| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Expression of human F8B, a gene nested within the coagulation factor VIII gene, produces multiple eye defects and developmental alterations in chimeric and transgenic mice
Introduction
Results
Production of chimeric mice expressing human F8B
Chimeric mice show growth retardation
Eye malformations also characterize chimeras
Attempts to establish a transgenic mouse line from chimeras
Construction of transgenic mouse lines
Growth was normal for transgenic lines
Transgenic lines showed eye malformations
Discussion
Human F8B produces developmental defects in mice
F8B encodes a factor VIII C2 domain, a motif implicated in cell adhesion
Materials And Methods
Construction of the recombinant vector
Generation of recombinant ES cell clones and chimeric mice
Detection of the neo and F8B transgenes
Monitoring of F8B transgene expression
Ultrasound scanning of chimeric embryos
Generation of transgenic mouse lines
Histological sections of eyes
Acknowledgements
References
Expression of human F8B, a gene nested within the coagulation factor VIII gene, produces multiple eye defects and developmental alterations in chimeric and transgenic mice
Received February 22, 1999; Revised and Accepted April 12, 1999
Factor VIII-associated gene B (F8B) is a small human gene of unknown function which is nested within the gene encoding coagulation foactor VIII (FVIII ) in chromosome band Xq28. The sequence of F8B includes the C2 cell adhesion motif of factor VIII, which has also been identified in numerous proteins known to play important roles during development. Here we have constructed both chimeric and transgenic mice expressing normal human F8B to investigate its possible developmental effects. The chimeras produced from embryonic stem cells transfected with normal F8B under control of a cytomegalovirus promoter and selected for neomycin resistance expressed readily detectable levels of F8B mRNA in multiple tissues. They showed growth retardation, microcephaly, reduced longevity and severe ocular defects, and although they were fertile, gave birth to no F8B heterozygous pups. Seven transgenic mouse lines, produced by injection of the transgene into fertilized oocytes, were viable and of normal size but expressed lower levels of F8B mRNA. Strikingly, they showed the same severe eye abnormalities as the chimeras. These defects included anterior segment dysgenesis, absent or abnormal lens, persistence of the primary vitreous, Harderian gland tumors and ectopic pigmented cells, suggesting that migration of neural crest cells might have been perturbed during eye development. In addition, dysplastic retinas and the absence of photoreceptors were observed, providing a mouse model for retinal degeneration.
INTRODUCTION
The F8B gene, denoted as factor VIII-associated gene B (OMIM 305424), is a small gene nested completely within the gene encoding human coagulation factor VIII (FVIII ) (1). The function of factor VIII, a trace plasma glycoprotein, is well known since it plays a major role in blood coagulation and produces hemophilia A, a recessive X-linked disorder, when deficient. The function of F8B, in contrast, remains unknown.
The structural organization of these genes, shown in Figure 1, indicates that the first exon of F8B is embedded within FVIII intron 22, and it is transcribed in the same direction as FVIII. This first `private' exon subsequently is spliced to exons 23-26 derived from FVIII, creating a final overlapping transcript of 2.5 kb with an open reading frame encoding 210 amino acids. F8B shares a common bidirectional promoter with a second gene nested completely within intron 22, denoted F8A (OMIM 305423), which is transcribed in the opposite direction to FVIII (1,2). The complexity of the relationship among these three genes is accentuated by the fact that F8A, the first exon of F8B and their common promoter are encompassed within a 9.5 kb interval, denoted as int22h, of which two additional homologous copies are present 300 and 400 kb upstream from human FVIII (2,3). In the mouse, only one homologous copy of the human int22hinterval is present and it is extragenic to the FVIII gene. No F8B gene is known to exist, therefore, and no transcript has been found in any murine tissue (1). The evolutionary origin of the F8B gene is relatively recent and has been proposed to result from the fortuitous transposition, in humans, of one extragenic copy of int22h into intron 22 of FVIII (2). In humans, F8B has been found to be transcribed in a number of tissues, notably brain, and also in cultured fibroblasts and in lymphoblastoid lines. The sequence of this RNA suggests that it is likely to be translated (1), and attempts have been made to identify a protein product by western blot analysis, in extracts from either human lymphocytes or the F8B-transfected mouse embryonic stem (ES) cells which we have produced here (1; and data not shown). Several different rabbit polyclonal antibodies directed against the C2 domain of factor VIII and antibodies directed against the peptide encoded by the first exon of F8B have not revealed a protein of the predicted size.
