Transgenic mice carrying the diphtheria toxin A gene driven by mouse γ2-crystallin promoter sequences manifest microphthalmia due to ablation of fiber cells in the ocular lens. Here we map ablation events in the lens by crossing animals hemizygous for the ablation construct with transgenic mice homozygous for the in situ lacT reporter gene driven by identical γ2 crystallin promoter sequences. By comparing the spatial distribution of tacZ-expressing cells and the profile of γ-crystallin gene expression in the lenses of normal and microphthalmic offspring, the contributions of specific cell types to lens development were examined. The results suggest that phenotypically and developmentally distinct populations of lens fiber cells are able to contribute to the lens nucleus during organogenesis. We also show that dosage of the transgene and its site of integration influence the extent of ablation. In those mice homozygous for the transgene and completely lacking cells of the lens lineage, we show that the sclera, cornea, and ciliary epithelium are reduced in size but, otherwise, reasonably well formed. In contrast, the anterior chamber, iris, and vitreous body are not discernible while the sensory retina is highly convoluted and extensively fills the vitreous chamber.

Two contrasting strategies control and determine cell fate during development. One strategy depends on extrinsic interactions of a cell with its local environment while the other determines cell fate intrinsically through inheritance of a lineage-specific genetic program. Assessing the relative contributions of each of these strategies to organogenesis and other embryological processes represents a fundamental challenge to the molecular understanding of development.

The ability to manipulate the mammalian genome by the generation of transgenic animals provides an experimental approach for genetically altering developmental processes and studying the molecular mechanisms governing the regulation of gene expression during ontogeny. In this regard, studies employing reporter genes in transgenic mice have provided considerable insight into genetic regulatory elements utilized by specific cell types during embryogenesis. However, such studies have not addressed the mechanisms governing cell fate, particularly the role cell–cell interactions ultimately play in contributing to this process. The ability to ablate selectively specific cell types during development provides an experimental strategy for perturbing these processes and investigating the complex mechanisms that influence organogenesis. One method whereby this can be achieved involves the targetted expression of a lethal protein to specific cell types by appropriate transcriptional regulatory elements (Breitman et al. 1987; Palmiter et al. 1987; Behringer et al. 1988). In an earlier study, we used this approach, termed genetic ablation, to generate transgenic mice with microphthalmia in which a subset of cells in the ocular lens were ablated as the result of programmed expression of the diphtheria toxin A (DTA) gene by γ2-crystallin promoter/enhancer sequences (Breitman et al. 1987). The mammalian lens is a biconvex structure consisting of a mitotically active layer of anterior epithelial cells surrounding a highly compacted population of terminally differentiated fiber cells. The latter cells can be divided into two distinct morphogenetic groups: primary fiber cells, which elongate from posterior cells of the lens vesicle to form the lens nucleus, and secondary fiber cells, which differentiate continuously from the anterior lens epithelium and are deposited in successive layers about the lens nucleus, forming the lens cortex. These developmental features of the lens in conjunction with differential regulation of the crystallin genes (Piatigorsky, 1984; McAvoy, 1978) result in a precise spatial distribution of different crystallin species across the lens.

In the present study, we have analysed microphthalmic animals carrying the y2DT-A transgene to determine how the ablation of specific lens cell types affects the process of organogenesis. We have also generated and characterized transgenic mice completely lacking cells of the lens lineage. The results obtained yield new insight into the role the lens plays in eye morphogenesis and the relative contributions that intrinsic and extrinsic factors play in determining cell phenotype with respect to crystallin gene expression in the lens.

Fusion gene constructs and production of transgenic mice

The construction of the γ2lacZ and γ2DT-A hybrid genes has been described previously (Goring et al. 1987; Breitman et al. 1987), as has the transgenic animals carrying these constructs.

Genotyping of transgenic animals

DNA was extracted from tail biopsies using the proteinase K/SDS method (Hogan et al. 1986). The presence of the transgenes was detected by Southern blot analysis (Southern, 1975) with probes labeled by either nick translation (Rigby et al. 1977) or by the random-priming procedure (Feinberg & Vogelstein, 1983).

