Hypothalamic programming of systemic ageing involving IKK-[bgr], NF-[kgr]B and GnRH
guo zhang,juxue li,sudarshana purkayastha,y tang,hai zhang,y yin,bo li,gang liu,dongsheng cai
DOI: https://doi.org/10.1038/nature12143
IF: 64.8
2013-01-01
Nature
Abstract:Ageing is a result of gradual and overall functional deteriorations across the body; however, it is unknown whether an individual tissue primarily works to mediate the ageing progress and control lifespan. Here we show that the hypothalamus is important for the development of whole-body ageing in mice, and that the underlying basis involves hypothalamic immunity mediated by IκB kinase-β (IKK-β), nuclear factor κB (NF-κB) and related microglia–neuron immune crosstalk. Several interventional models were developed showing that ageing retardation and lifespan extension are achieved in mice by preventing ageing-related hypothalamic or brain IKK-β and NF-κB activation. Mechanistic studies further revealed that IKK-β and NF-κB inhibit gonadotropin-releasing hormone (GnRH) to mediate ageing-related hypothalamic GnRH decline, and GnRH treatment amends ageing-impaired neurogenesis and decelerates ageing. In conclusion, the hypothalamus has a programmatic role in ageing development via immune–neuroendocrine integration, and immune inhibition or GnRH restoration in the hypothalamus/brain represent two potential strategies for optimizing lifespan and combating ageing-related health problems. Ageing is characterized by the gradual and overall loss of various physiological functions, leading to the end of lifespan. Although the search for resolution of ageing pathology is continuing1, 2, 3, 4, 5, 6, research has shown that certain neurons can mediate environmental influences on ageing in Caenorhabditis elegans and Drosophila, and neural manipulations of insulin or insulin-like growth factor 1 signalling or uncoupling protein 2 were shown to affect lifespan in animals7, 8, 9, 10, 11. In this study, we have focused on the hypothalamus, a key brain region that is crucial for the neuroendocrine interaction between the central nervous system and the periphery. We asked whether the hypothalamus may have a fundamental role in ageing development and lifespan control, in addition to its critical involvement in basic life-supporting functions such as growth, reproduction and metabolism. In tackling this bold question, we increasingly appreciated that an atypical collection of hypothalamic inflammatory changes can broadly and causally underlie the development of metabolic syndrome components including being overweight, glucose intolerance and hypertension12, 13, 14, 15, and of note, all of these disorders are often related to ageing. Furthermore, we have noted recent literature showing that microglia are involved in neurodegenerative diseases16, 17, 18, 19, 20, 21, 22, 23, 24, which aligns with the appreciated connection between systemic immunity and ageing25, 26. Here, through targeting hypothalamic immunity/inflammation, we designed to test whether the hypothalamus is fundamentally important for ageing and lifespan control. In studying the potential role of the hypothalamus in ageing, we developed a strategy of targeting hypothalamic immunity, and as shown in our recent work12, 13, 14, 15, infection-unrelated inflammatory changes in the mediobasal hypothalamus (MBH) contribute to the development of various metabolic syndrome components, and the molecular basis is mediated crucially by NF-κB and its upstream IKK-β. Indeed, using phosphorylation of NF-κB subunit RelA to report NF-κB activation, we observed that although hypothalamic NF-κB was barely active in mice of young age (3–4 months), it was activated in the hypothalamus of mice at middle–old ages (11–13 months), and the activities further increased as the mice became older (22–24 months) (Fig. 1a, b). Agreeing with this observation, messenger RNA levels of many cytokines and immune regulators increased in the hypothalamus of old mice compared to the young group (data not shown). To visualize NF-κB activity in the MBH directly, we used an NF-κB reporter that induces green fluorescent protein (GFP) after the binding of NF-κB to its transcriptional response element in a lentiviral vector (Fig. 1c). After in vitro assessment of this approach (Supplementary Fig. 1a, b), we performed animal experiments by delivering this lentiviral NF-κB reporter into the MBH of mice at young, middle–old and old ages. A prolonged recovery period was used to minimize the procedure-related nonspecific effects on NF-κB. We found that GFP was negligible in the MBH of young mice (Fig. 1c), but was evident in the MBH of middle–old mice and became