Prenatal Lens Development in Connexin43 and Connexin50 Double Knockout Mice

Abstract

purpose. To determine the roles of intercellular communication in embryonic eye growth and development, mice with a targeted deletion of the Cx43 gene were examined, and mice without both Cx43 and Cx50 were generated and analyzed.

methods. Embryonic eyes and lenses from wild-type mice, or mice deficient in Cx43, Cx50, or both Cx43 and Cx50 were collected and analyzed structurally by light and electron microscopy, immunohistochemically using connexin-specific antibodies, biochemically by Western blot analysis, and physiologically by measuring patterns of junctional communication revealed by iontophoretic injection of junction-permeable reporter molecules.

results. Cx50 expression was limited to the ocular lens and was not detected in either the cornea or the retina. Cx43 ^−/− embryos showed development of structurally normal lenses and eyes when examined by light and electron microscopy through embryonic day (E)18.5. In addition, Cx43 ^−/− lenses synthesized four different markers of lens differentiation: MIP26, αA-crystallin,α B-crystallin, and γ-crystallin. Double-knockout lenses were also histologically normal through E18.5 and synthesized the four lens differentiation markers. When assayed by intracellular injection with Lucifer yellow (Molecular Probes, Eugene, OR) and neurobiotin at E15.5, Cx43 ^−/−/Cx50 ^−/− lenses retained gap junction–mediated dye transfer between fiber cells. In contrast, dye transfer in double-knockout lenses was dramatically reduced between epithelial cells and was eliminated between epithelial cells and fibers.

conclusions. These data indicate that the unique functional properties of both Cx43 and Cx50 are not required for prenatal lens development and that connexin diversity is required for regulation of postnatal growth and homeostasis.

The vertebrate lens has two cell types. A simple epithelium lines the anterior surface, whereas elongated fibers constitute the bulk of the lenticular mass. At the lens equator, epithelial cells differentiate into lens fibers, losing intracellular organelles and accumulating high concentrations of crystallin proteins. ¹ These cellular properties, together with an elastic capsule and zonular fibers, result in the optical transparency, high refractive index, and elasticity necessary for accommodation. With the loss of cellular organelles, the fibers lose their ability to support oxidative phosphorylation and an active metabolism. Although the lens is avascular, every cell is joined to its neighbors by the intercellular channels that comprise gap junctions. These channels network the lens cells into a functional syncytium, allowing the metabolically inactive fibers to directly share ions, second messengers, and metabolites with the biosynthetically active epithelium. ²  

Structural proteins belonging to the connexin multigene family make up the intercellular channels present in gap junctions, and three distinct connexin (Cx) genes are expressed in the lens. ³ Cx43 is predominantly detected in gap junctions between epithelial cells. ⁴ During fiber differentiation, Cx43 expression is downregulated and junctional coupling is maintained by an upregulation of two additional connexins, Cx46 and Cx50. ^{5 6 7} Each of these connexins forms channels with distinctly different physiological properties of gating, permeation, and selective interaction with other members of the connexin family. ⁸  

An essential role for gap junctional communication in the lens has recently been clarified by the findings that mutations in either the human Cx46 or Cx50 genes cause cataracts ^{9 10} and that lens opacities develop in mice with targeted deletions of these connexins. ^{11 12} In contrast to Cx46 and Cx50, no human congenital diseases have been linked to the Cx43 gene to date, and targeted ablation of Cx43 in mice leads to perinatal lethality, ¹³ precluding any functional analyses in the postnatal and adult lenses. The genetic studies demonstrate an absolute requirement for gap junctions in the maintenance of the intracellular ionic conditions necessary for crystallin solubility, a role that had been previously hypothesized for the lens connexins. ^{14 15 16}  

The dependence of lens homeostasis on intercellular communication was not surprising in light of a major function that has been traditionally attributed to gap junctions: the provision of both metabolic continuity and the synchronization of function between differentiated cells within an organ. ¹⁷ A second major proposed function for gap junctions is distilled from numerous studies correlating changes in junctional communication with various developmental processes, invoking a model whereby intercellular channels influence the signaling of morphogens and help to establish embryonic communication compartments. ^{18 19 20 21} Although gap junctions are abundant throughout the lens and these structures are thought to play a significant role in the development of other organs, the precise role of intercellular channels in lens differentiation remains largely unknown. 