To identify animals transgenic for the γ2DT-A construct, a 2·1 kbp BamHI–EcoRI probe that includes the DT-A coding region and SV40 splicing and polyadenylation sequences was used (Breitman et al. 1987). To identify homozygous y2DT-A2DT-A transgenic animals, a larger 2·98 kbp Xbal-EcoRI probe was used that includes, in addition to the sequences contained with the 2·1 kbp fiamHI-EcoRI fragment described above, the y2 sequences located at –759 to +45 relative to the transcriptional start site of the γ2 crystallin gene. The inclusion of these y2 sequences in the probe allowed standardization of the transgene copy number relative to the endogenous γ2 gene detected with this probe.

RNA analysis

Total RNA was extracted from the eyes of 3-week-old normal and transgenic mice hemizygous or homozygous for the γ2DT-A transgene. RNA was electrophoresed through a 1·5% formaldehyde–agarose gel and transferred to a Zeta bind filter. Conditions for prehybridization and hybridization were as described previously (Breitman et al. 1987).

The probes used for the Northern blot analysis included an oA and γ2-crystallin cDNA, obtained from J. Piatigorsky. For the detection of specific γ-crystallin transcripts, γ-crystallin cDNA subprobes were used (Murer-Orlando et al. 1987). These included a 166bp XhoII–Stul fragment of γ2, a 275 bp Xholl–Xbal fragment of γ3, and a 139 bp HinfI-BgI II fragment of γt.

Staining for lacZ expression

Whole eyes were frozen in liquid nitrogen and mounted in OCT compound (Fisher) for cryosectioning. Sections (7 μm) were cut, air-dried, and incubated overnight in a buffer containing X-gal (2 mm-5-bromo-4-chloro-3-indolyl-β-D-galactoside and 3 mm-potassium ferrocyanide in phosphate-buffered saline, pH7·0. Sections were then counterstained with hematoxylin and eosin and mounted in Canada balsam.

Histologic studies

After killing mice with a lethal dose of intraperitoneal phenobarbital the eyes were enucleated and fixed in Zenker’s fixative for 4h. They were then transferred to 80% ethanol for 18 h before being dehydrated through graded alcohols and embedded in paraplast according to standard procedures. Sections (6μm thick) were stained with hematoxylin and eosin.

In situ mapping of lens ablation events

Development of the mouse eye lens is accompanied by the expression of three major families of crystallin genes, α, β, and γ, that are differentially regulated during development (Piatigorsky, 1984; McAvoy, 1978). The γ-crystallins, which are themselves differentially regulated (Murer-Orlando et al. 1987; Van Leen et al. 1987b), are encoded by six closely linked genes (Quinlan et al. 1987) that are expressed exclusively in the terminally differentiated lens fiber cells. To determine precisely which cells in the lens are ablated by the γ2DT-A transgene, we carried out comparative in situ studies on the lenses of wild-type and transgenic microphthalmic littermates. Because the mouse γ-crystallins are encoded by 6 highly related genes, conventional in situ techniques could not be used to analyze the expression of a single member of the family. Therefore, we made use of a line of transgenic indicator mice expressing the lacZ reporter gene driven by the identical γ2 promoter sequences that were used to drive expression of the DT-A gene in the microphthalmic animals (Goring et al. 1987). By crossing mice homozygous for the lacZ reporter gene with animals hemizygous for the γ2DT-A ablation construct, we generated microphthalmic and wild-type littermates all of which contained the lacZ indicator gene. Ablation events in the off-spring were then readily visualized by histochemical staining of ZacZ-expressing lens cells.

Fig. 1 shows the pattern of ZacZ-expressing cells in the lenses of 50-day-old wild-type and microphthalmic animals derived from mating the microphthalmic γ2DT-A mice with mice carrying the lacZ reporter gene. As noted previously (Goring et al. 1987), expression of the lacZ reporter gene was confined to the central fiber cells of normal lenses (Fig. 1A). The lens from a microphthalmic littermate was morphologically normal but reduced in size. This reduction in size was accompanied by the loss of most, but not all, blue-staining fiber cells expressing the lacZ transgene (Fig. IB). The residual blue-staining cells from the microphthalmic eye were present within the lens nucleus consistent with the location of ZacZ-expressing cells in normal lenses. No blue cells were detectable in serial histologic sections slightly displaced from the very center of the lens (data not shown), establishing that the ZacZ-expressing cells were confined to the central core of the ablated lens. Despite the marked reduction of lacZ expressing cells, genetic ablation did not appear to subvert the normal program of lens morphogenesis as evidenced by the formation of a prominent nuclear core.

Fig. 1.

