|Home | About | Journals | Submit | Contact Us | Français|
Although forming a heterodimer or heterooligomer is essential for MDM2 and MDMX to fully control p53 during early embryogenesis, deletion of either MDM2 or MDMX in specific tissues using the loxp-Cre system reveals phenotypic diversity during organ morphogenesis, which can be completely rescued by loss of p53, suggesting the spatiotemporal independence and specificity of the regulation of p53 by MDM2 and MDMX. In this study, we investigated the role of the MDM2-MDMX-p53 pathway in the developing lens that is a relatively independent region integrating cell proliferation, differentiation and apoptosis. Using the mice expressing Cre recombinase specifically in the lens epithelial cells (LECs) beginning at E9.5, we demonstrated that deletion of either MDM2 or MDMX induces apoptosis of LEC and reduces cell proliferation, resulting in lens developmental defect that finally progresses into aphakia. Specifically, the lens defect caused by MDM2 deletion was evident at E10, occurring earlier than that caused by MDMX deletion. These lens defects were completely rescued by loss of two alleles of p53, but not one allele of p53. These results demonstrate that both MDM2 and MDMX are required for monitoring p53 activity during lens development, and they may function independently or synergistically to control p53 and maintain normal lens morphogenesis.
Embryonic development of lens originates from the surface ectoderm, and is driven by a series of inductive signals. This process comprises regional specification events and coordinated differentiation, starting from the induction of the lens placode, a single layer of thickening cells, through the physical contact of optic vesicles with the overlying ectoderm at mouse embryonic day 9.5 (E 9.5), followed by invagination at E 10 to form a lens pit and further detachment from the rest of the ectoderm to become a lens vesicle. Next, the posterior epithelial cells of the vesicle elongate, denuclearize and differentiate into primary lens fibers to fill up the vesicle cavity at E 12.5, The secondary lens fibers eventually develop from differentiated anterior epithelial cells in the equatorial zone on the margins of the lens epithelia that undergo proliferation and inward migration while the anterior cells in the center of the epithelium remain largely quiescent (Graw, 1996; Pan et al., 2010; Wride, 1996). Once the lens epithelial cells initiate terminal differentiation, they withdraw from the cell cycle and elongate into lens fiber cells, undergoing denucleation, and loss of other cellular organelles. The mature lens retains a single layer of undifferentiated epithelial cells on its anterior surface (Graw, 1996). All the cellular events crucial for lens morphogenesis, including cell proliferation, migration and differentiation, are tightly controlled by the inductive interaction and sequential activation of transcription factors. As one of the most important lens-bias and lens-specification factors, Pax6 is expressed broadly in the head ectoderm including the presumptive lens ectoderm prior to contact with the optic vesicle (Ashery-Padan and Gruss, 2001; Klimova and Kozmik, 2014; Shaham et al., 2012), and it regulates the expression of other transcriptional factors or miRNAs during lens induction, such as Six3, Sox2, FoxE3, L-Maf, Prox1 and miR-204, which are implicated in the formation of the lens placode and vesicle and/or subsequent lens fiber differentiation (Ashery-Padan and Gruss, 2001; Kamachi et al., 2001; Ogino et al., 2012; Purcell et al., 2005; Shaham et al., 2013; Smith et al., 2009). In a feedback fashion, some of these transcription factors, such as Six3 and Sox2, also mediate Pax6 expression (Liu et al., 2006; Ogino et al., 2012; Smith et al., 2009), suggesting a coordinated interaction of the lens-specific proteins to determine the spatio-temporal order of the occurrence of cell proliferation and differentiation during lens morphogenesis. Cell apoptosis also plays a crucial role during lens development. Apoptosis assists the morphogenetic process of the lens vesicle and the detachment of the lens vesicle from the ectoderm (Yan et al., 2006). Activation of the apoptosis signals, such as Apaf-1, Bcl-2 family members, caspase family members and endonuclease, is necessary for lens denucleation and differentiation (Geatrell et al., 2009; Yan et al., 2006). Denucleation of lens fiber cells is an apoptosis-like process, but unlike apoptotic cells, denucleated fiber cells persist throughout life to form mature lens, rather than being destroyed and eliminated (Geatrell et al., 2009; Pan and Griep, 1995).
