The Stony Brook University Institutional Animal Care and Use Committee approved all animal experimentation. Mice with a global knockout of Cx50 (17) and lens-specific conditional knockouts of the p110α catalytic subunit of PI3K (16), or PTEN (60), were mated to generate double knockout animals (p110/Cx50α dKO, or PTEN/Cx50 dKO). Briefly, Cx50 KO and p110α floxed (65) mice were interbred to homozygosity. Cx50 KO and PTEN floxed mice (66) were also interbred to homozygosity. MLR10-Cre mice, whose Cryaa driven Cre expression is limited to the lens epithelium and fibers (67), were interbred with both strains to heterozygosity, maintaining the homozygous floxed p110α and PTEN alleles, in addition to the homozygous Cx50 KO alleles. Animal genotypes were confirmed by PCR of DNA from tail biopsies as previously described (16, 17, 60). Due to the mixed genetic background that resulted from the interbreeding of the different strains of original mice, littermate controls (Cx50 KO Cre negative) were used for all experiments. All animals used for breeding had homozygous Cx50 KO and homozygous floxed alleles for either PTEN or p110α, while one member of the breeding pair carried a heterozygous copy of the MLR10-cre transgene. The resulting litters contained half of the animals with no Cre (Cx50 KO controls) and half with a double knockout of p110α/Cx50, or a double knockout of PTEN/Cx50 (dKOs).

Littermate mice between birth and 24 weeks of age were euthanized by CO₂ asphyxia and weighed. Eyes were dissected, weighed, and transferred to 37°C Tyrode solution on a warmed stage. Lenses were removed, transferred to 35mm glass bottom culture dishes and photographed with a SZX16 dissecting microscope attached to a digital camera (Olympus, Waltham, MA). Diameters of lenses were measured from the images, and lens volume was calculated assuming a spherical shape. The incidences of the presence of blood vessels associated with the lens and lens rupture were also recorded (16, 17, 60).

Postnatal day 0 (P0) or postnatal day 2 (P2) mice were subcutaneously injected with 50 µg/gm EdU (Click-iT, Thermo Fisher Scientific, Waltham, MA) and then returned to their mothers for 2 hours. Subsequently, lenses were removed and fixed in a 4% formaldehyde in PBS for 1 hour at room temperature (~22°C) and photographed for measurement. Permeabilization and Click-iT staining were performed as described in the manufacturer’s instructions and previously published protocols (16, 68, 69). Z-stacks of fluorescent images were acquired on an Axiovert 200M microscope (Zeiss, Thornwood, NY) and processed using ImageJ. For line-scan analysis, the flattened fluorescent image of EdU staining was manually thresholded, and the plot profile function in ImageJ was used to measure fluorescent intensity across the entire lens diameter as described previously (16, 68). For central epithelial cell counts, a circular region with a diameter equal to the lens radius was drawn over the central region, fluorescent values were thresholded using the color threshold fuction in ImageJ, and EdU stained cells were quantified using the analyze particles function in ImageJ. The number of labeled cells was divided by the circular area to calculate the density of labeled cells.

Eyes from postnatal day 2 mice were dissected and fixed in 4% formaldehyde in phosphate-buffered saline (PBS) overnight at room temperature. Eyes were rinsed with PBS, dehydrated through increasing ethanol concentrations (50, 70, 70, 80, 95, 100%), and then embedded in paraffin. 2.5 µm sections were cut using a diamond knife, mounted on glass slides and deparaffinized. Slides were stained with hematoxylin-eosin as previously described (16, 60), and histological sections were observed on a BX51 microscope and photographed with a DP72 digital camera (Olympus).

Data were plotted as the mean ± SD, or SEM, as described in each figure. Statistical significance was determined using one-way ANOVA in the Origin software program (OriginLab Corporation, Northampton, MA) with Tukey’s post hoc test. P values less than 0.05 were considered significant.