Mapping of lens ablation events directed by the cytotoxic γ2DT-A transgene. Transgenic indicator mice homozygous for the γ2 fucZ reporter gene were crossed with microphthalmic line 1 animals hemizygous for the γ2 DT-A ablation construct. Resulting wild-type and microphthalmic progeny were sacrificed at 50 days and their eyes sectioned and stained histochemically for β-galactosidase activity. (A) Transverse section through center of the lens of the wild-type eye (×22-5). (B) Transverse section through center of the lens of a microphthalmic eye (×24).

Fig. 1.

Mapping of lens ablation events directed by the cytotoxic γ2DT-A transgene. Transgenic indicator mice homozygous for the γ2 fucZ reporter gene were crossed with microphthalmic line 1 animals hemizygous for the γ2 DT-A ablation construct. Resulting wild-type and microphthalmic progeny were sacrificed at 50 days and their eyes sectioned and stained histochemically for β-galactosidase activity. (A) Transverse section through center of the lens of the wild-type eye (×22-5). (B) Transverse section through center of the lens of a microphthalmic eye (×24).

Our initial characterization of transgenic mice carrying the γ2DT-A gene revealed considerable phenotypic variability in the size and structure of the lenses of animals within a single transgenic line (Breitman et al. 1987). The basis for this variability became evident during the course of in situ analysis of the lenses of a large number of microphthalmic littermates. Many of the lenses analysed were found to be completely devoid of ZacZ-expressing cells; a few retained a residual population of blue-staining cells as described above, whereas a very small minority contained a significant population of blue cells in the lens core (data not shown). Overall, a correlation was observed between the size of the lens and the degree of ablation as visualized by in situ staining of ZacZ-expressing cells. Based on these observations, we conclude that the phenotypic heterogeneity amongst ablated lenses of a single transgenic line results from variation in the number of cells that escape ablation by the γ2DT-A transgene.

γ -Crystallin mRNA levels in ablated lenses

The results described above established that the γ2DT-A construct usually ablates most, if not all, of the nuclear fiber cells expressing the γ2lacZ transgene and that the ablation of this early population of fiber cells is accompanied by the formation of a ‘new’ lens core during organogenesis. It was therefore of interest to investigate how the ablation of the lacZ-expressing cells affected the profile of γ -crystallin gene expression in the lens. Moreover, because the y-crystallins are encoded in a multigene family whose individual members are differentially regulated (Murer-Orlando et al. 1987; Van Leen et al. 1987b), we wanted to investigate whether ablation events mediated by the γ2DT-A construct were specific for those fiber cells expressing the γ2 gene.

Fig. 2 shows the relative amounts of γ2, γ3, and γ4 transcripts in wild-type and microphthalmic lens RNA preparations detected using probes that discriminate amongst these non-allelic genes. On average, the levels of total γ-crystallin mRNA, as detected with a broadly hybridizing γ-crystallin probe, were reduced 7-fold in the microphthalmic lenses. Considerable differences were observed in the relative reduction of specific γ-crystallin transcripts. Whereas there was about a 30- and 20-fold reduction in the levels of γ2 and γ4 transcripts, respectively, the amount of γ3 mRNA was down by only 2-fold. This differential reduction indicates that the ablation events preferentially eliminate y2- and γ4-producing cells and that approximately 50% of the γ3 transcripts are synthesized in fiber cells not subject to ablation. The preferential reduction of γ2 and γ4 mRNAs probably indicates that the majority of these transcripts are synthesized in a common population of cells. However, we cannot rule out the formal possibility that γ2 and γ4 are synthesized in distinct cell populations and that ablation of the γ2-producing population results in the failure of the γ4-producing population to form.

Fig. 2.

Relative amounts of γ-crystallin transcripts in the eyes of wild-type and microphthalmic mice. Total eye RNAs prepared from 3-week-old wild-type (WT) and microphthalmic (M) line 1 transgenic mice were electrophoresed in equal eye equivalents, or various fractions thereof, through 1·5% formaldehyde-agarose gels, blotted to Zetabind filters, and hybridized with a broadly reactive mouse γ2 cDNA probe (top panel), or gene-specific subprobes derived from mouse γ2, γ3 and γ4 cDNAs.

Fig. 2.

