Connexin Mediated Cataract Prevention in Mice

Cataracts, named for any opacity in the ocular lens, remain the leading cause of vision loss in the world. Non-surgical methods for cataract prevention are still elusive. We have genetically tested whether enhanced lens gap junction communication, provided by increased α3 connexin (Cx46) proteins expressed from α8(Kiα3) knock-in alleles in Gja8tm1(Gja3)Tww mice, could prevent nuclear cataracts caused by the γB-crystallin S11R mutation in CrygbS11R/S11R mice. Remarkably, homozygous knock-in α8(Kiα3/Kiα3) mice fully prevented nuclear cataracts, while single knock-in α8(Kiα3/−) allele mice showed variable suppression of nuclear opacities in CrygbS11R/S11R mutant mice. Cataract prevention was correlated with the suppression of many pathological processes, including crystallin degradation and fiber cell degeneration, as well as preservation of normal calcium levels and stable actin filaments in the lens. This work demonstrates that enhanced intercellular gap junction communication can effectively prevent or delay nuclear cataract formation and suggests that small metabolites transported through gap junction channels protect the stability of crystallin proteins and the cytoskeletal structures in the lens core. Thus, the use of an array of small molecules to promote lens homeostasis may become a feasible non-surgical approach for nuclear cataract prevention in the future.


Introduction
Cataracts, defined as any opacity in the lens, remain the leading cause of blindness in the world despite the success of surgical replacement with artificial lenses. A non-surgical method of cataract prevention remains an important and challenging topic. Age, mutated genes, radiation, smoking, chemical insults, physical injury and other systemic diseases can lead to cataract formation by initiating complicated pathological processes that result in abnormal protein aggregates and cellular disruptions in the lens [1,2]. The bulk of the lens consists of precisely organized elongated fiber cells that are coupled by intercellular gap junction channels for maintaining lens homeostasis [3,4,5]. Interior lens mature fiber cells have minimum metabolism without any intracellular organelles and mainly contain crystallin proteins [2,6]. Treatments using select small molecules, such as antioxidants, vitamins or ions, show conflicting results on the efficiency of cataract prevention in clinical trials [7,8,9,10,11]. This work has provided new evidence for the feasibility of cataract prevention via homeostasis regulation mediated by gap junction channels in animal models.
Crystallin proteins, classified as a, b and c groups, are the main structural components of a mammalian lens [2]. Cataract formation is directly associated with the stability and solubility of lens crystallin proteins [2,12,13,14]. Crystallin gene mutations are the most common cause of hereditary cataracts [15]. We have previously reported that cB-crystallin S11R mutation leads to a dominant congenital cataract, and homozygous Crygb S11R/S11R mutant mice develop severe nuclear cataracts regardless of the genetic background in the A/J, C57BL/6 and 129 strains [13].
Cataract formation was associated with abnormal degradation and aggregation of crystallins, disruption of membrane-cytoskeletal structures and the elevation of lens calcium level. Activated calcium-dependent proteases, such as calpains, are known to cleave crystallin proteins and to be associated with the degeneration of interior fiber cells in Gja3 tm1 knockout mice that lack intercellular gap junction communication consisting of a3 connexin (Cx46) proteins in the lens [16,17,18,19].
At least three members of the connexin gene family, a1 connexin (Cx43), a3 connexin (Cx46) and a8 connexin (Cx50), encoded by Gja1, Gja3 or Gja8, respectively, are utilized to form gap junction channels in the lens [5,20]. Connexin 23, encoded by Gjf1 and not verified to form gap junction channels, is also utilized in the embryonic lens [21,22]. Gap junction channels consisting of a3 connexin and/or a8 connexin are essential for the lens transparency while only a8 connexin is important for the lens growth [19,23,24]. Interestingly, the role of a8 connexin in lens transparency can be compensated by the knock-in a3 connexin in Gja8 tm1(Gja3)Tww mice [25]. Knock-in a3 connexins increase the cell-to-cell coupling between lens fiber cells [26]. We also reported that knock-in a3 connexin could suppress cataract formation caused by the a8 connexin G22R mutation in Gja8 Lop10 mice [27].

