Reproductive BioMedicine Online (2010) xxx, xxx–xxx

ARTICLE IN PRESS

www.sc iencedi rec t . comwww.rbmonl ine .com

ARTICLE

HOX A10 and HOX A11 mutation scan in congenitalmalformations of the female genital tract

Spiros A Liatsikos a, Grigoris F Grimbizis a,*, Ioannis Georgiou b,Nikolaos Papadopoulos a, Leandros Lazaros b, John N Bontis a,Basil C Tarlatzis a

a 1st Department of Obstetrics and Gynecology, Medical School, Aristotle University of Thessaloniki,Papageorgiou General Hospital, Thessaloniki, Greece; b Laboratory of Human Reproductive Genetics,Department of Obstetrics and Gynecology, Medical School, University of Ioannina, Ioannina, Greece* Corresponding author. E-mail addresses: grimbi@med.auth.gr, grimbi@auth.gr (GF Grimbizis).

Abstract Homeobox (HOXgenital tract during the em

1472-6483/$ - see front matdoi:10.1016/j.rbmo.2010.03

Please cite this article inBioMedicine Online (201

Dr Grigoris F Grimbizis received his MD and graduated as an obstetrician-gynaecologist from the Aristotle’sUniversity of Thessaloniki School of Medicine in Greece. After his PhD from the same University in 1996, heworked as a post-graduate fellow in infertility and endoscopic surgery at the Centre for ReproductiveMedicine of the Free University of Brussels under Professors Paul Devroey and Andre Van Steirteghem(1997). He is currently Assistant Professor in Obstetrics and Gynecology at the Aristotle’s University Schoolof Medicine. His clinical and research interests include gynaecological endoscopy and reproductivemedicine.

) genes encode a number of transcription factors, expressed along the developmental axis of the femalebryonic period. Because HOX A10 and HOX A11 genes are expressed in the embryonic paramesonephric

(Mullerian) ducts, abnormally low expression by mutant HOX A10 and HOX A11 genes might cause genital tract anomalies. This case–control study examined if one or more mutations in the HOX A10 and HOX A11 genes are included in the pathogenesis of the femalegenital tract anomalies. Blood samples were obtained from 30 women diagnosed with malformations of the genital tract (18 withseptate uterus, three with bicornuate uterus, two with didelphys uterus, two with unicornuate uterus and five with aplasia/dyspla-sia) and 100 normal controls. DNA samples prepared from blood leukocytes were used as templates for polymerase chain reactionamplification of DNA fragments from HOX A10 and HOX A11 genes. The gene fragments were tested for DNA sequence differencesusing single-strand conformation polymorphism analysis and sequenced when genetic variation was detected. No subject showed aplausible causative mutation in HOX A10 or HOX A11; the sole variant observed (P38R) found in a patient with septate uterus was also

present in her clinically normal mother. RBMOnline

ª 2010, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.

KEYWORDS: congenital uterine malformations, female genital tract, HOX genes, Mullerian duct, mutations

ter ª 2010, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved..015

press as: Liatsikos, SA et al., HOX A10 and HOX A11 mutation scan in congenital ..., Reproductive0), doi:10.1016/j.rbmo.2010.03.015

2 SA Liatsikos et al.

ARTICLE IN PRESS

Introduction

Mullerian and Wolffian ducts are the primordial for theinternal reproductive tracts of females and males, respec-tively. In the beginning, both ducts exist side by side inthe human embryo, until genetic sex triggers differentiationof either ovaries or testes. The persistence and normal dif-ferentiation of the Mullerian or paramesonephric ducts isimportant for the development of the female internalreproductive organs. Thus, in genetic female embryos(46,XX) in the absence of testes and, therefore, absenceof anti-Mullerian Hormone (AMH) secretion, the Mullerianducts, which have been formed by an invagination of thecoelomic epithelium, develop into the oviducts, uterus, cer-vix and the upper part of the vagina. Furthermore, in theabsence of testosterone secretion, the Wolffian ducts leadto atrophy (Guerrier et al., 2006; MacLaughlin et al.,2001; Simpson, 1999).

Congenital anomalies of the female genital system aremainly the result of four major disturbances during fetallife, concerning the development, formation or fusion ofthe Mullerian ducts: (i) failure of one or both Mullerian ductsto develop (agenesis or unicornuate uterus without arudimentary horn); (ii) failure of the ducts to canalize(unicornuate uterus with a rudimentary horn without propercavities); (iii) failure of the fusion or abnormal fusion of theducts (uterus didelphys or bicornuate uterus); and (iv) fail-ure of reabsorption of the midline septum of the uterus(septate uterus or arcuate uterus) [American FertilitySociety (AFS), 1988; Ashton et al., 1988; Acien, 1992;Grimbizis et al., 1998]. The factors underlying these distur-bances are not yet known.