Figure 1. Organization of the FVIII, F8A and F8B genes and molecular analysis of chimeric and transgenic mice. (A) Orientation of F8B with respect to the human FVIII gene. The FVIII gene (186 kb) is shown on the top line as a shaded box and its direction of transcription is indicated by an arrow originating at a vertical bar representing its promoter. Within intron 22 is shown, as a black box, theint22h interval which spans 9.5 kb and which contains F8A and the first exon of the F8B gene. Two additional copies of int22hare located 300 and 400 kb upstream of FVIII. Below is represented the splicing of the F8B transcription product. The primary transcript of F8B includes the first exon, which is shown in black and located within FVIII intron 22, plus the following FVIII gene sequence. The first exon subsequently is spliced to FVIII exons 23-26 of this transcript, shown shaded, to create the complete 2.5 kb F8B mRNA. The white boxes correspond to the 5[prime] and 3[prime] non-coding sequences of the F8B mRNA. (B) Expression of the F8B transgene in four independent recombinant ES cell clones. RNA was extracted from four recombinant ES cell clones and F8B mRNA was detected by RT-PCR, as described. A positive control in lane 1 shows the expected human F8B band of 420 bp amplified from RNA of normal lymphocytes. This band was not present in the negative control shown in lane 2, which corresponds to RNA from non-recombinant ES cells and confirms the absence of endogenous F8B mRNA in mouse cells. The four clones are shown in lanes 3, 5, 7 and 9, and the band of 420 bp, characteristic for the F8B transgene, was present in each. Control reactions in which reverse transcriptase was omitted (lanes 4, 6, 8 and 10, respectively) show no band of 420 bp, indicating the absence of contaminating DNA in the RNA preparations. Size markers ([phis]X174/HaeIII digest) are shown in lane M. (C) Detection of the neo transgene in tissues of a chimeric male mouse used in genetic crosses. DNA was isolated from tissues and subjected to PCR as described. The tissues analysed were lung (Lu), liver (Li), testis (T), spleen (S), heart (H) and kidney (K), as indicated above each gel channel, and each showed a band characteristic for the neo transgene. A negative control in which DNA was omitted from the PCR is shown in lane 7 (T-). (D) Detection of F8B mRNA in eye tissues from three chimeric mice. RNA was isolated and RT-PCR carried out as described using 35 PCR cycles. Lane 1 shows the expected amplified band from human lymphocytes. An amplified band diagnostic for F8B mRNA is observed for each eye sample. A control for DNA contamination, in which the reverse transcriptase reaction was omitted, is shown for each sample in the adjacent lane. (E) Detection of F8B mRNA in thymus and spleen from a chimera that died prematurely. RNA was isolated and RT-PCR carried out as described using 35 PCR cycles. Lane 1 shows the control band from human lymphocytes and diagnostic bands are present for thymus and spleen samples. A control in which the reverse transcriptase reaction was omitted is shown for each sample in the adjacent lane. (F) Detection of F8B mRNA in eye tissues from four transgenic founder mice. RNA was isolated and RT-PCR carried out as described using 45 PCR cycles. Lanes 1, 3, 5 and 7 show the diagnostic band, whereas no band is seen in corresponding control samples in which the reverse transcriptase reaction was omitted (lanes 2, 4, 6 and 8).
Although the function of F8B is not known, its sequence suggests a potential role for the encoded protein. The eight amino acids issuing from the first exon show no homology with any known proteins, but the remaining amino acids correspond to the functionally important C2 domain of factor VIII. It is via this domain that factor VIII adheres to the platelet surface before complexing in situ with coagulation factors IX and X (4-6). The C2 domain of factor VIII shows a high degree of amino acid homology with similar motifs found in a wide range of proteins known to be developmentally important in mammals and other species (7-13). These proteins with `factor VIII-like domains' have been found both as soluble factors, as is factor VIII, and as cell surface components, as are, for example, discoidin and neuropilin (7-13). They have been proposed to mediate, by this domain, cell surface interactions such as cell migration and adhesion, as we will discuss later.
Until now, a large volume of genetic information, generally from hemophiliacs, has provided no clue to the possible function of F8B (14-17). Furthermore, known rearrangements of the FVIII gene in hemophiliacs have impliedthat complete deletion of F8B results in no readily detectable, associated clinical abnormality, and this observation has led to the general assumption that F8B may be dispensable (1,2,15).
From a different viewpoint, however, it is interesting to note that the FVIII and F8B genes are located in the most distal band on the long arm of the X chromosome, Xq28, and are normally subjected to dosage compensation (18). Rare clinical cases in which translocation or duplication has caused failure of X inactivation for Xq28 only, including three well-defined X-Y translocations, show characteristic symptoms which have been attributed to overexpression of genes in this band (19-21). These symptoms include growth retardation, microcephaly, hypotonia, mental retardation and, for some families, intra-uterine growth retardation, early death and retinopathy (21-23). Such observations led us to consider the possibility that F8B might reveal its significance to development not when it is deficient, as does FVIII, but under conditions in which it is functionally overexpressed.
To test this possibility, we have constructed chimeric mouse cohorts and seven independent transgenic mouse lines which express the human F8B gene under control of a cytomegalovirus (CMV) promoter. Here we describe developmental abnormalities which have been observed in these animals.