As one of the most important tumor suppressors, p53 induces cell cycle arrest, apoptosis, autophagy and senescence through transactivation of many cell death and survival related protein-encoding genes, including p21, Bax, Puma, and so on (Boominathan, 2010; Lane and Levine, 2010), or transcription-independent mechanisms (Leung and Sharp, 2010). Genetic studies have shown that p53 is not essential for normal embryonic development (Donehower et al., 1992; Jacks et al., 1994). However, mice with inactive or deficient p53 displayed incomplete lens fiber cell denucleation (Jean et al., 1998; Pan and Griep, 1994; Wiley et al., 2011), and failed to induce the regression of the primary vitreous during eye development (Reichel et al., 1998), resulting in a high frequency of cataracts. Likewise, abnormal induction of apoptosis in lens epithelial cells promotes the cataractogenesis caused by oxidative stress (Mok et al., 2014), UV irradiation (Ayala et al., 2007; Godar, 1996; Xiao et al., 2012), diabetes (Takamura et al., 2003; Ye et al., 2013) and aging (Zheng and Lu, 2011), which could be p53-dependent (Yan et al., 2006). These studies indicate that tight control of p53 level and activity and subsequent cell death is crucial for lens development and pathology.
To prevent the detrimental effects of p53 on cell growth, cells recruit two physiological inhibitors of this protein, MDM2 (called HDM2 in human) and MDMX (also called MDM4, an MDM2 analog), so that p53 activity is restrained below an undetectable level in most of the normal cells under normal conditions (Chen et al., 1993; Kawai et al., 2007; Kruse and Gu, 2009; Linke et al., 2008; Momand et al., 1992; Oliner et al., 1993; Wade et al., 2010). Both MDM2 and MDMX directly inhibit p53’s transcriptional activity by binding to the transactivation domain of p53 with similar affinities (Shadfan et al., 2012). MDM2, but not MDMX, can also act as an E3 ubiquitin ligase to mediate p53 protein degradation (Kruse and Gu, 2009). Although both MDM2 and MDMX null mice died at the early stage of embryonic development due to p53 over-activation, loss of MDM2 in embryos caused the lethality at implantation, which occurred several days earlier than the lethality due to MDMX deletion (Montes de Oca Luna et al., 1995; Parant et al., 2001), suggesting the regulation of p53 by MDM2 and MDMX during embryogenesis may vary in a temporal and tissue-specific manner. It has been indicated that MDM2 forms a heterodimer or heterooligomer with MDMX through their C-terminal domains to inhibit p53 during early embryogenesis (Huang et al., 2011; Pant et al., 2011; Sharp et al., 1999; Tanimura et al., 1999). However during later developmental stages of organ morphogenesis and adult life, MDM2 and MDMX may independently or cooperatively regulate p53 to delicately control cell proliferation, differentiation and death so that each organ maintains unique shape and function, which is supported by the more severe phenotypes observed with conditional deletion of MDM2 in cardiomyocytes (Grier et al., 2006; Xiong et al., 2007), central nervous system (Xiong et al., 2006) and smooth muscle cells (Boesten et al., 2006), compared to those with loss of MDMX. Besides p53, pRB (retinoblastoma protein) is another major tumor suppressor that executes its function by controlling G1 phase and cell proliferation (Weinberg, 1995), and deletion of Rb gene or abnormal expression of pRB has been highly related to several human tumors including retinoblastoma (Clarke et al., 1992; Dyer et al., 2005; Jacks et al., 1992; Lee et al., 1992). Loss of Rb also causes lens developmental defects due to unchecked proliferation, impaired differentiation and p53-dependent apoptosis in lens fiber cells (Morgenbesser et al., 1994). It has been shown that MDM2 accelerates pRB protein degradation through both ubiquitin-dependent and -independent manners (Miwa et al., 2006; Uchida et al., 2005), but surprisingly, MDMX binds to pRB to inhibit MDM2 mediated pRB downregulation (Uchida et al., 2006). All these previous studies suggest that MDM2 and MDMX could be crucial during lens development and lens disease progression to control the lens cell proliferation, differentiation and cell death, by targeting either p53 or pRB, or both. To address this issue, in the current study we take advantage of the mice with MDM2 or MDMX conditional alleles and deleted MDM2 or MDMX gene by crossing them with the mice expressing Cre beginning at E9.5 under the control of the Pax6 enhancer specifically in both lens epithelial progenitors and pancreatic endocrine progenitors (LE-Cre) (Ashery-Padan et al., 2000; Ashery-Padan et al., 2004). Interestingly, we found that loss of MDM2 results in a more severe phonotype even though both MDM2 and MDMX are required to maintain normal lens development by controlling p53 activity, and provided evidence for the tissue specific difference in the roles of MDM2 and MDMX in development. Also, we found that homozygous loss of p53, but not Rb, alleles can completely rescue the defects of lens development caused by deleting either MDM2 or MDMX, indicating that p53, but not Rb, is the downstream target of MDM2 or MDMX, leading to lens dysmorphogenesis.