The AFS (1988), based on the previous work of Buttramand Gibbons (1979), classified the anomalies of the femalegenital tract according to the degree of failure of normaldevelopment into six major uterine anatomic types: (I)hypoplasia/agenesis; (II) unicornuate uterus (IIa: with acommunicating rudimentary horn; IIb: with a non-communi-cating rudimentary horn; IIc: with a rudimentary horn with-out a cavity; IId: without a rudimentary horn); (III) didelphysuterus; (IV) bicornuate uterus (IVa: complete; IVb: partial);(V) septate uterus (Va: complete; Vb partial); and (VI) arcu-ate uterus. The current AFS classification system seems tobe, until now, the most accepted way of categorizingfemale’s genital tract congenital anomalies, although itcannot incorporate effectively all the anatomic variationspublished in the literature. Some new proposals have beenpublished with limited acceptance (Acien et al., 2004;Oppelt et al., 2005).

The incidence of Mullerian anomalies in the generalpopulation is not yet accurately known. The lack of a stan-dard system of classification in the available studies, thedifferent diagnostic methods used and the fact that thepopulation studied is not often representative of the gen-eral population are serious limiting factors. Furthermore,some cases of uterine malformations remain undiagnosed,mainly because they are asymptomatic. It seems thoughthat uterine defects are not uncommon. In a review of fivestudies and about 3000 cases, the mean overall incidence inthe general population and in the population of fertilewomen was 4.3% (Grimbizis et al., 2001). More interesting

Please cite this article in press as: Liatsikos, SA et al., HOX A10BioMedicine Online (2010), doi:10.1016/j.rbmo.2010.03.015

is the prevalence of the different types of the anomalies.According to a review of nine studies from 1978 to 1999and a total of about 1400 cases, the mean incidence of sep-tate uterus was found to be about 35%, of bicornuate uterusabout 25%, of arcuate uterus about 18%, of unicornuateuterus about 9%, of didelphys uterus about 9% and of Mulle-rian agenesis about 3% (Grimbizis et al., 2001). It seems,therefore, that about 55% of the uterine malformationsare either septate or arcuate uterus.

Homeobox (HOX) genes encode a number of transcriptionfactors, which are expressed along the anterior–posterioraxis and control the embryonic patterning of the body.Patterning of the vertebrate hindbrain, the skeleton axisand the limb axis is also performed by HOX genes, expressedin a similar way. The embryonic female reproductive systemis considered to represent a developmental axis similar tothose described above, as the uniform paramesonephricduct will develop into the oviducts, the uterus, the cervixand the upper part of the vagina. HOX genes belonging toparalogue groups 9–13 seem to provide the axis of thedeveloping paramesonephric duct with a positional identity:HOX A9 is expressed in areas designated to become theoviduct, HOX A10 is mainly expressed in the developinguterus, HOX A11 is expressed in parts of the Mullerian ductforming the lower compartment of the uterus and the cervixand HOX A13 is expressed in the upper third of the vagina.There is no HOX A12 gene. It has probably been lost duringevolution (Du and Taylor, 2004).

The aim of this prospective study was to examine if adysfunction due to mutations or sequence polymorphismsof the HOX genes, which are expressed along the axis ofthe developing paramesonephric duct, may underlie thepathogenesis of the congenital malformations of the femalegenital tract. Since HOXA10 and HOXA11 are mainlyexpressed in the developing uterus and the cervix, thesetwo genes are chosen for DNA sequence analysis.

Materials and methods

Patient selection

Thirty Caucasian women, all ethnic Greeks, presenting withany type of congenital malformation of the internal repro-ductive organs and 100 normal control subjects withoutknown reproductive tract defects were included in thisstudy from June 2005 until December 2007. Local ethicalreview and consenting procedures were followed for eachvolunteer. The mean age was 28 years (range 15–40 years).The diagnosis and classification of the malformation wasperformed using any available method, such as ultrasound,hysterosalpingography, hysteroscopy, laparoscopy and insome cases laparotomy. All the women underwent a pyelog-raphy for the detection of coexistent anomalies of the uri-nary tract.

Patients were attributed according to AFS classification.Eighteen out of the 30 women presented with a septateuterus, indicating that is the most common uterine anom-aly. In three cases the diagnosis was bicornuate uterus, intwo cases didelphys uterus and in two cases unicornuateuterus. In five out of 30 women, some type of aplasia/hypo-plasia was diagnosed. In particular, one woman lacked the

and HOX A11 mutation scan in congenital ..., Reproductive

Table 1 American Fertility Society (AFS) classification and coexistent urinary tract anomalies of the patientsincluded in the study.

