Mechanical homeostasis is altered in uterine leiomyoma

MRFR

OtERwp

Slf

Rm

C4

LeahcmEafrct

FDLNUCAB

PMG

R

RM

TNH

0

AGOS Papers www.AJOG.org

4

echanical homeostasis is altered in uterine leiomyomaebecca Rogers, BS; John Norian, MD; Minnie Malik, PhD; Gregory Christman, MD; Mones Abu-Asab, PhD;aye Chen, PhD; Casey Korecki, MSME; James Iatridis, PhD; William H. Catherino, MD, PhD;ocky S. Tuan, PhD; Namisha Dhillon; Phyllis Leppert, MD, PhD; James H. Segars, MD

BJECTIVE: Uterine leiomyoma produce an extracellular matrix (ECM)hat is abnormal in its volume, content, and structure. Alterations inCM can modify mechanical stress on cells and lead to activation ofho-dependent signaling and cell growth. Here we sought to determinehether the altered ECM that is produced by leiomyoma was accom-anied by an altered state of mechanical homeostasis.

TUDY DESIGN: We measured the mechanical response of pairedeiomyoma and myometrial samples and performed immunogold, con-ocal microscopy, and immunohistochemical analyses.

ESULTS: Leiomyoma were significantly stiffer than matched myo-

002-9378/$34.00 • © 2008 Mosby, Inc. All rights reserved. • doi: 10.1016

74.e1 American Journal of Obstetrics & Gynecology APRIL 2008

he ECM, cell shape, and cytoskeleton in leiomyoma, comparedith myometrial samples from the same uterus. Levels of AKAP13,protein that is known to activate Rho, were increased in leio-yoma compared to myometrium. AKAP13 was associated with

ytoskeletal filaments of immortalized leiomyoma cells.

ONCLUSION: Leiomyoma cells are exposed to increased mechanicaloading and show structural and biochemical features that are consis-ent with the activation of solid-state signaling.

ey words: AKAP13, Brx, fibroid, mechanotransduction, RhoA,
etrium. The increased stiffness was accompanied by alteration of solid-state signaling
ite this article as: Rogers R, Norian J, Malik M, et al. Mechanical homeostasis is altered in uterine leiomyoma. Am J Obstet Gynecol 2008;198:474.e1-74.e11.

eiomyoma are stiff tumors that arecharacterized by the presence of

xcessive extracellular matrix (ECM)nd fibrosis.1 Our previous studiesave shown not only are genes that en-ode ECM proteins expressed abnor-ally in leiomyoma2,3 but also that the

CM of leiomyoma is altered.4 Alter-tions in the structure of the matrix af-ect its load-bearing properties and di-ectly influence the mechanical load onells that are growing within the ma-rix.5 Studies in other tissue types re-

veal that changes in mechanicalstresses that are applied to cells pro-foundly affect cell growth and prolifer-ation.6 Although it is generally ac-cepted that leiomyoma cells exhibitaltered growth and apoptosis,1,7,8 pre-vious studies have concentrated on theinvolvement of growth factors inleiomyoma development. Despite thestriking overproduction of ECM few, ifany, studies have characterized thecontribution of mechanical signalingto leiomyoma growth.

Mechanical stress is a well-docu-mented stimulus of muscle hypertrophy,but the few studies of mechanical signal-ing in uterine smooth muscle cells havefocused on pregnancy.9-11 The cellularprocess by which mechanical stress isconverted to cell signaling is known asmechanotransduction or solid-state sig-naling.12,13 Mechanical stresses that areimposed on the ECM are transmittedthrough integrins,14 mechanotransduc-ers, and caveoli15 to biochemical signalsin the cell. Downstream cellular signal-ing pathways of mechanotransductioninclude mitogen-activated protein ki-nases, phosphatidylinositol-3 kinase/Akt, Janus kinase/signal transducer,16

and nitric oxide.17 Rho family smallGTPases play a critical role in the pro-cess. Rho-GTPases function as molecu-lar switches that are on in the GTP-bound state and off in the GDP-boundstate.18 Rho family members are acti-vated by Rho-guanine nucleotide ex-change factors (GEFs) to influence cy-toskeletal rearrangement, contractility,and cell proliferation.19 The specific pro-teins that activate Rho in uterine musclecells have not been characterized fully.

In addition to physiologic activa-tion, pathologically altered states of

rom the Reproductive Biology and Medicine Branch, NICHD, NIH (Ms Rogers, Mshillon, Drs Norian, Christman, Catherino, Leppert, and Segars); the Pathologyaboratory, NCI, NIH (Dr Abu-Asab); Cartilage Biology and Orthopaedics Branch,IAMS, NIH (Drs Chen and Tuan); the Department of Obstetrics and Gynecology,niformed Services University of the Health Sciences, Bethesda, MD (Drs Malik andatherino); the Department of Obstetrics and Gynecology, University of Michigan, Annrbor, MI (Dr Christman); and the School of Engineering, University of Vermont,urlington, VT (Ms Korecki and Dr Iatridis).

resented at the 61st Annual Meeting of the American Society for Reproductive Medicine,ontreal, QC, Canada, Oct. 16-19, 2005, and the 26th Annual Meeting of the Americanynecological and Obstetrical Society, Chicago, IL, Sept. 26-29, 2007.

eceived May 5, 2007; accepted Nov. 27, 2007.

eprints: James H. Segars, Building 10, CRC, 1E-3140, 10 Center Dr, NICHD, NIH, Bethesda,D 20892. [email protected].

his research was supported in part by the intramural research programs of NCI, NIAMS, andICHD, by the NIH intramural program of the Reproductive Biology and Medicine Branch,D-Z01-008737-URE, and by the Howard Hughes Foundation (N.D.).

twamc

Clt

K

m/j.ajog.2007.11.057

echanical stress can cause fibrosis by

mailto:[email protected]

scaismmtuodEbtpfmiccati

Ctiu

mgmvwcpwRmdmRtraa

bmwawpmccmtfisw

MTTula

yom

www.AJOG.org AGOS Papers

timulating fibroblasts to deposit ex-essive ECM.6 Very little is knownbout the pathologic response of uter-ne smooth muscle cells to mechanicaltress, in particular whether altered

echanical stress might lead to leio-yocyte transformation and deposi-

ion of excessive ECM, for example as aterine fibroid tumor. One key featuref uterine fibroid tumors is the pro-uction of excessive amounts ofCM.1,3,4 In fact, growth of large fi-roid tumors (�5 cm) was largely dueo deposition of ECM,20 and the com-osition of this excessive ECM differed

rom smaller fibroid tumors and myo-etrium.21,22 Previous studies of ECM

n fibroid tumors have reported in-reased levels of total glycosaminogly-ans and a reduction in hyaluroniccid.21-24 Several groups have reportedhat the ECM in leiomyoma is altered

TABLE 1Patient characteristicsPatient # Age Ra

1 35 Ca...................................................................................................................

