Spatial differences in biologic activity of large uterineleiomyomataJian-Jun Wei, M.D., Xing-Min Zhang, M.D., Luis Chiriboga, Ph.D., Herman Yee, M.D., Ph.D,Mary A. Perle, Ph.D., and Khush Mittal, M.D.
Department of Pathology, Bellevue Hospital, New York, New York
terine leiomyomata (ULM) are the most common neo-lasms in women of the reproductive age, with a lifetimeccumulating risk of 70% in white and 80% in black women1). Although patients with large ULM account for a smallroportion of ULM in the general population, they oftenresent with unbearable symptoms and eventually seek sur-ical intervention. The annual hysterectomies in the Unitedtates due to large and symptomatic ULM number approx-
mately 200,000 (2). Medical costs and patients’ sufferingssociated with ULM are substantial (3).
ULM are benign smooth muscle tumors driven by sexteroid hormones. The size and tumor growth rate are asso-iated with the presence of the sex steroid hormones; theumor size, for instance, is reduced in postmenopause andith the use of GnRH agonist treatment. The large ULM,owever, do have several growth features that appear to bendependent of the sex steroid hormone status: 1) Most largeLM are found in women in the late, rather than early,
eproductive age (4); 2) large ULM exhibit a faster growth
eceived April 4, 2005; revised and accepted July 4, 2005.resented at the national meeting of United States and CanadianAcademy of Pathology, San Antonio, Texas, March 2005. This studywas also presented in part in Advances in Uterine LeiomyomaResearch, the 2nd NIH International Congress.
eprint requests: Jian-Jun Wei, M.D., Department of Pathology, NewYork University SOM, 462 First Avenue, New York, New York (FAX:
ate than small ULM (5); 3) large ULM present a lower levelf up-regulation of estrogen receptors (ER) and progesteroneeceptors (PR) than small ULM (6); and 4) patients withumerous ULM often have only one or a few large ULMour clinical observation).
It is poorly understood why some large ULM continue torow fast, despite their histologically presented disadvan-ages of high levels of hyaline degeneration, necrosis, andbrosis, and other ULM remain small. For those large ULM
reated by GnRH agonists, the hyaline degeneration andecroses are even broader than those of untreated ones (7).nce treatment is stopped, the tumors quickly resume theirriginal sizes (8). Apparently, the large ULM continue torow faster than small ULM (5), even in the presence of highegrees of these degenerative changes. We presume thathere are spatial differences of the tumor activity in the largeLM that may contribute to this growth behavior. Theighly active regions may be determined either by the localutritional status or by the genetic alterations.
In this study, we explored the tumor growth behaviors byxamining the spatial distributions of the ULM-associatedene expression, tumor proliferation rate, vessel density, andhe levels of hyalinized degenerations in large leiomyomata.
e found that the tumor region next to the peripheral zones the area with higher levels of the gene expression, higher
essel densities, higher proliferation rates, and lower hyaline
179Fertility and Sterility� Vol. 85, No. 1, January 2006iety for Reproductive Medicine, Published by Elsevier Inc.
di
MPSlamsehcw
TMWsswvfFsf
s
pa(Tc
UAtpesngeg[((Maf
ISut
egeneration. The spatial difference is one of the key factorsn determining growth behavior in large ULM.
ATERIALS AND METHODSatients and Specimenseven hysterectomies with large ULM (�10 cm) were col-
ected in fresh tissues. The patients’ ages, tumor sizes, cycles,nd ethnicities are summarized in Table 1. Tissues from normalatched myometrium, one large ULM (�10 cm), and two
mall ULM (�2 cm) were collected from each patient, and allxcept one large ULM received cytogenetic analysis. Threeysterectomies with large uterine leiomyosarcomas (LMS; �10m) were also collected as part of this study (Table 1). Thisork was approved by our IRB committee.
hree-dimensional Tissue Preparation and Tissueicroarray (TMA)ide strips of tissue from each large leiomyoma were dis-
ected in the three dimensional axes X, Y, and Z (because nopecific tumor orientation was defined, the X, Y, and Z axesere randomly chosen). Each axis was further orderly di-ided into 6 zones of approximately 1 cm in width (rangingrom zone 1 of the periphery to zone 6 of the center) (Fig. 1).rom each case, one section from each zone of each axis, 2ections from small leiomyomata (�2 cm), and 2 sectionsrom matched myometrium were prepared.