RESULTS
Production of chimeric mice expressing human F8B
Chimeric mice expressing the F8B gene were constructed, as described in Materials and Methods, using transfected ES cells. The recombinant vector contained normal human F8B cDNA under control of a CMV promoter (24). Four independent ES cell clones, which had been selected in G418, were found by our standard RNA extraction protocol and RT-PCR assay to express a readily detectable level of the F8B transgene, as shown by the amplified band in Figure 1B. All four clones were used successfully to generate mouse cohorts. Of a total of 68 pups born, 35 were considered to be chimeras due to mottled coat color. As shown in Figure 1C, the internal organs of a representative chimera were all found to be chimeric by PCR analysis of the associated neo transgene, and F8B expression was readily detectable in chimeric tissues using our standard protocol (Fig. 1D and E). The range of phenotypes displayed by chimeras was the same for each of the four ES cell clones.
Chimeric mice show growth retardation
The first major abnormality observed among the chimeras was growth retardation which was evident at birth, as shown in Figure 2A. Mice with low birth weight invariably developed chimeric coat color. Its severity was variable, however, since the 16 chimeras which were judged to be growth retarded showed weights, at 4 days of age, ranging from 1.1 to 2.8 g with an average weight of 2.0 0.22 g (n= 9). The average weight of their littermates at the same age was 3.6 0.08 g (n= 10). In addition, occasional litters included a small, dead pup which was not considered here.All chimeras were judged to show normal locomotion, and hypotonia was not evident. The neonatal growth retardation persisted as the mice aged, as seen in Figure 2B.
Figure 2. Morphological abnormalities of the chimeric mice. (A) External appearance of two littermates at 3 days of age. A pup of normal size and weight is shown on the right, and a small, underweight littermate is shown on the left. The growth-retarded pup later developed chimeric coat color. (B) External appearance of two chimeric littermates at 7 weeks of age. A mouse of chimeric coat color and normal weight is shown on the left. A littermate showing chimeric coat color, severe growth retardation and bilateral microphthalmy is shown on the right. (C and D) Thymus and spleen from a chimera which died prematurely. In (C), the small thymus of a chimera of normal weight which died prematurely is compared with a thymus of normal size dissected from a control C57BL/6 mouse of the same age. In (D), the small spleen of the chimera is compared with the spleen of normal size from the same control.
Most chimeras survived to adulthood, although seven small or moderately sized chimeras died prematurely of undetermined causes. Dissection revealed no abnormalities other than a spleen and thymus which were hypomorphic relative to other internal organs of the same mouse and to a normal control (Fig. 2C), and the transgene was found by RT-PCR analysis to be expressed in these organs (Fig. 1E).
Since we had thought, originally, that any expression of F8B might possibly lead to the death of all mouse embryos, we initiated, after blastocyst transfer, monitoring of intra-uterine embryonic development using high-frequency ultrasound scanning (25). Three gestational females that had received the blastocysts were scanned first on approximately day 10 of development and again 1-2 days before birth of the pups. On embryonic day 10, yolk sacs were found to be variable in diameter, as shown in Figure 3A and B. Within some of the smaller yolk sacs, the embryo showed no cardiac activity, indicating that it had died and was probably destined to be resorbed. One to two days before birth, the embryos varied in length. A small embryo (Fig. 3C) shows a cranio-caudal distance of 13 mm, while an embryo of normal size (Fig. 3D) shows a length of 19 mm. In these same embryos, the biparietal distance of the heads was approximately proportional to body length, being 3.88 and 5.50 mm, respectively, indicating microcephaly of the smaller embryo (Fig. 3E and F). Again, the absence of cardiac activity in some of the smallest embryos indicated that they, too, had died in utero after day 10. The large number of embryos per uterus and the relatively high mortality rate usually associated with any injected blastocysts prevented us from determining statistically from these scans whether the transgene had brought about an increase in embryonic lethality. We also chose not to sacrifice any of the gestational females and, thus, were unable to determine by PCR analysis of fetal DNA whether the presence of the transgene increased the probability that an embryo might be among those which died in utero. The small, growth-retarded embryos observed in utero1-2 days before birth, however, in all probability correspond to the small chimeric pups born to these mothers.
Figure 3. Intra-uterine ultrasound scanning of chimeric embryos. (A and B) Yolk sacs on approximately day 10 of development. In these sonograms, yolk sacs appear black and are indicated by black arrows. The embryo inside each sac appears white. The three yolk sacs in (A) are 5.26, 3.42 and 9.45 mm in diameter. The diameters (dotted lines) of the two yolk sacs shown in (B) are 7.26 and 3.88 mm. (C and D) Embryos 2 days before birth. The cranio-caudal length of each embryo is indicated by a dotted white line with white arrowheads. For the small embryo shown in (C) it is 13.3 mm and for the embryo of normal size shown in (D) it is 19.0 mm. (E and F) Heads of the same embryos 2 days before birth. The biparietal distance of the head was measured at the points indicated by arrowheads. The length was 3.88 mm for the small embryo shown in (C) and 5.50 mm for the embryo of normal size shown in (D).
