Loss of Nfkb1 leads to early onset aging

Loss of Nfkb1 accelerates observable age-related characteristics and leads to a decrease in lifespan

The importance of p50/NF-(B1 in mediating apoptotic signaling raised the question of whether loss of this subunit leads to a predisposition for the development of chronic disease. We therefore followed cohorts of Nfkb1^+/+ and Nfkb1^−/− animals over an extended period of time. While the most commonly used Nfkb1^−/− mouse is found on the original B6/129 cross background [14], it is well documented that animal strain plays a significant role in lifespan and disease predisposition [16]. Therefore, we obtained Nfkb1^−/− animals that have been backcrossed to C57BL6 mice for 12 generations and interbred them with wildtype (wt) C57BL6 mice to obtain single strain littermates. Because of the susceptibility of Nfkb1^−/− mice to infection when housed under normal conditions [14], animals were followed in a pathogen-free environment and sacrificed when they displayed signs associated with a terminal state.

Consistent with the original description of these animals, other than being slightly smaller than wt, Nfkb1^−/− mice are identical to their littermates and do not display any overt differences for the first six months of life. However, at 12 and 18 months, compared to age-matched Nfkb1^+/+ animals, Nfkb1^−/− mice have a higher incidence of several observable age-related characteris-tics including rough fur coat, alopecia, rectal prolapse and paraphimosis (Fig. 1A, Table 1).

These observations suggest that loss of Nfkb1 leads to premature aging. Consistent with this hypothesis, when Nfkb1^+/+, Nfkb1^+/− and Nfkb1^−/− littermate mice are followed for eighteen months, significantly fewer Nfkb1^−/− mice remain alive than Nfkb1^+/+ (Fig. 1B). While all Nfkb1^+/+ animals are alive at 1 year, only 70 % of Nfkb1^−/− mice remain alive at this time (P<0.01, Log rank). Moreover, heterozygote mice have an intermediate survival that is significantly greater than that of Nfkb1^−/− mice (p<0.05, Log rank). Notably, autopsy reveals no evidence of overwhelming infection or sepsis in the Nfkb1^−/− animals nor do they have evidence of increased tumor formation compared to controls.

Loss of Nfkb1 leads to premature age-related skeletal changes

Given the findings with loss of Nfkb1, we next examined whether Nfkb1^−/− animals also have premature age-related skeletal changes. Gross inspection suggests that compared to Nfkb1^+/+ mice, Nfkb1^−/− animals have an in-crease in kyphosis (Fig. 1A), a finding closely associated with advanced age [17]. To more objectively examine kyphosis, spiral CT was performed on 12-month old Nfkb1^−/− and Nfkb1^+/+ animals and Cobb's angle measured (Fig. 1C). Nfkb1^−/− mice have significantly higher Cobb's angle (p<0.05), confirming that loss of Nfkb1 is associated with increased kyphosis. Age-related kyphosis has been linked to osteoporosis and decreased cortical bone density [18]. We therefore examined cortical bone density using high-resolution micro–computed tomography (μCT) imaging of femurs. Although at 8 weeks there is no difference in cortical bone characteristics between Nfkb1^+/+ and Nfkb1^−/− animals, by 12 months Nfkb1^−/− mice have a significantly lower cortical thickness and cortical bone fraction (p<0.05) and higher cortical porosity (p=0.05) than Nfkb1^+/+ (Fig. 1D). These findings further support the hypothesis that loss of Nfkb1 leads to premature mouse aging.

Loss of Nfkb1 results in tissue inflammation and increased CNS gliosis

We next performed histological analysis of tissues, and consistent with a previous report that Nfkb1^−/− mice have increased inflammatory cell infiltrate in the liver [19], we note an increase in inflammation in the kidney (Fig. 2A) and skeletal muscles (data not shown) of Nfkb1^−/− compared to Nfkb1^+/+ mice. Examination of age-matched brain sections does not reveal differences in inflammation or other histomorphologic features (Fig. 2B). However, given the association between aging and CNS white matter glial fibrillary acidic protein (GFAP) staining [20], we examined GFAP reactivity in these animals. A significantly higher level of GFAP labeling is observed in brains of 18 month-old Nfkb1^−/− animals compared to Nfkb1^+/+ (p<0.01, Fig. 2C). These findings confirm the increase in inflammation in Nfkb1^−/− animals and also demonstrate that these mice have increased CNS gliosis relative to age-matched Nfkb1^+/+ littermates.