2 39 Afr...................................................................................................................

3 37 Ca...................................................................................................................

4 37 Afr...................................................................................................................

5 46 Afr...................................................................................................................

12 39 Afr...................................................................................................................

14 37 Ca...................................................................................................................

16 37 Afr...................................................................................................................

18 38 Afr...................................................................................................................

19 47 Ca...................................................................................................................

20 47 Afr...................................................................................................................

21 44 Afr...................................................................................................................

23 40 Afr...................................................................................................................

24 44 Ca...................................................................................................................

40 36 Ca...................................................................................................................

41 49 Asi...................................................................................................................

42 46 Ca...................................................................................................................

46 46 Afr...................................................................................................................

50 48 Afr...................................................................................................................

51 44 Afr...................................................................................................................

57 45 AfrRogers. Mechanical homeostasis is altered in uterine leiom

n both content3,21,24 and structure.4,24 b

ollectively, these reports suggest thathe mechanical properties of the ECMn leiomyoma may differ from normalterine tissue.In addition to causing fibrosis, alteredechanical stress can influence cell

rowth and tumorogenesis.25 Many tu-ors have been shown to have an ele-

ated elastic modulus or stiffness,26

hich is an attribute that is used clini-ally to detect metastatic nodules oneritoneal surfaces at laparotomy andith imaging modalities.27 Specifically,ho-GTPases were elevated in stiff tu-ors28 and activation of the Rho-depen-

ent kinase has been shown to cause tu-or cell dissemination.29 Activation ofho-dependent cytoskeletal tension led

o increased levels of extracellular signal-egulated kinase (ERK) and cell prolifer-tion30 and was a requisite step in thecquisition of a malignant phenotype of

Fibroid Size

sian 2 x 1.5 x 1.7.........................................................................................................................

American 7 x 7.........................................................................................................................

sian 5 x 4.........................................................................................................................

American 1.5 x 1.5 x 1.7.........................................................................................................................

American 2 x 1.3 x 1.4.........................................................................................................................

American 1 x 1 x 1.........................................................................................................................

sian 5 x 4 x 4.........................................................................................................................

American 1.5 x 1.2.........................................................................................................................

American 1.2 x 1.1 x 1.3.........................................................................................................................

sian 3.5 x 3.5 x 3.........................................................................................................................

American 7 x 7 x 6.........................................................................................................................

American 4.2 x 4.5 x 5.5.........................................................................................................................

American 3 x 3 x 3.........................................................................................................................

sian 7 x 8.........................................................................................................................

sian 5.5 x 4x5.........................................................................................................................

7 x 5 x 8.........................................................................................................................

sian 5 x 5 x 5.........................................................................................................................

American 4 x 4 x 4.........................................................................................................................

American 5 x 5 x 5.........................................................................................................................

American 5 x 3 x 3.........................................................................................................................

American 4 x 4 x 4a. Am J Obstet Gynecol 2008.

reast cancers.25 M

APRIL 2008 America

The 2 cardinal features of uterine fi-roid tumors, excessive fibrosis and tu-or formation, suggest that fibroblastsithin a leiomyoma may be exposed to

ltered mechanical stress. To determinehether altered mechanical stress maylay a role in fibroid tumor develop-ent, we characterized the state of me-

hanical homeostasis in matched surgi-al specimens of leiomyoma and normalyometrium. The findings indicate that

he production of an abnormal ECM inbroid tumors was accompanied by atate of altered mechanical homeostasisithin leiomyoma.

ATERIALS AND METHODSissue collectionissues were collected from women whonderwent hysterectomy for symptomatic

eiomyoma in institutional review board–pproved studies at the National Naval

Location Time of cycle

Submucosal Follicular..................................................................................................................

Intramural Periovulatory..................................................................................................................

Submucosal Periovulatory..................................................................................................................

Intramural Luteal..................................................................................................................



Intramural Ovulatory..................................................................................................................

Intramural Anovulatory..................................................................................................................


Subserosal NA..................................................................................................................

Subserosal NA..................................................................................................................

Intramural NA..................................................................................................................

Intramural Ovulatory..................................................................................................................

Intramural NA..................................................................................................................

Submucosal Ovulatory..................................................................................................................

Subserosal NA..................................................................................................................


Subserosal Ovulatory..................................................................................................................

Intramural Anovulatory..................................................................................................................


Intramural Luteal

ce

uca......... .........

ican......... .........

uca......... .........

ican......... .........

ican......... .........

ican......... .........

uca......... .........

ican......... .........

ican......... .........

uca......... .........

ican......... .........

ican......... .........

ican......... .........

uca......... .........

uca......... .........

an......... .........

uca......... .........

ican......... .........

ican......... .........

ican......... .........

ican

edical Center and the National Institutes

n Journal of Obstetrics & Gynecology 474.e2

os�cp

Eesa

MFtwTbccdlmcdmtso

GSmd(HrbtwfctHpd

F(S5fidppP0sSmFL0mr

AmpmtlmbsslTssR


4

f Health. For Western blots, tissue wasnap-frozen on dry ice and stored at

70°C until processed. For immunohisto-hemistry, samples were frozen in cryo-

FIGURE 1Mechanical testing in matched suof uterine leiomyoma (L) and myo

, Measurement of compressive resistance to 0.1 syometrial samples. The Y axis represents force

ressive strain � 0.1. Leiomyoma had significayometrium (P � .003). Similar results were obs

hat were tested in duplicate. B, Measurement ofeiomyoma and myometrial 3 paired samples anal

ethod for wet (left panel) and dry weight (right pars denote SEM) micrograms of glycosaminoglycaample myometrium or leiomyoma. Averages of 2hown. Differences were not significant. C, Hydeiomyoma and myometrial samples that were testhe Y axes represent the mean � SEM (error baample weight. Averages of 2 tests performed onignificant.ogers. Mechanical homeostasis is altered in uterine leiomyom

rotectant (OCT 4583; Tissue-Tek, t

74.e3 American Journal of Obstetrics & Gynecol

lkhart, IN) or were formalin-fixed andmbedded in paraffin. For immunogoldtudies, tissue was fixed in formalin. Char-cteristics of patients and leiomyoma in

cal specimenstrium (M)