Vertebrate lens development follows a similar pattern in many species, including rodents and humans. ^{1 22} The embryonic ectoderm is induced by the underlying optic vesicle and thickens to form the lens placode. The placode then invaginates and separates from the ectoderm, giving rise to the lens vesicle, a hollow epithelial sphere. At this stage of development, the lens is de facto a separate compartment and can only interact with surrounding tissues through extracellular signals. The posterior cells of the vesicle then differentiate into primary fiber cells that elongate to fill the lumen of the lens vesicle. Fiber cell differentiation is characterized by increased cell volume, the upregulation of fiber–specific proteins, including the crystallins and the plasma membrane protein MIP26 (aquaporin 0), and the loss of intracellular organelles. The lens continues to grow throughout life, because of the ongoing differentiation of epithelial cells into secondary fiber cells at the lens equator. Secondary fiber formation is initiated by the progressive morphologic and biochemical differentiation of a small population of epithelial stem cells near the lens equator. 

The equatorial region shows the greatest magnitude of intercellular coupling within the lens, ^{16 23} consistent with the upregulation of connexin expression coinciding with fiber differentiation. ^{24 25} In addition to the increases in expression of Cx46 and Cx50 taking place at the equator, immunohistochemical studies in the chick have shown that Cx43 levels dramatically increase in equatorial epithelial cells, ²⁶ although this increase in Cx43 expression is less prominent in the mouse (White et al., unpublished observations, 1998). It is unclear why synthesis of Cx43 increases immediately before the elimination of Cx43 as fiber differentiation progresses, although it has been hypothesized that this may reflect a need for gap junctions in the initiation or coordination of fiber differentiation. ^{27 28}  

More recently, an in vitro study has suggested that junctional communication is not required for the initiation of lens differentiation in a cell culture model. ²⁹ Although this study relied on nonspecific pharmacologic blockers of intercellular channels, the results clearly illustrated that several aspects of epithelial-to-fiber differentiation, including the expression of crystallins and MIP26, could occur in the absence of detectable gap junctional communication. The availability of mice with targeted deletions of the different lens connexin genes makes it possible to directly test in vivo whether gap junctional coupling is required for fiber differentiation. 

To directly address the role of connexin-mediated intercellular communication in lens development, we have examined lenses from mice without Cx43 and Cx50, and double-knockout mice without both Cx43 and Cx50. Cx43 ^−/− embryos showed development of histologically normal eyes and lenses and expressed four different markers of lens fiber differentiation: MIP26, αA-crystallin,α B-crystallin, and γ-crystallin. As assayed by intercellular transfer of neurobiotin, a gap junction–permeable reporter molecule, Cx43 ^−/−/Cx50 ^−/− lenses exhibited a substantive reduction in epithelial–epithelial and epithelial–fiber gap junctional communication, although differentiated fiber–fiber communication was unchanged from that in control lenses. Despite this genetic alteration of intercellular communication, double-knockout lenses were structurally normal as assayed by light and electron microscopy and synthesized normal levels of the four lens differentiation markers. These results suggest that the large increases in gap junctional intercellular communication that occur in the lens bow region are a consequence, rather than a cause, of lens fiber differentiation. 