more profound in old mice (Fig. 1c, d), confirming that ageing is associated with hypothalamic NF-κB activation. We also injected this lentiviral NF-κB reporter into various other brain regions, and comparatively, the MBH was most sensitive to ageing-related NF-κB activation (Supplementary Fig. 1c–e). Of interest, immunostaining with the neuronal marker NeuN revealed that NF-κB activation in neurons was relatively modest under middle–old ageing, but became prominent when age further increased (Fig. 1c, d). Thus, ageing development is characterized by chronic activation of NF-κB-directed innate immune pathway predominantly in the hypothalamus. We then tested our proposed involvement of IKK-β and NF-κB in the hypothalamic control of ageing, and our experiments focused on the MBH. Using MBH-directed lentiviral gene delivery as we previously established12, 13, we delivered dominant-negative IκB-α (DNIκB-α) to inhibit NF-κB, and constitutively active IKK-β (CAIKK-β) to activate NF-κB in MBH neurons; MBH delivery of GFP in the same lentiviral system was used as the control (Supplementary Fig. 2a). Middle–old C57BL/6 mice received bilateral MBH lentiviral injections; use of middle–old age mice helped to eliminate developmental concerns, and, indeed, ageing retardation can be achieved through intervention starting at a middle–old age27. These mice with MBH delivery of DNIκB-α, CAIKK-β and control GFP were named MBH–IκB-α, MBH–IKK-β and MBH–ctrl mice, respectively, and all mice were maintained under pair feeding of a normal chow so that they had similar daily food intake. Our longitudinal follow-up revealed that MBH–ctrl mice displayed a typical pattern of lifespan (Fig. 2a), which indicated that our approach of MBH injection was technically suitable. Importantly, we found that lifespan significantly increased in MBH–IκB-α mice but decreased in MBH–IKK-β mice compared to controls (Fig. 2a). In parallel with lifespan analysis, separate mice were generated to evaluate ageing-related physiology and histology. We assessed cognition and muscle endurance of mice at ~6 months after gene delivery, at which hypothalamic NF-κB remained overactivated in MBH–IKK-β mice but suppressed in MBH–IκB-α mice (Supplementary Fig. 2b). In cognitive tests, we found that compared to controls, MBH–IκB-α mice performed better but MBH–IKK-β mice performed worse (Fig. 2b), and all of these mice were technically eligible for the test (Supplementary Fig. 2c). These mice were also subjected to a grip test, showing that ageing-related muscle weakness was attenuated in MBH–IκB-α mice but worsened in MBH–IKK-β mice (Fig. 2c). Furthermore, these mice were examined for a panel of histological biomarkers including muscle size, skin thickness, bone mass, and tail tendon collagen cross-linking. As shown in Fig. 2d–g, ageing-related changes of these biomarkers were dampened in MBH–IκB-α mice but exacerbated in MBH–IKK-β mice. Finally, given that these data were based on males, we further generated female mouse models, and results from females agreed with the observations in males (Supplementary Fig. 3). In summary, the hypothalamus has a unique role in the development of systemic ageing, and hypothalamic IKK-β and NF-κB represents a driving force in this process. To understand ageing-related hypothalamic immunity/inflammation further, we profiled microglia in the hypothalamus. Using immunostaining, we found that numbers of microglial cells in the MBH increased in an age-dependent manner (Fig. 3a, b). Overproduction of tumour necrosis factor-α (TNF-α) (Fig. 3a, c) and activation of NF-κB (Supplementary Fig. 4) were both detected in these microglial cells, indicating that they were inflammatory. We noted that under early ageing, NF-κB activation was already evident in hypothalamic microglia (Fig. 3a, b); however, this change was still modest in hypothalamic neurons (Fig. 1c, d). Also as observed, TNF-α overproduction was mostly limited to hypothalamic microglia during early ageing, but became prevalent across the MBH, which affected other neural cells (such as neurons) in this region. We additionally measured hypothalamic Tnfa mRNA levels in mice of different ages, and data obtained (Supplementary Fig. 5a) well correlated with cell counting of TNF-α immunostaining (Fig. 3c). It should be mentioned that TNF-α is a gene product of NF-κB and also acts to activate IKK-β and NF-κB. Overall, our data indicate that TNF-α is generated mainly by microglia during early ageing, and the paracrine actions of this cytokine on neighbouring cells is predicted to lead to ageing-associated neuronal IKK-β and NF-κB activation. In the literature, TNF-α is known to be