Eye malformations also characterize chimeras
The second phenotype which was apparent externally in the chimeras was eye abnormalities. Approximately 35% of the chimeras either showed microphthalmia with or without corneal opacities from birth or developed corneal opacities by 10 months of age. Only in the most severe cases were the defects bilateral. Several chimeras, in fact, showed one severely affected and one normal eye, as shown in Figure 4B. Ten chimeras, including four with no external abnormalities, were killed for histological ocular examination at 4-8 months of age. RNA was extracted from the contralateral eye in most cases and tested by RT-PCR for expression of F8B, and in all cases the mRNA was present (Fig. 1D). In all 10 chimeras, eye abnormalities were found histologically even when they were not apparent externally.
Figure 4. Comparison of the left and right eyes from a chimera which showed abnormal histology only in the left eye. (A and C) Equatorial sections through the optic nerve from the right and left eyes, respectively, showing a normal lens in the right eye and the total absence of lens in the left eye. (B) External appearance of the two eyes of a chimera showing central corneal opacity of the left eye. (D), Normal cornea of the right eye. (E and F) Abnormal cornea of the left eye showing infiltration by pigmented cells and persistence of the corneal lenticular membrane from the cornea to the retina (indicated by a star). (G) Normal retina of the right eye. (H and I) High magnification of the retina, shown in (C), from the left eye. Folds are present in the retina, and abnormal pigmented deposits, indicated by arrows, can be seen within the folds. ah, aqueous humor; c, cornea; ce, corneal epithelium; cs, corneal stroma; gcl, ganglion cell layer; i, iris; inl, inner nuclear layer; l, lens; onl, outer nuclear layer; r, retina; rpe, retinal pigmented epithelium; v, vitreous. Scale bars: (A and C) 500 µm; (D) 20 µm; (E and F) 200 µm; (G, H and I) 50 µm.
Although quite variable in manifestation, the phenotype consistently included specific malformations, which are illustrated in Figures 4 and 5. Retinas exhibited dysplasia and/or degeneration. The dysplasia was indicated by disorganization of the neural retina with rosettes formed by abnormal infolding of the retina layers. In some folds, abnormal pigmented deposits were apparent (Fig. 4H and I). Retinal degeneration was indicated by reduced thickness of some portions of the neural retina (Fig. 5H). The patches of neural retinal atrophy involved the outer and inner nuclear layers and, in one chimera, the ganglion cell layer (Fig. 5I). Lens malformations varied in severity from severe cataract up to the total absence of lens (aphakia), as shown in Figures 4C and 5A. In one chimera, a rupture of the lens capsule was evident, leading to extravasation of lens material (Fig. 5C). Anterior segment defects were also observed and included reduction of the anterior chamber, thickening of the corneal stroma containing vessels and mastocysts (Fig. 5F), adhesion between corneal endothelium and iris (Fig. 5E), and folding and rosette formation of the corneal epithelium (Fig. 5D-F). Strikingly, pigmented cells were observed in the anterior chamber (Fig. 5A-D), in the vitreous cavity (Fig. 5A and B), in the corneal stroma (Fig. 4E) and, in one case, inside the lens (Fig. 5D). In the eye in which the lens was absent, these pigmented cells had infiltrated the corneal stroma and had migrated through the vitreous up to the retina, suggesting a persistence of the corneal-lenticular membrane (indicated by a star in Fig. 4E and F).
Figure 5. Abnormal histology found in eyes from three chimeric mice. (A) Equatorial section of a microphthalmic and disorganized eye showing a dysplastic retina with rosettes (indicated by a star), the presence of pigmented cells and cataract of the lens with extravasation of lens material. (B and C) Enlargements of the boxed regions in (A) showing pigmented cells and lens material outside the lens capsule, respectively. (D) Folding of the hyperplasic epithelium of the cornea and the presence of pigmented cells in the stroma and lens (indicated by a star), confined to an area opposite the abnormal cornea. (E) At the level of the pupilary hole, the iris adheres abnormally to the corneal endothelium, in contrast to that seen in a normal eye (A) where the two structures are separated. (F) Mastocysts (thick arrows) and capillaries (thin arrows) inside the corneal stroma, and the presence of a rosette of the corneal epithelium (indicated by a star). (G) A fold in the neural retina. (H) Disorganization and reduced thickness of three neuronal layers of the retina. (I) Neural retinal degeneration involving predominantly the ganglion cell layer. Abbreviations as in Figure 4; lc, lens capsule. Scale bars: (A) 200 µm; (B, C, H and I) 30 µm; (D and E) 100 µm; (F and G) 50 µm.
