Molecular Vision 2019; 25:535-545
Received 24 April 2019 | Accepted 03 July 2019 | Published 03 October 2019
Peter J. Minogue, Eric C. Beyer, Viviana M. Berthoud
Department of Pediatrics, University of Chicago, Chicago, IL
Correspondence to: Viviana M. Berthoud, Department of Pediatrics, University of Chicago, 900 East 57th St., KCBD-5150, Chicago, IL 60637; Phone: (773) 834-2115; FAX: (773) 834-1329; email: email@example.com
Purpose: CCAAT/enhancer-binding homologous protein (CHOP), a transcription factor that has been implicated in differentiation, apoptosis, and autophagy, is greatly elevated in lenses with cataracts due to mutations of several different lens proteins. To test the possible role of CHOP in the cataractous lens, we studied the effect of knocking out Chop in mice that were homozygous for the Cx50D47A mutation of the lens fiber gap junction protein connexin50 (Cx50).
Methods: Mouse lenses were examined by dark-field microscopy. Lens equatorial diameters and intensities of the opacities were quantified using ImageJ. Transcript levels were assessed by real-time quantitative PCR. Protein levels were determined by immunoblotting.
Results: Homozygous Chop knockout lenses were transparent. Deletion of Chop in Cx50D47A mice did not improve lens transparency and had no effect on lens size. In Chop null-Cx50D47A lenses, the protein kinase R–like endoplasmic reticulum kinase (PERK)-dependent pathway was activated similarly to Cx50D47A lenses. In Cx50D47A mice, Chop deletion did not improve connexin levels or lens fiber cell differentiation, and it did not decrease the levels of Trib3 or Irs2 transcripts to wild-type values. However, homozygous Chop knockout significantly diminished the increased levels of Cebpb transcripts of Cx50D47A lenses.
Conclusions: The results show that CHOP is not required for lens transparency. They also suggest that CHOP is not the critical etiological factor for the cataracts observed in homozygous Cx50D47A lenses, further supporting a major role for connexins in the disease.
Congenital cataracts are a major cause of visual impairment and blindness in infants and young children (reviewed in ). Often, they result from mutations in different genes, including those encoding crystallins, transmembrane proteins, transcription factors, and extracellular matrix proteins (compiled in Cat-Map) . Among the transmembrane proteins, mutations in the lens fiber cell gap junction proteins, connexin46 (Cx46) and connexin50 (Cx50), are common. We have been studying mice expressing one such mutant, Cx50D47A, as a prototype of connexin-linked cataracts. Both heterozygous and homozygous Cx50D47A mice develop cataracts [3-5]. The lenses of these mice are small, and fiber cell differentiation is impaired . In the mutant lenses, the protein kinase R–like endoplasmic reticulum kinase (PERK) transducing pathway of endoplasmic reticulum (ER) stress is activated . This response is most severe in homozygotes, as shown by phosphorylation of eukaryotic translation initiation factor‐2A (EIF2α) and increased protein levels of activating transcription factor 4 (ATF4) and its downstream target, CCAAT/enhancer-binding homologous protein (CHOP). We hypothesized that persistent activation of this pathway may contribute to the impaired differentiation and cataract formation in Cx50D47A mice. CHOP is also significantly increased in lenses containing cataracts, including those caused by mutations of other genes [7-10], suggesting that CHOP may be a general critical factor that contributes to these abnormalities.
CHOP (also known as DNA damage inducible transcript 3, DDIT3; C/EBPζ; and growth arrest and DNA damage protein, GADD153) is a transcription factor that can be induced by physiological conditions and a wide variety of cellular stresses (including ER stress; reviewed in [11,12]). Studies in various cell types suggest that CHOP has a critical role in the induction of cell cycle arrest and apoptosis in response to stress. CHOP has been implicated in regulating apoptosis, autophagy, and cell differentiation (reviewed in ). It can dimerize with other transcription factors and act as a negative or positive regulator of transcription, depending on its transcription factor partner, the cell type, and the stress condition (; reviewed in ). Our prior investigation of Cx50D47A lenses showed large increases in some transcripts that could result from CHOP-mediated transcriptional activity (including Trib3). Therefore, we sought to investigate the importance of CHOP in the lens by studying the effects of deleting CHOP from wild-type mice and homozygous Cx50D47A mice.