Hematoxylin and eosin (H&E)–stained slides from each
TABLE 1Summary of seven cases with large leiomyomata
Caseno.
Age(yrs) Racea
Size(cm)
Weight(g) Phaseb tu
ULM1 34 B 10 1550 SULM2 38 B 10 610 SULM3 47 B 12 1950 PULM4 45 W 11 1350 PULM5 52 W 13 1850 IULM6 44 A 13 1300 PULM7 46 W 10 1050 PLMS1 55 W 14 1350 ILMS2 53 O 12 1400 ILMS3 63 O 17 946 INote: ULM � uterine leiomyoma; LMS � uterine leiom
leiomyoma.a B � black; W � white; A � Asian American; O � other.b S � luteal phase; P � follicular phase; I � inactive endoc ND � not determined; NM � normal karyotype; Comple
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
ection (1.5 � 1.0 cm) were reviewed to: 1) calculate the t
180 Wei et al. Spatial differences in large leiomyomata
ercentage of hyaline degeneration; and 2) mark the cellularnd viable areas for tissue microarray selection. Tissue cores0.6 mm) from each tissue section were microarrayed into aMA recipient block. A total of 220 tissue cores wereollected for the study.
LM-Associated Gene Productsgroup of ULM-associated gene products were selected for
he study based on data from gene transcription profiles andrevious studies (6, 9). This group of genes are abnormallyxpressed in ULM and cover broad functions including: [1]ex steroid hormone receptors: ER� and PR-A; [2] otheruclear receptors: retinoid X receptor � (RXR�) (10, 11) andlucocorticoid receptor (GCR) (9, 12); [3] growth factors:pithelial growth factor receptor (EGFR) (9) and insulin-likerowth factor 2 (IGF2) (12–20); [4] apoptosis: Bcl-2 (21);5] angiogenesis: vascular endothelial growth factor (VEGF)18); [6] ULM-specific markers: human mobility group geneHMGA2) (15) and CD24 (9, 11–13, 22); and [7] HIF1,
IB1, and Factor VIII: used to evaluate the proliferativectivity, vessel density, and stress of hypoxia. The conditionsor immunostaining have been described previously (9).
mmunoscores and Statistical Analysistained TMA slides were graded jointly by two pathologistssing a visual semiquantitation method (optical density ofhe immunoreactivity). A one-score system for immunoin-
ensity (II) was used for the markers showing a diffuse
Vol. 85, No. 1, January 2006
iAups2sw3
nnttccfdus
RGTuisaob1fh(tml
SB
F
mmunoreactivity in tumors and the matched myometrium.two-score system for II and immunopositivity (IP) was
sed for those markers that were immunoreactive for onlyortions of the tested cells. The semiquantitation for inten-ity was scored on a scale of 0–4: 0 � negative; 1 � weak;
� moderate; 3 � strong; and 4 � very strong. Theemiquantitation for percentages of immunopositive cellsas scored on a scale of 1–4: 1 � 1%–10%; 2 � 1%1–50%;� 51%–80%; and 4 � �80%.