Among all the chimeras, the most constant feature was the presence of pigmented cells which had invaded the interior of the eye. Such cells could have been derived from either the neural crest cells (NCCs) or the neuroectoderm (26-28). During normal eye development, at 8.5 days post-coitum (d.p.c.), NCCs colonize the perioptic mesenchyme around the optic vesicule and then migrate, at 12.5 d.p.c., in three waves anteriorly and inside the eye, to form the corneal endothelium, corneal stroma, iris stroma and the primary vitreous (retrolenticular membrane), which normally disappears by 14.0 d.p.c. The neuroectoderm, in contrast, gives rise to the retinal pigmented epithelium, a structure which did not appear grossly abnormal in these chimeras. Since chimeras exhibited mainly defects in structures derived from neural crest periocular mesenchymal cells, we speculate that the pigmented cells were derived from NCCs. Their abnormal location inside the eye was correlated with various eye defects, and it appears likely that normal eye development had been disturbed by their ectopic location. The results suggest, therefore, that expression of F8B might first bring about the abnormal migration, into the interior of the eye, of pigmented cells derived from NCCs and that this abnormality initiates a cascade of secondary defects involving the anterior chamber, cornea, iris and lens.
Attempts to establish a transgenic mouse line from chimeras
We attempted to produce transgenic lines from the chimeras using usual protocols, and seven chimeric males gave rise to progeny when crossed with C57BL/6 females. Four of these males, derived from two different ES cell clones, proved to be germ cell chimeras, transmitting to their offspring the agouti coat color characteristic of strain 129 mice, from which the ES cells were derived. Of a total of ~330 pups born to these four chimeras, 89 were agouti and all of these were phenotypically normal. PCR analysis of tail biopsy DNA indicated that the F8B transgene was not present in any of the 89 pups while, statistically, 50% of all agouti progeny were expected to inherit the transgene. Subsequent PCR analysis of DNA extracted from tissues of one of the four chimeric fathers confirmed that F8B had been present in all tissues tested, including the testis (Fig. 1C). The fact that F8B was not found in any viable offspring might be explained in either of two ways. First, the relatively elevated expression level of F8B, characteristic of the ES cell clones and chimeric tissues, might have been detrimental to individual spermatozoa that carried the F8B gene. This possibility appears unlikely, however, since male germ cells develop in syncytia after meiosis and since later, in mature, individual spermatozoa, transcriptional activity has largely ceased (29). Alternatively, a relatively high level of F8B expression, if uniformly present in an embryo, might, in fact, have brought about its death.
Construction of transgenic mouse lines
In parallel, we successfully produced transgenic mouse lines expressing human F8B by injecting, into fertilized oocytes, a 2.2 kb fragment excised from the recombinant vector. Of a total of 57 mice born to foster mothers, 20 had incorporated the F8B transgene and were used as founders to establish transgenic lines. Since eye tissues had displayed a relatively high level of transgene expression as estimated from RT-PCR analysis of tissues of the chimeras (Fig. 1D), expression of F8B mRNA in each line was tested using total ocular RNA. Seven of the transgenic lines were found to express F8B (Fig. 1F) but, in all cases, the level of expression was remarkably reduced when compared with that previously assayed in ES cells and chimeras (Fig. 1B-D). In order to detect F8B expression consistently in the transgenics, we routinely used 10 additional PCR cycles and occasionally performed Southern hybridization to confirm the identity of the amplified band. Since chimeric tissues are composed of both normal and transgenic cells, the difference in the cellular expression level of the transgene between chimeras and transgenics theoretically might have been even greater than that estimated from these RT-PCR assays. Each of the seven lines was found to have integrated 5-20 copies of the F8B transgene, each at one site which was different for each line. The expression levels of the transgene did not appear to increase with the number of integrated copies (data not shown), and it has been found previously that a correlation between gene expression level and the number of integrated copies is quite difficult to predict (30-32).
Growth was normal for transgenic lines
In contrast to chimeras, no FI transgenics, or founders, from any of the seven lines showed any significant difference in weight from non-transgenic littermates. Among the later generations, the number of mice per litter was normal, and only very rarely were small, moribond pups or premature deaths observed. All major organs were of normal size.
Transgenic lines showed eye malformations
On the other hand, all mice tested to date (26 mice) from each of the seven transgenic lines have shown eye abnormalities strikingly similar to those of the chimeras. The externally apparent symptoms included progressive corneal opacity with tumors of the Harderian lacrymal gland which caused bulging of the eye, as seen in Figure 6B. In fact, the time of onset of these external symptoms was at 6 months in the founder mice but decreased to only 3 weeks of age in mice from later generations. As with the chimeras, eye abnormalities were found histologically even when no external abnormality was apparent. The two eyes could be affected to very different degrees, as shown in Figure 6C. Histological sections, illustrated in Figure 6, show dysgenesis of the anterior segment including reduction or absence of the anterior chamber with adherence between the iris and the corneal endothelium, and neovascularization of the corneal stroma with the presence of mastocysts, aphakia, retinal dysplasia and retinal degeneration. Retinal degeneration was even more severe in transgenics than in chimeras, since a total absence of photoreceptors was found in three lines as early as 15 days of age (Fig. 6A-H). Also, interruption of the retinal pigmented epithelium was clearly evident in some cases (data not shown). The presence of a retrolenticular membrane due to persistence of the primary vitreous was observed in four lines. The choroid was normal in some transgenics, but in others it showed either local agenesis or a marked thickening. Pigmented cells were also observed to have invaded the interior of the eye, as shown in Figure 6C-F, and were observed in the lens or in some folds of dysplasic retinas (Fig. 6D-I).