KCl, Ponceau S, 1,2-Bis(dimethylamino)ethane (TEMED), 2-mercaptoethanol, Trizma base and glycine were obtained from Sigma-Aldrich (St. Louis, MO). NaCl, KH2PO4, Na2HPO4*7 H2O, ethylenediaminetetraacetic acid disodium salt dihydrate, ProtoGel, sodium dodecyl sulfate and methanol were obtained from Thermo Fisher Scientific (Pittsburgh, PA). Ammonium persulfate was obtained from Bio-Rad (Hercules, CA). ProSieve QuadColor™ protein markers (4.6 kDa – 300 kDa) were obtained from Lonza Rockland, Inc. (Rockland, ME).
Mouse monoclonal anti-CHOP (9C8) antibody (MA1–250; lot # QG214389) was obtained from Thermo Fisher Scientific. Rabbit monoclonal anti-EIF2α (D7D3) (5324P; lot # 3), anti-P-EIF2α (Ser51) (D9G8) XP (3398P; lot#2), anti-histone H3 (D1H2) XP (4499S; lot # 9), and anti-calreticulin (CALR) (DE36) (12238S; lot # 4) antibodies were obtained from Cell Signaling Technology (Danvers, MA). Rabbit polyclonal anti-translocase of outer mitochondrial membrane 20 (TOM20) antibodies (FL-145; sc-11415; lot # D1613) were obtained from Santa Cruz Biotechnology (Dallas, TX). Affinity-purified rabbit polyclonal anti-Cx46 intracellular loop and anti-Cx50 C-terminus antibodies have been previously described [3,15]. Horseradish peroxidase–conjugated goat AffiniPure anti-rabbit immunoglobulin G (IgG; H+L; 111–035–144; lot # 123520 and 135724) and horseradish peroxidase–conjugated goat AffiniPure F(ab')2 fragment goat anti-mouse IgG, F(ab')2 fragment-specific (115–036–072; lot # 111686 and 135721) antibodies were obtained from Jackson ImmunoResearch (West Grove, PA).
Cx50D47A (No2, ENU-326) mice (originally identified by Favor by screening for the cataract phenotype following ethylnitrosourea (ENU-326) mutagenesis ) were maintained as previously described . Wild-type C57BL/6J mice (# 000664) and homozygous Chop knockout mice (# 005530) were obtained from the Jackson Laboratory (Genetic Resource Science at the Jackson Laboratory. 2008. Expression/Specificity Patterns of Cre Alleles, 2008 Direct Data Submission from Genetic Resource Science: MGI: J:137887). Lenses from some of these mice (on a C57BL/6J background) were examined between 7 and 8 months of age. The Chop knockout mice were bred into the C3H line for six generations before performing the experiments. Heterozygous Chop knockout (Chop+/−) mice were mated with homozygous Cx50D47A (Cx50D47A/D47A) mice to generate heterozygous Chop knockout mice that were homozygous for the Cx50 mutation (Chop+/−-Cx50D47A/D47A). Mice of this genotype from different litters were mated to obtain homozygous Cx50D47A mice that were wild type for Chop or heterozygous or homozygous for the Chop deletion. All the animal procedures followed the University of Chicago Animal Care and Use Committee guidelines and were conducted in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research and National Institutes of Health (NIH) regulations.