neurotoxic or neuroprotective28, 29, 30, which may reflect the differential functions of soluble versus transmembrane TNF-α (ref. 30). In our ageing model, soluble TNF-α seems to be involved in IKK-β and NF-κB-mediated microglia–neuron crosstalk that controls systemic ageing. Subsequently, we generated a mouse model with IKK-β knockout in the MBH microglia through bilaterally delivering microglia-specific (CD11b promoter-driven) lentiviral Cre into the MBH of Ikbkblox/lox mice, and control mice were Ikbkblox/lox mice injected with Cre-deficient lentiviruses. Our assessment confirmed that Cre was delivered specifically in ionized calcium binding adaptor molecule 1 (Iba-1)-expressing microglia, and most of these cells in the MBH were induced with Cre (Supplementary Fig. 5b). By profiling these IKK-β knockout mice and matched controls, both of which were generated at a middle–old age, we observed that IKK-β ablation in microglia prevented against the increase of microglial cells over ageing (Fig. 3d and Supplementary Fig. 5c). Moreover, IKK-β ablation prevented ageing from inducing TNF-α expression not only in microglia but also in neighbouring cells. Such ageing-related hypothalamic microglia–neuron crosstalk via IKK-β and NF-κB led us to predict that microglia-specific IKK-β ablation might slow down ageing. To test this prediction, we continued to use this IKK-β knockout mouse model generated at a middle–old age, maintained them until old ages, and assessed their ageing manifestations. After technical evaluation (Supplementary Fig. 5d–f), we tested these mice using the Morris water maze, and data showed that microglia-specific IKK-β ablation reduced ageing-related cognitive decline (Fig. 3e–g). Furthermore, IKK-β ablation resulted in improvements in ageing-related muscle weakness (Fig. 3h) and tail collagen cross-linking (Fig. 3i). Altogether, hypothalamic microglia can act via IKK-β and NF-κB to contribute to the role of the hypothalamus in ageing development. We further resorted to a genetic model of brain-specific IKK-β knockout mice, N/Ikbkblox/lox mice, which we generated by breeding nestin-Cre with Ikbkblox/lox mice as described previously13. Compared to wild-type littermates with matched Ikbkblox/lox background, these knockout mice were developmentally indistinguishable in terms of brain size and gross morphology (Supplementary Fig. 6). We also compared Ikbkblox/lox mice to additional types of control, and confirmed that all these mice were similar across a spectrum of ageing-related physiological and histological changes (Supplementary Fig. 7). In this context, we profiled ageing-related physiology and pathology in N/Ikbkblox/lox mice and littermate wild types. At an old age, after technical assessment (Supplementary Fig. 8a–c), we subjected mice to the Morris water maze, and found that N/Ikbkblox/lox mice outperformed wild types (Fig. 4a). This cognitive improvement was specific to ageing, because young N/Ikbkblox/lox mice and wild types performed similarly (Supplementary Fig. 8d–h). Thus, although NF-κB seems to have a role in the development of hippocampal synaptic plasticity31, 32, 33, the net effect from suppressing brain IKK-β and NF-κB under the ageing model is cognitively beneficial. Using a grip test, we further found that compared to wild type, N/Ikbkblox/lox mice had a reduced extent of ageing-related muscle weakness (Fig. 4b). Also, as shown in Fig. 4c–h, N/Ikbkblox/lox mice were protected against ageing-induced muscle and skin atrophy, bone loss and collagen cross-linking. In addition to males, female N/Ikbkblox/lox mice were studied, and the findings were consistent (Supplementary Fig. 9). Notably, we did lifespan analysis by following a cohort of male N/Ikbkblox/lox mice and wild-type littermates. As shown in Fig. 4i, wild-type mice had a typical pattern of median and maximal lifespan; by contrast, N/Ikbkblox/lox mice showed a pronounced phenotype of longevity, with median lifespan 23% longer (P = 0.0002) and maximal lifespan 20% longer (P < 0.05) than wild types. We recognize that the longevity phenotype of this genetic model could be a result of IKK-β inhibition jointly in neurons and glia, as nestin-Cre is known to target neural stem/progenitor cells and derived neurons and glia. To summarize, longevity in this genetic model considerably recapitulates ageing retardation from hypothalamic IKK-β and NF-κB inhibition, and technologically, ageing retardation can be achieved via IKK-β and NF-κB inhibition across the brain without evident side effects or compromised efficacy. To depict the hypothalamic control of ageing better, we focused on neuroendocrine pathways of the hypothalamus, and