Figure 6. Abnormal histology found in eyes from transgenic mice. (A) Equatorial section showing a dysplastic retina with the absence of the outer nuclear layer, which normally contains the photoreceptors. A tumor of the Harderian lacrymal gland is present behind the eye. (B) External view of a tumor of the Harderian lacrymal gland. (C) Equatorial section showing structural disorganization and the total absence of lens (aphakia). (D) Abnormal lens showing infiltration by pigmented cells, indicated by arrowheads, and reduction of the anterior chamber. (E) Abnormal cornea with a local thickening of the corneal epithelium with a probable rupture of the Bowman's membrane (thin arrows). The abnormal presence of capillaries in the corneal stroma is indicated by arrowheads. (F) High magnification of the anterior segment of the eye, seen in (C), showing infiltration by pigmented cells in the abnormal cornea. (G) Retina from an eye of an F1 transgenic mouse showing rosettes and folds. (H) Retina from an eye of an F2 transgenic offspring from the F1 transgenic mouse seen in (G), showing the total absence of the outer nuclear layer, which normally contains photoreceptors, and local interruption of the retinal pigmented epithelium. (I) Abnormal pigmented deposits, indicated by an arrow, present within the folds of the retina. Abbreviations as in Figure 4; ch, choroid; s, sclera; t, tumor. Scale bars: (A) 430 µm; (C) 333 µm; (D) 30 µm; (E) 55 µm.
These histological observations for transgenics support our previous hypothesis from the chimeras that aberrant migration of NCCs might have disrupted normal eye development. That the Harderian lacrymal gland, choroid and primary vitreous were also found to be affected reinforces this idea, since these additional structures are also colonized during embryogenesis by NCCs. The observed early retinal degeneration remains difficult to explain with this hypothesis alone, and we imagine that additional factors may be important in causing this severe, complex abnormality. In this respect, it will be of interest to use these transgenic mouse lines to study the genesis of the abnormalities during development.
DISCUSSION
Human F8B produces developmental defects in mice
Developmental abnormalities have been observed in chimeric mice expressing readily detectable levels of human F8B in multiple tissues, and these include intra-uterine growth retardation with microcephaly, postnatal growth failure, premature death and severe ocular malformations. Although the chimeras were fertile, none produced pups carrying the F8B transgene, suggesting that a relatively high expression level of F8B in all embryonic tissues might possibly lead to death of such heterozygous embryos. Seven independent lines of transgenic mice expressing F8B were generated by transgene injection into fertilized oocytes, but in all cases the level of F8B expression was apparently much lower than that assayed in chimeras. F8B mRNA was detected consistently in ocular tissues, and ocular defects were the only abnormalities which were readily evident in the transgenics. These results indicate that expression of F8B in mice is responsible for causing developmental defects and suggest that the level of F8B expression may play a role in the manifestation of the defects.
The observation that expression of human F8B produces deleterious developmental errors in mice raises the possibility that overexpression of F8B in humans might also cause developmental abnormalities. It remains problematic, however, for several reasons, to draw parallels between genetically altered mice and human patients. It is not uncommon for artificially generated mouse models to display phenotypes that do not reproduce exactly the human disease state and in some instances differ considerably(33,34). Here specifically, variations between species may play a role, since mice produce no endogenous F8B mRNA. The CMV promoter that was used here has been shown previously to direct abundant, widespread expression of transgenes, but with some tissue-specific variation (24). Such variations introduced by the promotermight explain the predominance of some defects, such as those in ocular tissues where expression from this promoter has been reported to be high, and, conversely, mask other possible abnormalities in tissues where low expression from the promoter has been reported (20). Differences in the temporal, spatial and quantitative expression of F8B in conjunction with genetic background might also influence the clinical manifestations (1).
It remains interesting, however, to review the rare Xq28-linked human disorders for which a gene dosage effect has been proposed. Chromosome duplications involving the distal long arm of the chromosome X, including Xq28, and failure of inactivation of this region have been observed in at least two independent families with intra-uterine growth retardation, microcephaly, postnatal growth failure, generalized hypotonia, rare or single seizures and mental retardation, with retinopathy in one family and miscarriage and premature death of an infant in the second family (22,23). Similar symptoms have also been observed in at least three unrelated boys with XYXq syndrome, a condition defined as disomy of the terminal portion of band Xq28, a region of 5 Mb which includes the FVIII gene (21). These phenotypes are intriguing, since they represent developmental defects and since several of the symptoms resemble the defects which we were able to monitor in mice expressing human F8B.