Taken together, the above results indicate that loss of Nfkb1 leads to the premature appearance of multiple age-related findings underlined by a decrease in overall lifespan.

Loss of Nfkb1 results in premature cellular senescence

Primary cells have a limited lifespan in culture and undergo senescence following serial passage [5]. This observation, coupled with the accelerat ed aging of Nfkb1^−/− mice raises the question of whether loss of Nfkb1 affects the lifespan of cells in vitro. We isolated primary mouse embryonic fibroblasts (MEFs) from Nfkb1^+/+ and Nfkb1^−/− animals and examined their growth in culture. A significant decrease in population doubling (PD) rate is seen in Nfkb1^−/− MEFs compared to Nfkb1^+/+ with increasing time in culture (Fig. 3A). Importantly, this observation is noted using multiple separate MEF isolates harvested at different times. The decrease in Nfkb1^−/− cell number is associated with an increase in cellular senescence, relative to Nfkb1^+/+, as demonstrated by a change in cellular morphology to an enlarged and flattened shape and an increase in senescence associated-β-galactosidase (SA-β-gal) staining (Fig. 3B). In addition, we studied expression of the cyclin-dependent kinase inhibitors, p21 and p16 (CDKN1a and CDKN2a, respectively), factors that are closely associated with senescence [21]. Although both cell types have relatively equal expression of p21 and p16 initially, at later passage the expression of both proteins is higher in Nfkb1^−/− MEFs than similarly passaged Nfkb1^+/+ cells (Fig. 3C). Finally, to compare the rate of proliferation, S-phase-dependent incorpora-tion of the thymidine analog 5-bromodeoxyuridine (BrdU) was examined. Nfkb1^−/− MEFs have less BrdU-positive cells than Nfkb1^+/+ cells at a similar passage (Fig. 3D). Together, these results suggest that loss of Nfkb1 leads to a decline in PD rate due to an increase in senescence.

Loss of Nfkb1 leads to reduced apoptosis and increased phospho-H2AX accumulation

An association between cellular senescence and loss of apoptosis is well described [22]. Given that loss of Nfkb1 leads to decreased apoptosis in response to DNA alkylation damage [15] and that alkylation damage is induced by environmental and endogenous processes [23, 24], we examined spontaneous apoptosis in cultured MEFs. Primary Nfkb1^+/+ and Nfkb1^−/− MEFs were harvested and apoptosis levels assessed by analysis of annexin V binding. Nfkb1^−/− MEFs have a significantly lower level of spontaneous apoptosis than Nfkb1^+/+ MEFs at the same passage (p<0.05; Fig. 4A), an observation noted with multiple separate MEF isolates. Cellular senescence and animal aging have been linked to expression of markers of activation of the DNA damage response including phosphorylation of histone H2AX (γH2AX) [25-27]. Therefore, we examined MEFs for γH2AX foci formation. Nfkb1^−/− MEFs accumulate significantly more γH2AX foci than Nfkb1^+/+ following serial passage (p<0.05, Fig. 4B, Fig. Supplemental S1), indicating that loss of Nfkb1 leads to an increase in damage signaling.

The difference in spontaneous apoptosis and γH2AX foci in cultured cells raises the question of whether such changes are also evident in vivo. To examine apoptosis in vivo, we looked at TUNEL and caspase 3 staining in brain sections of 12-month old Nfkb1^+/+ and Nfkb1^−/− mice. Consistent with the MEF data, Nfkb1^+/+ mice have substantially more apoptotic cells in the cortex and dentate gyrus than Nfkb1^−/− mice as demonstrated by both TUNEL analysis (Fig. 4C) and caspase 3 staining (Fig. Supplemental S2). In addition, examination of γH2AX foci in both newborn and 12-month old mouse brains demonstrates that, as with MEFs, significantly more γH2AX foci are seen in the brains of Nfkb1^−/− compared to age-matched Nfkb1^+/+ animals (p<0.01, Fig. 4D).