The MDM2FL/FL, MDMXFL/FL, p53FL/FL, RbFL/FL and Le-Cre transgene mouse lines were described previously (Ashery-Padan et al., 2000; Grier et al., 2006; Grier et al., 2002; Jonkers et al., 2001; Marino et al., 2000), and individually maintained with homozygotes. Briefly, the Le-Cre transgene expression starts at around E9, and the Cre-mediated recombination is evident since E9.5 in most cells of the surface ectoderm, and subsequently in the SE-derived eye structures including the developing lens, cornea, conjuctiva, skin of the eyelids and pancreatic endocrine precursors (Ashery-Padan et al., 2000). To obtain the mouse lines with MDM2 or MDMX specific deletion in developing lens, Le-Cre mice were crossed with either MDM2FL/FL or MDMXFL/FL homozygotes, then Le-Cre; MDM2FL/+ or Le-Cre; MDMXFL/+ heterozygous offsprings were further crossed with MDM2FL/FL or MDMXFL/FL homozygotes, respectively. Le-Cre; MDM2FL/FL and Le-Cre; MDMXFL/FL mice were compared with MDM2 FL/+ and Le-Cre; MDM2 FL/+ mice, or MDMX FL/+ and Le-Cre; MDMX FL/+, respectively. To obtain Le-Cre; MDM2FL/FL; p53FL/FL mice, Le-Cre; MDM2FL/+ were first crossed with p53FL/FL mice, then Le-Cre; MDM2 FL/+; p53FL/+ offsprings were further crossed with MDMX FL/+; p53FL/+ offspring. To obtain Le-Cre; MDMXFL/FL; p53FL/FL mice, Le-Cre; MDMXFL/FL mice were first crossed with p53FL/FL mice, then Le-Cre; MDM2 FL/+; p53FL/+ offsprings were further crossed with MDMX FL/+; p53FL/+ offsprings. The mice from these breeders were recruited in the experiments as shown in Fig. 6, including MDM2 FL/FL (or MDMX FL/FL); p53FL/FL, Le-Cre; MDM2 FL/FL (or MDMX FL/FL); p53FL/+, and Le-Cre; MDM2 FL/FL (or MDMX FL/FL); p53FL/FL mice. The Le-Cre; MDMXFL/FL (or MDM2FL/FL); RbFL/FL mouse line was obtained through a similar mating strategy as described above. Their genotypes were determined by PCR analysis of the genomic DNA extracted from their tails. For the embryo collection, noon of the day of vaginal plug observation was considered as E0.5 of embryogenesis. For BrdU staining, 200mg/kg body weight of BrdU was administrated to mice via intraperitoneal injection 2 hrs before their embryos were harvested.
All animal experiments were conducted in accordance with the National Institutes Health “Guide for the Care and Use of Laboratory Animals” and were approved by the Institutional Animal Care and Use Committee at Tulane University School of Medicine and our animal protocol number is 4257.
Antibodies used in the study include rabbit polyclonal anti-p53 (fl393, santa cruz, 1:100 for immunostaining), rabbit polyclonal anti-Pax6 (PRB-278P, Convance, 1:200 for immunostaining), mouse monoclonal anti-MDM2 (SMP14, Santa Cruz, 1:50 for immunostaining), rabbit polyclonal anti-BrdU (Santa Cruz, 1:100 for immunostaining) and mouse monoclonal anti-Ki67 (BD, 1:100 for immunostaining). The fluorescein In situ cell death detection kit was purchased from Roche.
Mouse embryos were isolated and rinsed with PBS, followed by fixation with 4% PFA overnight, serial dehydration and paraffin embedding. Paraffin blocks were then cut into 8μm-thick sections, and subjected to hematoxylin/eosin staining to determine possible morphological changes.
Basically, 8μM-thick tissue sections underwent regular deparaffinization and rehydration, and then were dipped into Hematoxalin solution for 3 min and Eosin solution for 45 sec, followed by dehydration in 70%, 95% and 100% ethanol and xylene. The slides were finally covered with permanent mounting medium (Vector Laboratories, CA) and stored at RT, and scanned using a microscope equipped with a digital camera (Zeiss 200) with a 10× objective.