Type of anomaly andmalformation description

Cases(n)

Coexistent urinary tractmalformation (n)

Developmental failure

IRight ovary aplasia 1 Gonadal dysgenesia

Uterine cervix agenesis 2 Left kidney agenesis (1) Failure of Mullerian ducts to developDouble renal pelvis of theright kidney (1)

Failure of Mullerian ducts to develop

Mayer–Rokitansky–Kuster–Hauser syndrome

2 Left kidney agenesis (1) Failure of Mullerian ducts to develop

IIUnicornuate uterus 2 Left kidney agenesis (2) Failure of Mullerian ducts to develop

or canalize

IIIDidelphys uterus 2 Right kidney agenesis (1) Failure of Mullerian ducts to fuse

IVBicornuate uterus 3 Abnormal fusion of Mullerian ducts

VSeptate uterus 18 Ectopic urinary bladder (1) Failure of reabsorption of the

midline uterine septumTotal 30 7

Hox gene mutations and congenital uterine malformations 3

ARTICLE IN PRESS

right ovary, while agenesis of the uterine cervix was foundin two women. Finally, two women presented with the mostextreme form of congenital aplasia of the genital tract, theMayer–Rokitansky–Kuster–Hauser syndrome, lacking theuterus, cervix and proximal vagina. Seven women had acoexistent urinary tract anomaly (Table 1).

Preparation of genomic DNA

Peripheral blood samples from each patient and the controlsubjects were sent to the laboratory. Total genomic DNAwas prepared from peripheral blood leukocytes (200 llwhole blood) according to a standard protocol, using theQIAamp DNA Blood Mini Kit (Qiagen, USA).

Polymerase chain reaction

Both HOX A10 and HOX A11 genes are located on humanchromosome 7. The HOX A10 DNA size is 9.63 kb and itstranscription unit includes two exons (579 and 1582 bp).The HOX A11 DNA size is 4.1 kb and its transcription unit alsoincludes two exons (776 and 1519 bp).

Human genomic DNA was used as a template for poly-merase chain reaction (PCR). The DNA fragments thatincluded HOX A10 (exon 1 and 2) and HOX A11 (exon 1and 2) coding sequences were amplified individually usingfive pairs (Forward and Reverse) of oligonucleotide primersfor each gene, conventionally called, sequentially, HOXA10 F1/R1 to HOX A10 F5/R5 and HOX A11 F1/R1 to HOXA11 F5/R5 (Table 2).

Please cite this article in press as: Liatsikos, SA et al., HOX A10BioMedicine Online (2010), doi:10.1016/j.rbmo.2010.03.015

The PCR reactions were carried out according to stan-dard procedures (Orti et al., 1997) in a total 25 ll mixture,containing 1–2 ll (100–200 ng) genomic DNA, 2.5 ll 10·PCR buffer, 2.5 ll of dNTP mix (2 mmol/l), 2 ll of eachprimer, 0.5 ll MgCl2 (50 mmol/l) and 0.2 ll (1 unit) ofPlatinum Taq DNA polymerase (Invitrogen, USA). In somePCR reactions (HOX A10 F1/R1, A10 F3/R3, A10 F4/R4 andHOX A11 F3/R3, A11 F4/R4), 1.25 ll 5% dimethyl sulphoxide(DMSO) was included to the total volume of the reactionbuffer. In the case of HOX A10 F5/R5 reaction, PCRx Enhan-cer System (Invitrogen) was included in a total volume of50 ll buffer, comprising of 5 ll 10· PCRx Enhancer Solutionand 5 ll 10· PCRx Amplification Buffer. Some DNAsequences, especially for HOX A10 gene are extremely GCrich. Amplifying those sequences using standard PCRreaction conditions is difficult, due to Platinum Taq DNApolymerase stalling. DMSO and PCRx Enhancer Systemreduce polymerase stalling.

PCR amplification was performed with an initial denatur-ation step at 94�C for 10 min, followed by 35 temperaturecycles each consisting of 94�C for 1 min (denaturation),1 min at the appropriate annealing temperature (58�C forthe HOX A10 F2/R2, A11 F1/R1 and A11 F3/R3 reactions,56�C for the HOX A10 F5/R5 reaction, 55�C for the HOXA11 F2/R2 and A11 F5/R5 reactions) and 72�C for 1 min(extension), and a final extension step at 72�C for 10 min,unless stated otherwise below. For the HOX A10 F1/R1reaction after the initial denaturation, 35 cycles werecarried out, consisting of 94�C for 1 min, 60�C for 30 s and72�C for 1 min. For HOX A10 F3/R3 and HOX A11 F4/R4

and HOX A11 mutation scan in congenital ..., Reproductive

Table 2 Primers used for PCR-mediated amplification of genomic DNA of HOX A10 and HOX A11 genes.