(Young’s modulus) of 5 matched leiomyoma andopascals [kPa]) at equilibrium required for com-

greater compressive resistance compared withd in 2 other specimens from patients 23 and 57

l sulfated glycosaminoglycan content in matchedby compression (A) with the use of the blyscan

el). The Y axes represent the mean � SEM (errorer microgram of sample. The X axes represent thatts that were performed on 3 paired samples areyproline measurement in the same 3 matchedn B, corrected for wet and dry weight, are shown.enote SEM) for hydroxyproline per microgram ofpaired samples are shown. Differences were not

m J Obstet Gynecol 2008.

he study are summarized in Table 1. a

ogy APRIL 2008

echanical testingrom fresh or snap-frozen leiomyomaissues, 4-5 mm cylindric pieces were cutith an 8-mm biopsy punch (Miltex,uttlingen, Germany) and a microtomelade. Height and diameter were re-orded, and tissue was subjected to a me-hanical testing protocol with the use ofisplacement control.31 Young’s modu-

us is defined as the longitudinal defor-ation in terms of strain (fractional

hange in length) in response to longitu-inal stress (force per unit area). Becauseost tissues exhibit nonlinear viscoelas-

ic changes in physiologic ranges oftress, we used equilibrium assessmentf Young’s moduli.

lycosaminoglycan assaymall pieces of frozen fibroid and myo-etrium specimens were weighed and

igested overnight in digestion buffer0.1 mol/L NaAc, 10 mmol/L cysteineCl, 50 mmol/L ethylenediaminetet-

aacetic acid, plus 20 �L papain/mLuffer). Glycosaminoglycan concen-ration of completely digested samplesas determined with the Blyscan Sul-

ated Glycosaminoglycan Assay (Bio-olor, Newtownabbey, UK), accordingo the manufacturer’s instructions.ydroxyproline content of each sam-

le was determined after proteinase-Kigestion, as described elsewhere.32

luorescein isothiocyanateFITC) phalloidin stainnap-frozen tissue blocks were cut into-�m sections and placed on glass slides,xed for 15 minutes in 3.7% paraformal-ehyde, washed 3 � 5 minutes in phos-hate-buffered saline solution (PBS),ermeabilized in 0.2% Triton X-100 inBS for 1 hour, washed in PBS with.05% Triton, and blocked in 1% bovineerum albumin (BSA) in PBS for 1 hour.lides were then placed in a dark hu-idity chamber for 1 hour in 5 �g/mL

ITC Phalloidin (Sigma-Aldrich, St.ouis, MO). After being rinsed in PBS/.05% Triton, slides were mounted inedia with DAPI (Vector Laborato-

ies, Burlingame, CA), cover-slipped,

rgime

train(Kilntlyervetotayzedann ptes

roxed irs d

3

a. A

nd visualized.

WPv(g

tmNh(

5oawuaBpatbaoBuC1TcctaCws4

MSs

FcmAaAttustaoR(ptENim

Rl


estern blotsrotein was isolated according to a pre-iously described protocol.33 Samples50 �g) were resolved in Tris-Glycineels (Invitrogen, Carlsbad, CA) and

FIGURE 2Structural changes that were asso

ransferred to Hybond nitrocelluloseembranes (Amersham, Piscataway,J). Membranes were blocked for 2-3ours at 4°C in Tris-buffered salineTBS) with 0.05% Tween-20 (TBST) and

ted with uterine leiomyoma

APRIL 2008 America

.0% evaporated milk and incubatedvernight at 4°C in TBST with 0.1% BSAnd primary antibodies. Membranesere rinsed 3 times in TBST for 5 min-tes between all incubations. Secondaryntibody was added in TBST with 0.1%SA at room temperature for 1 hour. Su-ersignal developing reagents were usedccording to the manufacturer’s direc-ions (Pierce, Rockford, IL), and mem-ranes were exposed to film. For eachntibody that was used, horseradish per-xidase– conjugated �-actin (Santa Cruziotechnologies, Santa Cruz, CA) wassed to confirm equal protein loading.os-7 lysates that were transfected with a70-kd form of AKAP13 (Brx; Upstate,emecula, CA) was used as a positiveontrol. Recombinant RhoA (Calbio-hem, San Diego, CA) was used as a con-rol. Antiphosphorylated p38 mitogen–ctivated protein kinase (SC 7973; Santaruz Biotechnology, Santa Cruz, CA)as used. Extracts were analyzed from

pecimens of patients 12, 14, 16, 20, 21,0-42, 46, 50, and 51.

icroarrayample collection and microarray analy-is were previously described.3,34 The ex-

ITC-phalloidin staining of actin (green) andell nuclei (blue) with DAPI of sections fromatched leiomyoma and myometrial samples.ctin structure was disordered in leiomyomand differed markedly from normal myometrium., Leiomyoma and B, myometrium were from

he same uterus; C, leiomyoma and D, myome-rium were matched specimens from a separateterus. Note the deformed nuclei (blue) and celltructure in A and C. Representative sectionsaken from samples obtained from patients 40nd 42. (Original magnification, �40.) Stainingf paired E, leiomyoma and F, myometrium forhoA (1:1000) suggested increased staining.