Preservation of lens interior fiber cells
The dense nuclear cataract caused by the cB-crystallin S11R mutation was associated with the deterioration of intracellular components in fiber cells of the lens core ( Figure 2). Enlarged extracellular spaces or disrupted fiber-to-fiber contacts were observed in deeper cortical fibers close to the degenerative mature fibers in the core of cB(S11R/S11R) lenses. In comparison, interior fiber cells of cB(S11R/S11R) a8(Kia3/Kia3) lenses appeared intact with appropriate cell organization without any histopathology observed in cB(S11R/S11R) lenses ( Figure 2).
Immunohistological staining results of F-actin and c-crystallin revealed that cB(S11R/S11R) lenses had substantial membraneassociated c-crystallin aggregates, a lack of F-actin and extremely low levels of cytosolic c-crystallin in interior fiber cells. But cB(S11R/S11R) a8(Kia3/Kia3) lenses showed normal distribution of F-actin and cyotosolic c-crystallin in interior fiber cells, similar to that of wild-type lenses ( Figure 3A). Punctuate a3 connexin staining was present at the plasma membranes of wild-type lens fiber cells ( Figure 3B). Although punctuate a3 connexin signals still appeared in altered fiber cells of cB(S11R/S11R) lenses, much stronger a3 signals were present in cB(S11R/S11R) a8(Kia3/Kia3) lens fiber cells ( Figure 3B). F-actin staining was absent in fiber cells from the cB(S11R/S11R) lens core, but F-actin signals appeared normal in cB(S11R/S11R) a8(Kia3/Kia3) lenses at the same age. Thus, knock-in a3 connexin protected the integrity of lens interior fiber cells probably by maintaining normal distribution and properties of F-actin and c-crystallin proteins.
Lens total calcium level was measured by an inductively coupled plasma-optical emission spectrometry (ICP-OES). In comparison to wild-type lenses, cB(S11R/S11R) lenses had substantially increased total calcium level while the calcium level in cB(S11R/ S11R) a8(Kia3/Kia3) lenses remained unchanged ( Figure 4B). All the differences were statistically significant (N = 5, P,0.004). Moreover, we examined calcium-dependent crystallin protein degradation and found that cleaved aA-, aB-, band c-crystallins Figure 2. Histology of P7 WT, cB(S11R/S11R) and cB(S11R/S11R) a8(Kia3/Kia3) lenses. Wide field view of lens sections are shown on the left panels. Scale bar, 100 mm. High magnification views of selective interior fiber cells of lens sections are shown on the middle panels (dashed boxes, about 500-580 mm distance from the lens capsule) and the right panels (solid boxes, about 800-880 mm from the lens capsule). Uniformly elongated and tightly packed interior fiber cells are present in wild-type lens sections (the top panels) while irregularly elongated and loosely packed interior fiber cells (indicated by arrowheads on the middle panel) and disintegrated fiber cells (indicated by arrows on the middle-right panel) appears in cB(S11R/S11R) lenses. However, cB(S11R/S11R) a8(Kia3/Kia3) lens section displays uniformly elongated and tightly packed interior fiber cells without noticeable disintegrated fiber cells in the lens core (the bottom panels). Scale bars, 20 mm. doi:10.1371/journal.pone.0012624.g002 were present in cB(S11R/S11R) lenses but were almost absent in cB(S11R/S11R) a8(Kia3/Kia3) lenses with a very small amount of cleaved aB-crystallin (aB-C1) and c-crystallin ( Figure 4C). These results suggest that the elevated level of a3 connexin suppressed calcium elevation to prevent the degradation of aand bcrystallins and nuclear cataract formation in cB(S11R/S11R) a8(Kia3/Kia3) lenses.

