Unfolded-protein response–associated stabilization of p27(Cdkn1b) interferes with lens fiber cell denucleation, leading to cataract

Abstract

Failure of lens fiber cell denucleation (LFCD) is associated with congenital cataracts, but the pathobiology awaits elucidation. Recent work has suggested that mechanisms that direct the unidirectional process of LFCD are analogous to the cyclic processes associated with mitosis. We found that lens-specific mutations that elicit an unfolded-protein response (UPR) in vivo accumulate p27(Cdkn1b), show cyclin-dependent kinase (Cdk)-1 inhibition, retain their LFC nuclei, and are cataractous. Although a UPR was not detected in lenses expressing K6W-Ub, they also accumulated p27 and showed failed LFCD. Induction of a UPR in human lens epithelial cells (HLECs) also induced accumulation of p27 associated with decreased levels of S-phase kinase-associated protein (Skp)-2, a ubiquitin ligase that regulates mitosis. These cells also showed decreased lamin A/C phosphorylation and metaphase arrest. The suppression of lamin A/C phosphorylation and metaphase transition induced by the UPR was rescued by knockdown of p27. Taken together, these data indicate that accumulation of p27, whether related to the UPR or not, prevents the phosphorylation of lamin A/C and LFCD in maturing LFCs in vivo, as well as in dividing HLECs. The former leads to cataract and the latter to metaphase arrest. These results suggest that accumulation of p27 is a common mechanism underlying retention of LFC nuclei.—Lei, L., Whitcomb, E. A., Jiang, S., Chang, M.-L., Gu, Y., Duncan, M. K., Cvekl, A., Wang, W.-L., Limi, S., Reneker, L. W., Shang, F., Du, L., Taylor, A. Unfolded protein response–associated stabilization of p27(Cdkn1b) interferes with lens fiber cell denucleation, leading to cataract.

Opacification of the eye lens, or cataract, is probably the most common malady of the elderly and still blinds 20 million people worldwide in 2010 (1). In addition to age-related cataracts, there are congenital cataracts (0.4% of all births) that present early and result from the expression of abnormal proteins (2). The lens is formed by epithelial cells. During differentiation, lens epithelial cells undergo extreme morphologic and biochemical changes as they become fiber cells. This process includes induction of expression of a large quantity of proteins called crystallins and the removal of virtually all organelles from the central fiber cells. Removal of the cell nuclei, mitochondria, lysosomes, endoplasmic reticulum (ER), and Golgi apparatus forms an organelle-free zone (OFZ) in the center of the lens. The process continues throughout life, and, although cell nuclei are destroyed, there is no detectable lens fiber cell (LFC) turnover (3–6). The formation of an OFZ is essential for the transparency and function of the lens. LFC denucleation (LFCD) appears to be initiated by cyclin-dependent kinase (Cdk)-1–induced phosphorylation of lamin A/C, with subsequent disassembly of the nuclear envelope. Failure to remove lens fiber cell nuclei is associated with various congenital cataracts (7–14). Despite significant progress in lens development research, the mechanisms underlying the LFCD process remain largely unknown.

The nuclear lamina is a meshwork of intermediate filament proteins that closely associates with several inner nuclear membrane proteins and interacts with portions of the chromatin (15–18). Initiation of the nuclear envelope disassembly process during mitosis is regulated by phosphorylation of lamins, particularly by Cdk1 (19–21). We recently demonstrated that Cdk1 is critical for nuclear envelope disassembly in the eye lens in vivo as well (22).

Ubiquitination is best known for targeting proteins for degradation via the proteasome (23). Studies have shown that expression of mutant ubiquitin (Ub) in the lens in vivo results in the accumulation of p27 (12), a Cdk inhibitor that plays important roles in regulating the cell cycle (24–27). The stability of p27 is determined by its phosphorylation status. When p27 is phosphorylated on Thr¹⁸⁷, it is degraded by the Ub proteasome system (UPS) via SCF^skp2-mediated ubiquitination (28–30). Several lines of evidence suggest that p27 is a common upstream regulator of the Cdk1-driven cyclic process of mitosis and the unidirectional LFCD process: 1) overexpression of mutant Ub in human lens epithelial cells (HLECs) also causes the accumulation of p27, concurrent with G₂/M phase arrest (12, 31); 2) a few reports indicate that p27 has roles in directing the G₂/M transition during the cell cycle via activation of Cdk1 (32–34); and 3) Cdk1 is necessary in directing LFCD (19–22).