found that IKK-β and NF-κB negatively regulated GnRH. The classical action of GnRH is to regulate sex hormones and reproduction, but whether GnRH is important for whole-body ageing has yet to be determined. We found that ageing was associated with reduced hypothalamic Gnrh1 mRNA, and this change was reversed by IKK-β and NF-κB inhibition but enhanced by their activation (Fig. 5a–c and Supplementary Fig. 10a). Using GT1-7 cells, a cell line of GnRH neurons, we confirmed that GnRH release from these cells decreased after IKK-β and NF-κB activation, but increased after IKK-β and NF-κB inhibition (Fig. 5d). To study whether NF-κB might inhibit the Gnrh1 gene, we introduced Gnrh1 promoter-driven luciferase into GT1-7 cells, and simultaneously activated or inhibited IKK-β and NF-κB in these cells. Results showed that Gnrh1 promoter activity reduced ~50% after IKK-β and NF-κB activation, but increased 4–5-fold by IKK-β and NF-κB inhibition (Fig. 5e, f). Moreover, IKK-β and NF-κB activation increased Fos (also known as c-fos), Jun (c-jun), Prkca (Pkca) and Prkcd (PKCδ) mRNA levels (Fig. 5g), and this finding was relevant because c-Fos and c-Jun overexpression and protein kinase C (PKC) activation were both able to inhibit the Gnrh1 promoter (Fig. 5h). Furthermore, IKK-β and NF-κB inhibition of the Gnrh1 promoter was attenuated by blocking c-Fos and c-Jun (Fig. 5i) or by suppressing the PKC pathway (Supplementary Fig. 10b). Altogether, the c-Fos, c-Jun and PKC pathways can work together to mediate the inhibitory effect of IKK-β and NF-κB on GnRH (Fig. 5j), and in conjunction with relevant literature34, transcriptional integration of NF-κB and c-Jun seems to account for downregulation of GnRH in the hypothalamus. On the basis of the known role of GnRH in regulating sex hormones, GnRH changes in our mouse models might correlate with changes in sex hormones, and this prediction was proved (Supplementary Fig. 10c, d). However, a sex hormone may not be a primary mediator for ageing phenotypes in our models, because hypothalamic IKK-β and NF-κB are important for ageing in both sexes. This context provoked us to propose that GnRH works as a primary mediator independently of a specific sex hormone. To explore whether GnRH exerts intra-brain actions to affect ageing, we delivered GnRH into the hypothalamic third-ventricle of old mice, and examined ageing-related changes in brain cell biology. A notable observation was that GnRH promoted adult neurogenesis despite ageing. Using BrdU tracking following a single BrdU injection to report neurogenesis12, we found that ageing is characterized by diminished neurogenesis, particularly in the hypothalamus and the hippocampus; however, this defect was substantially reversed by GnRH treatment (Fig. 6a–c). Thirty-day BrdU tracking (with seven days of daily BrdU injections) also confirmed that BrdU-labelled cells in GnRH-treated mice significantly survived (Fig. 6d, e). Of note, these effects were seen in not only the hypothalamus but also the hippocampus and other brain regions (data not shown), reflecting the fact that GnRH travels within the brain to promote neurogenesis. Therefore, given the leadership role of the brain in controlling whole-body physiology, the brain-wide neurogenesis induced by hypothalamic GnRH may provide an explanation about how the hypothalamus, a very small structure in the brain, could control systemic ageing. Finally, to study whether GnRH could affect ageing, we subjected old MBH–IKK-β mice and MBH-ctrl mice described in Fig. 2 to daily GnRH therapy for a prolonged period, and then examined their ageing physiology and histology. As we were also interested in testing whether GnRH could act peripherally to affect ageing, we treated mice with GnRH via peripheral injections. Notably, GnRH treatment reduced the magnitude of ageing histology in control mice and abrogated the pro-ageing phenotype in MBH–IKK-β mice (Fig. 6f–h). Interestingly, despite the peripheral administration, GnRH led to an amelioration of ageing-related cognitive decline (Fig. 6i and Supplementary Fig. 11). Thus, a prolonged increase of systemic GnRH can cumulatively yield actions on the brain; despite the mechanism remains to be studied, some GnRH-responsive brain regions outside of the blood–brain barrier, such as the median eminence, subfornical organ and area postrema, can have access to peripheral-delivered GnRH. These effects of GnRH were not specific to a sex, as similar outcomes were shown in males (Fig. 6f–i, Supplementary Fig. 11) and females (Supplementary Fig. 12). For comparison, we treated MBH–IκB-α mice with GnRH, and it turned out that GnRH did not further enhance