Role of HOXA7 to HOXA13 and PBX1 genes in various forms of MRKH syndrome (congenital absence of uterus and vagina)

The Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome refers to the congenital absence or severe hypoplasia of the female genital tract, often described as uterovaginal aplasia which is the prime feature of the syndrome. It is the second cause of primary amenorrhea after gonadal dysgenesis and occurs in ~1 in 4500 women. Aetiology of this syndrome remains poorly understood. Frequent association of other malformations with the MRKH syndrome, involving kidneys, skeleton and ears, suggests the involvement of major developmental genes such as those of the HOX family. Indeed mammalian HOX genes are well known for their crucial role during embryogenesis, particularly in axial skeleton, hindbrain and limb development. More recently, their involvement in organogenesis has been demonstrated notably during urogenital differentiation. Although null mutations of HOX genes in animal models do not lead to MRKH-like phenotypes, dominant mutations in their coding sequences or aberrant expression due to mutated regulatory regions could well account for it. Sequence analysis of coding regions of HOX candidate genes and of PBX1, a likely HOX cofactor during Müllerian duct differentiation and kidney morphogenesis, did not reveal any mutation in patients showing various forms of MRKH syndrome. This tends to show that HOX genes are not involved in MRKH syndrome. However it does not exclude that other mechanisms leading to HOX dysfunction may account for the syndrome.


Background
The most common cause of vaginal agenesis is congenital absence of the uterus and vagina which is also referred to as Müllerian aplasia, Müllerian agenesis or Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome [1]. The frequency of this syndrome is not yet entirely clear, although reported incidences vary from 1 in 4,000 to 5,000 female births [1][2][3]. Affected individuals are clearly phenotypic females with normally developed ovaries [4,5] and normal 46, XX karyotype [6,7]. Aetiology of the syndrome is poorly understood but it is often associated with other anomalies including renal defects, skeletal abnormalities and deafness (MURCS association [8]), suggesting the involvement of major developmental genes such as HOX genes [9][10][11][12].
The homeobox (HOX) genes belong to a large family of 39 genes organized in four clusters, HOXA, HOXB, HOXC and HOXD, each on a different chromosome. During organogenesis, the proteins encoded by these genes act through various and highly complex spatiotemporal combinations to trigger positional identity of embryonic cells. This determines the patterning and segment identity along the anterior-posterior axis of the skeleton and a variety of organ systems [13]. For instance, 30 HOX proteins participate to the elaboration of the spine, 12 for the digestive tract and 7 for the urogenital tract [14]. More precisely, Müllerian ducts (the primordia for oviducts, uterus, cervix and anterior vagina) development seems to involve relatively few HOX genes in the mouse model. Indeed, HOXA7 [15], HOXA9 to HOXA13 [16], as well as HOXD9 to HOXD13 [17], are expressed along the differentiating Müllerian duct. However, alteration of the female genital tract is only observed in HOXA10, -A11 and -A13 deficient mice (homozygotic inactivation of the gene): -in HOXA10 -/-mice, the upper part of the uterus is transformed into oviduct, the uterotubular junction is abnormal as well as the uterine epithelium and an anterior homeotic transformation of lumbar vertebrae has occurred [18]; -in HOXA11 -/-mice, the uterus is thinner and shorter than normal and endometrial glands have not developed [19]; in HOXA13 -/-mice, the distal Müllerian duct has not developed [20]. Finally HOXA10 to HOXA13 are also expressed in the developing kidney [21] and are both required for correct patterning of the skeleton [22]. HOX proteins share in common a highly conserved 60 amino acid DNA binding motif referred to as the homeodomain. Proteins containing this domain are regulatory factors that control expression of target genes [23]. Their high biological specificity comes from cooperation with specific cofactors that contribute to modulate DNA binding specificity. Members of the three amino-acid loop extension (TALE) class of homeodomain proteins that comprise the mammalian PBX proteins [24] and the MEIS-like TALE factors or MEINOX group (mammalian MEIS and PREP1 proteins) [25] are now considered as essential cofactors forming heterotrimeric complexes with HOX proteins that regulate specific target gene transcription [26]. Among these cofactors, PBX1 is of great interest in regards to malformations found in MRKH syndrome: it

