERα protein was almost twofold lower in thecervical portion of uterosacral ligaments of POP II patients(1.73±0.28) compared with POP III patients (0.96±0.18)(Fig. 2). The average percentage of ERα-positive cells ineach group is shown in Fig. 3. Expression of ERα in non-prolapsed ligaments was 2.5 times greater than in prolapsedligaments (64.2±15.1 % versus 26.1±11.5 %, p00.008).

Correlations of miR-221/miR-222 expression and expressionof the target gene, ERα, in the cervical portion of uterosacralligaments of POP patients

Statistically significant inverse correlations were observedbetween the expression of miR-221 and ERα protein expres-sion (Fig. 4, r0−0.8542, r200.7296, p00.000) in the 40specimens obtained from POP patients. ERα protein expres-sion levels were based on the Western blot analysis. Further-more, a statistically significant inverse correlation was alsoobserved between miR-222 expression and ERα protein ex-pression (Fig. 4, r0−0.861, r200.7411, p00.000). Theseresults suggest that ERα protein expressionmay be negativelyregulated by both miR-221 and miR-222 in the cervical por-tion of uterosacral ligaments of POP patients.

Fig. 1 Quantitative real-timePCR expression analysis ofmiR-221/222 in the uterosacralligaments of 40 patients withprolapse (POP) and 40 controlpostmenopausal women. a, bMiR-221 and miR-222 expres-sion levels are both upregulatedin POP patients compared towomen without prolapse(*p<0.05). c, d Expressionlevels of miR-221 and miR-222are both significantly increasedin stage III POP compared tostage II POP (*p<0.05;**p<0.01)

Fig. 2 Expression of ERα protein in the uterosacral ligaments from 22POP II and 18 POP III patients and 40 control postmenopausal women. aWestern blot analysis of ERα protein in the uterosacral ligaments of POPII–III patients and control subjects; β-actin was used as the loadingcontrol. b Expression of ERα protein is lower in the uterosacral ligamentsof the POP group compared with the control group, and the expression ofERα protein is almost twofold lower in the POP III group compared withthe POP II group. Bars represent ERα/β-actin ratio±SD, *p<0.01

932 Int Urogynecol J (2012) 23:929–934

Discussion

It is well known that polymorphisms in the ERα gene areassociated with an increased risk of POP [18], and manystudies have indicated that reduced ERα expression occursin POP [12, 13]; however, the mechanisms that regulatereduced ERα protein levels are not known. This studyconfirms the previous studies of reduced ER expression inPOP [12, 13], as we observed that ERα protein is alsodownregulated in the cervical portion of the uterosacralligament of POP patients.

Through bioinformatic analyses (http://www.targetscan.org/), we found that ERα is a validated target for miR-221/222. Both miR-221/222 are important regulators ofERα and are major determinants of ERα status in humanbreast cancer [15, 19, 20]. As ERα status and estrogen areboth related to the development of breast cancer and POP [21,22], we hypothesized that miR221/222 might also have animportant role in the regulation of ERα protein expression inPOP patients. To our knowledge, this is the first study tocompare miR-221/222 expression in the cervical portion ofuterosacral ligaments of women with POP and those with no

prolapse, and we observed that both miR-221and miR-222 areupregulated in the cervical portion of uterosacral ligaments ofwomen with POP. MiR-221 and miR-222 are highly con-served in vertebrates with the same seed sequence and areencoded in tandem on the same chromosome in human,mouse, and rat [23]. This indicates that miR-221 and miR-222 have an important role in biological processes [24].

ERα is a validated target of miR-221 and 222 [15];however, a relationship between ERα protein and miR-221/222 has not previously been demonstrated in the path-ogenesis of POP. We observed a significant negative corre-lation between expression of ERα and both miR-221 andmiR-222 in POP patients. We speculate that reduced ERαprotein expression observed in POP may be mediated by thealtered posttranslational control of ERα due to increasedmiR-221 and miR-222 expression.

In summary, our results suggest that miR-221 and miR-222 expression is significantly increased in the cervicalportion of uterosacral ligaments of women with POP com-pared to women without POP. This finding suggests thatboth miR-221 and miR-222 could participate in the patho-genesis of POP. Elevated miR-221 and miR-222 levels

Fig. 4 Correlation of miR-221/miR-222 expression with theprotein expression of their targetgene, ERα, in the uterosacralligaments of 40 prolapse (POP)patients. a Regression line(dotted line) of relative miR-221and ERα expression in POPpatients (r200.7296, p00.000).b Regression line (dotted line)of relative miR-222 and ERαexpression levels in 40 POPpatients (r200.7411, p00.000)

Fig. 3 ERα immunohistochemical staining of uterosacral ligamentssections from 40 prolapse (POP) patients and 40 control postmeno-pausal women. a Representative images of ERα immunohistochemis-try (brown) with hematoxylin counterstaining of uterosacral ligamentsections from control and POP groups. These indicate increased ERα-positive cells in POP, ×400. The positive cells are indicated by arrows.

b Quantification of mean percentage±SD of ERα protein expression inthe uterosacral ligament tissue sections from 40 prolapse (POP)patients and 40 control postmenopausal women. The level of ERαexpression in non-prolapsed ligaments is significantly higher than inprolapsed ligaments (**p<0.01)

Int Urogynecol J (2012) 23:929–934 933

might be responsible for reduced ERα expression in thecervical portion of uterosacral ligaments of POP women.Therefore, these results indicate that the knockdown of miR-221 and/or miR-222 might restore ERα expression andestrogen sensitivity; therefore, miR-221 and miR-222 mightserve as potential therapeutic targets for POP. However, asthe regulatory pathways of ERα protein expression in POPpatients are highly complex, further studies are necessary toelucidate the altered regulation of ERα protein in the path-ogenesis of POP.