the anti-ageing phenotype in MBH–IκB-α mice (Supplementary Fig. 13), suggesting that NF-κB inhibition and GnRH action may work in the same pathway to counteract ageing. Clearly, future studies are still needed to detail the central and peripheral roles of GnRH in hypothalamic control of ageing; regardless, this body of data can lead to the conclusion that the hypothalamus can integrate NF-κB-directed immunity and GnRH-driven neuroendocrine system to program ageing development. In this work, we conceived that the hypothalamus, which is known to have fundamental roles in growth, development, reproduction and metabolism, is also responsible for systemic ageing and thus lifespan control. Notably, through activating or inhibiting immune pathway IKK-β and NF-κB in the hypothalamus of mice, we were able to accelerate or decelerate the ageing process, leading to shortened or increased lifespan. Thus, in line with the literature that appreciated the effects of the nervous system on lifespan7, 8, 9, 10, 11, our findings provide a proof of principle to the hypothesis that ageing is a life event that is programmed by the hypothalamus. Indeed, brain change is an early ageing manifestation4, and we reasoned that some hypothalamic alterations may act to motivate ageing of the rest parts in the body, and this outreaching role of the hypothalamus aligns with the fact that it is the neuroendocrine ‘headquarters’ in the body. Along this line, we further revealed a direct link between IKK-β and NF-κB activation and GnRH decline, and also importantly, we discovered that GnRH induces adult neurogenesis broadly in the brain, and GnRH therapy can greatly amend ageing disorders. Thus, whereas the inhibition of GnRH by NF-κB may lead to the end of reproductive length—which seems necessary for species’ quality—it initiates systemic ageing at the same time. Questions remain about how hypothalamic IKK-β and NF-κB is activated in this process; speculatively, as deduced from some recent studies about sirtuins and NF-κB35, 36, age increase-induced epigenetic changes might be accountable, which calls for future investigations. To summarize, our study using several mouse models demonstrates that the hypothalamus is important for systemic ageing and lifespan control. This hypothalamic role is significantly mediated by IKK-β- and NF-κB-directed hypothalamic innate immunity involving microglia–neuron crosstalk. The underlying basis includes integration between immunity and neuroendocrine of the hypothalamus, and immune inhibition and GnRH restoration in the hypothalamus or the brain represent two potential strategies for combating ageing-related health problems. Nestin-Cre mice and Ikbkblox/lox mice were described in our previous publications13, 14, 15, 37, and maintained on C57BL/6 strain for more than 15 generations. C57BL/6 mice were obtained from Jackson Laboratory or the National Institute of Ageing, NIH. All mice were kept under standard and infection-free housing, with 12-h light/12-h dark cycles and 4–5 mice per cage. Pathogen-free quality was ensured with quarterly serology, quarterly histopathological examinations and routine veterinarian monitoring, and a bacteriological test was additionally included. All mice in this study were maintained on a normal chow from LabDiet (4.07 kcal g−1). For animal GnRH therapy, mice were subcutaneously injected with GnRH (Sigma) at the dose of 2 ng per mouse on a daily basis for a period of 5–8 weeks. The Institutional Animal Care and Use Committee at the Albert Einstein College of Medicine approved all the procedures. Body weight and food intake were measured regularly using a laboratory scale. We performed the grip test to measure muscle endurance using the method as similarly described in the literature38, 39, using a homemade square grid with a small mesh size to allow mice to hang for longer time. In brief, a mouse was lifted by the tail and placed on a homemade square grid (1-cm mesh size). The grid was then inverted 30.5 cm over a soft pad, and the mouse was allowed to hang by paws for a maximum of 5 min. The time that the mouse was able to hang was recorded during a 5-min test period. Synapsin promoter-directed lentiviral vector was used to drive neuron-specific gene delivery as previously established13, 14, 15, 37. These lentiviral vectors contain the cDNA of CAIKK-β or DNIκB-α or only GFP under the control of synapsin promoter. To create lentiviral NF-κB reporter vector, a target plasmid was constructed to have the GFP open reading frame controlled by a DNA cassette containing five tandem repeats of the NF-κB transcriptional response element, according to the approach established in