In eukaryotic cells, if the influx of secretory and transmembrane proteins into the ER exceeds the folding or processing capacity or mutant proteins enter and cannot be folded properly, ER stress is induced (35). In response to ER stress, cells activate a set of adaptive intracellular signal-transduction pathways, known as the unfolded-protein response (UPR) (35–37). The UPR is reportedly activated in the lens during embryogenesis, and moderate levels of ER stress are thought to play a physiologic role during lens development (14). A UPR is induced in the lenses of transgenic mice that express mutant collagen IV (Col IV), dominant negative fibroblast growth factor receptor (dnFGFR), and other mutant proteins (7, 8, 13). These mutant mice have cataracts and retain their LFC nuclei in the OFZ. The retention of LFC nuclei is similar to that observed in the lenses of mutant Ub transgenic mice (12) in which p27 accumulates. Recent reports have indicated that the UPR can stabilize p27 through the suppression of the expression of the E3 ligase S-phase kinase-associated protein (Skp)-2 (38, 39). In this study, we hypothesized that UPR-induced accumulation of p27 is a mechanism that links UPR to defects in the LFCD process.

To test this hypothesis, we used several in vivo and in vitro models of the UPR in transgenic mice [mutant Col IV, dnFGFR, and dominant negative nuclear receptor coactivator 6 (dnNCOA6)] (7, 8, 40) and HLECs, respectively. The UPR in both the cultured cells and lenses in vivo resulted in the accumulation of p27. We compared the findings to systems in which p27 accumulated but did not show a UPR. In all cases, accumulation of p27 was associated with inhibition of the phosphorylation of lamin A/C. In contrast, knockdown of p27 (p27KD) partially rescued the inhibition of lamin A/C phosphorylation and cell cycle arrest at the G₂/M phase in HLECs. The data suggest that the accumulation of p27 induced by the UPR mechanistically links inhibition of nuclear envelope disassembly in both the cell cycle and LFCD.

MATERIALS AND METHODS

Antibodies

The following primary antibodies were used: p27^kip1 (610241, 1:2,000; BD Biosciences, San Jose, CA, USA); BiP (ab21685, 1:5,000), phospho-lamin A/C (ab58528, 1:5,000 for Western blot analysis and 1:200 for immunohistochemistry), and Lamp1 (ab24170, 1:100) (Abcam, Cambridge MA, USA); phospho-serine CDKs substrate (P-S2-100, 1:1,000; Cell Signaling, Danvers, MA, USA); Skp2 (sc-7164, 1:1,000), DDB1 (sc-136180, 1:500), Kpc1 (sc-101122, 1:500), cyclin B1 (sc-245, 1:1,000), securin (sc-5839, 1:1,000), and cyclin A (sc-751, 1:1,000) (Santa Cruz Biotechnology, Dallas, TX, USA); Pirh2 (A300-357A, 1:1,000; Bethyl Laboratories, Montgomery, TX, USA); GAPDH (G9545, 1:10,000) and β-actin (A5441, 1:10,000) (Sigma-Aldrich, St. Louis, MO, USA); phospho-T2055- nuclear mitotic apparatus protein (NuMA; 1:1,000; a gift from Dr. P. Gönczy, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland); and DNAse IIβ (1:100; a gift from S. Nagata, Kyoto University, Toshida, Kyoto, Japan). Primary antibodies were detected with species-specific secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA) conjugated to IgG-horseradish peroxidase (HRP; 1:2000) for Western blot analysis and FITC or Cy3 (1:250) for immunohistochemistry.

Animals

The transgenic mouse line 520, which overexpresses Col IVα3 under the regulation of a modified αA crystallin promoter, was maintained on the FVB/N genetic background in the University of Delaware Animal Facility as has been described (7). The Cryaa-dnNcoa6 transgenic line was generated on the FVB/N background at the Albert Einstein College of Medicine (New York, NY, USA) and analyzed as has been described (40). The dnFGFR mouse was constructed as according to a published method (41). Transgenic mice expressing truncated FGFR1 (or tR1) in the ocular lens, driven by the αA-crystallin promoter, were maintained on the FVB/N genetic background in the animal facility of the University of Missouri Animal Science Research Center (ASRC). The homozygous tR1 female mice were mated to C57BL6/J male mice and the eyes were enucleated from newborn pups for experiments. The K6W-Ub mouse was maintained in the Human Nutrition Research Center on Aging (Tufts University, Boston, MA, USA) (12). All procedures used in the study were approved by the Institutional Animal Care and Use Committee at Tufts University and were conducted in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Histology and immunohistochemistry