Immunohistochemistry was performed with 6μM-thick cryosections. Briefly, mouse embryos were fixed in 4% PFA, put into 30% sucrose overnight, embedded with O.C.T, and cut into 6μm-thick sections. The cryosections were boiled in fresh citrate buffer (10mM Sodium Citrate, 0.05% Tween-20, pH 6.0) using a steamer for 40 min for antigen retrieval. After being blocked with blocking buffer (5% Goat serum, 0.3% TritonX-100 in 1×PBS) for 1h at RT, the sections were incubated with primary antibodies in a humid chamber at 4°C overnight followed by incubation with Alexa 488 or 594-conjugated goat-anti mouse or goat-anti rabbit IgG secondary antibodies (Biorad) at RT for 30 min, then the sections were washed with 1× PBS for 3 times, and covered with anti-fading buffer. For some immunofluorescence staining, the sections were incubated with HRP-conjugated goat-anti mouse or goat-anti rabbit IgG secondary antibody after incubation with a primary antibody, and then subjected to FITC tyramide signal amplification (TSA)-plus or cyanine 3 (Cy3) TSA-plus (PerkinElmer) incubation for 10 min to amplify the signal. Images were obtained under a fluorescence microscope (Zeiss 200) with a 20× objective. Cell proliferation was measured as the ratio of BrdU or Ki67-positive cells to DAPI-positive cells in lens epithelia by one-way ANOVA analysis. Four mice per group were recruited to obtain statistical significance.
Apoptosis was also determined by in situ TUNEL staining, using the Fluorescein In situ cell death detection kit (Roche) according to manufacturer’s instructions. Briefly, the cryosections were permeabilized in permeabilisation buffer (0.1% TritonX-100, 0.1% sodium citrate) at RT for 5min, followed by incubation with TUNEL reaction mixture at 37 °C for 1h, and mounted with anti-fading buffer after rinsing with PBS. Images were obtained under a fluorescence microscope equipped with a digital camera (Zeiss 200) with a 20× objective.
Data were reported as means ± SEM with N being the sample size. Comparisons among different groups were conducted by using One-way ANOVA. Probability values of P<0.05 were considered statistically significant.
To obtain mice with MDM2 deletion specifically in developing lens, we first crossed Le-Cre; MDM2FL/+ with MDM2FL/FL mice as described in the Methods section, and the genotypes of their offsprings were determined by PCR analysis as shown in Fig 1A. Four genetic types of pups including Le-Cre; MDM2FL/FL, Le-Cre; MDM2FL/+, MDM2FL/FL and MDM2FL/+, were obtained with a normal Mendelian ratio as shown in Fig. 1B, indicating the deletion of MDM2 in developing eyes and pancreas didn’t cause embryonic lethality. However, all the Le-Cre; MDM2FL/FL pups developed hyperglycemia with random blood glucose level higher than 400 mg/dl (data not shown), and died within one week after birth (Fig. 1B). In this study, we only focused on lens morphogenesis, as the pancreatic phenotypes will be reported separately. We found that deletion of MDM2 floxed alleles by Le-Cre results in severe lens dysmorphogenesis. At E9–9.5, the surface ectoderm competent for the presumptive lens was roughly normal in Le-Cre; MDM2FL/FL compared to MDM2FL/+, MDM2FL/FL or Le-Cre; MDM2FL/+ mice (bottom panels in Fig. 1C), as indicated by the immunofluoresence staining of Pax6. However, at E10 the lens placode was unable to develop into lens pit in the MDM2 deletion mice as indicated by DAPI staining (mid-lower panels of Fig. 1C). At E12.5, the lens vesicle was formed in control mice, but in the MDM2 conditional knockout mice by contrast, the lens tissues disappeared completely, and the eye space was filled up with twisted retina (mid-upper panels of Fig. 1C). After birth, intact eyes developed normally in the control group, but aphakia was seen in all the MDM2-null mice (top panels of Fig. 1C). Since there was no obvious lens defect observed in Le-Cre; MDM2FL/+, MDM2FL/+ and MDM2FL/FL (data not shown), we combined the three groups as a control afterward (Fig. 1C). Together, these results indicate that loss of MDM2 during embryogenesis causes irreversible defects of lens development as early as E9.5–10, leading to congenital aphakia.