Materials and Methods

Generation of Connexin-Deficient Mice

This study adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The generation of Cx50-deficient mice has been previously described. ¹² For this study, F1 hybrid (129Sv x C57BL/6) Cx50-knockout animals were backcrossed for six generations into the C57BL/6 genetic background. Cx43-knockout mice ¹³ were purchased as a fifth-generation backcross in the C57BL/6 genetic background from Jackson Laboratories (Bar Harbor, ME). Double-knockout mice were obtained by breeding Cx50 ^−/− females with Cx43 ^+/− males. Male offspring heterozygous for both Cx43 and Cx50 were bred with Cx50 ^−/− females to obtain viable animals of both genders with the Cx50 ^−/−/ Cx43 ^+/− genotype, which were then interbred to obtain Cx43/Cx50 double-knockout embryos. Animals were genotyped by PCR screening using separate protocols for the Cx50 and Cx43 alleles. To detect Cx50, a common 5′ flanking primer (pcr 1; 5′-GCC CCC TCC TGC TTA TTT CTG-3′) was paired with either a 3′ primer derived from vector sequences unique to the Cx50 replacement cassette (pcr 2; 5′-CGG GCC TCT TCG CTA TTA CG-3′) or a third primer derived from the Cx50 coding region (pcr 3; 5′-CTC CAT GCG AAC GTG GTG TAC-3′). Primers 1 plus 2 amplified a 1370-bp band from Cx50-knockout chromosomes. Amplification of wild-type chromosomes with primers 1 plus 3 produced a 1600-bp band. To detect Cx43, a common 3′ flanking primer (Cx-3′; 5′-ACT TTT GCC GCC TAG CTA TCC C-3′) was paired with either a 5′ primer derived from neomycin sequences unique to the Cx43 replacement cassette (Neo-5′; 5′-GCT TGC CGA ATA TCA TGG TGG A-3′), or a third primer derived from the Cx43 coding region (Cx-5′; 5′-CCC CAC TCT CAC CTA TGT CTC C-3′). Primers Cx-3′ plus Neo-5′ amplified a 1000-bp band from Cx43-knockout chromosomes. Amplification of wild-type chromosomes with primers Cx-3′ plus Cx-5′ produced a 500-bp band. DNAs isolated from tail biopsy specimens were amplified in a thermal cycler (GeneAmp 9600; Perkin Elmer, Foster City, CA), and amplified products were resolved by agarose gel electrophoresis. 

Light Microscopy

Adult mouse eyes were fixed in 1% formaldehyde, freshly prepared from paraformaldehyde, in phosphate-buffered saline (PBS) for 1 hour at room temperature. Fixed eyes were rinsed in PBS, embedded in optimal cutting temperature (OCT) compound (Miles, Elkhart, IN), and frozen in liquid nitrogen. Frozen sections 10 μm thick were prepared and processed as previously described. ⁶ Sections were incubated for 1 hour at room temperature with anti-Cx50 antiserum ⁷ diluted 1:100, or anti-Cx50 monoclonal 6-4-B2-C6 ⁵ (Zymed Laboratories, San Francisco, CA) diluted 1:100 with 1% normal goat serum and 2% BSA in PBS. Sections were washed in PBS and then incubated with rhodamine-conjugated goat anti-rabbit antiserum or Cy3-conjugated goat anti-mouse antiserum. For histologic analysis, embryonic eyes were fixed in 4% formaldehyde in PBS for 16 to 24 hours. After they were rinsed in PBS, samples were dehydrated and embedded in paraffin. Sections of 2 μm were cut on a diamond knife, deparaffinized, and stained with hematoxylin and eosin. 

Electron Microscopy

Embryonic lenses were fixed in 2.5% glutaraldehyde and 1% tannic acid in 0.1 M sodium cacodylate buffer (pH 7.4; cacodylate) for 4 hours at room temperature. After a wash in cacodylate, specimens were then cut into quadrants, postfixed in 1% OsO₄ in cacodylate, and washed in distilled water. The lenses were stained en bloc with 1% uranyl acetate in water, washed in water, dehydrated in ethanols and propylene oxide, embedded in Epon 812, and sectioned. 

Gel Electrophoresis and Western Blot Analysis

For crystallin and MIP26 analysis, lenses were dissected from embryonic day (E)18.5 animals and individually homogenized in 0.5 ml 0.1 M NaCl and 0.1 M Na₂HPO₄ (pH 7.4), containing 10 mM ascorbic acid (as an antioxidant). Tail biopsy specimens from individual embryos were genotyped by PCR, and equal volumes of lens fractions from animals of each genotype were electrophoresed on 15% polyacrylamide gels and transferred to nitrocellulose membranes. Western blot analyses were probed with antibodies to MIP26 and αA-, αB-, or γ-crystallin (generously provided by Joseph Horwitz, Jules Stein Eye Institute, University of California, Los Angeles). Primary antibodies were detected with alkaline phosphatase–conjugated goat anti-rabbit IgG (Roche Molecular Biochemicals, Indianapolis, IN), by using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate as substrates (Sigma Chemical Co., St. Louis, MO). 