HOX A10 HOX A11

Primer name Sequence (50–30) Primer name Sequence (50–30)

HOXA10 F1 atgtcagccagaaagggcta HOXA11 F1 ctacttcacggatccgcttcHOXA10 R1 cagctctgcagcccgtag HOXA11 R1 caatggcgtactctctgaaggHOXA10 F2 ggcggtggcggttactac HOXA11 F2 cgcccaatgacatactcctaHOXA10 R2 cgcgtctagccacaggtcta HOXA11 R2 gcccacggtgctatagaaatHOXA10 F3 gtggcgggggtctaggtc HOXA11 F3 agagctcggccaacgtctacHOXA10 R3 gccgagtcgtagaggcagta HOXA11 R3 ccgctgtccgaacttgaaHOXA10 F4 caggccacctcgtgctct HOXA11 F4 gacaagagcgccgagaagHOXA10 R4 ggtgccgtaggcctgaga HOXA11 R4 gatttccaactcccctttcaHOXA10 F5 gctacttccgcctttctcag HOXA11 F5 ctcaccccatgccttttctHOXA10 R5 tccttgtgtctgcctgtctg HOXA11 R5 gtcaagggcaaaatctgcat

F, forward; R, reverse.

4 SA Liatsikos et al.

ARTICLE IN PRESS

reactions, the initial denaturation at 94�Cwas followed by 20cycles of 94�C for 1 min, 56�C for 30 s, 72�C for 1 min and 10cycles of 94�C for 1 min, 56�C for 1 min, 72�C for 1 min.Finally, for the HOX A10 F4/R4 reaction 35 cycles were per-formed, each consisting of 93�C for 1 min, 64�C for 2 minand 72�C for 1 min. The PCR reactions with no DNA templatewere used as negative controls for all PCR reaction sets.

To ensure that a single band of the expected size wasamplified, 5 ll of each PCR product was admixed with 1 llloading buffer, electrophoresed in small 2% agarose gelscontaining ethidium bromide and visualized under ultravio-let light. The size of each PCR product was compared withthat of a TrackIt 100 bp DNA ladder (Invitrogen).

Single-strand conformation polymorphism analysis

Single-strand conformation polymorphism (SSCP) analysiswas used as a screening method to detect any geneticvariation of the PCR products from different individuals.The purpose of the method is to take PCR products, dena-ture them into single-stranded nucleic acids and electropho-rese them through a non-denaturing polyacrylamide gel. Asthe PCR products move into and through the gel, they willregain secondary structure that is sequence dependent.PCR products that contain substitutional sequence differ-ences, as well as insertions and deletions will have differentmobilities. The major advantage of SSCP analysis is thatmany individual PCR products may be screened for variationsimultaneously reducing the amount of sequencing neces-sary to detect new alleles at loci of interest, thus SSCP anal-ysis offers an inexpensive, convenient and sensitive methodfor determining genetic variation (Miterski et al., 2000).

SSCP analysis was performed under strict conditions. EachPCR product (2 ml) was admixedwith 2 ml formamide loadingbuffer (100% formamide, EDTA, bromophenol blue 2%, xylenecyanol 5%) and 6 ml dH2O. Denaturation of the DNA followedat 95�C for 10 min. Immediately after denaturation the prod-ucts were loaded on 8% polyacrylamide gels, made up of40% acrylamide/bis-acrylamide (37:1) polymerized by 10%ammonium persulphate and tetramethylethylenediamine.Samples were loaded side by side (so as to be easier to

Please cite this article in press as: Liatsikos, SA et al., HOX A10BioMedicine Online (2010), doi:10.1016/j.rbmo.2010.03.015

compare the results from different individuals simulta-neously) and electrophoresed at 80–100 V for 18–22 h,dependingon theDNA size of thePCRproduct, in a Biorad unit.

The gels were then transferred to nylon membranes andwere fixed with fixing solution, containing ethanol andacetic acid. SSCP alleles were detected using silver staining(0.1% AgNO3). Bands were developed in a developingsolution comprising of NaOH, formaldehyde and NABH4.Finally, the bands were compared with each other and thepattern was interpreted. It is clear that different bandscontain strands with different sequences: the more farapart the bands, the less similar the nucleotide sequences.