Acknowledgements This work was financially supported by theNational Natural Science Foundation of China (81000258,81100436),the Natural Science Foundation of Jiangsu Province (BK2010586), theKey Project of Nanjing Department of Health (YKK10035) and theFoundation of Nanjing Medical University (2010NJMUZ16) andNanjing Medical Science and Technique Development Foundation.

Conflicts of interest None.

References

1. Reisenauer C, Shiozawa T, Oppitz M et al (2008) The role ofsmooth muscle in the pathogenesis of pelvic organ prolapse—animmunohistochemical and morphometric analysis of the cervicalthird of the uterosacral ligament. Int Urogynecol J Pelvic FloorDysfunct 19:383–389

2. Dviri M, Leron E, Dreiher J et al (2011) Increased matrixmetalloproteinases-1,-9 in the uterosacral ligaments and vaginaltissue from women with pelvic organ prolapse. Eur J ObstetGynecol Reprod Biol 156:113–117

3. Vulic M, Strinic T, Tomic S et al (2011) Difference in expression ofcollagen type I and matrix metalloproteinase-1 in uterosacral liga-ments of women with and without pelvic organ prolapse. Eur JObstet Gynecol Reprod Biol 155:225–228

4. Goh JT (2003) Biomechanical and biochemical assessmentsfor pelvic organ prolapse. Curr Opin Obstet Gynecol 15:391–394

5. Suzme R, Yalcin O, Gurdol F et al (2007) Connective tissuealterations in women with pelvic organ prolapse and urinaryincontinence. Acta Obstet Gynecol Scand 86:882–888

6. Moalli PA, Shand SH, Zyczynski HM et al (2005) Remodeling ofvaginal connective tissue in patients with prolapse. Obstet Gynecol106:953–963

7. Connell KA, Guess MK, Chen H et al (2008) HOXA11 is criticalfor development and maintenance of uterosacral ligaments anddeficient in pelvic prolapse. J Clin Invest 118:1050–1055

8. Mokrzycki M, Mittal K, Smilen S et al (1997) Estrogen andprogesterone receptors in the uterosacral ligament. Obstet Gynecol90:402–404

9. Hall JM, McDonnell DP (2005) Coregulators in nuclear estrogenreceptor action: from concept to therapeutic targeting. Mol Interv5:343–357

10. Smith P, Heimer G, Norgren A et al (1990) Steroid hormonereceptors in pelvic muscles and ligaments in women. GynecolObstet Invest 30:27–30

11. Ewies AA, Thompson J, Al-Azzawi F (2004) Changes in gonadalsteroid receptors in the cardinal ligaments of prolapsed uteri:immunohistomorphometric data. Hum Reprod 19:1622–1628

12. Lang JH, Zhu L, Sun ZJ et al (2003) Estrogen levels and estrogenreceptors in patients with stress urinary incontinence and pelvicorgan prolapse. Int J Gynaecol Obstet 80:35–39

13. Bai SW, Chung DJ, Yoon MJ et al (2005) Roles of estrogenreceptor, progesterone receptor, p53 and p21 in pathogenesis ofpelvic organ prolapse. Int Urogynecol J Pelvic Floor Dysfunct16:492–496

14. Reinhart BJ, Slack FJ, Basson M et al (2000) The 21-nucleotidelet-7 RNA regulates developmental timing in Caenorhabditiselegans. Nature 403:901–906

15. Zhao JJ, Lin JH, Yang H et al (2008) MicroRNA-221/222 negativelyregulates estrogen receptor alpha and is associated with tamoxifenresistance in breast cancer. J Biol Chem 283:31079–31086

16. Zhao C, Dong J, Jiang T et al (2011) Early second-trimester serummiRNA profiling predicts gestational diabetes mellitus. PLoS One6:e23925

17. Wahab M, Thompson J, Hamid B et al (1999) Endometrial histo-morphometry of trimegestone-based sequential hormone replace-ment therapy: a weighted comparison with the endometrium of thenatural cycle. Hum Reprod 14:2609–2618

18. Chen HY, Chung YW, Lin WY et al (2008) Estrogen receptoralpha polymorphism is associated with pelvic organ prolapse risk.Int Urogynecol J Pelvic Floor Dysfunct 19:1159–1163

19. Di Leva G, Gasparini P, Piovan C et al (2010) MicroRNA cluster221-222 and estrogen receptor alpha interactions in breast cancer. JNatl Cancer Inst 102:706–721

20. Cochrane DR, Cittelly DM, Howe EN, et al (2010) MicroRNAslink estrogen receptor alpha status and Dicer levels in breastcancer. Horm Cancer 1:306–319

21. Zhang N, Moran MS, Huo Q et al (2011) The hormonal receptorstatus in breast cancer can be altered by neoadjuvant chemotherapy: ameta-analysis. Cancer Invest 29:594–598

22. Ismail SI, Bain C, Hagen S (2010) Oestrogens for treatment orprevention of pelvic organ prolapse in postmenopausal women.Cochrane Database Syst Rev 9:CD007063

23. Rao X, Di Leva G, Li M et al (2011) MicroRNA-221/222 confersbreast cancer fulvestrant resistance by regulating multiple signalingpathways. Oncogene 30:1082–1097

24. Terasawa K, Ichimura A, Sato F et al (2009) Sustained activationof ERK1/2 by NGF induces microRNA-221 and 222 in PC12cells. FEBS J 276:3269–3276

934 Int Urogynecol J (2012) 23:929–934