Dark-field photomicrographs of lenses from 1-month-old mouse littermates containing all the genotypes (referred to as a set) were obtained using a Zeiss Stemi-2000C dissecting scope (Carl Zeiss, München, Germany). The same settings (i.e., magnification, illumination, and exposure time) were used to photograph all lenses in each set. Cataracts (opacities) were quantified in photomicrographs by integrating the gray values over a circular region of interest using NIH ImageJ software, as previously performed . The circular region used to analyze all images was identical in size. These data are reported in arbitrary units. The dimensions of the equatorial diameters were determined from the photomicrographs using NIH ImageJ software. The opacities and the diameters were measured in both lenses of each mouse (unless one of the lenses was damaged during dissection); the reported values are the averages of the measurements of the two lenses for each mouse.
Real-time quantitative PCR (RT-qPCR) was performed similarly to our previous study . Lenses from 1-month-old homozygous Cx50D47A mouse littermates of different Chop genotypes were homogenized in QIAzol Lysis Reagent (Qiagen, Valencia, CA) using a glass–glass homogenizer, and the RNA was purified using the miRNeasy Mini Kit (Qiagen). The concentrations of RNA were determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific) and the quality of the RNA was analyzed on an Agilent Bio-Analyzer (Agilent Technologies, Santa Clara, CA). cDNA was synthesized using the QuantiTect Reverse Transcription Kit (Qiagen). RT-qPCR was performed on a 7500 Fast Real-time PCR System (Applied Biosystems, Foster City, CA) using a cDNA aliquot, a primer set for the RNA of interest, and Fast SYBR® Green Master Mix (Life Technologies, Grand Island, NY). The primer sequences are listed in Table 1. These primer sets had efficiencies within a 95% confidence interval, and they did not produce primer dimers as assessed by the melting curves. All reaction mixes contained equal amounts of RNA. Each sample was run in triplicate (i.e., technical replicates) for each primer set. Each variable was studied using at least three independent sets of littermates (i.e., biological replicates). The transcript for cyclophilin A was used as the internal control to normalize the relative levels of RNA among the different genotypes. The spliced/total Xbp-1 (sXbp-1/tXbp-1) ratio was calculated by dividing the 2-ΔCT of spliced Xbp-1 by the 2-ΔCT of total Xbp-1, where ΔCT = CT for spliced or total Xbp-1 in the sample – CT for cyclophilin A in that sample. Graphs were prepared using SigmaPlot 10.0 (Systat Software, Inc., San Jose, CA).
Whole-lens homogenates from 1-month-old homozygous Cx50D47A mice that were Chop wild-type or heterozygous or homozygous Chop knockout were prepared by homogenizing the two lenses from each mouse in PBS containing 4 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 20 mM NaF, 10 mM Na3VO4, pH 7.4 and cOmplete Mini EDTA-free Protease Inhibitor Cocktail (Roche Applied Science, Indianapolis, IN) using a glass–glass homogenizer followed by sonication. Protein concentrations were determined using the Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad) based on the Bradford method . Aliquots from lens homogenates containing equal amounts of total protein were loaded in each lane and resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The gels were then blotted onto Immobilon P membranes (Millipore, Bedford, MA) using a wet transfer apparatus. Then, the membranes were stained with Ponceau S (to verify equal electrotransfer of the proteins) before being subjected to immunoblotting as previously described [6,18]. The X-ray films were scanned on a flat-bed scanner (Epson Perfection V700 Photo; Epson America, Long Beach, CA) to obtain a digital image for quantification of the bands. The bands obtained in at least three independent experiments (i.e., biological replicates) were quantified by densitometry using a rectangular box that encompassed the immunoreactive band(s) using Adobe Photoshop CS3 (Adobe Systems Inc., San Jose, CA). The box size used to quantify the immunoreactive band was kept constant for each set of littermates containing all three genotypes. The same size box was used to obtain a background value in each lane and correct the integrated band density (Photoshop Extended: Evaluation of gel images; Photoshop Extended: Performing a comparative analysis of bands; Using Photoshop CC to quantify band intensity in a DNA gel). The results are reported in arbitrary units. Graphs were prepared using SigmaPlot 10.0 (Systat Software, Inc.).