The immunoscores from each tumor tissue core wereormalized by subtracting the immunoscores from the inter-al myometrial controls (the normalized values representedhe net gain or loss of immunoreactivity in the tumor sec-ions). The numbers of vessels in each tissue core wereounted with the aid of factor VIII immunostain. The per-entage of hyalinization/necrosis/fibrosis was estimatedrom the original tissue sections. The mean values and stan-ard errors for each zone were calculated. Paired t test wassed and P value less than .05 was considered as statistical
FIGURE 1
Scheme diagram of the tissue microarray preparationuterine leiomyomata (ULM). (A) Tissue sections werecenter to periphery). (B) H&E-stained slides from eacand arrayed on a recipient block (C). (D) Some examin large ULM (zones 1 to 6), myometrium (MM), and s
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
ignificance. d
ertility and Sterility�
ESULTSeneral Informationhe cases for the study were randomly selected based onlypon tumor size. The cases with gross necrosis were notncluded in the study. All ULM were reviewed on H&E-tained slides and diagnosed as the usual type. The patients’ges ranged from 34 to 52 years (mean 43.7). The patientsriginated from different ethnic backgrounds, includinglack, Asian, white, and other. The tumor sizes ranged from0 to 13 cm (mean 11.3), and the number of tumors rangedrom 3 to 12 in each case. Two large ULM were found toave nonrandom chromosome anomalies (46,XX,del7)(q21–31) and 47,XX,�1, add(1)(p10), �1). Three pa-ients with large LMS had a mean age of 57 years old and aean tumor size of 14.3 cm. All tissue sections were col-
ected and prepared under the same conditions.
patial Difference of the Gene Products in Large ULMecause the original tissue sections dissected from three
m three-dimensional tumor sections in the largeected from 3 axes (X, Y, and Z) and 6 zones (fromsue section were reviewed and tissue cores takenof immunoreactivity for Factor VIII (F8) and VEGF
ll ULM (S-ULM).
frodissh tisplesma
imensions of axes X, Y, and Z were randomly defined, to
181
imfmbAaftusst
(mlcZ(
rdrt
s
Rol
sUacag(
ropigid
V
tcc
ncrease the power of the statistical analysis, we analyzed theean scores from each zone by pooling scores of each zone
rom axes X, Y, and Z together from all samples. Further-ore, we normalized the immunoscores of each tumor cores
y subtracting the scores of internal matched myometrium.s illustrated in Figure 2, there was minimal change of ER
nd PR in all zones (�0.22, �0.27) of the large ULM exceptor zone 2, where a significant up-regulation of both recep-ors (�1.27, �1.33 respectively) was noted. The level ofp-regulation of ER and PR in zone 2 in large ULM is quiteimilar to that in small ULM (�2 cm). The overall expres-ion level of ER and PR in large ULM is significantly lesshan that in matched small ULM (P�.05) (Fig. 2).
Bcl-2 was previously reported to be up regulated in ULM23). We found an up-regulation of Bcl-2 gene product inore than 50% of ULM in our series of 120 ULM (unpub-
ished data). However, Bcl-2 expression was slightly de-reased in all zones of the large ULM except for zone 2.one 2 is the only area that showed an up-regulation of Bcl-2
Fig. 3).
The immunoreactivity for MIB1 in large ULM was up-egulated overall. The level of up-regulation of MIB1 variedramatically between the tumors and zones. There is a higherate of MIB1 immunopositive cells in the peripheral zoneshan in the center zones (Fig. 3).
RXR was up-regulated in nearly 50% of ULM in our
FIGURE 2
Differential expression of ER and PR in large (L-ULMline) and PR (solid line) expression in L-ULM (from zovalues in myometrium (MM) is set as zero. (B) The mstandard errors are indicated as vertical T bars. *P�.
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
eries of 120 ULM (unpublished data). Up-regulation of T
182 Wei et al. Spatial differences in large leiomyomata
XR was also present in the large ULM, in which the levelf up-regulation of RXR was higher in peripheral zones andower in central zones (Fig. 3).
The transcription profiles in ULM (12) and our TMAtudy (9) have indicated that GCR is down-regulated inLM. The role of GCR in the myometrium appears to be
ssociated with antifibrosis (24). We found that GCR wasonsistently down-regulated in all large ULM in the studynd that the level of down-regulation of GCR was muchreater (�5.5) in zone 6 than in zone 2 (�4.7) in large ULMP�.05) (Fig. 3).
