Efecto de la presión en el proceso de microfiltración tangencial

Rina M. Davison, Coiin J. Davis and Gerard S. Conway

Cobbold Laboratories, Division of Endocrinology, Department o f Medicine, University Coiiege London School of Medicine, London, UK

The X and Y chromosomes have taken very different paths in evolution. Starting as autosomes the Y chromosome shed 90% o f its genetic content while the X retains over 200 genes. H alf o f the Y chromosome genes have homologues on the X chromosome and are thought to be housekeeping genes. O f the 20 or so genes remaining on the Y, half have a clear male determining role as befits the male chromosome (Lahn & Page, 1997). To what degree then can we consider the X chromosome to be a ’female sex chromosome’ ?

The clinical models for studying ovary-determining genes are gonadal dysgenesis and premature ovarian failure (POF). Several pedigrees have been described in the literature with more than one affected member w ith ovarian dysgenesis or POF (Coulam et al., 1983; Mattison et a i , 1984; Portuondo et al., 1987; Aittomaki, 1994; Vegetti et al., 1998). Clinically, there are dividends in being able to predict POF prior to its onset in these families. A particular example involves a fam ily with POF in association with the karyotype 46X, Xdel(X)(q26) (Davison et al., 1998). The proband was found to share the deletion but managed to conceive before her early menopause. In this case the gross deletion was easily detectable. Thus a genetic marker for POF enables clinicians to predict whether a cytogenetically normal person was susceptible to POF. In this review we present various proposed candidate genes for ovarian failure on the X chromosome and discuss their relative merits (Fig. I).

around the time o f implantation. The choice o f X chromosome to be inactivated is random and once established the inactivity is clonally maintained. Several X-linked genes have been shown to escape X inactivation and the fact that an absent X chromosome is so deleterious to ovarian function, as compared to a largely inactivated X chromosome in normal females, suggests that two intact alleles are required for the normal function o f some genes on the X chromosome. Hence candidate genes for ovarian failure in Turner’ s syndrome are like ly to be those which escape X inactivation (Stratakis & Rennet, 1994).

Study o f naturally occurring defects o f the X chromosome provides further evidence o f an association between ovarian failure and deletions and translocations o f the X chromosome. It appears in general that abnormalities o f the long arm o f the X chromosome affect only ovarian function while those o f the short arm affect stature as well, resulting in a typical ’Turner’ phenotype (Sarto et a l , 1973; Therman et al., 1990). Sarto et al.

(1973) proposed the ’ critical region’ hypothesis whereby the region Xq21 — q25 must be intact for normal ovarian function. Since then the region has been refined to two loci; POF 1 at Xq26-q27 (Krauss et a l, 1987) and POF 2 at X q l3 -q 2 1 (PoweU et al., 1994). Deletions within this region as well as terminal deletions involving distal X q have been associated with varying degrees o f diminished reproductive capacity (Fitch e ta l., 1982; Veneman e ta l., 1991; Tharapel e ta l., 1993; Powell et al., 1994). Yet further detail was reported by Sala et al. (1997) who mapped 11 balanced X/autosome translocations associated with POF to a Yeast A rtificial chromosome (YAC) contig, spanning most o f Xq21 corresponding to a 15Mb region. A region o f this size was estimated to contain at least 8 different genes in Xq21 involved in ovary development, and interruption o f such genes could be the cause o f POF.

Evidence for X chromosome genes causing ovarian failure

Complete absence o f one X chromosome, as in Turner’ s syndrome, results in short stature, ovarian dysgenesis and primary amenorrhoea. In female mammals, dosage compensa­ tion o f X-linked genes between males and females occurs by inactivation o f one o f the two X chromosomes. Both the paternally and maternally inherited X chromosomes are active in embryos prior to implantation, and X inactivation occurs

Correspondence: D r R. M . Davison, Cobbold Laboratories, The Middlesex Hospital, M ortim er Street, London W IN 8AA, UK. E-mail: [email protected]

Candidate genes for premature ovarian failure on the X chromosome

Drosophila fat facets related X-iinked gene (DFFRX)

Jones et al. (1996) reported that an expressed sequence tag derived from human adult testis shares homology with the Drosophila fat facets (faf) gene, and related sequences are on both the human X and Y chromosomes. The human X-linked homologue was termed DFFRX and the corresponding Y- specific locus designated DFFRY. DFFRY maps to Y q ll.2 and is expressed in a wide range o f adult and embryonic tissues, including testis. Three azoospermie males have been found to harbour deletions removing the entire coding sequence o f

674 R. M. Davison et al. 22.3 22.1 21.3 21.1 11.4 11.3 11.2 11 12 13 21.1 21.2 21.3 22.1 22.3 23 24 25 26 27 28 ZFX DFFRX XIST AT2 DIA FSHPRH1 P0F2

ü

)

Fig. 1 Candidate genes for premature ovarian failure on the X chromosome.

DFFRY (Brown et al., 1998), confirm ing a role for DFFRY in spermatogenesis.