Since previous studies demonstrated that loss of MDM2 or MDMX causes developmental defects with tissue specificity (Boesten et al., 2006; Grier et al., 2006; Hilliard et al., 2011; Xiong et al., 2007; Xiong et al., 2006), we also wanted to determine whether lens would undergo dysmorphogenesis when the MDMX gene is knocked out in the developing lens by using the same genetic approach as described above. After MDMXFL/FL mice were crossed with Le-Cre; MDMXFL/+ mice, 4 genetic types of pups, including Le-Cre; MDMXFL/FL, Le-Cre; MDMXFL/+, MDMXFL/FL, and MDMXFL/+were obtained with a normal Mendelian ratio (Fig. 2A and B). No aberrant eye-related phonotypes were observed in adult Le-Cre transgenic or Le-Cre; MDMXFL/+ mice (Fig. 2B and C). Interestingly, unlike Le-Cre; MDM2FL/FL mice, Le-Cre; MDMXFL/FL mice were viable, but eyeless (Fig. 2B). Histological analysis showed that the lens pit is formed normally in conditional MDMX deletion mice at E10. However, the development of lens vesicle was delayed at E10.5, and the shape of their lens vesicle cells was dramatically narrower and thinner than that of the control group (Fig. 2C). Even though the development of their retinae was relatively normal, lens development was completely blocked at E12.5 and finally progressed into aphakia at adulthood (Fig. 2C). In addition, loss of MDMX might exert some effect on the development of other eye structures such as eyelids, suggested by the closed eyelids in the MDMX knockout mice (Fig. 2C). These results demonstrate that loss of MDMX during embryogenesis causes retardation of lens morphogenesis at around E10.5, leading to adult aphakia similar to that of mice with MDM2 deletion.
As described above, MDM2 and MDMX are two physiological inhibitors of p53 and work together as a complex and/or independently to control p53 level and activity by directly blocking p53’s transactivational activity and promoting its ubiquitination-dependent proteasomal degradation, in a spatial-temporally specific manner. Thus, we wanted to determine if p53 is induced in the lens tissue lacking either MDM2 or MDMX. Indeed, the p53 level was markedly increased in the lens epithelial cells of E10 embryos without MDM2 gene, whereas p53 was undetectable in the control mice, as shown by the immunofluorescence staining in Fig. 3A. Likewise, loss of MDMX also induced p53 protein level dramatically in the lens epithelial cells as early as E10 when the Le-Cre was shown to be expressed at a high level (Ashery-Padan et al., 2000). The expression of Pax6, which is highly expressed in the ectoderm and developing lens, didn’t display any change upon MDM2 or MDMX deletion (Fig. 3A and B), indicating that the lens induction and differentiation that are mediated by this transcriptional factor during early embryogenesis was not affected by loss of either MDM2 or MDMX. Of note, we used FITC TSA-plus (at E10) or Cy3 TSA-plus fluorescence system (at E10.5 and E11) to enhance the p53 signal induced by MDMX deletion (Fig. 3B right), otherwise the signal would be undetectable by our regular immunofluorescence staining procedure. On the contrary, p53 signal was easy to be detected in the MDM2 deleted lens epithelial cells, implying that a much higher level of p53 was induced by loss of MDM2 than that by loss of MDMX.
To further investigate if p53 activation in MDM2 and MDMX deficient lens might lead to any cellular consequences, such as cell death and proliferation, we then assessed cell apoptosis by TUNEL assay and the cell proliferation rate by BrdU incorporation assay. As shown in Fig. 4A and B, cell apoptosis was remarkably induced in the defective lens epithelial cells of both MDM2 and MDMX knockout mice, but could not be detected in the control lens. Similarly, the cell proliferation rate decreased from around 75% to 28% in the MDM2 deleted mouse lens at E10, and from 68% to 50% in the MDMX deleted mouse lens at E10.5 (Fig. 5A and B). These results indicate that p53 is over-activated in both MDM2 and MDMX deficient lens tissues, consequently leading to aberrant cell apoptosis and impaired cell proliferation during lens development.