CD24, EGFR, HMGA2, IGF2, and IGF1R were up-egulated in ULM based upon the gene transcription data andur TMA data (9). The expression patterns of these generoducts in the large ULM in this study were similar to thosen our previous study. Although the expression levels of theene products tended to be slightly lower in the center regionn large ULM, there were no significant differences betweenifferent zones (Fig. 3).
Spatial Differences of the Vessel Density, HIF1 andEGF Expression, and Hyalinized Degeneration
Factor VIII was used to visualize the capillary vessels inissue cores. The net loss of the vessels in tumor cores wasounted by subtracting the numbers of vessels in each tumorore from the vessels in the matched myometrial cores.
small (S-ULM) ULM. (A) The levels of ER (dashedto zone 6) and in S-ULM. The baseline of ER/PR
scores of ER and PR in S-ULM and L-ULM. The
) andne 1ean05.
here were reductions of the tumor vessels in all spatial
Vol. 85, No. 1, January 2006
rzTp
tstvl(
eairicrpcau(piw
F
egions in large ULM. The level of the vascular reduction inone 2 was �14 in comparison to �25 in zone 6 (P�.05).here was also a steady decline in vessel density from theeripheral zones to the center areas (Fig. 4).
The percentage of the degenerative changes (hyaliniza-ion, necrosis, and fibrosis) was counted in all H&E-tained slides. The overall level of the hyaline degenera-ion accounted for approximately 20% of the total tumorolume. The percentage of hyaline degeneration wasowest (13%) in zone 2 and highest (21%) in zone 5Fig. 4).
To evaluate whether the spatial difference of the genexpression, vessel density, and hyaline degeneration weressociated with levels of hypoxia stress, we examined themmunoreactivity for HIF1 and VEGF. HIF1 was up-egulated in all 6 zones of the large fibroids, whereas HIF1n small ULM (Fig. 5) showed only minimal change whenompared to matched myometrium. The level of up-egulation in the large ULM was primarily lower in theeripheral zones (immunoscore �0.4 in zone 2) than in theenter zones (score �1.0 in zone 6) (Fig. 5). VEGF showed
slight down-regulation in zones 1 and 2, was steadilyp-regulated in zone 3, and reached its peak in zone 6central zone) (Fig. 5). The levels of HIF1 and VEGF ex-ression were positively associated with the level of hyalin-zation (r � 0.54 and r � 0.86) and negatively associated
FIGURE 3
Spatially differential expression of GCR and MIB1 (leand Bcl2 (right panel) in the large ULM (from zone 1baselines in myometrium (MM) for each marker are sgiven as vertical T bars. *P�.05.
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
ft panel) and CD24, EGFR, IGF1R, HMGA2, RXR, IGF2,to zone 6) in comparison to small ULM (S-ULM). The
et as zero. The standard errors for GCR and MIB1 are
ith the vessel density (r � �0.78).
ertility and Sterility�
FIGURE 4
Spatial distribution of the vessel density (A) andpercentage of hyaline degeneration (B) in largeULM (from zone 1 to zone 6) and in small ULM(S-ULM). The baselines of the vessel density andhyalinization in myometrium (MM) are set as zero.*P�.05.
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
183
SLTsfwzomflsltwgogV
DWtolh(is
pd
acUfcpli
loaactweldasIsgm
patial Distribution of the Gene Products in Uterineeiomyosarcomashe same group of gene products selected for the large ULMtudy was examined in the large LMS study. Tumor sectionsor the large LMS were collected and examined in the sameay as the large ULM, i.e., tissue cores from 3 axes and 6
ones were arrayed on the same recipient blocks. The levelf expression for each gene product was normalized by theatched myometrium. In general, there were minimal dif-
erences in the gene expression between the zones in thearge LMS (Fig. 6). Unlike ULM, however, LMS demon-trated large losses of ER and PR, presenting the greatestevel of down-regulation of ER and PR in zone 4. In contrast,he immunoreactivity for MIB1 was higher in zone 4. Whene reviewed the tumor necrosis spatially, LMS showedeographic necrosis and no spatial pattern. An examinationf the immunoreactivity for HIF1 and VEGF indicated alobal up-regulation of HIF1 and ascending up-regulation ofEGF from peripheral to center zones in LMS.