These findings indicate that loss of Nfkb1 leads to a decrease in spontaneous apoptosis and an increase in DNA damage signaling, findings that are consistent with the increase in cellular senescence and premature aging seen in Nfkb1^−/− cells and animals.

Aging is associated with a relative decrease in p50 DNA binding

The premature aging of Nfkb1^−/− mice raises the question of whether loss of Nfkb1 is associated with physiological aging. To study this question, we looked at NF-(B subunit expression in young and old Nfkb1^+/+ brains. Although there is no change in p50 or p105 protein levels, a substantial increase in p52, the other non-rel NF-(B subunit, is noted in 18-month old compared to 1-month old brain tissue (Fig. 5A). The age-dependent increase in p52 is a progressive process as 6-month old animals have an intermediate level of p52 expression (Fig. Supplemental S3A). Minimal increase in p100 and p65 level is also noted in the aged brains (Fig. 5A). Interestingly, as with Nfkb1^+/+ tissue, an increase in p52 is also seen in aged compared to young Nfkb1^−/− brains (Fig. Supplemental S3B).

NF-(B mediates its action by binding DNA, therefore to examine whether the increase in p52 is associated with a change in the composition of the DNA bound NF-(B dimer, gel shift assay was used. As previously reported with other tissue [28-30], a general increase in NF-(B DNA binding is evident in brains of old mice compared to young (Fig. Supplemental S3C). Remarkably, however, supershift analysis consistently demonstrates that while anti-p50 shifts the NF-(B band in young brains (Fig. 5B), in old brain anti-p50 has no effect and instead anti-p52 induces supershift. These results indicate that p52 replaces p50 in the DNA bound dimer in aged brain compared to young. We next examined whether the change in NF-(B dimer composition is also present in other tissue. Consistent with the findings in the brain, an increase in p52, but not p50, protein is seen in the liver of aged compared to young Nfkb1^+/+ mice (Fig. 5C). Similarly, aged Nfkb1^−/− mice have higher p52 protein levels in their liver than young animals (Fig. Supplemental S3D). Moreover, supershift analysis recapitulates the findings in brain tissue showing that while p50 binds DNA in young liver, it is replaced by p52 in the aged tissue (Fig. 5D). Supershift analysis also demonstrates that p65 remains part of the NF-(B dimer in both young and old tissue, consistent with previous observations in cells demonstrating that in the absence of p50, p52 dimerizes with p65 [15, 31].

Finally, we examined whether cultured cells also demonstrate changes in subunit levels with increasing passage. Although no change in p50 or p105 expression is noted in Nfkb1^+/+ MEFs over serial passage, a robust increase in p52, but not p100, is seen at late passage (Fig. 5E). Similarly, in Nfkb1^−/− MEFs an increase in p52, but not p100, is seen in late compared to early passage (Fig. Supplemental S3E).

Taken together, the findings in old compared to young tissue indicate that p50 is replaced by p52 in aged tissue and that the overall increase in NF-(B binding with age is related to an increase in p52. These findings suggest that loss of DNA bound p50 is a prominent feature of physiological aging.

Animals

Nfkb1^−/− animals that have been backcrossed for 12 generations with C57BL/6J mice (B6.Cg-Nfkb1^tm1Bal/J), and the corresponding C57BL/6J Nfkb1^+/+ control, were obtained (Jackson Laboratory) and kept in a pathogen-free environment with food and water available ad libitum and a 12-hr light/dark cycle. Mice were interbred and the resultant Nfkb1^+/+, Nfkb1^+/− and Nfkb1^−/− littermates used for long-term studies. All animals were maintained in accordance to a protocol approved by the Institutional Animal Care and Use Committee of the University of Chicago.