Lenses dissected from wild-type (WT) and transgenic mice (7, 8, 40) were fixed with 4% paraformaldehyde for 2 h, embedded in optimal cutting temperature compound (Sakura Finetek, Torrance, CA, USA) frozen, and subsequently sectioned according to standard protocols. HLECs were grown on 4-well chamber slides (Thermo Scientific, Waltham, MA, USA) and fixed with 4% paraformaldehyde for 15 min. Before they were stained, the lens cryosections (10 μm) and cells grown on slides were permeabilized with 0.05% Triton X100 for 5 min at room temperature. The slides were blocked with 5% normal donkey or goat serum (Jackson ImmunoResearch) in 5% bovine serum albumin (BSA)/PBS for 1 h at room temperature. Primary antibodies were incubated overnight at 4°C and secondary antibodies for 45 min at room temperature. Nuclear DNA was stained with 1 μg/ml DAPI. Fluorescent images were obtained with an inverted fluorescence microscope (Axiovert 200 and AxioVision V 4.5 software; Zeiss, Jena, Germany). The detection and localization of p27, p-lamin A/C, and DNase IIβ, as well as their respective semiquantitations, were repeated in more than 2 lenses, each examining multiple sections.

Cell culture and synchronization

HLECs (cell line SRA 01/04) were grown in DMEM (Thermo Scientific) and supplemented with fetal bovine serum (10% v/v; Thermo Scientific), penicillin (100 U/ml) and streptomycin (100 μg/ml) (Thermo Scientific).

For synchronization, the HLECs were grown to 60% confluence and synchronized in medium containing 2 mM hydroxyurea (HU; Sigma-Aldrich). After 21 h, the cells were washed twice with PBS and fresh medium was added to allow the G₁/S phase–synchronized cells to enter the cell cycle.

UPR induction in cultured HLECs

HLECs grown to 70% confluence were incubated with 1 μg/ml tunicamycin (TM), 1 μM thapsigargin (Thg), or 5 mM homocysteine (HC) (all from Sigma-Aldrich) for times indicated in the figure legends.

Generation of p27KD HLECs

Short hairpin (sh)RNAs (TRCN0000039930 and TRCN0000295863) and a nontarget shRNA control virus (SHC016-1EA) were purchased from Sigma-Aldrich. HLECs were cultured in 6-well plates to 80% confluence. p27 shRNA lentiviruses [10⁴ titer units (TU)] or nontarget control (NC) virus was added in the presence of 8 μg/ml hexadimethrine bromide (Sigma-Aldrich). The medium containing the lentiviral particles was removed 24 h later, and the cells were grown in fresh medium overnight. Stable cell lines were selected with 2 μg/ml of puromycin (Santa Cruz Biotechnology).

MTS assay

NC and p27KD cell lines were plated in 96-well plates at 5000 cells/well. Cell counts were assessed with the MTS assay kit (Promega, Madison, WI, USA), as soon as they adhered to the plate, or after 24 and 48 h of growth, according to the manufacturer’s instructions. Absorbance at 495 nm was read with a microplate reader (ELx808; BioTek China, Beijing, China).

Quantitative real time-PCR

Total RNA was extracted from cells with the RNeasy Mini Kit (Qiagen, Redwood City, CA, USA), and cDNA was synthesized from 500 ng of RNA with SuperScript III First-Strand Synthesis System for RT-PCR (Thermo Scientific). Quantitative real-time PCR reactions were performed in duplicate. Each reaction contained 1.5 μl cDNA, target-gene–specific primers, and SYBR Green (GoTaq qPCR Master Mix; Promega), according to the manufacturer’s instructions. Forty cycles of qPCR were performed with the 7300 Real Time PCR System (Thermo Scientific). The sequences of primers used were the following: p27, forward, gagtggcaagaggtggagaa, and reverse, gtccgacggatcagtctttg; GAPDH, forward, acccagaagactgtggatgg and reverse, tttctagacggcaggtcagg. The results were analyzed by the comparative C_t (^δδC_t) method. The quantified values for p27 were normalized by GAPDH.

Western blot analysis

Cells were washed with cold PBS and lysed in SDS-PAGE loading buffer at various time points. Lenses from WT and transgenic mice were also homogenized directly in loading buffer. Equal quantities of samples were separated by SDS-PAGE. Proteins were transferred by electroblotting onto nitrocellulose membranes (0.2 μm; Bio-Rad, Hercules, CA, USA), after which the membranes were probed with primary antibodies overnight at 4°C and HRP-conjugated secondary antibodies for 45 min at room temperature. Signals were detected by chemiluminescence (SuperSignal West Pico Chemiluminescent Substrate; Thermo Scientific).

p27 turnover assay

HLECs were grown to 70% confluence and incubated with 1 μg/ml TM (Sigma-Aldrich) or with the same amount of DMSO for 24 h and then treated with 25 μg/ml cycloheximide (CHX) (Sigma-Aldrich). The cells were collected at various times after the addition of CHX, for immunoblot analysis of p27.