Original magnification, �40.) Staining ofaired G, leiomyoma and myometrial H, sec-ions for alpha-smooth muscle actin (1:1500).xpression appeared greater in leiomyoma cells.ote the increase in ECM (clear area) surround-

ng the cells in (G) compared with (H). (Originalagnification, �40.)

ogers. Mechanical homeostasis is altered in uterineeiomyoma. Am J Obstet Gynecol 2008.

cia


iatmw

IPcsAalwucapqtLowt4

ITaopdPot(fiCbatbpsbmjPwdsua2

CC(wwSbgtcB1ob1(iwcZwvS

SGac2aTwC

RTsvmttbimpcsfdcpogo

abst

ccttagmrlwolop


4

sting database was queried for genes thatre known to be involved in mechano-ransduction. Results were from 4 paired

yometrial and fibroid samples thatere collected from patients 2-4.

mmunohistochemistryaraffin-embedded tissue samples wereut into 5-�m sections. Sections weretained according to the Vectastain EliteBC kit protocol (Vector Laboratories),s previously described,34 with the fol-owing modifications: Antigen retrievalas performed with boiling water andnmasking solution in a pressureooker. Slides were incubated in primaryntibody either for 3 hours at room tem-erature or 4°C overnight. All subse-uent steps were carried out according tohe manufacturer’s instructions (Vectoraboratories). Sections with no primaryr no secondary antibodies or incubatedith preimmune antisera served as con-

rols. Studies included patients 1-5, 24,0-42, 46, and 57.

mmunogoldissue pieces were removed from a par-ffin block and deparaffinized in xylener taken directly from formalin andlaced in absolute ethanol and embed-ed in LR White (SPI, West Chester,A). Ultrathin sections were mountedn 150-mesh uncoated nickel grids. Sec-ions were hydrated in blocking bufferPBS, 0.1% Tween-20, 0.5% cold-watersh gelatin [Ted Pella, Inc., Redding,A]) for 20 minutes, then incubated inuffer that contained AKAP13 primaryntibody (1:125) for an hour. Sampleshat were not incubated in primary anti-ody served as a control. Samples werelaced in blocking buffer with 2% goaterum for 5 minutes, washed twice inlocking buffer without serum for 2inutes, incubated in 10-nm gold-con-

ugated secondary antibody (1:1000; Tedella Inc, Redding, CA) for 1 hour, andashed 2 times in PBS and twice ineionized water. Dried grids weretained with uranyl acetate for 3 min-tes, rinsed 3 times, and allowed to drygain. Paired samples from patients 18-
0, 21, 24, and 40-42 were studied. t

onfocal microscopyells were grown on glass chamber slides

Nalge Nunc Int, Rochester, NY), fixedith cold methanol, and permeabilizedith 0.2% Triton-X 100 (Sigma-Aldrich,t. Louis, MO). Non-specific sites werelocked using 1% BSA and 10% normaloat serum and slides were incubated ei-her with antibody to alpha smooth mus-le actin and 2665 affinity-purified anti-rx (AKAP13) polyclonal antibody35 at:50 dilution overnight at 4°C. For sec-ndary antibody either FITC (anti-rab-it, 1:100) or Alexa-594 (anti-mouse,�g/ml; Invitrogen) was used. DAPISigma-Aldrich) was added at 0.1�g/mLn PBS before the cells were examinedith an Axiovert 405 mol/L epifluores-

ence inverted light microscope (Carleiss, Oberkochen, Germany). Imagesere acquired with a charge couple de-ice (CCD) camera (Hamamatsu Orca,hizuoka, Japan).

tatisticsroups and sample values are reported

s mean � SEM. Groups and values wereompared with the use of an unpaired,-tailed Student t test. Significance wasssigned at a probability value of �.05.ests were performed with Stata soft-are (version 9.2; Stata Corporation,ollege Station, TX).

ESULTShe Young’s modulus, a measure of tis-

ue stiffness, was quantified with averagealues of 4.9 kPa and 18.6 kPa for myo-etrial and leiomyoma samples, respec-

ively (Figure 1). Variation was noted be-ween specimens and within specimens,ut uterine leiomyoma tissues were typ-

cally 2- to 3-fold stiffer than matchedyometrium (P � .05). Because the

resence of an excessive mechanical loadan influence ECM formation, we mea-ured 2 components of the ECM, sul-ated glycosaminoglycans and hy-roxyproline content (reflective ofollagen content), in the subset of sam-les that were exposed to testing. Previ-us investigators had reported levels oflycosaminoglycan content in leiomy-ma,21-23 and our results also suggested
hat an increase of glycosaminoglycans
ogy APRIL 2008

nd collagen was present in leiomyoma,ut the differences were not statisticallyignificant for the number of samplesested.

Cells that are exposed to increased me-hanical loading often exhibit structuralhanges and altered expression of struc-ural support proteins, such as actin;herefore, we examined smooth musclelpha actin and F-actin in matched sur-ical specimens of leiomyoma and myo-etrium. Phalloidin staining (for actin)

evealed a striking difference betweeneiomyoma and myometrial tissues,hich was consistent with our previousbservation of disordered ECM in

eiomyoma.4 Cell shape within leiomy-ma was angular with a compressed ap-earance in comparison to myometrium

TABLE 2Microarray results for stress-related genes in leiomyomaand myometriumGene L/M ratio

Alpha-cardiac actin 5.2...........................................................................................................

Alpha–smooth muscle actin 1.2...........................................................................................................

Alpha 1 actinin -2.02...........................................................................................................

Alpha 1 actinin, skeletalmuscle

-1.79

...........................................................................................................

myosin 8.3...........................................................................................................

Tropomyosin 1 & 4 -3.6...........................................................................................................

Smooth muscle MHC -3.3...........................................................................................................

Alpha-MHC consensus 2.37...........................................................................................................

Fibrillin 2 3...........................................................................................................

Vinculin -1.15...........................................................................................................

Talin 1.39...........................................................................................................

ROCK (Rho-kinase) -1...........................................................................................................

Rho-GEF FERM -1.34...........................................................................................................

Rho-GEF 12 -1.06...........................................................................................................

Rho-GEF p115 1...........................................................................................................

Rho-GEF 4 1.2...........................................................................................................

Rho-GEF AKAP13 4...........................................................................................................

Endothelin receptor type B -2.13...........................................................................................................

Endothelin receptor -1.57...........................................................................................................

Samples from patients 2-5 were used for microarray.L/M, ratio of leiomyoma to myometrial signal; MHC,myosin heavy chain; GEF, guanine nucleotide ex-change factor.