As mentioned before, earlier studies showed that MDMX regulates RB protein level by blocking the MDM2-mediated RB degradation (Uchida et al., 2006), suggesting the possible involvement of RB in the MDM2 or MDMX deletion-caused lens defects. To test this possibility, we generated MDM2/RB and MDMX/RB lens-specific double knockout mice. As shown in Fig. S1, loss of RB exerted no evident effects on either MDMX or MDM2 deletion-mediated lens defects (Fig. S1). MDMX/RB double knockout mice developed the same eyeless phenotype as that observed in MDMX single knockout mice. Also, MDM2/RB double knockout mice died within 1 week after birth with defective lens. Although we didn’t further investigate if RB deletion could affect the onset of lens defects during embryogenesis or not, these results at least indicate that RB deregulation is not the causative factor for lens dysmorphogenesis observed in MDM2 or MDMX deficient mice.
Because p53 is the most important target of MDM2 and MDMX (Kruse and Gu, 2009; Wade et al., 2010), and activated in the MDM2- or MDMX-deleted lens epithelial cells (Fig. 3), we tested if the activated p53 might be the cause for the cell death, proliferation abnormalities and subsequent developmental defects of lens in the MDM2- and MDMX-deleted mice by introducing p53 floxed alleles into MDM2 or MDMX conditional knockout mice and deleting p53 gene by Le-Cre. The genotype of p53 floxed allele was identified by PCR (Fig. 6A), and the loss of p53 protein expression was confirmed by the immunofluorescence staining (Fig. 7A and Fig. 8A). As expected, the lens defect caused by either MDM2 or MDMX knockout was completely rescued by p53 deletion. Similar to the control mice, the lens pit was formed at E10 in MDM2 deleted mice and the lens vesicle was formed and detached from the ectoderm at E12.5 in MDMX deleted mice, which finally developed into normal eye structure in the adult mice with MDM2/p53 or MDMX/p53 double knockout (Fig. 6B, C, D, Fig. 1B, C, Fig. 2B and C). However, loss of one copy of p53 seemed not to be enough to counteract the MDMX or MDM2 deletion caused eye defects (Fig. 6B, C and D). In addition, loss of two copies, rather than one copy, of p53 completely rescued the MDM2 mediated neonatal lethality from pancreatic development defects (Fig. 6B), which will not be discussed in this study. Consistently, the MDM2/p53 or MDMX/p53 double knockout lens epithelial cells did not display any detectable cell apoptosis (Fig. 7A and and8A),8A), and the cell proliferation rate restored to around 75% in both MDM2/p53 and MDMX/p53 double knockout lens epithelia at E10 or E12.5, similar to that of the control mice, as shown in Fig. 7B and and8B.8B. Taken together, these results demonstrate that p53 activation, leading to induction of cell death and inhibition of proliferation, is the main reason for the lens developmental abnormality in either MDM2 or MDMX conditional null mice.
The fact that systemic deletion of MDM2 or MDMX causes p53-dependent embryonic lethality at very early stages indicates the indispensability and non-redundancy of both MDM2 and MDMX and their tight regulation on p53 during embryogenesis (Montes de Oca Luna et al., 1995; Parant et al., 2001). Even though the interaction or complex formation between MDM2 and MDMX has been shown to be essential for controlling p53 level and activity during early embryogenesis (Huang et al., 2011; Pant et al., 2011), more subtle studies using loxp-Cre system demonstrate that difference may exist between MDM2 and MDMX on controlling p53 to maintain the balance of cell death and proliferation, and ensure the normal development and function of different tissues under different physiological and pathological conditions (Boesten et al., 2006; Grier et al., 2006; Xiong et al., 2007; Xiong et al., 2006). Lens development is a relatively independent process integrating cell proliferation, differentiation and apoptosis, which provides an ideal system to investigate more clearly the regulation of MDM2 and MDMX on p53, and the delicate cooperation and different responsibilities of these two p53 regulators during development. In the current study, we demonstrate that both MDM2 and MDMX are indispensable to maintain the early embryonic development of lens. Deletion of either MDM2 or MDMX in lens epithelium by the loxP-Cre system results in abnormal cell apoptosis and impaired cell proliferation, and consequently irreversible lens defects and aphakia which can be completely rescued by the loss of p53. Notably, compared to the lack of MDMX, MDM2 deletion-caused lens developmental deficiency happens earlier. This difference might be because that though both MDM2 and MDMX have been shown to bind the transactivation domain of p53 with similar affinities and to inhibit p53 activity equivalently (Bottger et al., 1999), the former also possesses an intrinsic E3 ubiquitin ligase, and thereby can mediate p53 proteasomal turnover perhaps independently of MDMX in vivo in a tissue-specific and embryonic stage-specific fashion (Boesten et al., 2006; Grier et al., 2006; Haupt et al., 1997; Kubbutat et al., 1997; Michael and Oren, 2003; Pant et al., 2011). The different actions between MDM2 and MDMX on p53 might also account for their differential roles in organogenesis during development as further discussed below.