Dye Injection

Embryonic lenses (E15.5) were dissected at 37°C from the enucleated globe in M199 culture medium buffered with 10 mM HEPES (pH 7.4; M199). After a brief incubation (1–3 minutes) in 1% collagenase and 1% hyaluronidase in M199, the tunica vasculosa lentis was grasped with fine forceps and removed. Lenses were then mounted on their equatorial edges in 35-mm polylysine-coated dishes in M199. High-resistance microelectrodes were filled at the tip with a solution containing 2% Lucifer yellow (Molecular Probes, Eugene, OR) and 2% neurobiotin (Vector Laboratories, Burlingame, CA) in 100 mM LiCl₂ by capillary action. Electrodes were then back-filled with 100 mM LiCl₂. Centrally located epithelial cells were impaled from the anterior lens surface, and fibers were impaled from either anterior (through the epithelium) or posterior surfaces. Dyes were iontophoretically injected for 10 minutes using 5-nA current pulses of 500-msec duration at a frequency of 1 Hz. Injected lenses were photographed, fixed in 4% formaldehyde freshly made from paraformaldehyde, and processed for paraffin embedding, after which they were serially sectioned, deparaffinized, and rehydrated. Sections were incubated with rhodamine-conjugated avidin (Pierce, Rockford, IL) diluted 1:500 with PBS containing 1% normal goat serum and 2% BSA for 1 hour at room temperature. Sections were washed in PBS and photographed by epifluorescence. 

Results

Distribution of Cx50 in the Murine Eye

Determining the functional consequences of gene deletion requires an unambiguous knowledge of the pattern of expression of the gene in question. Cx50 was originally described as a lens-specific gap junction protein, ^{5 7} although numerous recent immunohistochemical studies using a commercially available antibody (6-4-B2-C6) have identified Cx50 as a component of gap junctions between corneal epithelial cells, ^{30 31 32} as well as between glial cells in the retina. ³³ To resolve these conflicting data and aid our analysis of the effect of the targeted deletion of Cx50 on embryonic eye development, we re-examined the distribution of Cx50 in the murine eye, comparing the immunoreactivity of the monoclonal antibody 6-4-B2-C6 ⁵ with that of a previously characterized polyclonal serum 9496. ⁷  

A comparison of the staining pattern of the two antibodies on lens sections is shown in Figure 1 . Consistent with reports in the literature of studies conducted in numerous laboratories, ^{5 6 7 11 12 34 35 36 37} both 6-4-B2-C6 and 9496 labeled abundant punctate structures in the elongated fiber cells of wild-type lenses (Figs. 1A 1C) . The availability of genetically engineered mice, with only the gene encoding the Cx50 antigen missing, ¹² provided a rigorous source of control tissue to test for spurious cross-reactivity of the antibodies. Neither 6-4-B2-C6 nor 9496 exhibited staining in fiber cells of lenses from Cx50-knockout animals (Figs. 1B 1D) . Thus, both antibodies recognized bona fide Cx50 epitopes in lens tissue, as originally characterized. 