The raw data obtained from heterozygotes and homozygotes were compared to the raw data obtained from wild-type (or control) littermates to assess statistical significance using the Student t-test. A p value <0.05 was considered significant. The number of sets of littermates containing all genotypes (statistical “n”) was at least three for each type of data presented.
Germline deletion of Chop has been reported to induce no major phenotypic effects on mouse organs, although histomorphometric analysis showed a decreased bone formation rate at 1–12 months of age [19-21]. However, the lens was not specifically examined in these studies. To test whether expression of Chop was required for lens transparency in normal mice, we studied the lenses from heterozygous and homozygous Chop null mice and compared them with those of wild-type animals. Microscopic examination under dark-field illumination showed that the lenses from 1-month-old heterozygous and homozygous Chop null mice were transparent and appeared indistinguishable from those of their wild-type littermates (Figure 1A). At 2 months of age, lenses from wild-type and Chop null heterozygous and homozygous littermates were almost completely transparent and appeared identical, but they had a few of the typical micro-opacities of this mouse line (not shown). Moreover, wild-type and homozygous Chop null lenses from mice of the C57BL/6J background appeared similar and mostly transparent at >7 months of age (not shown).
We found a small variation in the equatorial diameters of wild-type lenses from different litters (Figure 1B). Similar variations were found in the lens diameters of heterozygous and homozygous Chop knockout mice. However, we consistently found that the equatorial diameters of homozygous Chop null lenses were slightly larger (1.7%) than those of their wild-type littermates, while the lens diameters from heterozygous Chop knockout mice were not significantly different from those of the wild-type littermates (Figure 1B). There were no detectable differences in lens weights.
To elucidate whether increased CHOP played a unique role for the cataract phenotype of homozygous Cx50D47A lenses, we generated and studied litters of mice that were homozygous for the Cx50D47A mutation and contained all three Chop genotypes (wild-type and heterozygous and homozygous Chop knockout). These animals were tested for the expression of Chop (Figure 2A). The levels of Chop mRNA decreased to about half in Chop+/−-Cx50D47A/D47A lenses as compared with the values in Chop+/+-Cx50D47A/D47A lenses (Figure 2A; p<0.05). RT-qPCR revealed that Chop mRNA was reduced by 1,164 times on average in Chop−/−-Cx50D47A/D47A lenses (Figure 2A; p<0.05). These results are consistent with the absence of Chop expression in Chop−/−-Cx50D47A/D47A samples and confirm the deletion of Chop in knockout animals, as previously reported . Levels of CHOP protein were weakly detectable in Chop+/−-Cx50D47A/D47A lenses, and they were undetectable in Chop−/−-Cx50D47A/D47A lenses (Figure 2B).
Having verified the expected reductions of Chop RNA and protein, we examined the lenses of these mice under dark-field illumination. All lenses contained cataracts, regardless of the Chop genotype (Figure 3A). Quantitation of the lens images using ImageJ showed that there were no significant differences in lens opacities, nor did the equatorial lens diameters differ between littermates of Cx50D47A/D47A mice with different Chop genotypes (Figure 3B,C).
We previously showed a significant increase in the phosphorylated form of EIF2α (P-EIF2α) in homozygous Cx50D47A lenses. In these lenses, the P-EIF2α/EIF2α ratio was 3.3 times the value in wild-type lenses . Because Chop deletion in mouse models of other diseases alters phosphorylation of EIF2α [22,23], we tested whether deletion of Chop affected the levels of P-EIF2α or the P-EIF2α/EIF2α ratio. The P-EIF2α levels and P-EIF2α/EIF2α ratio in lenses from Chop+/−-Cx50D47A/D47A and Chop−/−-Cx50D47A/D47A mice did not significantly differ from those in Chop+/+-Cx50D47A/D47A animals as detected by immunoblotting (Figure 4A).