ISCUSSIONe found that the tumor cell activities are not evenly dis-
ributed spatially in the large ULM. This is illustrated by anrderly decrease in the levels of gene expression, cell pro-iferation activity, and vessel density and an increase inyaline degeneration from peripheral to center zones. Zone 2next to the periphery) in the large ULM is the most biolog-cally active region, a conclusion established by its con-
FIGURE 5
Spatially differential expression of VEGF (solid line) azone 6) and in small ULM (S-ULM). The baselines ofzero. The standard errors were given as vertical T ba
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
tantly high levels of the gene product, a high rate of cell a
184 Wei et al. Spatial differences in large leiomyomata
roliferation, a high vessel density, and a low level of theegeneration compared to the other tumor regions.
Therefore, the high potent areas driving the tumor growthre in the peripheral zones, particularly in zone 2. Whatauses the regional difference of cell activity in the largeLM? Studies have shown that there is spatial heterogeneity
or hypoxic stress in malignant solid tumors. The hypoxicondition induces a response of HIF1 and VEGF gene ex-ressions. The level of hypoxia is linearly correlated to theevel of HIF1 and VEGF expression, which can be evaluatedn a spatial manner (25, 26).
To test if the zonal difference of the molecular and patho-ogic findings in the large ULM are associated with ischemicr hypoxia change, we examined the spatial pattern of HIF1nd VEGF expression. We found a steady increase of HIF1nd VEGF gene products from the peripheral zones to theenter zones in the large ULM. The findings support theheory that there is a spatial difference in the hypoxic stress,hich further suggests that there are high levels of hypoxic
nvironment in the center areas of the large ULM. However,ower vessel density in the central zones of large ULMemonstrates that, unlike the situation in malignant tumors,n up-regulation of HIF1 and VEGF induced by hypoxictress fails to promote the vascular production in large ULM.t is apparent that lower vessel density and higher hypoxictress make central zones of the large ULM an unfavorablerowth environment. The tumor growth in the large ULMay simply rely upon the peripheral regions where there are
IF1 (dashed line) in the large ULM (from zone 1 toF and HIF1 values in myometrium (MM) are set as(P�.05); **(P�.01).
nd HVEGrs. *
dvantages of vessel distribution and nutrition supplement.
Vol. 85, No. 1, January 2006
podpacstUih2prcfdp
s(oi
sEtmfwiflcashUts
lbadwd
F
Another striking feature of the large ULM, when com-ared to the matched myometrium, is a universal reductionf vessel density. Reduction of the vessels in ULM may beue to a generally lower level of angiogenesis activity. Thishenomenon is partially supported by a down-regulation ofngiogenetic genes, including VEGF in ULM (18, 20). Be-ause the large ULM appear to grow much faster than themall ULM (5), the peripheral zones in the large ULM haveo possess a higher potent growth activity than the smallLM. We previously found a positive correlation of MIB1
mmunoreactivity with the tumor sizes (6). We also found aigher antiapoptotic activity (up-regulation of Bcl-2) in zone
in this study. It would be interesting to know if theeripheral zones accumulate more genetic alterations that areesponsible for the higher level of the gene expression andell proliferation activity. Additional study is needed tourther clarify whether the rapid tumor growth in zone 2 isependent upon local circulation or upon geneticrogression.
We also found substantial differences in expression of theelected gene products between large ULM and small ULMFig. 7). One of the important findings is that the mean levelf up-regulation of ER and PR in the large ULM is signif-
FIGURE 6
Spatially differential expression of the selected geneleiomyosarcomas (LMS) (from zone 1 to zone 6) andmarker value in myometrium (MM) are set as zero.