In contrast to the lens, staining of other ocular tissues with the two antibodies yielded highly divergent results. In sections of wild-type cornea, 6-4-B2-C6 produced a very bright punctate staining pattern (Fig. 2A) that was not observed in wild-type corneal sections stained with 9496 (Fig. 2B) . The positive staining of corneal epithelial cells with the 6-4-B2-C6 antibody was probably due to cross reactivity of the antibody with an epitope other than the Cx50 protein in this ocular tissue, in that an identical staining pattern was apparent in corneal sections from Cx50-knockout mice (Fig. 2C) . In a similar fashion, staining of wild-type retina with 6-4-B2-C6 produced a bright staining pattern that was particularly strong surrounding the cell bodies of photoreceptors in areas consistent with the location of Müller glial cells (Fig. 2D) . As was the case in the cornea, the 9496 antibody produced no staining in wild-type retina (Fig. 2E) , and the staining derived from 6-4-B2-C6 persisted in retinal sections derived from mice without Cx50 (Fig. 2F) . These data suggested that murine Cx50 was expressed exclusively in the lens, as originally reported, and that analysis of the embryonic effects of the targeted deletion of Cx50 could be focused on this ocular organ. 

Generation of Cx43/Cx50 Double-Knockout Mice

To generate double-knockout animals in a common genetic background for this study, Cx50-knockout animals ¹² were first backcrossed for six generations into the C57BL/6 genetic background. Cx43-knockout mice ¹³ were purchased as a fifth-generation backcross in the C57BL/6 genetic background from Jackson Laboratories. Because homozygous Cx43-deficient animals are not viable, ¹³ double-knockout embryos were produced by mating Cx50 ^−/−/Cx43 ^+/− parents. The resultant embryos were genotyped by PCR screening using separate protocols for the detection of wild-type and knockout alleles of Cx43 and Cx50. Figure 3 shows representative genoptypes obtained from tail DNAs isolated from pups derived from matings using Cx50 ⁺/⁺ or Cx50 ^−/− parents that were heterozygous for Cx43. Cx43 ^+/+, Cx43 ^+/−, and Cx43 ^−/− embryos were obtained in a Mendelian ratio through E19. 

Embryonic Eye Development in Cx43/Cx50 Double-Knockout Embryos

One feature of normal lens development is the highly ordered differentiation of new fiber cells in the bow region. Epithelial cells divide and migrate toward the equatorial region of the lens where they differentiate into fiber cells. This process continues throughout life, and older fibers are displaced toward the center of the lens, resulting in an ever-enlarging lens and a densely layered inner core of fibers, the nucleus. To examine this process at different developmental stages, wild-type, Cx43, or Cx43/Cx50-knockout eyes were fixed, serially sectioned, and stained with hematoxylin and eosin. Through E18.5, there were no differences in the size or integrity of lenses from either wild-type (Fig. 4A) or Cx43-knockout embryos (Fig. 4B) . In addition, deletion of both Cx43 and Cx50 failed to perturb the apparently normal cytodifferentiation of lens fibers (Fig. 4C) . Lens cytology in Cx43/Cx50 double-knockout lenses was also examined by electron microscopy, which failed to detect irregular cell-to-cell appositions between epithelial cells and/or fibers (Fig. 5) , a structural abnormality that has been previously reported in Cx43-knockout lenses. ³⁸ Thus, multiple connexin gene depletion did not adversely affect the normal embryonic pattern of lens cellular differentiation. 

Effect of Connexin Genotype on Synthesis of Lens Fiber Differentiation Markers

A second characteristic feature of normal lens fiber differentiation is the synthesis and accumulation of high levels of soluble crystallins, ^{1 39} and of MIP26 (aquaporin 0), the principal membrane protein of differentiated fibers. ^{3 40} To determine whether the normal pattern of lens-specific gene expression was influenced by connexin-mediated intercellular communication, wild-type, Cx43-knockout, Cx50-knockout, and Cx43/Cx50 double-knockout embryonic lenses were biochemically analyzed for the expression of crystallins and MIP26. Equal volumes of homogenates prepared from the lenses of single embryos were examined by Western blot with anti-crystallin or anti-MIP26 antibodies. All the crystallin proteins examined were abundant in the lens fractions from wild-type, knockout, or double-knockout lenses. MIP26 was also normally expressed, regardless of the connexin genotype (Fig. 6) . Therefore, the typical pattern of gene expression associated with lens fiber differentiation was not compromised by single or double lens connexin knockout. 