Flow cytometry analysis

Synchronized HLECs treated with 1 μg/ml TM or DMSO were harvested as single cells with 0.25% trypsin/EDTA and fixed in 70% ethanol overnight at 4°C. The cells were then washed with PBS and treated with 0.5 μg/ml RNase A (Sigma-Aldrich) for 1 h at room temperature. Propidium iodide (10 μg/ml; Sigma-Aldrich) was added for DNA staining. The cells were analyzed on an Accuri C6 Cytometer (BD Biosciences). The fluorescence activated cell sorting (FACS) data regarding cell cycle distribution were analyzed with FlowJo v7.6.1 (FlowJo, LLC, Ashland, OR, USA).

Statistical analysis

All data from at least 3 independent experiments are presented as averages ± se. Statistical analysis and comparisons were performed with Student’s t test. Western blot images were plotted and analyzed with ImageJ software (National Institutes of Health, Bethesda, MD, USA).

RESULTS

The UPR is associated with stabilization of p27, failed LFCD, and cataracts in vivo

To investigate the mechanism underlying the relationship between the UPR and failure of LFCD, we first examined several transgenic mouse lines that were reported to induce a UPR and showed failed LFCD. Elevated levels of the ER chaperone BiP in mutant Col IV lenses, one of the hallmarks of the UPR (42), indicated that a UPR was induced, as has been reported (7) (Fig. 1A). We then asked whether p27 was increased in Col IV-expressing lenses by UPR induction. As shown in Fig. 1A and Supplemental Fig. S1B, lenses that expressed mutant Col IV contained ∼3-fold higher p27, compared with WT lenses. Immunohistochemical analyses showed a gradual decrease in nuclear p27 (red) from the bow region to the center of the WT lenses , but retention of higher levels of nuclear p27 toward the center of the Col IV lenses (Fig. 1B). Semiquantitative analysis indicated that 3.5-fold more (P = 6 × 10⁻⁵) nuclei retained p27 in the Col IV lens than in the WT lens. To further explore the etiologic relationship of the UPR with accumulation of p27 and aberrant LFCD, we examined 2 additional in vivo models that exhibit lens fiber nuclei retention and cataract. dnFGFR transgenic mice, a confirmed UPR model (8), and the dnNCOA6 model (40) elicited a UPR, as evidenced by the increase in BiP (Fig. 1C). They also expressed (>5-fold higher levels of p27, compared with their WT littermates (Fig. 1C; Supplemental Fig. S1B).The expression of K6W-Ub in the lens also elicited >4-fold p27 accumulation and a denucleation defect (12), but did not induce the UPR (Supplemental Fig. S1). These findings imply that the accumulation of p27 is a common mechanism for impaired denucleation.

Figure 1.

Expression of cataract-inducing proteins induces a UPR and sustained p27 expression. A) Lenses from Col IV transgenic mice (P1) were lysed with SDS-gel loading buffer. Levels of BiP, p27, and GAPDH were determined by Western blot analysis. GAPDH was used as a loading control. B) P1 lenses from WT (left) and mutant Col IV (right) transgenic mice were cryosectioned and immunostained with anti-p27 (red) and counterstained with DAPI (blue). Note that p27 staining diminishes more peripherally in the WT than in the Col IV lens. Bow, bow region; epi, epithelium. C) Lens lysates from dn-FGFR (P0) and dnNCOA6 (P1) mice were also examined for levels of BiP and p27 by Western blot analysis, as described in (A). Scale bars, 50 μm.