Biochemical studies have demonstrated that MDM2 and MDMX interact with each other directly through their RING domains (Tanimura et al., 1999). Through mechanisms that still remain debated, MDMX is shuttled from the cytoplasm into the nucleus (Elias et al., 2005; Gu et al., 2002; Li et al., 2002; Migliorini et al., 2002), where it binds to and works together with MDM2 to inactivate p53 (Li et al., 2003; Wang and Jiang, 2012). The physiological importance of this MDM2-MDMX binding in monitoring p53 level and activity is elegantly validated by two genetic studies. These studies show that disruption of the MDM2-MDMX interaction or dissociation of the complex by knocking in a mutant MDMX that is unable to bind to MDM2, but retains its ability to bind to p53, leads to high p53 activation during mouse embryogenesis, causing vast apoptosis, impaired cell proliferation, and consequent lethality at E9.5, which could be completely rescued by homozygous deletion of TP53 (Huang et al., 2011; Pant et al., 2011). In contrast to this general role in embryogenesis, MDM2 and MDMX appear to play an independent role in specific organ formation and function maintenance. This is revealed by several conditional knockout studies. For example, conditional inactivation of MDM2, but not MDMX, in embryonic heart by α-MHC Cre led to cardiac development abnormalities and embryonic lethality (Grier et al., 2006). By contrast, loss of MDMX in heart was not embryonic lethal, instead resulted in dilated cardiomyopathy in adult mice and shorter longevity (Xiong et al., 2007). Also, deletion of MDM2, but not MDMX, in some quiescent cells, such as terminally differentiated smooth muscle cells (SMCs), contributed to severe cell loss and rapid death (Boesten et al., 2006). Deletion of MDM2 or MDMX in the central nervous system led to hydranencephaly at E12.5 or proencephaly at E17.5 (Xiong et al., 2006). Furthermore a latest study indicates that the regulation of MDM2 on p53 is tissue-specific in adult mice and becomes blunt with aging, suggesting the difference of p53 controlling by its inhibitors exists between embryogenesis and adulthood (Zhang et al., 2014). Regardless of the phenotype diversity, all the defects caused by MDM2 or MDMX deletion can be completely rescued by concomitant deletion of TP53, indicating that the functional difference and tissue specificity of MDM2 and MDMX are p53-dependent. Our study as presented here adds another example of these differences in terms of the role MDM2 or MDMX plays in embryonic organogenesis. Specifically, even though both MDM2 and MDMX deletion caused lens developmental defects, the phenotype occurred earlier in MDM2-null eyes than that of MDMX-null eyes (Figs. 1–2). Again these defects of lens development were completely rescued by the loss of p53, demonstrating that a tight control of p53 level and activity by both MDM2 and MDMX is also essential for normal lens morphogenesis. Also, MDMX possesses non-redundant function of MDM2 on p53 regulation during lens development, and neither of them can compensate the loss of the other. The observation of more severe and earlier lens defect resulting from lack of MDM2 than that from lack of MDMX suggests that MDM2 and MDMX are able to suppress p53 by different ways, separately or synergistically, more than forming a complex, which is also indicated by the much stronger p53 signal induced by MDM2 deletion than that induced by MDMX deletion, as mentioned in the result section (Fig. 3). Further studies using conditional knock-in strategy of MDMX or MDM2, which eliminates the binding of MDM2 and MDMX, meanwhile maintains their function on p53, would help further illustrate the independence and interaction of MDM2 and MDMX in regulation of p53 during development and function maintenance of all specific tissues.
Previous studies showed that MDM2 and MDMX can also regulate pRB protein level (Miwa et al., 2006; Uchida et al., 2006; Uchida et al., 2005). Lack of Rb expression in lens fiber cells induced p53 dependent apoptosis and consequent lens defects (Morgenbesser et al., 1994). These studies suggest that pRB may be one of the factors that could link MDM2 or MDMX to lens development. We thus also tested this possible link in vivo. However, we did not observe any obvious effect of Rb deletion on the defects of lens formation caused by either MDMX or MDM2 deletion (Fig. S1). Our study demonstrates that p53, but not pRB, is the reason for why MDM2 or MDMX deletion causes dysmorphogenesis of lens. Our study also suggests that pRB might not interplay with these two p53 inhibitors during embryonic lens development, even though we didn’t investigate if either MDM2 or MDMX deletion could influence the expression of pRB in developing lens.