Epithelial Dye Transfer in Cx43/Cx50 Double-Knockout Lenses

Normal embryonic lenses exhibit three distinct pathways of gap junctional coupling, between epithelial cells, between fiber cells, and between the epithelial cells and fibers. ^{2 12 16 41 42} Intercellular communication was assayed by iontophoretic injection of Lucifer yellow and neurobiotin into lenses dissected from E15.5 embryos to determine whether the patterns of junctional communication were altered in the connexin double-knockout mice. The primary advantage of this age lens is that the anterior suture is not extensively developed, allowing independent analysis of junctional communication to be performed between all the different cell types. Microelectrodes delivering tracer were placed into epithelia anteriorly (Fig. 7A) or into primary fibers anteriorly by passing through the epithelium (Fig. 7B) . After injection, lenses were fixed and serially sectioned and neurobiotin visualized by incubation of sections with rhodamine-avidin. Injection of single epithelial cells in double-knockout lenses resulted in the retention of Lucifer yellow within the injected cell, with no evidence of transfer to neighboring epithelial cells or to the underlying fibers (Fig. 7C) . 

Consistent with this complete absence of Lucifer yellow transfer, the distribution of neurobiotin was restricted to the immediate neighbors of the injected epithelial cell and was not detectable in the underlying lens fibers (Fig. 7E) . In contrast, extensive transfer of neurobiotin to many neighboring fiber cells was observed in double-knockout lenses after impalement of a single fiber cell (Fig. 7D 7F) . Notably absent in the double-knockout lenses was any accumulation of neurobiotin in the overlying epithelium, which normally results from fiber-to-epithelial cell junctional transfer. ^{12 43} These results demonstrate that the targeted ablation of both Cx43 and Cx50 resulted in a large decrease in intraepithelial dye transfer and a complete loss of detectable epithelial-fiber communication in embryonic lenses at E15.5. 

Discussion

In the current study, targeted deletions of Cx43 and Cx50 alone or in combination had negligible effects on the normal differentiation of embryonic mouse lenses. Synthesis of the major cytoplasmic and membrane proteins that characterize lens fiber differentiation was completely unaffected by connexin genotype, as was the histologic structure of the lens at both the light and electron microscopic level. Double connexin deletion severely compromised two of the three possible gap junctional pathways between lens cells when assayed by dye transfer, underscoring the apparent unimportance of junctional communication per se in embryonic lens development. In contrast, deletion of Cx50 results in two postnatal phenotypes: microphthalmia and pulverulent cataracts. ¹² Taken together, these data demonstrate that connexin diversity in the lens is required for postnatal growth and homeostasis, but is not necessary for prenatal development. 

Cx43 and Cx50 represent two of the three connexins known to be present in the lens. Although Cx43 is broadly expressed in many different tissues, Cx50 has a much more limited distribution, having been reported as a component of lenticular, ^{5 7} retinal, ³³ and corneal gap junctions. ^{30 31} In the present study, spurious immunostaining in the cornea and retina resulted when a commercially available Cx50 antibody (6-4-B2-C6) was evaluated, even though this antibody recognized a true Cx50 epitope in the lens. Thus, it is concluded that Cx50 is expressed exclusively in the lens in the murine eye. It was of great advantage to have available control tissues from animals with targeted deletions of Cx50 when probing the distribution of this antigen by immunohistochemistry. Using other connexin antisera, similar kinds of spurious cross reactivity have been demonstrated in murine retinas, by using control animals with the appropriate connexin genes absent. ⁴⁴  

In general, although mutations and deletions of more than half a dozen of the connexin genes in mice or humans often lead to specific functional failures in a variety of differentiated tissues, ^{45 46} disruptions in early embryonic patterning are rare. Notable examples of developmental consequences of connexin loss are provided by the Cx43 ^−/− mouse, which shows abnormal cardiac development possibly associated with neural crest migration, ^{13 47} and a possible role for gap junction–mediated signaling in the specification of the left–right axis in the chick. ⁴⁸ Although gap junctional intercellular communication has been implicated in specification of the dorsal–ventral axis in Xenopus, ^{49 50 51 52} these data have recently been questioned. ⁵³ In light of the finding that developmental patterning and differentiation of the lens proceeded normally in the absence of both Cx43 and Cx50, and without two of the three normal junctional communication pathways, a role for these two connexins in the development of the lens was not supported by our experimental data. 