Wei. Spatial differences in large leiomyomata. Fertil Steril 2006.
cantly lower (P�.05) than in the small ULM. This is con- i
ertility and Sterility�
istent with our previous study, in which the expression ofR and PR was shown to be inversely associated with the
umor size (6). The findings indicate that the large ULM areuch less dependent on the high level of ER/PR. This is
urther supported by the studies from transcription profiles,here ER and PR are not picked up as the up regulated genes
n most studies. When we compared the other genes asiderom ER and PR, the large ULM showed substantially higherevels of gene expression than the small ones, higher rates ofell proliferation, higher degrees of degenerative changes,nd lower densities of the vessels. It is still unknown why themall ULM, despite their advantages of higher ER/PR andigher densities of vessels, grow more slowly than the largeLM (5). The tumor progression in large ULM caused by
he gaining of additional genetic alterations may be a plau-ible explanation of this growth behavior.
Unique spatial patterns between the large ULM and thearge LMS were also noticed. The spatial tumor activity andehavior was found to be different between the large ULMnd LMS. The cell activity in the large ULM appears to beependent upon the status of vascularity and hypoxia,hereas the cell activity in the large LMS does not show thisependence. Therefore, the spatial pattern of tumor activity
ucts (listed on the right) in large uterineched small ULM (S-ULM). The baselines for each
prodmat
n the large ULM will be a unique tumor growth model. The
185
mta
bopbamtss
R
1
1
1
1
1
1
1
odel will be valuable to us in both further characterizinghe growth behavior and evaluating the spatial sensitivitynd its effect on current noninvasive treatment modalities.
It has been noted that an extensive tumor necrosis inducedy treatment with GnRH agonists will not prevent regrowthf the same fibroids when the treatment stops (27). The samehenomenon is also observed in some large fibroids treatedy uterine artery embolization (28). Therefore, it is reason-ble to believe that failure of the current nonsurgical treat-ent modalities in some large fibroids may be in part due to
he spatial difference of the tumor activity. Our findings alsouggest that central areas of hyalinization or infarcts may notlow down the tumor growth if zone 2 is still viable.
EFERENCES1. Day Baird D, Dunson DB, Hill MC, Cousins D, Schectman JM. High
cumulative incidence of uterine leiomyoma in black and white women:ultrasound evidence. Am J Obstet Gynecol 2003;188:100–7.
2. Varol N, Healey M, Tang P, Sheehan P, Maher P, Hill D. Ten-yearreview of hysterectomy morbidity and mortality: can we change direc-tion? Aust N Z J Obstet Gynaecol 2001;41:295–302.
186 Wei et al. Spatial differences in large leiomyomata
7. Colgan TJ, Pendergast S, LeBlanc M. The histopathology of uterineleiomyomas following treatment with gonadotropin-releasing hormoneanalogues. Hum Pathol 1993;24:1073–7.
8. Friedman AJ, Daly M, Juneau-Norcross M, Gleason R, Rein MS,LeBoff M. Long-term medical therapy for leiomyomata uteri: a pro-spective, randomized study of leuprolide acetate depot plus eitheroestrogen-progestin or progestin “add-back” for 2 years. Hum Reprod1994;9:1618–25.
9. Wei J, Chiriboga L, Mizuguchi M, Yee H, Mittal K. Expression profileof tuberin and some potential tumorigenic factors in 60 patients withuterine leiomyomata. Mod Pathol 2005;18:179–88.
0. Gamage SD, Bischoff ED, Burroughs KD, Lamph WW, Gottardis MM,Walker CL, et al. Efficacy of LGD1069 (Targretin), a retinoid Xreceptor–selective ligand, for treatment of uterine leiomyoma. J Phar-macol Exp Ther 2000;295:677–81.
1. Arslan AA, Gold LI, Mittal K, Suen TC, Belitskaya-Levy I, Tang MS,et al. Gene expression studies provide clues to the pathogenesis ofuterine leiomyoma: new evidence and a systematic review. Hum Re-prod 2005;20(4):852–63.