Generation and culture of MEFs

MEFs were harvested from 13-14 day Nfkb1^+/+ and Nfkb1^−/− embryos. Briefly, head and visceral organs were removed and remaining tissue chopped in cold, sterile PBS. The tissue was digested in 0.25% trypsin for 5 minutes at 37°C and then dispersed and spun down with fetal bovine serum (FBS). The supernatant was collected and spun down with high glucose DMEM supplemented with 10% FBS and 5% of a penicillin and streptomycin (Pen/Strep) cocktail. The cell pellet was then plated with DMEM and 20% FBS on a 10 cm dish. MEFs were slowly weaned off FBS with daily media changes over the next few days to reach a final concentration of 10% FBS. Subsequently, cells were passaged with each passage considered to be approximately 2 PDs. To examine growth curves, cells were plated starting between passage 2 and 3 (estimated to be PD5) at 25% confluency and passaged when they achieved approximately 75% confluency. Cells were passaged until they reached crisis and looked unhealthy.

SA-β-gal and γH2AX analysis

Primary MEFs at the indicated passage were plated onto 35 mm glass bottom plates (MatTek) and at 60 % confluency were fixed with formaldehyde for 20 minutes at 4°C. SA-β-gal staining was performed in the standard fashion. Cells were incubated overnight with freshly made staining solution and the percentage of cells positive for SA-β-gal was calculated from a total of 200 cells per plate. For γH2AX analysis, cells were incubated in rabbit anti-γH2AX antibody (Cell Signaling) for one hour followed by goat anti-rabbit FITC secondary antibody (Vector Laboratories Inc., Burlingame, CA, USA). Glass bottom coverslips were removed and mounted on slides using Fluoro Gel-II mounting medium containing DAPI (Electron Microscopy Sciences). Experiments for SA-β-gal and γH2AX analysis were repeated in triplicate.

Flow cytometry for annexin V binding and BrdU analysis

Annexin V and BrdU were analyzed using flow cytometry. Briefly, for detection of annexin V, cells were incubated with FITC-conjugated anti-annexin V, followed by propidium iodide (BD Biosciences), and immediately analyzed. For detection of BrdU, MEFs were incubated with 0.03 mg/ml BrdU for 1 hour, fixed in ice-cold methanol for 5 minutes and treated with 1.5 M HCl for 30 minutes at room temperature. Following blocking, cells were incubated with anti-BrdU monoclonal antibody (Santa Cruz) in 0.1% BSA for 1 hour. Cells were then incubated in donkey anti-mouse APC-conjugated secondary antibody (Millipore) for 30 minutes. Flow cytometry was performed using the BD LSR II (BD Biosciences) equipped with a 635-nm laser. Experiments were performed using at least 3 individually harvested cell populations per group, and each sample was run in triplicate. Data was analyzed using FlowJo software (TreeStar Inc.).

Cobb's angle measurement

To measure kyphosis, lateral images from a spiral CT scan of the spine of 12-month-old mice were acquired. Cobb's angle was calculated in the standard fashion by measuring the angle formed by the juncture of two parallel lines drawn between the last two vertebra found at the most rostral and caudal ends of the curvature. Images were analyzed using ImageJ (National Institutes of Health).

μCT analysis of femurs

Femurs were harvested from 8-week and 12-month old mice, cleaned of soft tissue and placed in a 20.5 mm tube. Samples were scanned horizontally, approximately 300 slices per sample, with a desktop cone-beam μCT scanner (SCANCO Model μCT40; SCANCO Medical). Femurs were then rotated to transverse slices and analyzed. Cortical bone was measured at the midshaft (50 % of the length of the femur), and then 50 slices were included on either side (100 slices total). Contours for cortical bone were drawn using the midshaft contour script (#16), and each was inspected & modified for best fit before analysis. Cortical bone was analyzed using μCT Evaluation Program V6.0 (SCANCO Medical) with the midshaft cortical bone script (#23) with multiple thresholds to gauge best fit, and a threshold of 350 was selected to be the most appropriate.