Patient 3
This 20-year-old white woman was evaluated for primary amenorrhea. She had normal secondary sexual development. There was no cyclic abdominal pain. Family history was unremarkable. The MRKH diagnosis was confirmed by laparoscopy. Absence of right ovary and fallopian tube was noticed during surgery. However, ultrasound examination showed normal kidneys.

Patients 4 to 6
These patients are three Turkish sisters already described [35] (patients III2, III3, III5 of pedigree). Interestingly, in this family, the fourth sister (III4) was not affected but two paternal aunts (II6 and II7), among 8 siblings, were sterile and were told they had no uterus. This three sisters case corresponds to typical MRKH syndrome with primary amenorrhea, normal sexual secondary development and absence of the vagina at physical evaluation. The Müllerian agenesis was confirmed by ultrasound examination and magnetic resonance imaging of pelvis. Their karyotypes were normal. Intravenous pyelogram and spine radiograms were normal in each case.

Results and discussion
The pattern of malformations observed in MRKH patients was, in our hypothesis, in favour of a HOX gene dysfunction. However no mutation as well as length/nucleotide polymorphism was found in the coding sequences of HOXA7 to -A13 genes of the patients we investigated. Post-transcriptional regulations also take place in the overall mechanisms of HOX gene expression and participate to the elaboration of the code referred to as "combinatorial HOX code". In this way, normal and alternative splicing of HOX pre-messengers [45,46] often results in two isoforms that putatively can antagonize each other [50,51]. In our experimental approach, we designed PCR/ sequencing primers so that we were able to verify the correct splicing acceptor and donor sites sequences of all exons for every gene investigated (including PBX1). No mutation was found in these sites.
PBX1 is one of the HOX genes' partners the most likely to be involved in the MRKH syndrome. Heterozygotic (+/-) inactivation of this gene does not provoke any congenital malformation in the mouse model whereas homozygotic (-/-) mice embryos die before birth due to multiple and severe malformations [27]. Therefore haploinsufficiency will probably not cause MRKH phenotype although mono-allelic mutations in a coding region of the gene may well lead to a dominant and deleterious effect such as titrating of HOX proteins clustered in non functional complexes. We carefully sequenced the overall exons of PBX1 in every patient and did not observe any mutation.

Conclusion
Investigation of candidate genes in biomedical research has often been unsuccessful unless target genes were obvious (for instance, see [52-54]). HOX genes, which play numerous roles during development, were good candidates for MRKH syndrome, based on deduction from their expression pattern during mouse development and from the phenotype of mice with a targeted disruption or overexpression of a specific HOX gene. Similar hypotheses were assumed for others congenital malformations or syndromes and revealed the involvement of these genes [55,56]. We based the present work on the investigation of MRKH patients showing various malformations associated with uterovaginal aplasia. This choice was based on the probable multigenic origins of the syndrome, assuming that at least one case would lead to evidence mutation of either a coding sequence of a HOX gene or part of the HOXA cluster (HOXA7 to -A13). Amongst the various MRKH cases analysed, we did not find any mutation in the coding sequences or in the [HCR-HOXA7] region. However, we did not sequence the whole HOXA cluster in every patient as this would have been a tremendous work but rather targeted genomic regions (coding sequences, splicing sites, regulatory sequences). Our negative results therefore do not mean that HOX genes are not involved in the syndrome. Additional investigation is necessary to settle or not the HOX hypothesis. This requires performing genetic linkage analysis of familial cases and wholegenome scan to seek for candidate chromosomal loci. -DG initiated the study in IP's group and has been leading this research program since then.