Gap junctional communication pathways may be important in the development and maintenance of cellular compartments. ^{54 55 56} In the case of the lens, it has been shown that mouse lenses without Cx46 and Cx50 undergo normal embryonic development, as do the double-knockout Cx43/Cx50 lenses. The lens becomes structurally a separate compartment at the moment of separation from the ectoderm, possibly obviating any further participation of gap junctions in the differentiation process. Indeed, because the lens vesicle is able to form in a temporally and spatially correct pattern in the absence of Cx43 and because cells in the lens placode and pit abundantly express Cx43, ⁵⁷ connexins may not participate in the initial formation of the lens vesicle as well. Cx50 is required for the lenticular stem cells to maintain an appropriate postnatal mitotic rate, ¹² although the loss of this connexin does not hamper fiber cytodifferentiation. Analysis of the phenotype resulting from a loss of all three lens connexins may provide additional insights in the future. 

With regard to Cx43-knockout mice, a previous study reported lens structural abnormalities not detected in our present work. ³⁸ In the prior analysis, grossly dilated extracellular spaces and large intracellular vacuoles were reported in neonatal Cx43 ^−/− lenses. We did not observe any abnormal cell-to-cell appositions in both Cx43 ^−/− and Cx43/Cx50 double-knockout lenses through E19. Because Cx43 null mice die at birth from anoxia, it is possible that postnatally recovered Cx43 ^−/− lenses were affected by oxygen deprivation. Alternatively, the structural abnormalities observed may have resulted from trauma endured during passage through the birth canal. Although we did not note any lenticular structural abnormalities in the present study resulting from Cx43 deletion, our other findings are in good agreement with the previous work of Gao and Spray. ³⁸ In particular, proper lens fiber differentiation was observed in both studies, including the normal expression of MIP26 and α-crystallin. Assimilation of the earlier studies of single-lens connexin knockouts with the present work on Cx43 ^−/− and Cx43/Cx50 double knockouts, leads to the conclusion that connexin diversity, and most likely junctional communication itself, is dispensable with regard to normal lens fiber differentiation. However, the large increases in gap junctional communication that accompany normal fiber differentiation are critically important for postnatal growth and the maintenance of lens clarity. Thus, connexin diversity and junctional coupling are important consequences, rather than active causes, of lens differentiation. 

 Supported by National Institutes of Health Grants EY13163 (TWW), EY02430 (DAG), and GM37751 (DLP).

 Submitted for publication May 2, 2001; revised June 22, 2001; accepted July 18, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Immunohistochemical analysis of Cx50 expression in control and knockout lenses. Wild-type lenses stained with the 6-4-B2-C6 monoclonal antibody showed typical macular labeling of fiber cell membranes (A) that was not present in Cx50-knockout lenses (B). Similarly, gap junctions were labeled in wild-type (C), but not in knockout (D) lenses with polyclonal antibody 9496. Therefore, both antibodies recognized a genuine Cx50 epitope in lens fibers, as originally characterized.

Figure 2.

 

Analysis of Cx50 distribution in other ocular tissues. In wild-type cornea, the monoclonal antibody 6-4-B2-C6 produced a punctate staining pattern (A) that was not observed in wild-type corneal sections stained with polyclonal antibody 9496 (B). The positive staining of corneal epithelial cells with 6-4-B2-C6 was due to cross-reactivity with an epitope other than Cx50, in that an identical staining pattern was produced in Cx50-knockout cornea (C). In a similar fashion, staining of wild-type retina with 6-4-B2-C6 produced a bright signal surrounding the cell bodies of photoreceptors (D). In contrast, the 9496 antibody produced no staining in wild-type retina (E) and the staining derived from 6-4-B2-C6 persisted in retinal sections derived from mice without Cx50 (F). Thus, murine Cx50 is expressed exclusively in the lens.

Figure 3.