DNA sequencing

DNA sequencing was performed on the PCR products whengenetic variation was detected via the SSCP analysis. Ini-tially, remaining primers were digested and unincorporatednucleotides were inactivated. Then bidirectional sequenc-ing of the PCR products was performed using the ABI PRISM3130 sequencer (Applied Biosystems, USA). Sequences wereanalysed and compared with sequences downloaded fromGenatlas (www.geneatlas.com). SIFT software (Ng andHenikoff, 2003) was used to predict if the possible muta-tions have any effect on protein function.

Results

Genetic variation was detected via the SSCP analysis in onesample, coming from a woman presenting with septateuterus and an ectopic urinary bladder. In particular, a shiftwas found in the sequence of HOX A11 F1/R1 PCR productwhen its band was compared with the bands of the corre-sponding products from other women, after electrophoresison a polyacrylamide gel (Figure 1). PCR reaction wasrepeated, using a new pair of HOX A11 F1/R1 primers, forthe specific sample, along with some more samples servingas controls. A new SSCP analysis was performed, leading tothe same shift in the HOX A11 F1/R1 product. The shiftwas not found in the SSCP analysis of the HOX A11 F1/R1products of 100 normal controls.

and HOX A11 mutation scan in congenital ..., Reproductive

Figure 1 Single-strand conformation polymorphism analysisrevealed a shift in the sequence of HOX A11 F1/R1 PCR productin one sample. As it can be seen, an extra band is producedafter the electrophoreses of the product below the arrow,which is not apparent in other samples when compared witheach other.

Hox gene mutations and congenital uterine malformations 5

ARTICLE IN PRESS

DNA sequencing of the specific PCR product was carriedout to determine the type of the genetic variation.Sequencing revealed a 113 C>G (P38R) mutation, concern-ing the heterozygotic substitution of a cytosine by a guaninein location 113 of the HOX A11 gene (Figure 2). That muta-tion causes the substitution of a proline by an arginine incodon 38. Nevertheless, it does not impair the homeodo-main. According to SIFT software, the 113 C>G (P38R)mutation has no effect on protein function (score = 0.61).

After that finding, a blood sample was asked from each ofthe woman’s parents in order to examine parental geno-type. A sample was obtained from her mother, but it wasimpossible to obtain a sample from her father due to familyreasons (divorced and no communication). Nevertheless,the maternal blood sample was analysed and the same 113C>G (P38R) mutation was also detected. An ultrasoundand a pyelography were performed on the woman’s mother,which showed a normal uterus, without fusion defects orothers congenital malformations of the reproductive andthe urinary tracts.

None of the other women with septate uterus or othermalformation, as well as none of 100 control subjects,presented the same mutation. Finally, no other mutation

Figure 2 DNA sequencing of the specific PCR product. Thedouble peak (arrow) represents a heterozygotic substitution ofa cytosine by a guanine, leading to a 113 C>G (P38R) mutation.The mutation causes the substitution of a proline by an argininein codon 38.

Please cite this article in press as: Liatsikos, SA et al., HOX A10BioMedicine Online (2010), doi:10.1016/j.rbmo.2010.03.015

or length/nucleotide polymorphism was detected in thecoding sequences of HOX A10 and HOX A11 genes of the pa-tients and the control subjects investigated.

Discussion

In humans, HOX genes are clustered in four genomic loci,HOX A–D, located on chromosomes 7, 17, 12 and 2, respec-tively, each of them containing a subset of 9–13 genes. HOXgenes in every cluster are classified into groups, calledparalogues, depending on their sequence. Repeated dupli-cations in a number of ancestral genes have led to a seriesof closely related HOX genes, which are separated by shortintergenic regions. During evolution some of the paralogueswere lost. As a result, all mammalians today contain a totalof 39 HOX genes. It is believed that HOX genes that are lo-cated more at the 30 end in the chromosome are expressedearlier and in more anterior positions of the body than their50 neighbours. As a result, both the time and the space ofexpression depend on the position of the genes within acluster (Du and Taylor, 2004).

Each homeobox gene contains a common 183-bpsequence, which encodes a 61-amino acid domain. This iscalled a homeodomain and is usually localized at a subtermi-nal position of the corresponding homeoprotein. The role ofthe homeodomain is to recognize and bind DNA sequenceswith a typical core of TTAT or TAAT. According to structuralanalyses, the homeodomain has the ability to shape into acompact globular structure, consisting of three helicesand one amino-terminal arm. The homeodomain’s three-dimensional structure is highly conserved and is veryimportant for DNA binding. Homeoproteins are essentialtranscription factors having the ability to regulate theexpression of other genes, mainly developmental.