2. Tsibris JC, Segars J, Coppola D, Mane S, Wilbanks GD, O’Brien WF,et al. Insights from gene arrays on the development and growth regu-lation of uterine leiomyomata. Fertil Steril 2002;78:114–21.
3. Skubitz KM, Skubitz AP. Differential gene expression in uterineleiomyoma. J Lab Clin Med 2003;141:297–308.
4. Ahn WS, Kim KW, Bae SM, Yoon JH, Lee JM, Namkoong SE, et al.Targeted cellular process profiling approach for uterine leiomyomausing cDNA microarray, proteomics and gene ontology analysis. Int JExp Pathol 2003;84:267–79.
5. Quade BJ, Wang TY, Sornberger K, Cin PD, Mutter GL, Morton CC.Molecular pathogenesis of uterine smooth muscle tumors from tran-scriptional profiling. Genes Chromosomes Cancer 2004;40:97–108.
6. Wang H, Mahadevappa M, Yamamoto K, Wen Y, Chen B, WarringtonJA, et al. Distinctive proliferative phase differences in gene expressionin human myometrium and leiomyomata. Fertil Steril 2003;80:266–76.
low the bars), vessel density, and hyaline(S-ULM) ULM. The standard errors are shown as
LK, et al. Strategy for elucidating differentially expressed genes inleiomyomata identified by microarray technology. Fertil Steril2003;80:282–90.
8. Chegini N, Verala J, Luo X, Xu J, Williams RS. Gene expressionprofile of leiomyoma and myometrium and the effect of gonadotropinreleasing hormone analogue therapy. J Soc Gynecol Investig 2003;10:161–71.
9. Weston G, Trajstman AC, Gargett CE, Manuelpillai U, VollenhovenBJ, Rogers PA. Fibroids display an anti-angiogenic gene expressionprofile when compared with adjacent myometrium. Mol Hum Reprod2003;9:541–9.
0. Hoffman PJ, Milliken DB, Gregg LC, Davis RR, Gregg JP. Molecularcharacterization of uterine fibroids and its implication for underlyingmechanisms of pathogenesis. Fertil Steril 2004;82:639–49.
1. Bodner K, Bodner-Adler B, Kimberger O, Czerwenka K, MayerhoferK. Bcl-2 receptor expression in patients with uterine smooth muscletumors: an immunohistochemical analysis comparing leiomyoma, uter-ine smooth muscle tumor of uncertain malignant potential, and leiomy-osarcoma. J Soc Gynecol Investig 2004;11:187–91.
2. Catherino W, Salama A, Potlog-Nahari C, Leppert P, Tsibris J, SegarsJ. Gene expression studies in leiomyomata: new directions for research.
Semin Reprod Med 2004;22:83–90.
ertility and Sterility�
3. Kovacs KA, Lengyel F, Kornyei JL, Vertes Z, Szabo I, Sumegi B, et al.Differential expression of Akt/protein kinase B, Bcl-2 and Bax proteinsin human leiomyoma and myometrium. J Steroid Biochem Mol Biol2003;87:233–40.
4. Guller S, Wozniak R, Leibman MI, Lockwood CJ. Negative regulationof placental fibronectin expression by glucocorticoids and cyclic aden-osine 3=,5=-monophosphate. Ann N Y Acad Sci 1994;734:132–42.
5. Tse V, Xu L, Yung YC, Santarelli JG, Juan D, Fabel K, et al. Thetemporal-spatial expression of VEGF, angiopoietins-1 and 2, and Tie-2during tumor angiogenesis and their functional correlation with tumorneovascular architecture. Neurol Res 2003;25:729–38.
6. Serganova I, Doubrovin M, Vider J, Ponomarev V, Soghomonyan S,Beresten T, et al. Molecular imaging of temporal dynamics and spatialheterogeneity of hypoxia-inducible factor-1 signal transduction activityin tumors in living mice. Cancer Res 2004;64:6101–8.