Relative amounts of γ-crystallin transcripts in the eyes of wild-type and microphthalmic mice. Total eye RNAs prepared from 3-week-old wild-type (WT) and microphthalmic (M) line 1 transgenic mice were electrophoresed in equal eye equivalents, or various fractions thereof, through 1·5% formaldehyde-agarose gels, blotted to Zetabind filters, and hybridized with a broadly reactive mouse γ2 cDNA probe (top panel), or gene-specific subprobes derived from mouse γ2, γ3 and γ4 cDNAs.

Gene dosage and extent of ablation

Our observation that phenotypic variability amongst the lenses of microphthalmic animals was due to differences in the extent of ablation prompted us to investigate whether the penetrance of ablation events could be enhanced by increasing the dosage of the transgene in the microphthalmic founder lines. We therefore crossed two hemizygous animals from one of the γ2DT-A founder lines (line 1) to generate DT-A/DT-A homozygotes. Of the offspring resulting from this cross, a proportion had normal wild-type eyes, some were microphthalmic like their parents, while others had a much more severe microphthalmia, which resembled clinical anophthalmia. Offspring representative of these 3 phenotypes are shown in Fig. 3.

Fig. 3.

Ocular phenotypes among progeny of microphthalmic animals hemizygous for the γ2DT-A transgene. From left to right are wild-type, microphthalmic, and severely microphthalmic mice.

Fig. 3.

Ocular phenotypes among progeny of microphthalmic animals hemizygous for the γ2DT-A transgene. From left to right are wild-type, microphthalmic, and severely microphthalmic mice.

To determine the relationship between the phenotype and genotype of these animals, DNA prepared from tail biopsies was subjected to Southern blot analysis using the γ2DT-A transgene as probe. Comparison of the intensities of the endogenous γ2-crystallin gene with the transgene provided an internal standard for distinguishing between animals hemizygous or homozygous for the γ2DT-A construct (Fig. 4). The results of this analysis are summarized in Table 1. The data show that the frequency of all 3 genotypes was close to the 1:2:1 ratio theoretically predicted; hence, inheritance of 2 copies of the ablation construct in line 1 animals was not deleterious to the developing embryo. The data also show that there was a striking correlation between the three ocular phenotypes (wild-type, microphthalmic, severely microphthalmic) and the number of inherited copies of the transgene. All of the animals hemizygous for the transgene were microphthalmic and, with one exception, all of the homozygotes had severe microphthalmia. The one homozygous animal which did not display the severe eye phenotype was microphthalmic. The homozygosity of this animal was verified by back-crossing to normal CD-I mice; all the resulting offspring were microphthalmic, as predicted.

Table 1.

Transmission of ablation phenotypes in hemizygous matings + /γ2DT-A × + /γ2DT-A

Transmission of ablation phenotypes in hemizygous matings + /γ2DT-A × + /γ2DT-A
Transmission of ablation phenotypes in hemizygous matings + /γ2DT-A × + /γ2DT-A
Fig. 4.

Dosage of transgene among representative progeny of line 1 microphthalmic mice hemizygous for the γ2DT-A ablation construct. DNA prepared from tail biopsies was digested with SacI and analysed by Southern blotting using the γ2DT-A transgene as probe. The intensity of the endogenous γ2-crystalin gene (upper band) was used as an internal standard for distinguishing between animals hemizygous and homozygous for the γ2DT-A construct. W, wild type; H, homozygous for the transgene; h, hemizygous for the transgene.

Fig. 4.

Dosage of transgene among representative progeny of line 1 microphthalmic mice hemizygous for the γ2DT-A ablation construct. DNA prepared from tail biopsies was digested with SacI and analysed by Southern blotting using the γ2DT-A transgene as probe. The intensity of the endogenous γ2-crystalin gene (upper band) was used as an internal standard for distinguishing between animals hemizygous and homozygous for the γ2DT-A construct. W, wild type; H, homozygous for the transgene; h, hemizygous for the transgene.

To explore further the relationship between the ablation phenotype and dosage of the transgene, two additional microphthalmic founder lines were mated to homozygosity and the progeny of these crosses were scored phenotypically and genotypically, as described above. The results are summarized in Table 1. For line 3, considerable variability in gross eye phenotype was observed, independent of gene dosage. For example, of 6 progeny that were hemizygous for the transgene, 5 were microphthalmic, while the sixth contained one microphthalmic and one severely microphthalmic eye, illustrating at the macroscopic level, variability in the extent of ablation events within a single animal. Moreover, of the 5 animals that were homozygous for the transgene, only 1 animal was severely microphthalmic; 3 others were microphthalmic, while 1 contained a microphthalmic and a severely microphthalmic eye. For line 4, all 3 hemizygous animals were microphthalmic. Of the 7 homozygotes, 4 had the severe eye phenotype while 3 were microphthalmic.