Rogers. Mechanical homeostasis is altered in uterine
leiomyoma. Am J Obstet Gynecol 2008.

(wim

tcsoiodivpTRbf

AtAaitWmBtCsTtcammwmmliFQaqaS(Ls

Rl


APRIL 2008 America

Figure 2). These findings are consistentith the finding that tissue stiffness was

ncreased in leiomyoma, compared withyometrium.The finding of increased stiffness led to

he question: did leiomyoma exhibit bio-hemical evidence of activation of solid-tate signaling? To characterize the levelsf factors that are known to be involved

n mechanotransduction, we queriedur complementary DNA microarrayata of 33,000 transcripts.3,34 As shown

n Table 2, several factors that were in-olved in solid-state signaling were ex-ressed differentially in leiomyoma.ranscripts of 1 protein, AKAP13 (aho-GEF), were over expressed 4-foldased on microarray results. We chose toocus on this factor because (1) RhoA is a

, Western analysis of leiomyoma and myome-rial lysates for expression of the Rho-GEFKAP13 with the use of 2665 affinity-purifiedntisera.36 A 220-kd band was present that wasncreased in leiomyoma compared with myome-rium. B, Beta actin control for lysates in A. C,

estern analysis for AKAP13 with the use ofonoclonal antibody that was directed againstrx. A 220-kd band is identified. Positive con-

rol (�) consisted of lysates prepared fromos-7 cells that were transfected with a con-truct that expressed a 170-kd form of AKAP13.he negative control (-), lane 12, were lysateshat were prepared from untransfected Cos-7ells that did not express AKAP13. D, Westernnalysis for AKAP13 in matched leiomyoma andyometrial lysates from 8 patients are shown. Inost but not all pairs, expression of AKAP13as greater in leiomyoma compared with myo-etrium. Results were repeated in 3 experi-ents. E, Western analysis for RhoA in the same

ysates. Readily extractable levels of RhoA werencreased in some, but not all, fibroid samples., Beta actin control for protein loading. G,uantification of Western analysis for AKAP13nd Rho. Signal intensity of Western blots wasuantified, corrected for beta-actin, and aver-ged for 8 paired samples (error bars denoteEM). AKAP13 staining was significantly greater

P � .05) in leiomyoma than in myometrium.evels of total Rho were similar (NS � notignificant).

ogers. Mechanical homeostasis is altered in uterineeiomyoma. Am J Obstet Gynecol 2008.

FIGURE 3Altered expression of factors that were involvedin mechanical transduction in matched surgical specimensof uterine leiomyoma (L) and myometrium (M)


kicota

bab

As

sbabotApprawfftccsm

p

I(tiempiEftaapdtpidw�mtwIpcR4o

Ro


4

nown target of AKAP13,36 and RhoA isnvolved in the cellular response to me-hanical stress19,25,28-30; (2) it was thenly Rho-GEF that is expressed differen-ially, based on the microarray results;nd (3) we had cloned this factor on the

FIGURE 4Subcellular distribution of AKAP13leiomyoma and myometrial specim


asis of its ability to augment estrogenction,35,37 and leiomyoma are known toe estrogen-responsive.To confirm the microarray data for

KAP13, we performed Western analy-es on paired fibroid and myometrial

matcheds

ogy APRIL 2008

amples. Western analysis with the use ofoth a rabbit polyclonal antisera directedgainst AKAP13 and a monoclonal anti-ody to AKAP13 revealed the presencef a 220-kd band (Figure 3). Quantifica-ion of Western results indicated thatKAP13 was over significantly over-ex-ressed in leiomyoma (P � .05), in sup-ort of the microarray (messenger RNA)esults. Levels of nonmembrane-associ-ted RhoA have been noted to correlateith activation of RhoA.15 We per-

ormed immunohistochemical stainingor Rho of matched fibroid and myome-rial lysates to determine whether in-reases in AKAP13 were coupled with in-reased levels of RhoA. There were noignificant differences in total cytoplas-

ic RhoA (Figure 3, G).Because mechanical homeostasis ap-

eared to be altered on the basis of direct

mmunohistochemical localization of AKAP131:500) in A, leiomyoma and B, myometriumissues. (Original magnification, �63.) C, Pos-tive control, breast tissue; D, negative controlquals fibroid tissue that is stained with preim-une antisera. Staining for AKAP13 often ap-

eared increased with a perinuclear appearancen fibroid tumors. Immunogold study of matched, leiomyoma and F, myometrium specimens

rom patient 20 with the use of 2665 antiserahat was directed against AKAP13. Note thengular cell shape, reduced cytoplasm (Cyt),nd notched nucleus in the leiomyoma (E) com-ared with myometrium (F). The round blackots indicate localization of protein; the black

riangle points to nucleus. The white arrow (E)oints to the irregular bundles of collagen fibersn leiomyoma, compared with the regular bun-les of collagen fibers seen en-face noted by thehite arrow in F. (Original magnification,21,000.) Immunogold staining from anotheratched pair of G, leiomyoma and H, myome-

rial tissues from patient 24 that were stainedith antisera that was directed against AKAP13.

n this view, the arrows point to AKAP13 ex-ression in the nucleus, nuclear envelope, andytoplasm. (Original magnification, �15,500.)esults were repeated in matched samples frompatients. Immunogold studies were conducted

n specimens from patients 18-21.

ogers. Mechanical homeostasis is altered in uterine leiomy-ma. Am J Obstet Gynecol 2008.

inen

mAqc

len

b4tiicsiiAptcrtm

ftchsmrlwmcfpaelwtwlaaswpsc

CTpitccc

AatCsacctsAfcR


easurement of tissue, and levels ofKAP13 were elevated in leiomyoma, weuestioned whether the subcellular lo-