Deletion of Chop could also affect other aspects of the ER response to unfolded proteins. For example, Chop deletion decreases the spliced form of X-box-binding protein 1 (XBP-1) in the retina of T7M rhodopsin mice , but it does not affect Xbp-1 splicing induced in retinal ganglion cells after optic nerve crush . To test whether the deletion of Chop affected Xbp-1 splicing in lenses of homozygous Cx50D47A mice, we used RT-qPCR to quantify the ratio of spliced versus total Xbp-1 mRNA (sXbp-1/tXbp-1). We found that the sXbp-1/tXbp-1 ratio was not significantly different in lenses from Chop+/−-Cx50D47A/D47A, Chop−/−-Cx50D47A/D47A, and Chop+/+-Cx50D47A/D47A littermates (Figure 4B).
Among other roles, CHOP is a key regulator of autophagy [25-27], a pathway implicated in the degradation of connexins . We previously showed that homozygous Cx50D47A mice have decreased levels of the lens fiber cell connexins, Cx46 and Cx50 (27.7% and 3.1% of the wild-type values, respectively, at 3 weeks of age) , likely resulting from degradation of the proteins. To determine the effect of Chop deletion on levels of Cx46 and Cx50, we performed immunoblots for Cx46 and Cx50 on whole-lens homogenates of Chop+/+-Cx50D47A/D47A, Chop+/−-Cx50D47A/D47A, and Chop−/−-Cx50D47A/D47A mice. Levels of these connexins were not significantly different among littermates of the different Chop genotypes (Figure 5).
We previously showed that the mRNA levels of Irs2 and Trib3 are increased by 7.3-fold (range: 3.7–14.4) and 1,002-fold (range: 299–1,873) in homozygous Cx50D47A lenses compared with wild-type lenses . Because these transcripts were identified as prosurvival genes in early stage erythroblasts , we hypothesized that this increase depended on the increase in CHOP or represented a compensatory response to it. To test this hypothesis, we performed RT-qPCR on lens RNA from Chop+/+-Cx50D47A/D47A, Chop+/−-Cx50D47A/D47A, and Chop−/−-Cx50D47A/D47A mouse littermates. Average levels of Irs2 decreased by 1.4-fold in Chop+/−-Cx50D47A/D47A lenses and increased by 1.2-fold in Chop−/−-Cx50D47A/D47A lenses (Figure 6A). Average levels of Trib3 increased by 1.2-fold in Chop+/−-Cx50D47A/D47A lenses and decreased by 1.5-fold in Chop−/−-Cx50D47A/D47A lenses (Figure 6B). These differences were not statistically significant compared with Chop+/+-Cx50D47A/D47A lenses, and they did not represent a substantial return toward the levels previously observed in wild-type mice (Chop+/+-Cx50+/+).
Because CHOP has been implicated in modulating differentiation in other tissues , and expression of Cx50D47A impairs lens cell differentiation with retention of organelles in fiber cells , we investigated whether deletion of Chop altered levels of proteins from different subcellular compartments that are increased in homozygous Cx50D47A lenses. In these lenses, histone H3 (a chromatin component) is increased more than 10-fold, TOM20 (a mitochondrial protein) is increased nearly 3-fold, and calreticulin (a protein that binds calcium and is mainly localized in the ER) is increased 2.6-fold as compared with levels in wild-type lenses [3,6]. Immunoblots of lens homogenates from Cx50D47A/D47A littermates that were wild type for Chop or heterozygous or homozygous for the deletion of Chop showed that levels of histone H3, TOM20, and calreticulin were not significantly different (Figure 7).
CCAAT/enhancer-binding protein β (CEBPβ) is a major dimerization partner of CHOP that contributes to transcriptional regulation and is involved in several processes, including proliferation, differentiation, and cell death. We found that Cebpb mRNA was significantly increased (by ~6.4-fold) in Chop+/+-Cx50D47A/D47A lenses relative to its levels in wild-type mice (Figure 8A). Homozygous (but not heterozygous) deletion of Chop resulted in a significant decrease (by ~4.3-fold) in Cebpb mRNA levels in homozygous Cx50D47A lenses (Figure 8B). This change would return Cebpb transcripts close to wild-type levels.