Abnormal p53 expression has been linked to pathogenesis of certain eye diseases. Excess p53 levels induced by ultraviolet (UV) radiation or galactose exposure result in improper apoptosis of lens epithelial cells (LECs), which is the major mechanism underlying the UV-caused, or galactosemic and diabetic cataracts (Ayala et al., 2007; Godar, 1996; Takamura et al., 2003; Xiao et al., 2012; Ye et al., 2013). Deficiency of some genes required for normal lens development, such as OREBP and Ncoa6, induces DNA strand breaks and cell cycle arrest, resulting in incomplete elongation and differentiation of lens fiber cells and nuclear cataract, due to the activation of p53 (Wang et al., 2010; Wang et al., 2005). On the other hand, lack of p53 in lens leads to the accumulation of the fiber cells in the posterior plaques, and then the fiber cells continue to proliferate and fail to withdraw from the cell cycle to form proper terminal differentiated fiber cells (Wiley et al., 2011). Consistently, p53 may play a role in preventing the formation of steroid- or radiation-induced posterior subcapsular cataracts by suppressing the improper proliferation of fiber cells and promoting the death of the fiber cells that enter the cell cycle (Wiley et al., 2011). Besides the cataracts, over-activated p53 and subsequent apoptosis of LECs have also been suggested to be associated with microphthalmia/anophthalmia caused by some genetic deficiency (Hettmann et al., 2000; Jaramillo-Rangel et al., 2013). In line with these previous studies, our study as presented here also identified the essential role of the p53 pathway in aphakia. Although the exact mechanism underlying the aberrant activity of p53 during the progression of some eye diseases remains unknown, it is likely that p53 is uncontrollable by its most important regulators MDM2 and MDMX under some pathological conditions, as suggested by this study. Binding to MDM2 and/or MDMX is the most essential step for p53 to be regulated. Disrupting the interaction between p53 and these two inhibitors has become an intriguing strategy for developing cancer therapy (Issaeva et al., 2004; Reed et al., 2010; Shangary et al., 2008; Vassilev et al., 2004), and it has been implicated in many pathological progressions of normal tissues as well (He et al., 2014; Hilliard et al., 2011; Liu et al., 2014; Xiong et al., 2007). For example, as identified by our previous studies and other groups, DNA damage causes phosphorylation of MDMX, followed by enhanced binding of 14-3-3 to phosphorylated MDMX, leading to the release of p53 from MDMX and subsequent activation of p53 (He et al., 2014; Jin et al., 2006; Wang et al., 2009). Also, ribosomal stress could be another outcome of DNA damage (Antoniali et al., 2014; Lim et al., 2013; Nalabothula et al., 2010), which further induces the binding of ribosomal proteins to MDM2, resulting in the disassociation of MDM2 and p53, and subsequent p53 activation (Dai and Lu, 2004; Sun et al., 2007; Zhang and Lu, 2009; Zhang et al., 2003). Inability of MDM2 or MDMX to restrain p53 through the above-mentioned pathways might be the underlying mechanisms by which DNA damage or other lesions activates p53 during the progression of some p53-related ophthalmic disorders, such as some types of cataract and microphthalmia. Further studies are needed to elucidate these possibilities and hypotheses.
In summary, this study demonstrates for first time that both MDM2 and MDMX are necessary to maintain the normal morphogenesis of lens, and loss of either of them causes aberrant activation of p53 and subsequent cell apoptosis, leading to lens defects, which cannot be compensated by the existence of any single one of the two p53 inhibitors. This suggests that MDM2 and MDMX function independently or synergistically during lens development to control p53, more than forming a hetero-complex. The significance of the MDM2/MDMX/p53 pathway could also be highlighted in some p53-related ophthalmic disorders, including microphthalmias and cataracts, but deserve further investigation in the near future.
We thank Guillermina Lozano for providing us with the mouse MDM2FL/FL and MDMXFL/FL lines, Shelya X. Zeng for general lab supports, and the members in the Lu laboratory for active discussion. H.L. was partly supported by NIH-NCI grants CA095441 and CA172468, and the Reynolds and Ryan Families chair fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.