Phosphorylation of lamin A/C by Cdk1 drives LFCD and establishment of a clear lens (12, 22). In our K6W-Ub mutant, the stabilization of p27 was associated with inhibition of Cdk1activation, delayed phosphorylation of lamin A/C, and delayed entry of DNAse IIβ into the nucleus. To determine whether the accumulation of p27 in UPR-inducing transgenic lenses is also associated with diminished Cdk1 activation in vivo, we monitored the phosphorylation of lamin A/C. In WT mice, the level of phosphorylated lamin A/C at the bow region was very low (Fig. 2A, left), but increased dramatically at the nuclear envelope of the fiber cells (appearing as green rings) that are scheduled to denucleate to form the OFZ (Fig. 2A). In contrast, phosphorylated lamin A/C was barely detectable surrounding the nuclei in the Col IV lenses (Fig. 2B, compare detail, right), and the fiber cells in these lenses retained their nuclei at a rate that was more than 70% higher (P = 0.0002) than in WT lenses (Fig. 2A, B, compare enlargements of dashed boxes, right). We also did not observe phosphorylation of lamin A/C in dnFGFR lenses (Supplemental Fig. S2A). These data suggest that disassembly of the LFC nuclear envelope in vivo is interrupted by the UPR-elicited accumulation of p27 and the associated inhibition of Cdk1 activation.

Figure 2.

Phosphorylation of lamin A/C in lens fiber cells is inhibited by the UPR. P1 lenses from WT (A) and mutant Col IV transgenic mice (B) were cryosectioned and immunostained with anti-phospho-lamin A/C (green) and counterstained with DAPI (blue). Right: higher magnifications of boxed areas. Scale bars, 50 μm.

DNAse IIβ or DLAD, generally a lysosomal enzyme, is a lens-specific acidic DNAse that enters the nucleoplasm to degrade chromatin at the conclusion of LFC differentiation (11). Because elevated p27 correlates with impaired nuclear envelope disassembly and decreased entry of DNAse IIβ into the nuclei in the K6W transgenic lenses, we predicted that elevated p27 in the UPR-inducing transgenic mice would also lead to failed disassembly of the nuclear membrane and exclusion of DNAse IIβ from the nucleus. In WT lenses, DNAse IIβ entered into the nuclei in the fiber cells in the denucleating region (Fig. 3A, red circles surrounding nuclei or red within nuclei) (9, 12). However, in lenses that expressed mutant Col IV, more DNAse IIβ was observed in the cytoplasm and less translocation into nuclei was noted (Fig. 3B). Moreover, DNase IIβ in Col IV transgenic lens was found more peripherally in the bow region and colocalized with lysosomes (Lamp1) that were also retained in what should be an OFZ (Fig. 3B, enlargement, right). Semiquantitative analysis suggested that the proportion of Col IV lens fiber cell nuclei that contained DNase IIβ was 6 times less (P = 0.0007) than that in the WT. In addition, because of the poor retention of structure during processing of these tissues, there was more nonspecific staining in Col IV lens, which was likely to have caused an overestimate of the DNase IIβ in the Col IV lenses. The results suggest that the proportion of Col IV fiber cell nuclei in which DNase IIβ has been delivered is significantly lower. The distribution of DNase IIβ was similar in dnFGFR lenses, with no staining observed in the inner cell nuclei (Supplemental Fig. S2B). These data are consistent with the hypothesis that UPR-associated elevation of p27 inhibits disassembly of the nuclear membrane that is associated with entry of DNase IIβ.

Figure 3.

UPR blocks the nuclear translocation of DNase IIβ in lens fiber cells. P1 lenses from WT (A) and mutant Col IV transgenic mice (B) were cryosectioned and immunostained with anti- DNAse IIβ (red) and counterstained with DAPI (blue). Lysosomes are stained with anti-Lamp1 (green). Right: higher magnification of boxed areas. Scale bars, 50 μm.

The UPR inhibits lamin A/C phosphorylation and leads to G₂/M arrest in proliferating cells through a p27-dependent mechanism

LFCs do not progress though mitosis in vivo. However, during denucleation, they appear to appropriate some of the mitotic machinery to terminally remove their nuclei, an obligatory process for establishing a clear lens (12, 43). Thus, to further explore the role of UPR-induced p27 expression in nuclear envelope breakdown, we asked whether induction of a UPR causes p27 accumulation in cultured HLECs. If a UPR elicits stabilization of p27, then that should also cause a delay in cell cycle progression and events associated with mitosis. Exposure of HLECs to TM, an established ER stressor that inhibits protein glycosylation, elicited a UPR, as indicated by a dramatic increase in levels of BiP (Fig. 4A). Exposure to TM also resulted in a 2-fold increase in p27 (Fig. 4A; Supplemental Fig. S3A) (38, 39).

Figure 4.