DFFRX maps to X p ll. 4 , escapes X-inactivation and is expressed in both human adult and embryonic tissues. In Drosophila, the/a/gene has been shown to be important in eye function and oocyte development. The location o f DFFRX on proxim al Xp coincides w ith the region for the major stigmata associated w ith Turner’ s syndrome, as defined by partial X chromosome deletions. This raises the possibility that DFFRX is a candidate fo r the defects o f oogonia proliferation and subsequent gonadal degeneration observed in Turner’ s syn­ drome. However a recent study by James et al. (1998) suggests that this is not the case. They tested 11 patients w ith breakpoints in proximal Xq fo r the presence o f one or two copies o f the DFFRX gene, and found that two patients w ith normal ovarian function had a single copy o f DFFRX. Therefore from these

small numbers it seems that haplo-insufficiency fo r DFFRX may not be responsible fo r the ovarian failure in Turner syndrome. The alternative possibility that these women possessed an occult 46,X X cell line w ithin the ovaries cannot be excluded.

Zinc finger protein Z F X

The Z FX gene is the homologue o f ZFY , which encodes a zinc finger protein form erly thought to represent a testis-determining factor (Page eta l., 1987,1990). Schneider Gadicke etal. (1989) showed that Z FX escapes X inactivation in humans, and Luoh

et al. (1995) provided evidence o f profound evolutionary conservation across species using comparative nucleotide sequencing o f human and mouse Z FX genes, and suggested a I

fundamental developmental role fo r this gene. A female ZFX knockout mouse was found to have a reduced number o f oocytes resulting in diminished fe rtility and shortened repro­ ductive lifespan m im icking POF in humans (Luoh et al., 1997). Male mutant mice also had fewer germ cells and both sexes were smaller and less viable. In a recent study by Avey and Conway (unpublished observations), mutation screening in 52 women w ith fa m ilia l and sporadic forms o f POF revealed only 3 sequence changes none o f which were predicted to affect translation. Hence whilst alterations o f the Z FX gene may contribute to ovarian failure in some women, they are unlikely to be im portant in the development o f an early menopause.

X-inactivation-specific transcript; X IS T

As previously mentioned in the case o f Turner’s syndrome, some genes required fo r ovarian function may be required in double dosage and hence escape X inactivation. It may be therefore, that an abnormality in the mechanism o f X inactivation causes POF. X IS T is a gene exclusively expressed from the inactive X, is located w ith in the X-inactivation centre at band X q l3 and is thought to be intricately involved in X inactivation (Brown et al., 1991). X IS T shows significant homology w ith Xist, the murine homologue that is located at the mouse X inactivation centre region and is also expressed from the inactive X chromosome. It has been shown that Xist knockout mice failed to inactivate an X chromosome (Penny et al., 1996). Panning & Jaenisch (1998) suggested that there are factors that firstly stabilize X IS T transcripts at the inactive X. then block the stabilization at the active X , as well as a mechanism that silences low-level X IS T expression from the active X, by demonstrating variable X IS T expression in embryonic stem cells.

Although it is commonly believed that the initiation o f X inactivation is random, there is significant variation in the proportion o f cells w ith either X inactive both in mice and 1999 Blackwell Science Ltd, Clinical Endocrinology, 51,673-679

The X chromosome and ovarian failure 675

among normal human females in the population. Famihes in which multiple females demonstrate extremely skewed inacti­ vation patterns that are otherwise quite rare in the general population are thought to reflect possible genetic influences on the X-inactivation process. Plenge e t al. (1997) reported a mutation in the XIST minimal promoter in 9 females from 2 unrelated famihes. A ll females demonstrated preferential inactivation o f the X chromosome carrying the mutation, suggesting that there is an association between alterations in the regulation o f XIST expression and X-chromosome inactivation. Mutations in human XIST might cause skew inactivation patterns resulting in haploinsufficiency o f vital ovarian developmental genes. Plenge e t al. (1997) screened a further 1666 independent unrelated X chromosomes and revealed only one more case o f this particular XIST promoter mutation ruling it out as a common polymorphism.

Angiotensin AT2 receptor

Angiotensin I I is a potent regulator o f cardiovascular haemostasis, whose action is mediated through the type 1 receptor A T I. The angiotensin I I type 2 (AT2) receptor is expressed abundantly in fetal tissues and decreases rapidly after birth (Daud e t al.,1988; Grady e t al.,1991). The mouse, rat and human AT2 receptor has been mapped to the X chromosome, the latter at Xq22 (Koike e t al., 1995).

A n ovarian role for the AT2 receptor was suggested by studies reporting that high levels o f angiotensin I I AT2 receptors were expressed in the granulosa cells o f rat atretic ovarian follicles, whereas only A T I receptors were present in other ovarian, structures (Tanaka et al., 1995). Further, stimulation o f AT2 receptors may contribute to the physiolo­ gical process o f atresia of the ovary and indeed it has been demonstrated that AT2 receptor induces apoptosis in several cell lines (Yamada e t al., 1995, 1996). Greater than 99 9% o f the follicles present at birth are destined to die by apoptosis during reproductive life. Accelerated ovarian follicular apop­ tosis may therefore cause POF (Hsueh e t al., 1994). Having cloned the human AT2 receptor gene, Katsuya e t al. (1997) searched for AT2 receptor mutations as a contributory factor to the early onset o f atresia in two POF families, but no changes were found in nucleotide sequences. Since only four subjects were examined for mutations, the possibility remains that AT2 receptor abnormalities eause POF in other women.