Our results demonstrate that expression of Chop is dispensable for the maintenance of transparency in wild-type lenses, since we found that lenses of mice in which only Chop was deleted did not exhibit increased opacities. This finding agrees with the initial report on the Chop knockout mouse line describing the absence of gross abnormalities when the animals are maintained on a regular diet [19-21]. However, deletion of Chop slightly increased lens size, implying that CHOP contributes directly or indirectly to the regulation of lens size in the normal animal. CHOP has previously been implicated in increased general body size, since female mice given a regular or a high-fat diet increase in body weight and fat mass . The increase in lens size in our mice is unlikely to result from increased fat mass, since the animals were maintained on a regular diet and there is no fat mass associated with the lens. Rather, it is more likely that the slight increase in lens size results from disrupting other CHOP roles (such as cell death or promotion of oxidizing conditions in the ER) [21,32].
We also found that deletion of Chop did not significantly improve the opacities of homozygous Cx50D47A mice. Thus, although Chop mRNA levels are significantly increased in these mice, this transcript is not the critical factor leading to cataract formation and progression. We speculate that deletion of Chop would also have little or no effect in animals with cataracts caused by different etiologies, including mutations of other major lens proteins like αA-crystallin and Aquaporin0 [9,10], where increased levels of Chop mRNA and CHOP protein are also found. While deletion of Chop ameliorates the effects induced by stresses in some systems, it is ineffective at mitigating the pathology in others. In Chop−/− mice, survival of retinal ganglion cells following optic nerve crush is increased as compared with wild-type mice . Deletion of Chop delays the onset of diabetes in a mouse model of type 2 diabetes . It also decreases the cardiac hypertrophy, fibrosis, and myocardial dysfunction induced by transverse aortic occlusion . In contrast, deletion of Chop accentuates skeletal muscle atrophy in a mouse model of spinal and bulbar muscular atrophy . In addition, Chop deletion does not rescue photoreceptors and worsens the pathology in a mouse model of autosomal dominant retinitis pigmentosa .
Our results imply that the Cx50 mutation has a strong effect on several parameters that cannot be overcome by deletion of Chop. Homozygous Cx50D47A mice have extremely small lenses . Although this may have resulted from increased cell death, deletion of Chop (an inducer of cell death in other systems) in homozygous Cx50D47A mice did not increase lens size. In addition, Chop deletion did not increase the levels of lens Cx46 or Cx50, decrease transcript levels of Trib3 or Irs2, improve differentiation (degradation of nuclei and mitochondria), or ameliorate the cataract phenotype.
Of all the parameters studied, a decrease in the levels of Cebpb transcripts was the only significant change detected in homozygous Cx50D47A lenses after homozygous deletion of Chop. This result implies that, in the lens, Cebpb is either a direct or indirect target of CHOP. Because mouse CEBPβ can autoregulate its transcription, it is possible that the lack of CHOP (one of the CEBPβ dimerizing partners) led to a decrease in CEBPβ levels, leading to an attenuation of derepression of its transcription . In contrast, in the livers of lipopolysaccharide-treated rats, CEBPβ increases transcription of Chop , implying that Chop is a target of CEBPβ.
The lack of a major effect of deletion of Chop on the phenotype and several of the parameters studied in homozygous Cx50D47A lenses may be due to alterations in the expression of other transcription factor(s) that compensate for the absence of CHOP. In addition, the effects of CHOP in homozygous Cx50D47A lenses may have been diminished by αB-crystallin, which is stress responsive [35-38] and has increased levels in these lenses . This crystallin has been implicated in attenuating the response of different cell types to strong cellular stresses [39,40]. αB-crystallin can bind to partially processed caspase-3 and inhibit its autocatalytic activation .
Taken together, our results suggest that CHOP influences some aspects of the diseased lens (like lens Cebpb transcript levels), but it is not a unique critical etiological factor for the cataracts and impaired differentiation of Cx50D47A lenses.
This work was supported by National Institutes of Health Grant R01EY08368 (ECB).