Taken together, these results indicate that the extent of genetic ablation is influenced, at least in part, by the dosage of the transgene in microphthalmic founder lines. However, the correlation between the severity of the ablation phenotype and homozygosity of the γ2DT-A construct varied among lines, suggesting that chromosomal position effects may contribute to the overall activity of the transgene.

Lenses of severely microphthalmic animals

Homozygous animals showing severe microphthalmia were analyzed histologically to gain further insight into the effects of ablation on eye development. The histological analyses revealed that the small eyes of these animals were almost entirely filled with convoluted retina. Most of the eyes analysed were completely devoid of any detectable lens tissue, although some contained trace remnants of the lens (Fig. 5). In either case, the anterior chamber, iris and vitreous body were not recognizeable in the severely microphthalmic eyes. Most other ocular structures, including the sclera, cornea, conjunctiva, eyelid, choroid, and retinal pigment epithelium appeared to be reasonably well formed, by light microscopy. The ciliary epithelium was apparent and extended anteriorly, where it was adherent to the posterior surface of the cornea. Some rudimentary ciliary muscle was discernible. Particularly noteworthy was the presence of a sensory retina containing well-formed photoreceptors and other retinal layers (outer nuclear, outer plexiform, inner nuclear, inner plexiform and ganglion cell). The nerve fiber layer was evident in the retina, but it was rudimentary in the specimens examined. However, relative to normal eyes, the sensory retinas were extensively convoluted, presumably resulting from the extensive proliferation of this lineage relative to the other cell types of the eye.

Fig. 5.

Histological analysis of severely microphthalmic eyes. (A) Histological section of severely microphthalmic right eye of a 46-day-old transgenic mouse secondary to genetic ablation of lens. The eye is almost entirely filled by convoluted retina. Arrow denotes rudimentary lens (× 40). (B) Higher magnification of A of anterior portion of microphthalmic eye showing convoluted retina, cornea, ciliary epithelium and fragments of rudimentary lens fibers (× 100). (C) Histological section of severe microphthalmic left eye from same 46-day-old transgenic mouse as in A. This eye is also almost entirely filled by convoluted retina, but no lens structures are apparent (× 40). (D) Higher magnification of retinal segment in 46-day-old transgenic mouse secondary to genetic ablation of lens. The layers of the retina are well-demarcated (×l50). C, cornea; CE. ciliary epithelium; R. retina; RPE. retinal pigment epithelium; OS, outer segments of photoreceptors; IS, inner segments of photoreceptors; ONL. outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer.

Fig. 5.

Histological analysis of severely microphthalmic eyes. (A) Histological section of severely microphthalmic right eye of a 46-day-old transgenic mouse secondary to genetic ablation of lens. The eye is almost entirely filled by convoluted retina. Arrow denotes rudimentary lens (× 40). (B) Higher magnification of A of anterior portion of microphthalmic eye showing convoluted retina, cornea, ciliary epithelium and fragments of rudimentary lens fibers (× 100). (C) Histological section of severe microphthalmic left eye from same 46-day-old transgenic mouse as in A. This eye is also almost entirely filled by convoluted retina, but no lens structures are apparent (× 40). (D) Higher magnification of retinal segment in 46-day-old transgenic mouse secondary to genetic ablation of lens. The layers of the retina are well-demarcated (×l50). C, cornea; CE. ciliary epithelium; R. retina; RPE. retinal pigment epithelium; OS, outer segments of photoreceptors; IS, inner segments of photoreceptors; ONL. outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer.

It was next of interest to examine crystallin gene expression in the rudimentary lenses of the severely microphthalmic animals. The results, shown in Fig. 6, demonstrate that the severely microphthalmic eyes had considerably lower levels of γ2-specific transcripts than were found in microphthalmic eyes, or at most, 1/200 that found in wild-type lenses. However, since the γ2-specific probe used in this analysis shows about 1 % cross-hybridization with other members of the γ-crystallin gene family (Murer-Orlando et al. 1987), the residual signal detected in Fig. 6 is probably not entirely due to γ2 RNA transcripts, and therefore corresponds to a maximum estimate of the average levels of γ2 mRNA in severely microphthalmic eyes.