Histology, TUNEL staining and immunohistochemistry

For general brain histology, GFAP and caspase-3 immunohistochemistry, mice were sacrificed and brains and other tissue harvested and fixed in formalin. The tissue was processed for paraffin embedding and sectioned at 4 μm. Tissue slides were deparaffinized for hematoxylin and eosin staining, or incubated in either rabbit anti-GFAP antibody (DAKO) or rabbit anti-caspase-3 antibody (Millipore) overnight. GFAP and caspase-3 slides were then incubated with goat anti-rabbit secondary antibody (DAKO), and counterstained with hematoxylin. For γH2AX immunohistochemistry, fixed brain tissue from newborn or 12-month old mice was frozen and sectioned at 20 μm with a cryostat. Sections were incubated in rabbit anti-γH2AX (Cell Signaling) overnight followed by incubation with a goat anti-rabbit FITC secondary antibody (Vector Laboratories). Slides were coverslipped with Fluoro Gel-II mounting medium containing DAPI and analyzed using the Axiovert 200M inverted widefield fluorescence microscope (Zeiss). Counting for γH2AX-positive cells in brain tissue was performed in blinded fashion on a total of 9 sections per animal separated by a minimum distance of 150 μm. All γH2AX-positive cells were counted in each section under analysis and the total number of cells from all 9 sections reported. For TUNEL analysis, the Apoptosis Detection Kit was used according to the manufacturer's instructions (Millipore). Sections were counterstained with hematoxylin and the ratio of TUNEL-positive cells to TUNEL-negative cells was calculated from 3 random areas within each specified brain region (internal capsule, hippocampus and thalamus) under bright microscopy at a magnitude of 40x. Data represent the average number of positive cells counted in sections from three separate animals in each group.

Immunoblotting and electrophoretic mobility shift assay (EMSA)

For immunoblot analysis, harvested tissue was flash frozen and manually triturated upon thawing. Both tissue and MEFs were then lysed with RIPA buffer and the resulting protein boiled with Laemmli buffer and loaded on a 4-15% polyacrylamide gel. Following electrophoresis (1.5 h at 150V), protein was electrotransferred onto nitrocellulose membranes overnight. The following day, membranes were blocked in 1% milk with TBS and 0.1% Tween 20 for 2 hrs and then incubated overnight with the indicated antibodies (GAPDH from Cell Signaling; all other antibodies are from Santa Cruz). Membranes were then rinsed with buffer and incubated with fluorescent-tagged secondary antibodies (Thermo Scientific) for 2 hours. After rinsing, membranes were analyzed by computer-assisted image analysis (LI-COR Biosciences), with comparisons made between samples on the same gel/immunoblot.

For EMSA, frozen mouse tissue was rinsed in PBS buffer and spun at 1,500 x g for 5 min. Tissue was homogenized in extraction buffer (20 mM HEPES, pH 7.5, 100mM NaCl, 0.05% Triton X-100, 1 mM dithiothreitol (DTT), 5 mM sodium b-glycophosphate, 0.5 mM sodium orthovanadate, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg/mL aprotinin, 5 μg/mL leupeptin, 2 μg/mL pepstatin) with pestle on ice. Lysate was cleared by centrifugation at 12000g for 15 min at 4°C. EMSA was performed using 5 μg of extract and a ³²P-labeled double-stranded 30-mer oligonucleotides containing a single central κB-site (GGGACTTTCC) with ten flanking base pairs on either side as a probe. EMSA reactions contained 10 mM Tris-HCl (pH 7.5), 1 mM MgCl₂, 50 mM NaCl, 0.5 mM EDTA, 0.5 mM DTT, 4% glycerol, 50 ng/μl poly(dI·dC). Reaction mixtures were incubated at room temperature for 20 min prior to resolving on a 5% polyacrylamide gel (0.5x TBE) at ~10 V/cm for 2-4 hours at 4^oC. Supershift assays were performed using antibody cocktails specific to the indicated NF-κB subunits. Competition was performed by pre-incubating the mixture with cold specific and non-specific probe.

Statistical analysis

Results are expressed as mean ±SEM or SD where indicated. Statistical significance was taken as p<0.05 using a 2-tailed Student's t-test. For observable age-related changes, Pearson's chi-square test was used. Kaplan-Meier survival curves were analyzed by the Log-rank method.