Diaphanous

Bione e t al. (1998) demonstrated that a balanced X ; 12 translocation, t(X; 12)q21; p i. 3) in a fam ily suffering from premature ovarian failure (Sala e t al., 1997) produced a breakpoint in the last intron o f the D IA gene. This gene is a

human homologue o f the d ro so p h ila gene diaphanous (dia),

mapped to Xq22 by Banfi e t al. (1997). D ia is ubiquitously expressed and conserved across species from yeast upwards. The protein encoded by the human D IA gene was the first member o f the FH1/FH2 fam ily o f proteins, which are involved in cytokinesis and other actin-mediated morphogenetic pro­ cesses that are required in early stages o f development. Mutant alleles o f drosophila d ia affect spermatogenesis and oogenesis and lead to sterility, with alteration in follicular cell division in the female. In humans mutations in D IA may interfere with mechanisms leading to follicle cell proliferation.

FSH primary response rat homologue 1 FSHPRH1

A mutation in the FSH receptor gene (FSHR), has been shown to cause hereditary hypergonadotrophic ovarian failure (A itto­ maki e t al.,1995). The histological appearance o f the ovaries o f women with this mutation showed hypoplasia with scant primordial follicles; none had the appearance o f complete ovarian dysgenesis with streak ovaries suggesting that early ovarian development is not FSH dependent (Aittomaki e t al.,

1996). Two subsequent screening studies o f women with both sporadic and fam ilia l POF have failed to identify any FSH receptor gene deletions or mutations and hence this is a very rare cause o f POF (Whitney e t al., 1995; Conway e t al., 1998a; Layman e t al., 1998). It may be however, that genes downstream o f the FSH receptor take part in ovarian development. One such FSH response gene is found on the X chromosome.

Roberts e t al. (1996) hypothesized that mutations in FSH response genes might be responsible for defect in female and male gonadal development. One such gene described in rats, leucine-rich primary response gene 1 (LR P R l) is transcription­ ally activated in response to FSH stimulation o f testicular Sertoli cells both in vitroand in vivo(Slegtenhorst-Eegdeman et al., 1995). Furthermore LR PR l mRNA is expressed in the ovary even before FSHR mRNA, suggesting that LR PR l may play an FSH independent role during ovarian development (Slegtenhorst-Eegdeman e t a l , 1998). Roberts e t al. (1996) characterized a human gene (FSHPRHl), which encodes a 756 amino acid polypeptide with a 72% identity to the rat LR PR l at the amino acid level. This gene maps to Xq22 which is adjacent to areas critieal for ovarian development and is therefore a potential candidate for human X-linked disorders o f gonadal function.

S 0X3

The mammalian genome contains a fam ily o f genes that are related to SRY (sex determining region Y), the putative testis determining gene. The homology is restricted to the region o f

676 R. M. Davison et al.

promoter _{4 AUG FMR1 gene}

B I AUG Normal 6-50 ► i AUG Premutation 50-200 ► D Full mutation 200 - 2000 I AUG

Fig. 2 A , promoter, 5' untranslated region in exon 1 with (A U G ) start of transcription, F M R l gene structure; B, normal trinucleotide repeat in exon 1; C, F R A X A premutation; D, F R A X A full mutation with methlyation o f the promoter and silencing of gene transcription.

SRY that encodes a D N A -binding m o tif o f the H M G -box class, and the various genes have been named SOX, fo r SRY related HM G-box. The S 0X 3 gene has been mapped to Xq26 — q27 (Stevanovic et al., 1993) by use o f a panel o f somatic cell hybrids. Foster & Graves (1994) identified a sequence on the marsupial X chromosome that shares homology w ith SRY and shows near-identity w ith the mouse and human S 0X 3 gene. They suggested that the highly conserved X chromosome- linked S 0X 3 represents the ancestral SOX gene from which the sex-determining SRY gene was derived. The close homology between SRY and S0X 3 m ight suggest that each is responsible fo r its respective gonadal development: SRY fo r the testis and S0X 3 fo r the ovary.

Analysis o f the distribution o f S0X 3 R N A shows that its main site o f expression is in the developing nervous system and the urogenital ridge where S0X3 protein products bind the same D N A sequence m o tif as SRY in vitro (Collignon et al., 1996). A deletion o f this gene was detected in a male patient with a contiguous gene syndrome o f haemophilia, mental retardation and primary testicular failure (Rousseau et a l, 1991), and since S 0X 3 is expressed in human fetal brain it is possible that its deletion causes mental retardation. S0X 3 deletions in females however, have yet to be reported. Screening o f 164 women with POF did not reveal any mutations in this gene (Conway 1999 unpublished observations).