Very few studies are available regarding HOX gene muta-tions. It was not until 1996 that certain mutations in a HOXgene, in particular HOX D13, were found to be the cause of ahuman malformation syndrome, called synpolydactyly(Muragaki et al., 1996). A year later, hand-foot-genitalsyndrome, which among other defects causes Mullerian ductfusion defects in females, was first proved to be caused by amutation in the HOX A13 gene (Mortlock and Innis, 1997). Itwas the first correlation between congenital malformationsof the female genital tract and a HOX gene mutation. In thefollowing years, more mutations of HOX A13 gene werefound in cases of hand-foot-genital syndrome (Devriendtet al., 1999; Goodman et al., 2000; Utsch et al., 2002). Untiltoday, no other mutation or polymorphism in HOX genes hasbeen confirmed in cases of women presenting with any typeof uterine malformation.

On the other hand, sex steroids as well as some syntheticoestrogens, such as diethylstilboestrol (DES), seem to alterHOX gene expression (Du and Taylor, 2004). In particular,HOX genes at the 50 end of a cluster, which among othersdetermine the development of the reproductive tract, arebelieved to be regulated mainly by sex steroids. HOX A10and HOX A11 expression is up-regulated by 17b-oestradioland by progesterone. The regulation is direct and is achievedby the oestrogen or the progesterone receptor binding toregulatory regions of the genes. In mice, after exposure toDES the localization of HOX A9 expression is shifted from

and HOX A11 mutation scan in congenital ..., Reproductive

6 SA Liatsikos et al.

ARTICLE IN PRESS

the oviducts to the uterus and both HOX A10 and HOX A11expression of the uterus is dramatically reduced. A new phe-notype of the uterus is generated, the T-shaped uterus, pre-senting with a tube-like form. DES probably impairs theconformation of the oestrogen receptor. As a result, thereceptor interacts preferably with atypical coactivators andcorepressors rather than those bound after treatment withoestradiol. A similar role of DES is also suggested in humans,as it has been proven that it shifts the expression pattern ofHOX genes in human uterine cell cultures (Goodman, 2002).

This study’s hypothesis was in favour of the pathogenesisof congenital malformations of the female internal repro-ductive organs being a dysfunction of the HOX genes, whichare expressed along the axis of the developing parameso-nephric duct. However, no mutations or sequence polymor-phisms were found in the protein-coding region of HOX A10andHOXA11genes in thecasesofMayer–Rokitansky–Kuster–Hauser syndrome examined, which is the most extremeparamesonephric duct developmental abnormality. This isin accordance with the findings of other investigators, whosestudies were limited only to cases of this specific syndrome.Burel et al. (2006) did not find any mutation or length/nucleotide polymorphism in the coding sequences of HOX A7to HOX A13 genes in six cases of Mullerian aplasia (MRKHsyndrome). Ina studyof Lalwani et al. (2008), noHOXA10genemutations were found in 26 patients with MRKH syndrome.

It is interesting that a single mutation, not previously an-nounced, was found in a coding sequence of the HOX A11gene, causing the substitution of a neutral amino acid(proline) by a basic one (arginine) in the correspondingprotein. More interesting is the fact that the specificmutation concerned just one woman presenting withseptate uterus, an anomaly caused by failure of reabsorptionof the midline uterine septum, while it was not apparent inthe other cases studied, presenting with more severe or eventhe same uterine malformation.

It becomes obvious that HOX A10 and HOX A11 genemutations do not participate in the pathogenesis of the firsttwo types of developmental failure (formation and canaliza-tion of the Mullerian ducts), thus they do not lead to aplasiaof the uterus or uterus unicornuate. However, the particulargenes seem to have a role in the fusion of the ducts and thereabsorption of the midline uterine septum. This is usuallydue to an alteration in the expression of the genes, underthe control of certain internal and external factors, mainlyhormonal. It remains doubtful if the isolated mutation dis-covered in the present study is a coincidental event or itshould be the subject for further investigation. It wouldbe worthwhile to check the penetration of the mutationby studying a larger sample of women presenting with sep-tate/bicornuate uterus.

Concerning the role of HOX gene mutations in the patho-genesis of uterine malformations, despite the fact that theresults of the studies have so far been disappointing, HOXgenes remain good candidate genes for the aetiology of suchanomalies. The present study searched for mutations andpolymorphisms in rather targeted genomic regions (pro-tein-coding and regulatory sequences) of the HOX A10 andHOX A11 genes. However, the HOX A10 and A11 clusterswere not fully scanned in every patient, for example, the in-trons were not examined. It is possible that mutations couldhave been present in these regions in some patients.