Induction of a UPR in HLECs increases p27 through reduced degradation and increased mRNA expression. A) Cells were treated with 1 μg/ml TM and collected at 16, 20, and 24 h. Levels of BiP and p27 were determined by Western blot. β-Actin was used as a loading control. B) Cells were treated with TM or left untreated for 16 h, and CHX was then added to stop protein synthesis. The p27 level was determined by Western blot. MG132 was used to inhibit proteasome-dependent degradation. C) The effect of UPR induction on the levels of Ub ligases, Skp2, DDB1, Kpc1, and Pirh2, which are related to p27 stability, was determined by Western blot analysis. Cells were treated with 1 μg/ml TM and collected at 16, 20, and 24 h. D) mRNA levels of p27 and BiP in the presence or absence of TM were determined by real-time quantitative PCR, with GAPDH as the control. *P < 0.001, vs. 0 h after TM treatment.

Levels of p27 are known to be controlled by the UPS (28–30); thus, we determined the effects of UPR induction on proteasomal degradation of p27. In the absence of TM (DMSO control), the level of p27 decreased rapidly such that it was barely detectable by 2 h (Fig 4B). In contrast, the level of p27 in cells treated with TM decreased more slowly: ∼70% of p27 protein remained 2 h after protein synthesis was arrested by addition of CHX in TM-treated cells. A similar delay in p27 degradation was observed in the presence of the bona fide proteasome inhibitor, MG132 (Fig. 4B;Supplemental Fig. S3B).

We then asked whether induction of the UPR affected the levels of E3 ligases known to be involved in p27 degradation (44–50). We found that the level of Skp2, the substrate recognition subunit of the E3 that is usually associated with p27 degradation, SCF^skp2 (27, 39), was significantly reduced upon treatment with TM, whereas the levels of DDB1, KPC1, and Pirh1, other E3s known to target p27 for proteasomal degradation, were not affected (Fig. 4C).

Quantitative RT-PCR analysis showed that TM treatment significantly increased both BiP and p27 mRNA in HLECs (Fig. 4D). At the 24-h time point, TM induced 3-fold and >25-fold elevations in the levels of p27 and BiP mRNA, respectively. These data indicate that reduced UPS-dependent degradation, likely via a Skp2-mediated pathway, and increased p27 synthesis contribute to the UPR-induced stabilization of p27.

Next, we determined whether decreases in p27 are necessary to set in motion a sequence of events involving activation of Cdk1, phosphorylation of lamin A/C, and disassembly of the nuclear envelope. These processes are required for both LFCD and transition through the M phase in cycling cells (22, 51, 52). To investigate the effect of UPR-associated accumulation of p27 in mitosis, HLECs were synchronized at the G₁/S phase boundary using HU (Fig. 5A). HU itself does not induce a UPR or p27 accumulation (data not shown). When the majority of the cells (80%) were in the S phase (3 h after the removal of HU), a UPR was induced with TM. p27 stabilization (Supplemental Fig. S4) was similar to that observed in the nonsynchronized cells (Fig. 4A). Consistent with the reduced level of phosphorylated lamin C in TM-treated HLECs (Supplemental Fig. S4A, middle right), the percentage of phosphorylated lamin A/C⁺ cells in the TM-treated group was significantly less than those in the control group (Fig. 5B, red arrows), indicating that a UPR is associated with decreased lamin A/C phosphorylation in HLECs.

Figure 5.

UPR induction prevents nuclear lamin phosphorylation and leads to cell cycle arrest at the G₂/M phase. A) Experimental scheme. Cells were synchronized by HU treatment. TM was added when most of the cells were in the S phase. B) Cellular localization of phosphorylated lamin A/C was determined with immunostaining after 8 h of TM treatment. The bar graph shows the percentage of cells with phosphorylated lamin. C) Cell cycle profiles were determined with propidium iodide staining and FACS analysis. D) The level of multiple Cdk substrates indicated as phospho-Ser substrates (top) and the level of specific Cdk1 product phospho-NuMA (T2055) (middle) were blotted after 24 h, in the presence or absence of TM, as the indicator of Cdk1 activation. *P < 0.01; ** P < 0.001 vs. the control cells at the specified time points.

To further explore the role of the UPR and accumulation of p27 in cycling cells, progression through the cell cycle was monitored. The majority of the cells (∼70%) were at G₁ phase after treatment with HU (t = −3 h; Fig. 5C, top, blue bars). Eighty percent of the cells were in S phase 3 h after the removal of HU (middle, t = 0, blue bars). In the absence of TM, most of the cells progressed from S phase to G₂/M phase within 8 h (bottom, blue bars), and the cells started to transition from G₂/M phase to the next G₁ phase within 12 h (Fig. 5C, top, blue bars). TM slowed, but did not block, the transition from the S phase to the G₂/M phase (Fig. 5C, bottom, 8 h). However, TM treatment significantly inhibited the transition beyond the G₂/M phase into the next G₁ phase (Fig 5C, top, red bars). By 16 to 24 h after TM treatment, there were ∼20% more cells at the G₂/M phase in the presence of TM than in its absence. Because TM treatment increases p27 levels and p27 inhibits Cdk1, we hypothesized that the activity of Cdk1 would in turn be decreased. Consistent with this notion, total phosphoserine Cdk1 substrates decreased during TM treatment (Fig. 5D, bar graph, Supplemental Fig. S4B). We also examined the level of phosphorylated (NuMA), an established Cdk1 substrate (53). The decreased levels of phosphorylated NuMA after 24 h of TM treatment also indicate that Cdk1 activity is suppressed during UPR (Fig. 5D).