Fig. 6.

Average relative levels of αA and γ2-crystallin transcripts in the eyes of wild-type, microphthalmic, and severely microphthalmic mice. Total eye RNAs prepared from 3-week-old wildtype (WT) and line 1 microphthalmic (M) and severely microphthalmic (SM) transgenic mice were run through 1·5% formaldehyde-agarose gels, blotted to Zetabind filters and hybridized with the mouse αA cDNA or the gene-specific subprobe derived from mouse γ2 cDNA. Lanes 1 to 4; fractions (1/2, 1/5, 1/10 and 1/20) of one eye equivalent of wild-type RNA. Lanes 5 and 7: 1 and 10 eye equivalents of microphthalmic RNA. Lanes 6 and 8: 1 and 10 eye equivalents of severely microphthalmic RNA.

Fig. 6.

Average relative levels of αA and γ2-crystallin transcripts in the eyes of wild-type, microphthalmic, and severely microphthalmic mice. Total eye RNAs prepared from 3-week-old wildtype (WT) and line 1 microphthalmic (M) and severely microphthalmic (SM) transgenic mice were run through 1·5% formaldehyde-agarose gels, blotted to Zetabind filters and hybridized with the mouse αA cDNA or the gene-specific subprobe derived from mouse γ2 cDNA. Lanes 1 to 4; fractions (1/2, 1/5, 1/10 and 1/20) of one eye equivalent of wild-type RNA. Lanes 5 and 7: 1 and 10 eye equivalents of microphthalmic RNA. Lanes 6 and 8: 1 and 10 eye equivalents of severely microphthalmic RNA.

The levels of α-crystallin mRNA, on the other hand, were only reduced on average about 30-fold (Fig. 6). The lower relative reduction in α-crystallin mRNA indicates that, even in eyes containing only vestigial lens tissue, there was preferential ablation of γ2-producing fiber cells.

In this paper, we have shown how the introduction of two different transgenes into the mouse germ line - the neutral lacZ indicator gene and the cytotoxic DT-A gene – can be combined to gain insight into development processes. Using this approach, it was possible to examine the contributions of specific cell types to lens organogenesis and the developmental strategies involved in determining lens cell fate. The results described here also provide insight into the role the lens plays in the development of other ocular structures during eye morphogenesis.

Cell fate during lens organogenesis

Through crosses with transgenic mice carrying the γ2lacZ indicator gene, it was possible to map ablation events directed by the γ2 promoter and to demonstrate that expression of the DT-A gene results in the specific ablation of central fiber cells that normally constitute the nuclear core of the lens. The ablation of lacZ- expressing cells correlated with a preferential reduction in γ2 mRNA, establishing that expression of the DT-A gene, as well as the neutral lacZ indicator gene, was specific for those fiber cells committed to expressing the endogenous γ2 gene. These results therefore provide direct evidence that mouse γ2 sequences −759 to +45 contain sufficient regulatory information to direct specific expression of reporter genes to γ2-producing fiber cells in the lenses of transgenic mice.

The marked reduction in γ2 transcripts relative to other members of the γ-crystallin gene family was consistent with current information concerning the spatial distribution of expression of the different members of the γ-crystallin gene family across the eye lens. In this regard, the greater reduction in γ2 transcripts can be explained in terms of differences in the spatial distribution of these mRNAs in the normal lens, with γ2 transcripts being most abundant in central nuclear fiber cells that are subject to ablation and γ3 transcripts being present in both nuclear and cortical fiber cells of the mature lens. This conclusion is supported by recent observations on the differential distribution of the corresponding γ-crystallins in the rat lens (Siezen et al. 1988). In addition, a different spatial distribution of the γ2 and γ3 transcripts is consistent with observed differences in the regulation of these genes during development. Although both γ2 and γ3 are transcriptionally activated at about day 10–11 of gestation, γ2 transcripts reach peak levels in the lens at about age 20 days and then decline rapidly, whereas γ3 transcripts are maintained at a relatively uniform level until at least 60 days after birth (Murer-Orlando et al. 1987). Taken together, these results suggest that following ablation of γ2-expressing nuclear fiber cells in microphthalmic animals, lens organogenesis is able to proceed relatively normally with a new lens core being generated primarily from a later population of (normally cortical) fibre cells that efficiently expresses the γ3, but not the γ2 gene. In other words, phenotypically and developmentally distinct populations of lens fiber cells would appear to be able to spatially and/or structurally compensate for one another during lens organogenesis. Of course, direct visualization of γ2- and γ3-expressing cells would be required to confirm this possibility.