Discussion
We have demonstrated that elevated a3 connexin expression can prevent hereditary nuclear cataracts caused by the cBcrystallin S11R mutation. Morphological evidence indicates that the prevention of this nuclear cataract is correlated with the inhibition of inner fiber cell degeneration and the maintenance of membrane-cytoskeleton structures and crystallin stability and/or solubility. Presumably, enhanced exchange of metabolites or ions transported by gap junction channels composed by a3 connexin is responsible for preventing calcium level elevation, crystallin aggregation and degradation in the lens. Therefore, this work proves the principle that cataract prevention can be achieved by promoting intercellular gap junction communication in the lens.
Cataracts are ultimately associated with aggregation and/or degradation of crystallin proteins regardless of the primary causes, such as age, genetic disorders or radiation damage. Nuclear cataracts are associated with disturbed membrane-cytoskeletal structures as well as with crystallin protein degradation and/or modifications in the lens core. Our work supports the notion that nuclear cataract prevention may require an array of metabolites or reagents, transported by gap junction channels, to prevent aberrant changes in lens proteins and to maintain lens homeostasis needed for lens transparency [29]. Calcium homeostasis is known to play important roles in the regulation of lens transparency [16,17,18]. The two main mammalian calpains, 1 and 2, are heterodimers of a large 80 kDa subunit and a small 28 kDa subunit that together bind multiple calcium ions during enzyme activation. Calpains seem to be important in the regulation of lens transparency and cataract prevention [18,30,31]. The activation of calpains in mouse lens is known to be related to the cleavages of aAand aB-crystallins [32]. Presumably, the level of cleaved forms of aAand aB-crystallins was dependent on the degree of calpain activation in the lens. Mutant cB-crystallin S11R proteins induce abnormal protein aggregation that probably disrupts membranecytoskeleton structures of inner fiber cells. Subsequently, increased calcium influx and activation of calcium-dependent protein degradation lead to the degeneration of inner mature fiber cells and a dense nuclear cataract. Probably by enhancing cell-cell communication, knock-in a3 connexin prevents calcium level elevation and the activation of calpains, which in turn prevents the degradation of aand b-crystallins in the lens. The cleavage of c-crystallin in connexin a3(2/2) knockout mice is related to the activation of calpain 3 (or Lp82 protease) [17]. However, the protease that is involved in the cleavage of c-crystallin in Crygb S11R/S11R mutant mice remains unknown.
This work demonstrates that it is feasible to prevent nuclear cataract formation by directly promoting lens homeostasis via increased gap junction communication. The mechanism for how a3 connexin leads to a dosage-dependent inhibition of nuclear cataract induced by cB-crystallin S11R mutation needs to be further investigated. It is very interesting that the presence of a low level of a3 connexin, expressed from one copy of the knock-in a8(Kia3) allele, shows incomplete suppression of this nuclear cataract. This result supports the observations of previous studies that other genetic modifiers play a significant role in the severity of nuclear cataracts when gap junction communication is insufficient in the lens and that the heterogeneity of nuclear cataracts was associated with the 129 and C57B6 mouse strain backgrounds [33,34]. Thus, identification of any of these genetic modifiers will be valuable for understanding the mechanisms of connexinmediated cataract formation and prevention. It is possible that the level or the amount of small metabolites or ions that pass through gap junction channels composed of a3 connexin is the key to understand a3 connexin-mediated prevention of nuclear cataracts. Although it is difficult to increase the numbers of gap junction channels in any mature lenses, it may be practical and effective to supplement key metabolites and ions transported through existing gap junctions to enhance lens homeostasis to prevent or delay age-related cataracts. This notion is supported by recent studies that vigorous physical activities, which elevate systemic metabolism, reduced the risk for agerelated cataracts as well as age-related macular degeneration [35]. Thus, the identification of an array of small molecules, transported by gap junction channels and needed for lens transparency, will be the next important step not only for understanding the lens Figure 5. Correlation between cleaved crystallin proteins and the severity of nuclear cataracts. (A) Lens photos from three cB(S11R/S11R) a8(Kia3/2) a3(2/2) littermates at the age of one month. Scale bar, 1 mm. (B) Western blotting images of lens total aA-, aB-, band c-crystallins. Arrowheads indicate the cleaved forms of crystallins. Two major cleaved forms of aAand aB-crystallins are marked as aA-C1, aA-C2, aB-C1, and aB-C2, respectively. doi:10.1371/journal.pone.0012624.g005 homeostasis but also for developing a non-surgical method for cataract prevention [8].

Generation and genotyping of compound and triple mutant mice
Mouse care and breeding were performed according to an animal protocol (protocol#: R280-1210) approved by the Animal Care and Use Committee at University of California, Berkeley and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Mouse pupils were dilated with 1% atropine and 1% phenylephrine before the eyes were examined for lens clarity by a slit lamp.
The Crygb S11R mutant and wild-type Crygb alleles can be genotyped by using PCR with satellite marker D1Mit156 (left: TCTGCTGCCACTTCTGAGAA; right: TGTGTGTCTATG-GACATGGATG). The Crygb S11R mutant allele originated from the AJ strain background displayed an 112 bp PCR fragment while the wild-type Crygb allele from the C57BL/6J strain background yielded a 143 bp PCR fragment. The genotyping of Gja3 tm1 and Gja8 tm1 mutant alleles was assessed by PCR as described previously [19,24].

Fresh lens imaging and light scattering quantification
Fresh lenses, dissected from enucleated eyeballs, were immediately immersed in PBS at 37uC and imaged under a Leica MZ16 dissecting scope using a digital camera. The intensity of light scattered by lenses was measured using the HR 2000CG_UV_ NIR High Resolution Spectrometer and a QP400-2-UV-VIS fiber optic cable (Ocean Optics, Dunedin, FL, USA). Lenses were illuminated by a white light source perpendicular to the equator of the lens. Intensity of scattered light was captured by the optical fiber with a whole acceptance of angle of 24.8u. Spectrums were recorded and saved for later comparison. Each lens was measured twice in succession to show repeatability. The intensity of the illuminating light was kept constant between measurements. The measurements were represented as graphs with wavelength on the x-axis and intensity of scattered light on the y-axis. Data were stored as ASCII files, and the area under the curve was calculated by a Matlab program. Student's t-test was used for statistical analysis. P values ,0.001 were considered significant.