FIGURE 5Altered focalization of AKAP13 toin immortalized cultured leiomyom

, Confocal laser microscopy of leiomyoma cegainst AKAP13 (green) and alpha smooth musco cytoskeletal filaments. DAPI-stained nucleusonfocal laser microscope image of myometrialmooth muscle actin (red). Staining for AKAP1ppears blue. Results were repeated in 3 indepenell simultaneously stained for AKAP13 (green)ytoskeletal filaments. D, Same cell, stained for ahat shows the nucleus (blue). F, Same cell withuggests close association of AKAP13 with actinKAP13 (green) and actin (red), but treated with

or AKAP13 localized to regions in the peripheryytoskeleton.ogers. Mechanical homeostasis is altered in uterine leiomyom

alization of AKAP13 might be altered in h

eiomyoma relative to myometrium. Toxamine the subcellular localization inative tissues, we performed immuno-

oskeletal structuresand myometrial cell lines

at was stained with antisera that were directedctin (red). AKAP13 protein appeared to localizeears blue (original magnification, �40, oil). B,that was stained for AKAP13 (green) and alpharotein was less robust. DAPI-stained nucleust experiments. C, Confocal image of leiomyomaF-actin (red). Note green staining localized to(red) filaments. E, Same cell, stained with DAPI

erlay of AKAP13 (green) and actin (red), whichucture. G, Overlay of leiomyoma cell stained forchalasin D to dissolve actin filaments. Staininghe cell (white arrow) in the absence of the actin

m J Obstet Gynecol 2008.

istochemical analysis of matched fi- s

APRIL 2008 America

roid and myometrial samples (Figure). These studies suggested that the pat-ern of staining for AKAP13 was alteredn fibroid tumors, with increased stain-ng often near the cell membrane or nu-lear envelope, but resolution was notufficient to define the subcellular local-zation accurately. To more preciselydentify the subcellular localization ofKAP13 in uterine fibroid tumors, weerformed immunogold studies. Elec-ron microscopy substantiated the in-reased localization of AKAP13 in pe-inuclear region and further suggestedhat the protein may associate with fila-

ents in the cytoplasm.To examine this possibility, we per-

ormed confocal microscopy on immor-alized uterine fibroid and myometrialells (Figure 5). These immortalized cellsave been characterized and closely re-embled the native leiomyoma and myo-

etrium from which they were de-ived.38 In cells that were derived fromeiomyoma, the staining for AKAP13as present robustly and localized pri-arily to cytofilaments (Figure 5, A) in

omparison with cells that were derivedrom myometrium (Figure 5, B). Theossible association with cytoskeletal fil-ments led us to perform dual-labelingxperiments on the immortalizedeiomyoma cells. Overlay experimentsith confocal microscopy suggested that

he pattern of AKAP13 staining and actinere strikingly similar and often over-

apped (Figure 5, F). Disruption of thectin cytoskeleton with cytochalasin Dlso disrupted and altered localization oftaining for AKAP13 (Figure 5, G),hich suggested that a significant pro-ortion of AKAP13 was associated withtructural actin filaments in leiomyomaells.

OMMENT

he findings reveal that cells that com-rise uterine leiomyoma are exposed to

ncreased mechanical loading. Consis-ent with the increased physical force,ells in leiomyomata exhibited structuralhanges that were associated with me-hanical stress, such as an angular cell

cyta

ll thle aappcell3 pdenandctinovstr

cytoof t

a. A

hape, altered actin organization, distor-


te

hsls0Ietcomtut(mpslasoelhnn

tsutrwoaompgTaatastimbmh

hPcrctuBmstbcsnrct

coo

gwrc(mrseAlacseplRta

ls

LoitcmR


4

ion of the nuclear envelope, and alteredxpression of the Rho-GEF, AKAP13.

The results imply that leiomyoma cellsave defective perception of mechanicaltress. It is reasonable to assume that theeiomyoma cells themselves have a ten-ile elastic modulus of approximately.52 kPa39 (the value for chondrocytes).n contrast, normal myometrium had anlastic modulus of 5 kPa, which indicateshat 90% of the stiffness of myometriumould be attributed to structural aspectsf the tissue. For leiomyoma, 97% of theeasured stiffness would be attributable

o structure of the ECM if the cell mod-lus was similar to chondrocytes. We in-

erpret the findings to suggest that either1) leiomyoma cells do not sense appliedechanical stress properly (have in ap-

ropriately low signaling) and, as a re-ult, produce excessive ECM; or (2)eiomyoma cells produce excessivemounts of load-bearing ECMs to stress-hield themselves, because the cells areverly sensitive to mechanical stress. Inither scenario, mechanical sensing byeiomyoma cells appears to be abnormal;owever, additional studies will beeeded to confirm and characterize theature of the abnormality.Our previous studies of uterine fibroid

umors with the use of microarrays haveerved to focus attention on the ECM ofterine fibroid tumors and its similarity

o tissues that undergo remodeling andepair.40 We noted that fibroid tumorsere composed of an ECM with alteredrganization of structural elements, suchs collagen fibers.4 In addition, we3 andthers21,24,41 have noted that fibroid tu-ors express altered levels of key com-

onents of the ECM, such as the proteo-lycans dermatopontin and versican.he current study builds on the role of anltered ECM in leiomyoma formationnd begins to address the following ques-ion: Why do leiomyoma cells producen excessive and altered ECM? The datauggest that mechanical signaling is al-ered in leiomyoma cells. It is worth not-ng that excessive mechanical stress of

uscle cells and abnormal repair haveeen shown to induce fibrosis in otheruscular tissues, for example in the

eart.42 c


Activated states of solid-state signalingave also been shown to affect cell growth.roliferation of breast cancer cells was in-reased significantly when the tensioneached 1.2 kPa. However, breast cells withhromosomal anomalies that are consis-ent with metastasis did not metastasizentil mechanical force was �5 kPa.25

reast tumors exhibited an average elasticodulus of 4.0 kPa, but can differ from

urrounding tissue by 90-fold.27 Myome-rium stiffness averaged 5 kPa, although fi-roid tumors from the same uterus typi-ally exhibited a 2- to 3-fold increase intiffness. The 2- to 3-fold increase in stiff-ess that was observed relative to sur-ounding muscle may be sufficient to in-rease proliferation or reduce apoptosis ofhe leiomyoma cells.