Cx50 KO mice (17) containing homozygous floxed p110α, or PTEN alleles (65, 66) were interbred with MLR10-Cre transgenic mice (67) to obtain double knockout animals. All comparisons in this manuscript are made between animals lacking Cx50. As described in the introduction, loss of Cx50 reduced postnatal lens cell proliferation, causing deficient lens growth, micropthalmia and cataract (13, 15, 17, 35). Please see Table 1 , and the original knockout papers (16, 17, 60) for a comparison of the single KO mouse phenotypes with wild-type lenses. Lenses isolated from homozygous floxed p110α/Cx50, or PTEN/Cx50 animals lacking the MLR10-Cre transgene (i.e. Cx50 KO) remained intact at all ages examined ( Figures 1A–D ), with a mild nuclear cataract as previously described (70). By contrast, lenses dissected between 1 and 24 weeks of age from homozygous floxed p110α/Cx50 animals expressing the MLR10-Cre transgene (i.e. p110α/Cx50 dKO) showed a distinct dual pathology. Intact lenses formed the first group, which were noticeably smaller than their Cx50 KO littermates, and no longer displayed the mild nuclear cataract phenotype of the Cx50 KO ( Figures 1E–H ). The second group of p110α/Cx50 dKO lenses displayed a partial posterior rupture, leaving a small transparent anterior lens fragment attached to posterior lens debris ( Figures 1I–L ). Lenses dissected between 1 and 24 weeks of age from homozygous floxed PTEN/Cx50 mice expressing the MLR10-Cre transgene (i.e. PTEN/Cx50 dKO) showed a progressive pathology that also included lens rupture. A dense central cataract, the presence of vacuoles, and progressively worsening cortical opacities, characterized the PTEN/Cx50 dKO lenses that remained intact ( Figures 1M–P ). Lens rupture was not detected at 1 week of age for PTEN/Cx50 dKO lenses, however from 5 weeks of age onward, ruptured lenses were observed with an increasing frequency leaving behind a dense nuclear cataract attached to cortical fragments ( Figures 1Q–T ). Thus, lens size, clarity, and integrity were all differentially impacted by the double deletion of Cx50 with p110α, or PTEN. p110α/Cx50 dKO lenses were prone to lens rupture, however, intact p110α/Cx50 dKO lenses were much smaller than Cx50KOs, and had lost the central cataract. PTEN/Cx50 dKO lenses had more severe cataracts that either Cx50, or PTEN single KOs, and displayed frequent lens rupture.

Previous studies reported that individual knockouts of either Cx50, or p110α, resulted in the lens volume being reduced by 44% and 27%, respectively with a corresponding decrease in eye mass (13, 16, 17). By contrast, single knockout of PTEN resulted in a small increase in lens volume and eye mass (60). To determine the effect of double knockout of Cx50 and p110α or PTEN, changes in animal mass, eye mass and lens volume were quantified over time by weighing mice, weighing eyes, and photographing lenses between postnatal day 0 (P0) and 24 weeks of age in Cx50 KO, p110α/Cx50 dKO and PTEN/Cx50 dKO animals. Lens diameters were measured, and then used to calculate lens volumes as previously described (16, 17, 60, 68). There were no significant differences in body mass between Cx50 KO, p110α/Cx50 dKO and PTEN/Cx50 dKO animals between birth and 24 weeks of age ( Figure 2A ). By contrast, eye mass in the p110α/Cx50 dKO was significantly reduced (p < 0.05, one-way ANOVA) compared to Cx50 KO and PTEN/Cx50 dKO mice ( Figure 2B ). At 1 week old, the mass of p110α/Cx50 dKO eyes was reduced 23% and by 12 weeks of age, the eyes were 40% smaller than Cx50 KO mice. By contrast, eyes from PTEN/Cx50 dKO animals were 8-15% larger than Cx50 KO littermates at 5 and 12 weeks of age (p < 0.05). The eye mass data includes all eye samples, regardless of lens rupture status. Similar to eye mass, lens volume in the p110α/Cx50 dKO animals was significantly reduced (p < 0.05) compared to either the Cx50 KO or PTEN/Cx50 dKO ( Figure 2C ). At P0, the volume of p110α/Cx50 dKO lenses was reduced 35% and at 12 weeks of age the lenses were 37% smaller than those of Cx50 KOs. The lenses from PTEN/Cx50 dKO animals were 8-18% larger than Cx50 KO littermates between 12 and 24 weeks of age (p < 0.05). The lens volume data was calculated from lenses that had not ruptured. These data show that the eye and lens growth deficiencies induced by individual knockout of Cx50 or p110α were compounded in the p110α/Cx50 dKO. The modest effect of single PTEN KO on lens and eye size (60) was not notably altered in the PTEN/Cx50 dKO animals.