Penetrance of ablation events and role of the lens in eye development

An intriguing feature of the γ2DT-A transgenic mice was the variable penetrance of ablation events which was most strikingly evident in the relationship between transgene dosage and increased severity of ocular phenotype. Although the molecular basis for this phenomenon is unknown, the results of the gene dosage studies served to exclude or render unlikely two possible mechanisms whereby variable penetrance might be achieved. First, the variable penetrance of ablation events seen in animals hemizygous for the γ2DT-A transgene cannot be due to somatic mutations in a cellular gene conferring resistance to DT, such as the gene for elongation factor 2 which is the target of the toxin ribosyltransferase; otherwise there would not have been an increase in the severity of the ocular phenotype with an increase in copy number of the transgene. Second, it is unlikely that variable penetrance is the result of random mutational inactivation of the transgene; otherwise one would have to postulate the unlikely event that animals homozygous for the transgene that lacked the severe microphthalmic phenotype contained lens cells that had sustained an independent somatic mutation in each of their two DT alleles. Since one molecule of the DT-A chain has been reported to be cytotoxic (Yamaizumi et al. 1978), it is also improbable that occasional cells fail to accumulate threshold levels of toxin. We therefore favor the idea that cells escape ablation through failure of the transgene to become transcriptionally activated. In this case, the correlation between gene dosage and extent of ablation would result from cells homozygous for the transgene having a significantly greater likelihood of activating at least one of their two DT-A alleles.

Our observation that numerous eyes of severely microphthalmic animals lacked detectable lens tissue was unexpected since in situ studies conducted on rat lenses have indicated that expression of the γ-crystallin genes is restricted to nucleated fiber cells (Van Leen et al. 1987a). Moreover, a recent study of transgenic mice carrying a ricin gene driven by a αA-crystallin promoter/enhancer sequences revealed incomplete ablation of the lens lineage (Landel et al. 1988). One possible explanation for our finding is that the normal program of lens development may be aborted when ablation of a threshold number of nuclear fiber cells is achieved. This might occur if extensive or complete ablation of γ2-expressing fiber cells triggers a reprogramming of developmental processes such that there is a chronic and futile attempt to generate such cells. In this case, subsequent populations of fiber cells that do not express the γ2 gene would never be formed. In any event, our analysis of severely microphthalmic eyes suggested that even in the complete absence of a lens most other ocular cell types are able to form during eye morphogenesis. However, it would appear that, with the exception of the vitreous body and retina, growth of the lens is necessary to cue proportional growth of the other ocular structures. Thus, the sclera, cornea, and ciliary epithelium were reduced in size but, otherwise, reasonably well formed. In contrast, severely microphthalmic animals contained extensively convoluted retina, suggesting that the extent of cell proliferation within this lineage is not strictly coordinated with lens growth and hence must be governed by additional factors. The finding of a well-defined sensory retina in these eyes supports the concept, based on observations in other situations, that the formation of a neural retina is not dependent upon the development of a lens (Coulombre, 1964). The absence of a vitreous body in eyes with genetically ablated lenses supports the possibility raised by Coulombre (1964) that vitreous body development requires the formation of both a neural retina and lens. Further studies conducted at early developmental stages should provide additional insight into the nature of these interrelationships during ocular morphogenesis.

This work was supported by grants from the Medical Research Council of Canada and the National Cancer Institute of Canada to M.L.B. and A.B. and by National Eye Institute grants R01-EY-00146 and P30-EY-05722 to G.K.K.