We were somewhat surprised that mi-roarray results revealed increased levelsf the Rho-GEF, AKAP13 (Brx) becauseur previous studies of AKAP13 had fo-

FIGURE 6Working model of mechanical stre

eiomyoma cells are under increased mechanicaf the ECM is altered in leiomyoma. The increa

n cell structure, nuclear shape, and actin organhe Rho-GEF AKAP13 appeared to be increasedytoskeleton. Future studies are needed to deteright alter the balance of growth/apoptosis thr

ogers. Mechanical homeostasis is altered in uterine leiomyom

used on its ability to augment estro- p

ogy APRIL 2008

en35,37 and glucocorticoid action.43 Weere intrigued by the fact that sex ste-

oids influence leiomyoma growth, butonversely excessive glucocorticoidssuch as Cushing’s syndrome) causeuscle atrophy. Because the microarray

esults focused attention on this tran-cript, we proceeded to characterize thexpression of AKAP13 in leiomyoma.ltered expression of AKAP13 in

eiomyoma was supported by Westernnalysis, immunohistochemical stains,onfocal microscopy, and immunogoldtudies. We postulate that the increasedxpression of AKAP13 may result, inart, from the altered state of mechanical

oading (Figure 6) and, through alteredho signaling, may contribute to the al-

ered state of growth and apoptosis char-cteristic of leiomyoma.7,8

Any physical measurement of tissues isimited by the fact that physiologic tis-ues are nonhomogeneous in viscoelastic

signaling in leiomyoma cells

ding compared with myometrial cells. Structuren mechanical load is associated with alterationon as shown in Figures 2 and 4. Expression ofleiomyoma cells and was associated with thee whether increased levels of Rho-GEF activityh stress-activated kinases.m J Obstet Gynecol 2008.

ss

l loase iizati

inminoug

a. A

roperties and anisotropic. Although we

acmadtsiwoflcettuns

tsistlevlcua

AWfPpDpf

R1e12all23saF4Ws

m25CuM6mm97Latn8JS9Mtfc21BmsP1mt1mC1ti1na1aaC1snd1esm1ct1kp2pLC22M

d12ntoo22MgG2sohB2st2Snt22ml52o12tcR3RaddN3tts13mo3Hort53SgF3at


re not aware of any previous mechani-al assessment in fibroid tissue, skeletaluscle was softer along fibers than

cross fibers.27 Nonetheless the 3-foldifference that was observed was greaterhan differences the were attributable topecimen orientation alone. The changesn glycosaminoglycans and collagen thate measured were consistent with previ-us reports21-23,44; however, because ourocus was on the biochemical intracellu-ar mechanisms that accompany the in-reased mechanical loading, we did notxpand the number of samples that wereested and were satisfied that the direc-ion of changes resembled published val-es, although our results for glycosami-oglycan and hydroxyproline were nottatistically significant.

In summary, these results are consis-ent with the activation of mechanicalignaling in leiomyoma. The results aren agreement with the general under-tanding that transforming growth fac-or beta signaling is augmented ineiomyoma that leads to the secretion ofxcessive amounts of ECM. These obser-ations indicate that cells within aeiomyoma exist in a state of altered me-hanical homeostasis that may contrib-te to the production of excessivemounts of ECM. f

CKNOWLEDGMENTSe thank Drs Paul Driggers and Lars Rackwitz

or critical assistance and instruction; Dr Markayson for his work on the tissue procurementrotocol; Drs George Chrousos and AlaneCherney for their critical support for the com-letion of the project; and Ms Joy Britten-Webb

or technical assistance.

EFERENCES. Walker CL, Stewart EA. Uterine fibroids: thelephant in the room. Science 2005;308:589-92.. Catherino WH, Leppert PC, Stenmark MH, etl. Reduced dermatopontin expression is a mo-

ecular link between uterine leiomyomas and ke-oids. Genes Chromosomes Cancer 2004;40:04-17.. Tsibris JC, Segars JH, Coppola D, et al. In-ights from gene arrays on the developmentnd growth regulation of uterine leiomyomata.ertil Steril 2002;78:114-21.. Leppert PC, Baginski T, Prupas C, Catherino, Pletcher S, Segars JH. Comparative ultra-

tructure of collagen fibrils in uterine leiomyo- n

as and normal myometrium. Fertil Steril004;82:1182-7.. Tomasek JJ, Gabbiani G, Hinz B, Chaponnier, Brown RA. Myofibroblasts and mechano-reg-lation of connective tissue remodeling. Nat Revol Cell Biol 2002;3:349-63.. Paszek MJ, Weaver VM. The tensionounts: mechanics meets morphogenesis andalignancy. J Mam Gland Biol Neopl 2004;:325-42.. Bourlev V, Pavlovitch S, Stygar D, Volkov N,indblom B, Olovsson M. Different proliferativend apoptotic activity in peripheral versus cen-ral parts of human uterine leiomyomas. Gy-ecol Obstet Invest 2003;55:199-204.. Martel KM, Ko AC, Christman GM, StribleyM. Apoptosis in human uterine leiomyomas.emin Reprod Med 2004;22:91-103.. Mitchell JA, Shynlova O, Langille L, Lye SJ.echanical stretch and progesterone differen-

ially regulate activator protein-1 transcriptionactors in primary rat myometrial smooth mus-le cells. Am J Physiol Endocrinol Metab004;287:E439-45.0. Oldenhof AD, Shynlova OP, Liu M, LangilleL, Lye SJ. Mitogen-activated protein kinasesediate stretch-induced cfos mRNA expres-

ion in myometrial smooth muscle cells. Am Jhysiol Cell Physiol 2002;283:C1530-9.1. Lye SJ, Mitchell J, Nashman N, et al. Role ofechanical signals in the onset of term and pre-