To determine whether the UPR-associated G₂/M arrest requires p27, we stably knocked down p27. Western blots indicate that ∼80–90% of the p27 was depleted in the p27 shRNA-infected cells (p27KD) as compared to the NC shRNA-infected cells (Fig. 6A, lane 9 vs. 3). Diminishing p27 did not affect the TM-induced UPR, as the level of BiP increased to an extent similar to that in the control cells (lanes 10–12 vs. 4–6). Consistent with the role of p27 as a cell cycle inhibitor, p27KD was associated with an increase in cell proliferation (Fig. 6B).

Figure 6.

p27KD abolishes the inhibition of phospho-lamin A/C and G₂/M arrest caused by the UPR. A) p27 was knocked down in HLECs using lentivirus-mediated shRNA. The NC and p27KD HLECs were synchronized by HU and treated with TM (Fig. 5A). The levels of p27 and BiP, in the absence or presence of TM, were detected at the indicated time points with Western Blot. B) The growth rates of NC and p27KD cells were determined by the MTS assay, as described in experimental procedures. C) The percentages of phospho-lamin A/C⁺ cells in the synchronized NC, and p27KD cells were detected by immunostaining 8 h after TM treatment. D) The synchronized NC and p27KD cells were treated with TM or left untreated (Fig. 5A). The cell cycle progressions were determined by FACS analysis. *P < 0.001 vs. TM-treated NC cells at the indicated time points.

We then asked whether p27KD could reverse the UPR-induced decrease in lamin A/C phosphorylation and delay progression through the G₂/M phase. TM treatment significantly reduced the phosphorylation of lamin A/C in NC cells (Fig. 6C). However, in p27KD cells, lamin A/C phosphorylation was not decreased upon TM treatment. Rather, it was more than twice that of TM-treated NC cells. These data are consistent with the hypothesis that accumulation of p27 plays a critical role in UPR-induced mitotic arrest, apparently by the inhibition of Cdk1 and subsequent lamin A/C phosphorylation. We then directly tested whether p27KD has an effect on TM-induced cell cycle arrest. Results from 2 independent shRNA clones indicate that p27KD abolished the G₂/M arrest induced by TM (Fig. 6D, Supplemental Fig. S5). This finding is of additional interest, because p27 is usually associated with inhibiting the G₁/S transition in cell cycle.

To gain additional corroboration that an increase in p27 and cell cycle arrest is a common feature of UPR induction, we asked whether 2 other inducers of the UPR, Thg, and HC, also cause an increase in p27 and cell cycle arrest (39, 54) (Supplemental Fig. S6B, C). In the control cells, the presence of these UPR inducers significantly delayed progression of the cell cycle at G₂/M phase, consistent with the results of the addition of TM (Fig. 5C), whereas depletion of p27 enhanced the rate at which the cells advanced to the G₁ phase, compared with treated NC cells (Supplemental Fig. S6C).

We were interested to know at which point in the M phase UPR induction causes arrest. Thus, we examined the levels of mitotic regulators cyclin A, cyclin B1, and securin (55–57). Although these substrates are all degraded in mitosis, cyclin A is degraded in the prometaphase, whereas cyclin B and securin are degraded in the metaphase. We found that cyclin A did not accumulate in the TM-treated cells (Supplemental Fig. S6A, B) whereas cyclin B1 and securin were stabilized in cells exposed to these UPR inducers when p27 was stabilized (Supplemental Fig. S6A, lane 2 vs. 1; Supplemental Fig. S6B, lanes 2–3 vs. lane 1). Together these results suggest that the arrest occurred during the metaphase. TM, Thg, and HC had no effect on cyclin B1 and securin levels in p27-depeleted cells (Supplemental Fig. S6A, lane 4 vs. lane 3; Supplemental Fig. S6B, lanes 5, 6 vs. lane 4). These data suggest that in cultured cells, a UPR induced during the S phase results in the accumulation of p27 and that this accumulation of p27 results in metaphase arrest, via decreased Cdk1 activation and subsequent inhibition of nuclear membrane disassembly.