Please cite this article in press as: Liatsikos, SA et al., HOX A10BioMedicine Online (2010), doi:10.1016/j.rbmo.2010.03.015

Furthermore, mutations located outside the protein-codingsequences could affect the transcription of the genes, themRNA processing or the mRNA stability. Additional investi-gation is necessary to elucidate those doubtful points.

Obviously HOX gene polymorphisms and mutations occurrarely, suggesting that most sequence shifts in those genesmight have lethal consequences during fetal development,thus they do not usually lead to living women with uterineanomalies. Nevertheless, other mechanisms could beassumed, like a misregulation of some paralogues of theHOX A cluster, post-transcriptional malfunction, impairedinteraction with activators and suppressors and failed regu-lation of HOX target genes, such as b3-integrin and EMX2(Daftary et al., 2002; Troy et al., 2003).

The pathogenesis of the congenital malformations of theuterus, as that of most developmental disorders, seems tobe multifactorial. The phenotype might be the result of theadditive effects of miscellaneous proteins under the influ-ence of hormonal factors. The role of sex steroids has alreadybeen mentioned. Additionally, unknown environmentalteratogens may modify the developmental fate of theMullerian duct progenitor cells (Taylor, 2008). Apart fromDES, several other oestrogen-like endocrine disruptors(xeno-oestrogens) are capable of altering the expression ofessential developmental genes, such as HOX genes. Such axeno-oestrogen is methoxychlor, which represses HOX A10expression in the uterus of mice, and bisphenol A, which in-creasesHOXA10expression in adultmice. Both cause a severereduction in the reproductive performance of the mice (Feiet al., 2005; Markey et al., 2005; Smith and Taylor, 2007).Further investigation is necessary for the identification ofmore exogenous oestrogenic factors, which shift HOXgene expression in women, leading possibly to congenitalanomalies.

Finally, other genes that seem to be important in devel-opmental procedures of the female reproductive system arebeing studied. Such genes are PBX1 and HOX A13, the onlygene that has so far been related to Mullerian duct fusiondefects in hand-foot-genital syndrome. Apart from those,genes of the oestrogenic receptors seem to have a similarrole in genital tract development, thus are good candidategenes for future study.

In conclusion, no subject showed a plausible causativemutation in HOX A10 or HOX A11; the sole variant observed(P38R) found in a patient with septate uterus was also pres-ent in her clinically normal mother.

Acknowledgements

The authors would like to thank Dr Ioanna Bouba for labora-tory assistance and Dr Anthoula Chatzikyriakidou for expertadvice (Laboratory of Human Reproductive Genetics, Medi-cal School, University of Ioannina).

References

Acien, P., 1992. Embryologic observations of the female genitaltract. Hum. Reprod. 7, 437–445.

Acien, P., Acien, M., Sanchez-Ferrer, M., 2004. Complex malfor-mations of the female genital tract. New types and revision ofclassification. Hum. Reprod. 19, 2377–2384.

and HOX A11 mutation scan in congenital ..., Reproductive

Hox gene mutations and congenital uterine malformations 7

ARTICLE IN PRESS

American Fertility Society, 1988. The American Fertility Classifi-cations of adnexal adhesions, distal tubal occlusion, tubalocclusion secondary to tubal ligation, tubal pregnancies,Mullerian anomalies and intrauterine adhesions. Fertil. Steril.49, 944–955.

Ashton, D., Amin, H.K., Richart, R.M., Neuswirth, R.S., 1988. Theincidence of asymptomatic uterine anomalies in women under-going transervical sterilization. Obstet. Gynecol. 72, 28–30.

Burel, A., Mouchel, T., Odent, S., et al., 2006. Role of HOXA7 toHOXA13 and PBX1 genes in various forms of MRKH syndrome(congenital absence of uterus and vagina). J. Negat. Res.Biomed. 5, 4.

Buttram Jr., V.C., Gibbons, W.E., 1979. Mullerian anomalies: aproposed classification (an analysis of 144 cases). Fertil. Steril.32, 40–48.

Daftary, G.S., Troy, P.J., Bagot, C.N., et al., 2002. Direct regula-tion of beta3-integrin subunit gene expression by HOXA10 inendometrial cells. Mol. Endocrinol. 16, 571–579.

Devriendt, K., Jaeken, J., Matthijs, G., et al., 1999. Haploinsuffi-ciency of the HOXA gene cluster, in a patient with hand-foot-genital syndrome, velopharyngeal insufficiency, and persistentpatent ductus botalli. Am. J. Hum. Genet. 65, 249–251.