Histology
Enucleated eyeballs opened at the anterior chamber or posterior vitreous were immersed in a fixative solution containing 2% glutaraldehyde and 2.5% formaldehyde in 0.1 M cacodylate buffer (pH 7.2) at room temperature for 5 days. Samples were postfixed in 1% aqueous OsO4 and then dehydrated through graded acetone. Samples were embedded in eponate 12-araldyte 502 resin (Ted Pella, Redding, CA, USA). Lens sections (1 mm thick) across the equatorial plane were collected on glass slides and stained with toluidine-blue. Bright-field images were acquired using a light microscope (Axiovert 200; Carl Zeiss, Oberkochen, Germany) with a digital camera.

Immunohistochemistry and confocal microscopy
Mouse eyes were fixed with fresh 4% formaldehyde in phosphate-buffered saline (PBS) for 30 minutes, then washed with PBS and soaked overnight in 30% sucrose in PBS. Afterward, the samples were processed and sectioned with a Cryostat 1900 (Leica, Germany) using a standard frozen-section method [19]. Tissue sections were washed with three exchanges of sterile PBS (10 minutes each time), followed by blocking (3% BSA, 3% NGS, 0.01% Triton X-100 in PBS) for 1 h at room temperature. Antigens were labeled using primary antibodies (see detail information in western blot analysis) for 1 hour at room temperature (or overnight at 4uC). After three more washes with sterile PBS, samples were incubated with secondary antibodies (Invitrogen, Carlsbad, CA, USA) for 1 hour at room temperature. Slides were then mounted with Vectashield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA, USA) after washing in PBS. The distribution of antigens was analyzed by a laser confocal microscope (Leica, Wetzlar, Germany).

Western blot analysis
Enucleated fresh lenses were weighed and homogenized directly in the sample buffer (60 mM Tris, pH 6.8, 2% SDS, 10% glycerol, 5% b-mercaptoethanol, and 0.001% bromophenol blue) to prepare lens total proteins. To prepare the NaOH-insoluble protein samples, two lenses were homogenized in 0.1 M NaCl and 50 mM Na 2 HPO 4 (pH 7) at the ratio of 40 mg lens wet weight/ml solution. The insoluble material was collected after centrifugation at 15,000 rpm for 15 min and washed once with the same solution. The insoluble pellet was further homogenized in 0.5 ml of 20 mM NaOH and 1 mM NaHCO 3 solution. Again, an insoluble pellet was collected after centrifugation at 15,000 rpm for 15 min and washed once with 20 mM NaOH and 1 mM NaHCO 3 solution. This insoluble pellet was dissolved in sample buffer. Equal volumes of samples were loaded on a 12.5% SDS-PAGE gel for separation, and separated proteins were transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA). Lens crystallin and connexin proteins were detected by Western blotting with rabbit polyclonal antibodies against aand c-crystallins (generously provided by Dr. Joseph Horwitz at the University of California, Los Angeles), b-crystallin (generously provided by Dr. J. Samuel Zigler at the National Eye Institute), a3 connexin (generated previously in the lab), a8 connexin (generously provided by Dr. J. Mario Wolosin Mount Sinai School of Medicine, New York) and a mouse monoclonal antibody against b-actin (Sigma, St. Louis, MO, USA). More than three sets of lens protein samples from different mice were examined, and representative data were shown.

Ion concentration measurement
Lens ion concentration was measured by inductively coupled plasma-optical emission spectrometry (ICP-OES) from a core facility at the University of California, San Diego. The method was described previously [18]. Twenty lenses were dissected from each mouse line and then immediately subjected to vacuum drying for 48 hours. Dry lenses were weighed and solubilized in 500 mL nitric acid (33.5,35%; Fisher Scientific, Pittsburgh, PA, USA) for 12 hours at 37uC with shaking and then diluted with water into 3 ml. Samples were further diluted to reach ion concentrations of 20 ppb to 1 ppm for measurement. According to the estimated sample ion concentration, a series of standards were made. Ion concentrations were determined by their intensities acquired by the instrument. Measurement error was approximately 3%. Ion concentrations were normalized by lens dry weight.