DISCUSSION

Timely removal of nuclei and other cytoplasmic organelles in maturing lens fibers is crucial for establishing a transparent lens. Failure to remove LFC nuclei is a common phenotype of congenital cataracts caused by various mutations (7, 8, 11–13, 58). Initial reports suggested that, to accomplish denucleation, lens fiber cells adapt a mechanism that is involved in the disassembly of the cell nucleus that occurs during the anaphase of the cell cycle (12, 22). Data from this study also support this hypothesis. Furthermore, our data indicate that reduced p27 levels facilitate Cdk1-dependent LFCD and suggest that p27 accumulation is a common pathobiological mechanism that underlies various LFCD defects.

p27 is a critical Cdk inhibitor and cell cycle regulator. p27 is usually associated with stabilization of the G₁/S transition (27, 47, 59), but p27 control of mitotic progression has also been reported (32–34). We have shown that lenses expressing mutant Ub exhibit increased p27 and a denucleation defect (12), but do not show evidence of a UPR (Fig 1C). In this study, we demonstrate that a UPR associated with expression of mutant proteins in lenses or induced by TM treatment in lens cells in culture also causes the accumulation of p27 (Figs. 1, 5), thus mechanistically linking increased p27 to denucleation defects in lenses that do and do not elicit a UPR.

Moreover, we observed that the stabilization of p27 involved enhanced transcription, as well as delayed degradation, the latter apparently occurring via down-regulation of Skp2 (Fig. 4C, D). This finding is consistent with previous observations that the UPR stabilizes p27 by decreased Skp2-mediated degradation and increased transcription (39). In our study, other E3s that are known to catalyze degradation of p27 appeared not to be involved (Fig. 4C).

We used multiple in vivo models of UPR and found corroboration in in vitro systems. In each of the in vivo UPR models—mutant Col IV, dnFGFR, and dnNCOA6 transgenic mice—as in the K6W-Ub transgenic mouse lenses, we found accumulation of p27, retention of nuclei in LFCs, smaller lenses, disorderly lens fiber alignment, and cataracts (Fig. 1) (7, 8, 12, 40). These features are in keeping with delayed synthesis of major lens proteins and impaired replication, thus causing a smaller lens. Taken together, these data suggest that when p27 accumulates, Cdk1 activity is markedly reduced. This decrease results in arrest at the G₂/M phase in cells in culture and multiple developmental abnormalities, including delayed disassembly of the nucleus in LFCs in vivo (Fig. 2). Consistent with the hypothesized role of p27 in this process, in our study, p27KD rescued the effects of the UPR on lamin A/C phosphorylation and cell cycle arrest (Fig. 6; Supplemental Fig. S5).

The findings enhance our appreciation of p27 biology by demonstrating in new in vivo systems that p27 is a major regulator of lens development and that a UPR is associated with p27 stabilization. The data also corroborate observations (12, 22) that the spatiotemporal activation of Cdk1 and subsequent phosphorylation of lamin A/C is necessary for LFCD. The down-regulation of p27 in the center of the lens appears to play an important role in Cdk1 activation near the OFZ. Analogies between events associated with LFCD and p27–Cdk1-catalyzed control of the cell cycle are striking. Specifically, we observed that nature exploits a major control node (p27 levels) of the cyclical pathway of cell proliferation for the unidirectional control of LFCD. It appears that mutations or stresses that interfere with the degradation or expression of p27 in the LFCs may impede LFCD and cause cataracts.

Glossary

Cdk

cyclin-dependent kinase

CHX

cycloheximide

Col IV

collagen IV

dnFGFR

dominant negative fibroblast growth factor receptor

dnNCOA6

dominant negative nuclear receptor coactivator 6

ER

endoplasmic reticulum

FACS

fluorescence-activated cell sorting

HC

homocysteine

HLEC

human lens epithelial cells

HRP

horseradish peroxidase

HU

hydroxyurea

LFC

lens fiber cell

LFCD

LFC denucleation

NC

nontarget control

NuMA

nuclear mitotic apparatus protein

OFZ

organelle-free zone

p27KD

p27 knockdown

shRNA

short hairpin RNA

Skp

S-phase kinase-associated protein

Thg

thapsigargin

TM

tunicamycin

Ub

ubiquitin

UPR

unfolded-protein response

UPS

ubiquitin proteasome system

WT

wild-type