Behringer
,
R. R.
,
Mathews
,
L. S.
,
Palmiter
,
R. D.
&
Brinster
,
R. L.
(
1988
).
Dwarf mice produced by genetic ablation of growth hormone expressing cells
.
Genes & Development
2
,
453
461
.
Breitman
,
M. L.
,
Clapoff
,
S.
,
Rossant
,
J.
,
Tsui
,
L.-C.
,
Glode
,
L. M.
,
Maxwell
,
I. H.
&
Bernstein
,
A.
(
1987
).
Genetic Ablation: Targeted expression of a toxin gene causes microphthalmia in transgenic mice
.
Science
238
,
1563
1565
.
Coulombre
,
A. J.
(
1965
).
The Eye in Organogenesis
,
Dehaan
,
R. L.
&
Ursprung
,
H.
(ed.)
Proceedings of International Conference on Organogenesis, Baltimore, Maryland, September 1-12, (1964), Holt, Rhinehart and Winston
,
New York
.
Feinberg
,
A. P.
&
Vogelstein
,
B.
(
1983
).
A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity
.
Analyt. Biochem
.
132
,
6
13
.
Goring
,
D. R.
,
Rossant
,
J.
,
Clapoff
,
S.
,
Breitman
,
M. L.
&
Tsui
,
L.-C.
(
1987
).
In situ detection of β-galactosidase in lenses of transgenic mice with a γ-crystallin/lacZ gene
.
Science
235
,
456
458
.
Hogan
,
B.
,
Costantini
,
F.
&
Lacy
,
E.
(
1986
).
Manipulating The Mouse Embryo: A Laboratory Manual
(
Cold Spring Harbor, New York
:
Cold Spring Harbor Laboratory
).
Landel
,
C. P.
,
Zhao
,
J.
,
Bok
,
D.
&
Evans
,
G. A.
(
1988
).
Lensspecific expression of recombinant ricin induces developmental defects in the eyes of transgenic mice
.
Genes and Development
2
,
1168
1178
.
Mcavoy
,
J. W.
(
1978
).
Cell division, cell elongation and distribution of α, β- and γ-crystallins in the rat lens
.
J. Embryol. exp. Morph
.
44
,
149
165
.
Murer-Orlando
,
M.
,
Paterson
,
R. C.
,
Lok
,
S.
,
Tsui
,
L.-C.
&
Breitman
,
M. L.
(
1987
).
Differential regulation of γ-crystallin genes during mouse lens development
.
Devi Biol
.
119
,
260
267
.
Palmiter
,
R. L.
,
Behringer
,
R. R.
,
Quaife
,
C. J.
,
Maxwell
,
F.
,
Maxwell
,
I. H.
&
Brinster
,
R. L.
(
1987
).
Cell lineage ablation in transgenic mice by cell-specific expression of a toxin gene
.
Cell
50
,
435
443
.
Piatigorsky
,
J.
(
1984
).
Lens crystallins and their gene families
.
Cell
38
,
620
621
.
Quinlan
,
P.
,
Oda
,
S.
,
Breitman
,
M. L.
&
Tsui
,
L. C.
(
1987
).
The mouse eye lens obsolescence (Elo) mutant: studies on crystallin gene expression and linkage analysis between the mutant locus and the γ-crystallin genes
.
Genes and Development
1
,
637
644
.
Rigby
,
P. J. W.
,
Dieckmann
,
M.
,
Rhodes
,
C.
&
Berg
,
P.
(
1977
).
Labeling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I
.
J. molec. Biol
.
113
,
237
251
.
Siezen
,
R. J.
,
Wu
,
E.
,
Kaplan
,
E. D.
,
Thomson
,
J. A.
&
Benedek
,
G. B.
(
1988
).
Rat lens y-crystallins. Characterization of the six gene products and their spatial and temporal distribution resulting from differential synthesis
.
J. molec. Biol
.
199
,
475
490
.
Southern
,
E. M.
(
1975
).
Detection of specific sequences among DNA fragments separated by agarose gel electrophoresis
.
J. molec. Biol
.
98
,
503
515
.
Van Leen
,
R. W.
,
Breuer
,
M. L.
,
Lubsen
,
N. H.
&
Schoenmakers
,
J. G. G.
(
1987a
).
Developmental expression of crystallin genes: in situ hybridization reveals a differential localization of specific mRNAs
.
Devi Biol
.
123
,
338
345
.
Van Leen
,
R. W.
,
Van Roozendeal
,
K. E. P.
,
Lubsen
,
N. H.
&
Schoenmakers
,
J. G. G.
(
1987b
).
Differential expression of crystallin genes during development of the rat eye lens
.
Devi Biol
.
120
,
457
464
.
Yamaizumi
,
M.
,
Mekada
,
E.
,
Uchida
,
T.
&
Okada
,
Y.
(
1978
).
One molecule of diphtheria toxin fragment A introduced into a cell can kill the cell
.
Cell
15
,
245
250
.