Previous studies have shown that single Cx50, or p110α, KO lenses never displayed lens rupture (16, 17, 68). By contrast, single PTEN KO lenses displayed an increasing propensity to rupture with age (60). To quantify changes in lens integrity over time, dissected lenses from Cx50 KO, p110α/Cx50 dKO and PTEN/Cx50 dKO animals between birth and 24 weeks were scored for the incidence of lens rupture ( Figure 3 ). Reduced viability or anophthalmia were never observed in either dKO strain. Lenses from p110α/Cx50 dKO mice were intact between P0 and P2, but displayed 65% to 81% rupture between 1 week and 24 weeks of age. Lenses from PTEN/Cx50 dKO mice were intact between P0 and 1 week of age. By 5 weeks of age, PTEN/Cx50 dKO animals had an 18% incidence of lens rupture, which increased to 32% at 12 weeks, and 73% at 24 weeks of age. These data show that loss of either p110α or PTEN from Cx50 KO lenses produced a loss of lens integrity and high incidences of lens rupture. However, the kinetics of lens ruptured differed dramatically between the p110α/Cx50 dKO and PTEN/Cx50 dKO mice.

In the early mouse eye, the hyaloid artery grows out from the optic nerve head ensheathing the lens as the tunica vasculosa lentis (71, 72). This structure persists during the first postnatal week before regressing prior to eye opening on P14 (73, 74). In PTEN/Cx50 dKO mice, this process appeared to be compromised in a subset of lenses dissected between P2 and P7. All of the postnatal lenses from Cx50 KO and p110α/Cx50 dKO, and the majority of lenses from PTEN/Cx50 dKO animals could be easily dissected away from the tunica vasculosa on P2 ( Figures 4A–C ). However, 29% (n = 14) of the P2 lenses ( Figures 4D–F ) and 13% (n = 16) of the P7 lenses ( Figures 4G–I ) from PTEN/Cx50 dKO could not be separated from highly adherent/penetrant blood vessels that presumably originated from the tunica vasculosa lentis. This phenomenon was not previously observed in wild-type, single Cx50 KO, or single PTEN KO eyes (13, 17, 60).

The pattern of mitosis in p110α single KO lenses was altered on P0, with a large reduction in cell division in the equatorial germinative zone of the lens epithelium (16, 68). To examine postnatal mitosis in p110α/Cx50 mice, lenses from Cx50 KO and p110α/Cx50 dKO mice were labeled with EdU on P0. Cx50 KO lenses showed strong EdU labeling, with the highest level of fluorescence in the germinative zone near the lens equator ( Figures 5A, B ). By contrast, p110α/Cx50 dKO lenses lacked the ring of increased labeling near the equatorial germinative zone, and showed a more uniform pattern of EdU incorporation across the lens epithelium ( Figures 5D, E ). Line scans of the mean fluorescent intensity taken across the lens diameter (black lines, ± the SEM green bars) confirmed this change in mitotic pattern between EdU labeled images from P0 Cx50 KO ( Figure 5C ) and p110α/Cx50 dKO lenses ( Figure 5F , arrowheads). All P0 Cx50 KO lenses displayed large peaks of EdU fluorescent intensity in the equatorial germinative zone (n = 5). By contrast, all P0 p110α/Cx50 dKO lenses lacked the equatorial peaks, and had maximum mean values of fluorescence that ranged from 35% to 39% lower than Cx50 KO lenses (n =4). These data show that further deletion of Cx50 from p110α knockout lenses did not restore normal P0 epithelial cell proliferation.