F8B encodes a factor VIII C2 domain, a motif implicated in cell adhesion
How expression of the human F8B gene might produce such a range of developmental defects in the mice is not yet known. Since F8B includes a factor VIII C2 domain, a motif implicated in cell adhesion, we speculate that the elusive F8B protein might interfere, when functionally overproduced, with a large number of different cell-cell interactions during development. An instructive example might be the neuropilins, a family of neuronal cell surface receptors (13,35,36). The extracellular portion of this protein is composed of three distinct domains, each of which shares homology with a motif found in serum proteins involved in cell surface interactions, one of these being a doubled factor VIII C2-like domain. Overexpression of neuropilin in chimeric mice has been shown to lead to severe deformities and embryonic lethality(37). It has been proposed that each of the three domains of neuropilin interacts with a different partner to constitute, on the cell surface, an active multiprotein complex, as has been found for factor VIII on the platelet surface (38). Expression or overexpression of the F8B protein, with its factor VIII C2 domain, might be imagined to participate in the assembly of this type of multiprotein complex, either by binding directly or by sequestering proteins normally destined to bind. At least one specific example drawn from the results presented here is consistent with this interpretation. Certain eye malformations found in the chimeras and in the transgenic mice suggest that the migration route of pigmented cells derived from the NCCs which colonize the mesenchyme surrounding the optic vesicule might have been perturbed during development. Interestingly, a high expression level of neuropilin-2 of mouse has been observed at 9.5 d.p.c. in the same region (39). We propose that expression of F8B in mice might disrupt the function of certain proteins which employ a C2-like domain and which are implicated in specific cell-cell interactions during embryogenesis.
In conclusion, human F8B is a nested gene whose relatively recent origin and normal function remain, as yet, unexplored. Results obtained here indicate that its expression in mice is responsible for causing severe developmental defects, including ocular malformations, a deleterious effect heretofore unsuspected for this gene.
MATERIALS AND METHODS
Construction of the recombinant vector
Human RNA was extracted from lymphocytes of a normal male using RNA Now (Biogentex, Montigny-le-Baux, France). Total RNA (1-2 µg) was subjected to reverse transcription using AMV reverse transcriptase with random hexamer primers, and then 20% of the reverse transcriptase reaction was amplified through 35 PCR cycles with 20mer primers, F8B1 and F8B2. Primer F8B1 began at nucleotide 22 of the F8B cDNA sequence (GenBank accession no. M90707) within the first exon, 19 bp 5[prime] from the protein initiation codon. The F8B2 primer was located within FVIII exon 26 and began at nucleotide 1019 of the F8B cDNA sequence, 328 bp 3[prime] from the protein termination codon. The primers included KpnI and ApaI restriction sites, respectively, at their 5[prime] ends. The sequence of the F8B1 primer was 5[prime]-CGCGGTACCCAAAATCGAGGGTCTCGGGGA-3[prime]. The sequence of the F8B2 primer was 5[prime]-CGCGGGCCCGAGGAAGTGGTGACTCTGAG-3[prime]. The length of the resulting PCR fragment was 1016 bp, and its nucleotide sequence was confirmed by direct sequencing. It was cloned into the pCR3 vector (Invitrogen, Groningen, The Netherlands) using KpnI and ApaIsites of the polylinker. The CMV promoter of the vector is located 5[prime] to the insert, and the bovine growth hormone polyadenylation signal is located 3[prime] to the insert. The vector included the neo selection marker.
Generation of recombinant ES cell clones and chimeric mice
ES cell culture and mouse embryo manipulations were carried out as described (40,41). The non-linearized recombinant vector was electroporated into ES cells (line D3) and clones resistant to 250 µM G418 (Gibco BRL, Cergy-Pontoise, France) were collected. Four different ES cell clones expressing F8B mRNA were identified by RT-PCR. Southern blot analysis of DNA from each of these clones revealed that each had integrated one copy of the transgene, each at a different site (42). Cells from each of the four clones were microinjected into C57BL/6 blastocysts (10-15 cells/blastocyst), and the blastocysts were then transferred to pseudopregnant females (~10 blastocysts/female).
Detection of the neo and F8B transgenes
DNA was prepared by incubating tail biopsies overnight with proteinase K (43). The supernatant was then phenol-chloroform extracted, and the DNA was ethanol precipitated and redissolved in TE (10 mM Tris, pH 7.8, 1 mM EDTA) at 0.5 mg/ml. To detect neo, a DNA aliquot of 1 µg was subjected to 35 PCR cycles using primers 5[prime]-GTGTTCCGGCTGTCAGCGCA-3[prime] and 5[prime]-GTCCTGAGCGGTCCGCCA-3[prime]. To detect the F8B transgene, 35 PCR cycles were run using primers F8B6 at nucleotide 28: 5[prime]-CAAATCGAGGGTCTCGGGGA-3[prime] and F8B7 at nucleotide 190: 5[prime]-AATCACAGCCCATCAACTCC-3[prime], giving an amplified fragment of 170 bp.