erm labor. Front Horm Res 2001; 27:165-78.2. Huang S, Ingber DE. Cell tension, matrixechanics, and cancer development. Cancerell 2005;8:175-6.3. Alenghat FJ, Ingber DE. Mechanotransduc-ion: all signals point to cytoskeleton, matrix andntegrins. Sci STKE 2002;119:PE6.4. Geiger B, Bershadsky A. Exploring theeighborhood: adhesion-coupled cell mech-nosensors. Cell 2002;110:139-142.5. Kawamura S, Miyamoto S, Brown JH. Initi-tion and transduction of stretch-induced RhoAnd Rac1 activation through caveolae. J Biolhem 2003;278:31111-7.6. Ruwhof C, van der Laarse A. Mechanicaltress-induced cardiac hypertrophy: mecha-isms and signal transduction pathways. Car-iovascular Res 2000;47:223-37.7. Petroff MGV, Kim SH, Pepe S, et al. Endog-nous nitric oxide mechanisms mediate thetretch dependence of Ca�2 release in cardio-yocytes. Nat Cell Biol 2001;3:867-73.8. Schmidt A, Hall A. Guanine nucleotide ex-hange factors for the Rho GTPases: turning onhe switch. Genes Dev 2002;16:1587-609.9. Wettschureck N, Offermanns S. Rho/Rho-inase mediated signaling in physiology andathophysiology. J Mol Med 2002;80:629-38.0. Davis BJ. The NIEHS Uterine Fibroid Study:reliminary results. Presented at: Advances ineiomyoma Research: 2nd NIH Internationalongress, Bethesda, MD, February 24-25,005.1. Wolanska M, Sobolewski K, Drozdzewicz, Bankowski E. Extracellular matrix compo-

ents in uterine leiomyoma and their alteration 1

APRIL 2008 American

uring the tumour growth. Mol Cell Biochem998;189:145-52.2. Mitropoulou TN, Theocharis AD, Stagian-is KD, Karamanos NK. Identification, quan-ification and fine structural characterizationf glycosaminoglycans from uterine leiomy-ma and normal myometrium. Biochimie001;83:529-36.3. Hjelm A, Ekman-Ordeberg G, Barchan K,almstrom A. Identification of the major proteo-lycans from human myometrium. Acta Obstetynecol Scand 2001;80:1084-90.4. Berto AGA, Sampaio LO, Franco CRC, Ce-ar RM, Michelacci YM. A comparative analysisf structure and spatial distribution of decorin inuman leiomyoma and normal myometrium.iochim Biophys Acta 2003;1619:98-112.5. Paszek MJ, Zahir N, Johnson KR, et al. Ten-ional homeostasis and the malignant pheno-ype. Cancer Cell 2005;8:241-54.6. Beil M, Micoulet A, von Wichert G, et al.phingosylphosphorylcholine regulates keratinetwork architecture and visco-elastic proper-ies of human cancer cells. Nat Cell Biol003;5:803-11.7. Greenleaf JF, Fatemi M, Insana M. Selectedethods for imaging elastic properties of bio-

ogical tissues. Annu Rev Biomed Eng 2003;5:7-78.8. Fritz G, Just I, Kaina B. Rho GTPases arever-expressed in human tumors. Int J Cancer999;81:682-7.9. Croft DR, Sahai E, Mavria G, et al. Condi-ional ROCK activation in vivo induces tumorell dissemination and angiogenesis. Canceres 2004;64:8994-9001.0. Wang F, Weaver VM, Petersen OL, et al.eciprocal interactions between �1-integrinnd epidermal growth factor receptor in three-imensional basement membrane cultures: aifferent perspective in epithelial biology. Procatl Acad Sci USA 1998;95:14821-6.1. Mauck RL, Soltz MA, Wang CC, et al. Func-ional tissue engineering of articular cartilagehrough dynamic loading of chondrocyte-eeded agarose gels. J Biomech Eng 2000;22:252-60.2. Reddy GK, Enwemeka CS. A simplifiedethod for the analysis of hydroxyproline in bi-logic tissues. Clin Biochem 1996;29:225-9.3. Tsibris JCM, Porter KB, Jazayeri A, et al.uman uterine leiomyoma express higher levelsf peroxisome proliferator–activated receptor �,etinoid receptor � and all-trans retinoic acidhan myometrium. Cancer Res 1999;59:737-44.4. Catherino WH, Prupas C, Tsibris JCM, et al.trategy for elucidating differentially expressedenes in leiomyomata identified by microarray.ertil Steril 2003;80:282-90.5. Rubino D, Driggers P, Arbit D, et al. Char-cterization of Brx, a novel DBL family memberhat modulates estrogen action. Oncogene
998;16:2513-26.
Journal of Obstetrics & Gynecology 474.e10

3owNv3ptp3cmo1

3SlC4nbm44Gmi

4if24atc24Rth


4

6. Park B, Nguyen N, Dutt P, et al. Associationf Lbc Rho guanine nucleotide exchange factorith �-catenin-related protein, �-catulin/CT-NAL1, supports serum response factor acti-ation. J Biol Chem 2002;277:45361-70.7. Driggers PH, Segars JH, Rubino DM. Theroto-oncoprotein Brx activates estrogen recep-or � by a p38 mitogen-activated protein kinaseathway. J Biol Chem 2001;276:46792-7.8. Malik M, Catherino WH. Novel method toharacterize primary cultures of leiomyoma andyometrium with the use of confirmatory bi-marker gene arrays. Fertil Steril 2007;87:
166-72. I
74.e11 American Journal of Obstetrics & Gyneco

9. Iatridis JC, Wu J, Yandow JA, Langevin HM.ubcutaneous tissue mechanical behavior is

inear and viscoelastic under uniaxial tension.onnect Tissue Res 2003;44:208-17.0. Leppert PC, Catherino WH, Segars JH. Aew hypothesis about the origin of uterine fi-roids based on gene expression profiling withicroarrays. Am J Obstet Gynecol 2006;195:15-20.1. Chegini N, Verala J, Luo X, Williams RS.ene expression profile of leiomyoma and myo-etrium and the effect of gonadotropin releas-

ng hormone analogue therapy. J Soc Gynecol
nvestig 2003;10:161-71. m
logy APRIL 2008

2. Sano M, Minamino T, Toko H, et al. p53-nduced inhibition of Hif-1 causes cardiac dys-unction during pressure overload. Nature007;446:444-8.3. Kino T, Souvatzoglou E, Charmandari E, etl. Rho family guanine nucleotide exchange fac-or Brx couples extracellular signals to the glu-ocorticoid signaling system. J Biol Chem006;281:9118-26.4. Rein MS, Barbieri RL, Welch W, GleasonE, Caulfield JP, Friedman AJ. The concentra-

ions of collagen-associated amino acids areigher in GnRH agonist-treated uterine myo-
as. Obstet Gynecol 1993;82:901-5.

Documents

Mechanical homeostasis is altered in uterine leiomyoma