Single PTEN KO lenses were previously shown to display robust postnatal EdU labeling with a pattern similar to wild-type animals with the highest level of fluorescence in the germinative zone near the lens equator (68). By contrast, single Cx50 KO lenses previously showed decreased DNA replication on P2, especially within the central anterior epithelium (15). To examine mitosis on P2 in in p110α/Cx50 and PTEN/Cx50 dKO animals, lenses from Cx50 KO, p110α/Cx50 dKO and PTEN/Cx50 dKO mice were labeled with EdU. On postnatal day 2, Cx50 KO ( Figures 6A, B ), p110α/Cx50 dKO ( Figures 6E, F ), and PTEN/Cx50 dKO ( Figures 6I, J ) lenses all displayed the greatest levels of proliferation in the equatorial germinative zone. Similar to p110α KO mice (16), P2 p110α/Cx50 dKO lenses had recovered the deficit of proliferation in this region that was evident on P0. The mean values of line scans (black lines, ± the SEM green bars) of EdU fluorescence taken across the lens diameter confirmed that the p110α/Cx50 dKO lenses had recovered peak proliferation in the equatorial germinative zone on P2 ( Figure 6H ). All three of the Cx50 deficient lenses also showed decreased EdU incorporation in the central lens epithelium ( Figures 6M–O ), consistent with prior studies (15, 35), with the lowest levels observed in the Cx50 KO and PTEN/Cx50 dKO animals ( Figure 6P ). The Hoechst staining of nuclei in the PTEN/Cx50 dKO lenses ( Figures 6J, K ) further illustrated a loss of organization of nuclei in the equatorial bow region, compared with the dense band of nuclear staining seen in both Cx50 KO and p110α/Cx50 dKO mice where new fiber cells become internalized ( Figures 6B, C, F, G , white arrowheads). Lens nuclei in the P2 PTEN/Cx50 dKO animals failed to internalize in the bow region and were diffusely distributed and displaced toward the posterior pole. These data suggest that the independent additive effects of transient loss of proliferating epithelial cells in the germinative zone on P0 and in the central zone on P2, as seen previously in the single p110α and Cx50 KO mice respectively could cause the significant reduction in lens size following p110α/Cx50 double deletion. They further show that double deletion of PTEN/Cx50 disrupted the spatial organization of lens nuclei.

To compare the histology of Cx50 KO, p110α/Cx50 dKO and PTEN/Cx50 dKO lenses, eyes from mouse pups on P2 were dissected, fixed, sectioned and stained with hematoxylin and eosin. Sagittal sections through the central region of Cx50 KO lenses ( Figure 7A ) contained a central zone where the lens fibers displayed reduced eosinophilic staining as previously described (17). The bow region of Cx50 KO lenses was normal ( Figure 7B ), although there was evidence of delayed denucleation of differentiated fiber cells as formerly reported (13). Consistent with the absence of nuclear cataract shown in Figure 1 , the p110α/Cx50 dKO lenses lacked the central zone of reduced eosinophilic staining ( Figures 7C, D ). In addition, p110α/Cx50 dKO lenses often displayed posterior defects that may represent early stages of lens rupture. By contrast, PTEN/Cx50 dKO lenses showed severe histological anomalies ( Figures 7E, F ). The bow region failed to form properly, with shortened lens fibers and their associated nuclei displaced toward the posterior lens, consistent with the Hoechst staining described above ( Figures 6J, K ). This equatorial area also contained large vacuoles, and additional histological defects were present in the anterior and posterior regions of PTEN/Cx50 dKO lenses. Thus, double deletion of PTEN/Cx50 from the lens induced several histological defects in the lens.