Du, H., Taylor, H.S., 2004. Molecular regulation of mulleriandevelopment by Hox genes. Ann. N. Y. Acad. Sci. 1034, 152–165.

Fei, X., Chung, H., Taylor, H.S., 2005. Methoxychlor disruptsuterine Hoxa10 gene expression. Endocrinology 146, 3445–3451.

Goodman, F.R., 2002. Limb malformations and the human HOXgenes. Am. J. Med. Genet. 112, 256–265.

Goodman, F.R., Bacchelli, C., Brady, A.F., et al., 2000. NovelHOXA13 mutations and the phenotypic spectrum of hand-foot-genital syndrome. Am. J. Hum. Genet. 67, 197–202.

Grimbizis, G.F., Camus, M., Clasen, K., et al., 1998. Hysteroscopicseptum resection in patients with recurrent abortions or infer-tility. Hum. Reprod. 13, 1188–1193.

Grimbizis, G.F., Camus, M., Tarlatzis, B.C., Bontis, J.N., Devroey,P., 2001. Clinical implications of uterine malformations andhysteroscopic treatment results. Hum. Reprod. Update 7, 161–174.

Guerrier, D., Mouchel, T., Pasquier, L., Pellerin, I., 2006. TheMayer–Rokitansky–Kuster–Hauser syndrome (congenitalabsence of uterus and vagina) – phenotypic manifestationsand genetic approaches. J. Negat. Res. Biomed. 5, 1.

Lalwani, S., Wu, H.H., Reindollar, R.H., Gray, M.R., 2008. HOXA10mutations in congenital absence of uterus and vagina. Fertil.Steril. 89, 325–330.

Please cite this article in press as: Liatsikos, SA et al., HOX A10BioMedicine Online (2010), doi:10.1016/j.rbmo.2010.03.015

MacLaughlin, D.T., Teixeira, J., Donahoe, P.K., 2001. Perspective:reproductive tract development – new discoveries and futuredirections. Endocrinology 142, 2167–2172.

Markey, C.M., Wadia, P.R., Rubin, B.S., et al., 2005. Long-termeffects of fetal exposure to low doses of the xenoestrogenbisphenol-A in the female mouse genital tract. Biol. Reprod. 72,1344–1351.

Miterski, B., Kruger, R., Wintermeyer, P., Epplen, J.T., 2000. PCR/SSCP detects reliably and efficiently DNA sequence variations inlarge scale screening projects. Comb. Chem. High ThroughputScreen. 3, 211–218.

Mortlock, D.P., Innis, J.W., 1997. Mutation of HOXA13 in hand-foot-genital syndrome. Nat. Genet. 15, 179–180.

Muragaki, Y., Mundlos, S., Upton, J., Olsen, B.R., 1996. Alteredgrowth and branching patterns in synpolydactyly caused bymutations in HOXD13. Science 272, 548–551.

Ng, P.C., Henikoff, S., 2003. SIFT: predicting amino acid changesthat affect protein function. Nucleic Acids Res. 31, 3812–3814.

Oppelt, P., Renner, S.P., Brucker, S., et al., 2005. The VCUAM(Vagina Cervix Uterus Adnex-associated Malformation) classifi-cation: a new classification for genital malformations. Fertil.Steril. 84, 1493–1497.

Orti, G., Hare, M.P., Avise, J.C., 1997. Detection and isolation ofnuclear haplotypes by PCR-SSCP. Mol. Ecol. 6, 575–580.

Simpson, J.L., 1999. Genetics of the female reproductive ducts.Am. J. Med. Genet. 89, 224–239.

Smith, C.C., Taylor, H.S., 2007. Xenoestrogen exposure imprintsexpression of genes (Hoxa10) required for normal uterinedevelopment. FASEB J. 21, 239–246.

Taylor, H.S., 2008. Endocrine disruptors affect developmentalprogramming of HOX gene expression. Fertil. Steril. 89, 57–58.

Troy, P.J., Daftary, G.S., Bagot, C.N., Taylor, H.S., 2003. Tran-scriptional repression of peri-implantation EMX2 expression inmammalian reproduction by HOXA10. Mol. Cell. Biol. 23, 1–13.

Utsch, B., Becker, K., Brock, D., et al., 2002. A novel stablepolyalanine [poly(A)] expansion in the HOXA13 gene associatedwith hand-foot-genital syndrome: proper function of poly(A)-harbouring transcription factors depends on a critical repeatlength? Hum. Genet. 110, 488–494.

Declaration: The authors report no financial or commercial conflictsof interest.

Received 11 July 2009; refereed 4 August 2009; accepted 3 February2010.

and HOX A11 mutation scan in congenital ..., Reproductive