Monitoring of F8B transgene expression
Total RNA was isolated from either recombinant ES cell clones or tissues of chimeric mice using RNA Now (Biogentex). The RNA was treated with RQ1 DNase I (Boehringer Mannheim, Meylan, France), extracted with phenol-chloroform and precipitated with ethanol. DNA-free RNA (1-2 µg) was subjected to reverse transcription using AMV reverse transcriptase (Gibco BRL) with random hexamer primers, and then 20% of the reverse transcriptase reaction was amplified through 35 PCR cycles with 20mer primers, F8B55 and F8B470. F8B55 was located at nucleotide 55 in the first exon of the F8B cDNA sequence, while F8B470 was located at nucleotide 470 in FVIII exon 25, giving an amplified F8B fragment of 420 bp. For the transgenic mice, the reverse transcriptase reaction was submitted to 45 PCR cycles with primers F8B6 (see above) and F8B765: 5[prime]-AGCACAAAGGTAGAAGGCAA-3[prime], giving an amplified band of 763 bp. PCR products were visualized after electrophoresis on a 2% agarose gel containing ethidium bromide. Their identity was confirmed by Southern blot of the gel and hybridization with primer F8B7 (see above).
Ultrasound scanning of chimeric embryos
Intra-uterine scanning of mouse embryos was performed using a 20 MHz annular probe with a theoretical axial and lateral resolution of 20 and 100 µm, respectively (Esaote, Milan, Italy) which is used principally for monitoring skin tumors (25). Appropriate images including length measurements were frozen on the monitor screen and printed directly. Three gestational females were scanned on day 10 after blastocyst transfer, which corresponds approximately to day 10 of embryogenesis. The same females were scanned a second time on day 17 after blastocyst transfer, 1-2 days before the pups were born.
Generation of transgenic mouse lines
The recombinant vector (see above) was digested by AflIII and DraIII at nucleotides 1024 and 5056, respectively, producing a linearized fragment of 2.2 kb which included the CMV promoter, human F8B cDNA and the bovine growth hormone polyadenylation signal. It was purified from an agarose gel by electroelution, dissolved in 10 mM Tris-HCl, pH 7.4, 0.1 mM EDTA at 5 µg/ml, and microinjected into male pronuclei of fertilized oocytes obtained by breeding (C57BL/6×CBA)F1 mice as described (41). The injected oocytes were then transferred to pseudopregnant females (~20 oocytes/female). Progeny were screened for the presence of the F8B transgene by PCR analysis of tail biopsy DNA using primers F8B6 and F8B7 (see above). F1 founder mice were first crossed to strain C57BL/6, and F8B transgene expression was monitored in progeny by RT-PCR analysis of tissue extracts (see above). Seven lines expressing the transgene were established by brother-sister matings.
Histological sections of eyes
Eyes were dissected from mice, fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 4 h, washed with PBS, dehydrated and embedded in Historesin (Leitz, Reuil-Malmaison, France). Sections of 3 mm were cut with an Autocut RM 2055 Jung microtome (Leitz). They were mounted on gelatin-coated slides and stained with 0.05% toluidine blue (Serva, Paris, France). Photographs were made with an Aristophan microscope (Leitz).
ACKNOWLEDGEMENTS
We wish to thank Dr J. Jami for valuable suggestions and the support of his laboratory throughout the course of this work. We thank Professors J.C. Kaplan and A. Kahn for critically reading the manuscript. We acknowledge also the help of Dr E. Julian and S. De Thomasseau for histological sections, and H. Coët for photographs. Financial support was provided by the French Ministry of Public Health (Programme Hospitalier de Recherche Clinique no. AOB94050). S.V. received grants from the French Association for Hemophilia, INSERM and the Bayer Foundation.
REFERENCES
*To whom correspondence should be addressed. Tel: +33 1 42 34 19 20; Fax: +33 1 44 41 15 22; Email: delpech{at}icgm.cochin.inserm.fr
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  W. Wang, J. Zhang, C. Alvarez, A. Llopart, and M. Long The Origin of the Jingwei Gene and the Complex Modular Structure of Its Parental Gene, Yellow Emperor, in Drosophila melanogaster Mol. Biol. Evol., September 1, 2000; 17(9): 1294 - 1301. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R G F Gray, M A Preece, S H Green, W Whitehouse, J Winer, and A Green Inborn errors of metabolism as a cause of neurological disease in adults: an approach to investigation J. Neurol. Neurosurg. Psychiatry, July 1, 2000; 69(1): 5 - 12. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||