 

Genotyping of connexin-deficient mice. Embryo tail DNAs were genotyped by PCR screening using separate three primer protocols for the Cx43 and Cx50 alleles. (A) To detect Cx43, a common 3′ flanking primer was paired with either a 5′ primer derived from the Cx43 coding region to detect wild-type alleles (lanes 1, 3, 5, 7), or a second 5′ primer derived from neomycin sequences unique to the Cx43 replacement cassette to detect knockout alleles (lanes 2, 4, 6, 8). Amplification of wild-type Cx43 chromosomes produced a 500-bp band, and Cx43-knockout chromosomes amplified a 1000-bp band. (B) For genotyping Cx50, a common 5′ flanking primer was paired with either a 3′ primer derived from the Cx50 coding region (lanes 1, 3, 5, 7) or a primer derived from specific sequences in the replacement cassette (lanes 2, 4, 6, 8). Amplification of wild-type Cx50 chromosomes produced a 1600-bp band, and Cx50-knockout chromosomes amplified a 1370-bp band. Thus, embryo 1 (lanes 1, 2) was a wild-type; embryo 2 (lanes 3, 4) was a Cx43 knockout; embryo 3 (lanes 5, 6) was a Cx50 knockout; and embryo 4 (lanes 7, 8) was a double knockout.

Figure 4.

 

Histologic analysis of connexin-deficient eye development. Tissues were fixed, serially sectioned, and stained with hematoxylin and eosin. At E18.5, there were no differences in the size or integrity of lenses from either wild-type (A) or Cx43-knockout embryos (B). In addition, deletion of both Cx43 and Cx50 failed to perturb the apparently normal cytodifferentiation of lens fibers (C).

Figure 5.

 

Ultrastructural analysis of the epithelium–fiber interface in Cx43/Cx50 double-knockout lenses. Lens cytology in Cx43/Cx50 double-knockout lenses was also examined by electron microscopy that failed to detect irregular cell-to-cell appositions or vacuoles between epithelial cells (E-E), between the epithelium and fibers (E-F), or between fibers (F-F).

Figure 6.

 

Biochemical analysis of connexin-deficient lens development. Wild-type (lane 1), Cx43-knockout (lane 2), Cx50-knockout (lane 3), and double-knockout (lane 4) embryonic lenses were homogenized and equal volumes of the homogenate were used in Western blots with anti-αA-crystallin, anti-αB-crystallin, anti-γ-crystallin, or anti-MIP26 antibodies. All the crystallin proteins examined and MIP26 were abundant in the lens fractions, regardless of the connexin genotype. Thus, the expression of lens fiber differentiation markers was not compromised by single- or double-lens connexin knockout.

Figure 7.

 

Targeted ablation of Cx43 and Cx50 eliminates gap junctional communication between the epithelium and the fibers. Intercellular communication in E15.5 embryonic lenses was assessed by intracellular injection of low-molecular-mass tracers. During the injections into whole lenses, the location of Lucifer yellow revealed the type of cell injected. Microelectrodes delivering tracer were placed into epithelia anteriorly (A) or into primary fibers anteriorly by passing through the epithelium (B). After injection, lenses were fixed and serially sectioned, and neurobiotin was visualized by incubation of sections with rhodamine-avidin. Injection of single epithelial cells in double-knockout lenses resulted in the retention of Lucifer yellow within the injected cell, with no evidence of transfer to neighboring epithelial cells or to the underlying fibers (C, and higher magnification inset). The distribution of neurobiotin was restricted to the immediate neighbors of the injected epithelial cell and did not transfer to the lens fibers (E and higher magnification inset). In contrast, neurobiotin was transferred to many neighboring fiber cells in double-knockout lenses after impalement of a single fiber cell (D, F). Also absent in the double-knockout lenses was any accumulation of neurobiotin in the overlying epithelium (F), which normally results from fiber-to-epithelial cell junctional transfer. ^{12 43} Thus, double deletion of Cx43 and Cx50 resulted in a decrease in epithelial coupling and a loss of epithelial–fiber communication in embryonic lenses.