the literature40. The lentiviruses were produced from HEK293T cells via co-transfecting a target plasmid with two package plasmids (VSVg and delta 8.9) using CaCl2. Lentiviruses were purified through ultracentrifugation. Intra-MBH viral injections were performed as we previously established13, 14, 15, 37. In brief, under an ultraprecise stereotactic instrument (resolution: 10 μm) (Kopf Instruments), lentiviruses were bilaterally injected at the coordinates of 1.5 mm posterior to the bregma, 5.8 mm below the skull, and 0.2 mm lateral to the midline. All mice were tested for general health, sensorimotor reflexes and motor responses before the onset of all behavioural testing. Mice were maintained on a 12-h light/12-h dark schedule in an isolation unit located inside the behavioural testing room. An Anymaze video tracking system (Stoelting) equipped with a digital camera connected to a computer was used to videotape the whole course of animal activities in training and experimental sessions of behavioural tests. Locomotor activities were assessed using the open field test. The open field arena consisted of a clear Plexiglas chamber that was 40 cm × 40 cm, with walls that were 35 cm high. The arena was placed in a brown box to reduce visual cues. Mice were placed in the arena and allowed to explore for 5 min, and measured for distance and time travelled and mean speed. The maze was filled with 22–23 °C water that was made opaque with Crayola non-toxic paint, and was located in the centre of a small square room with numerous extra-maze cues (various black shapes on white background, a cabinet and an experimenter). The diameter of the maze was 90 cm and divided into four quadrants (northwest, northeast, southwest and southeast). A circular platform with a diameter of 10 cm was placed 25 cm from the wall in the centre of the northwest quadrant. Visual platform test: the visual platform test was performed on a single day. There were six trials with 30-min inter-trial intervals. In the test, a visible flag was placed on the top of the platform to increase the visibility, and the platform was placed on a random location for each trial. A mouse was placed on water, at the same starting location for all trials, and was measured for latency, distance and mean speed travelled to the platform. Hidden-platform training: mice were first required to swim to and sit on a circular visible platform at 0.5 cm above water level for 10 s. If mice could not find the platform within 60 s, they were gently guided to the platform using a glass stirring rod. Mice were then subjected to five consecutive days of training, consisting of two trials per entry location (entry locations were north, south, east and west) for a total of eight trials per day. The platform was made invisible by submerging it 1 cm below the surface of the water. Mice were expected to find the location of the invisible platform, and measured for latency to reach the platform, distance travelled to reach the platform, path efficiency, time spent in and distance travelled in each quadrant as well as total distance and mean swim speed. Probe trial: on day 6, mice were subjected to a single probe trial, in which the platform was removed and mice were allowed to swim for 60 s. Mice were measured for the amount of time spent in all quadrants, distance and number of times that mice crossed the location of the former platform, and total distance and mean swim speed. Mice were tested for reward (1:1 water/full-fat sweetened condensed milk) (Nestle) on a forced-choice alternation test in a T-maze with an opaque floor and plastic sides. Mice first received food restriction to reduce body weight by 5–10%, and then a 4-day adaptation to the apparatus with the reward. After that, mice were given six pairs of training per day for 12 days, and tests of every 2 days were designated as a trial block. On the first trial of each pair, a mouse was placed in the start arm, forced to choose one of two goal arms in the T (the other is blocked by a removable door), and received the reward at the end. The mouse was kept in this goal arm for 15–20 s and subsequently returned by the experimenter to the start arm. The animal was then given a free choice between two goal arms, rewarded for choosing the ‘novel’ arm (the one that was not chosen in the first trial of the pair), but punished for choosing the other goal arm (the one that was chosen on the first trial of the pair) using a 20 s-blocking without the reward. The location of the sample arm (left or right) was varied across trials so that mice received equal numbers of left and right presentations, and no more than two consecutive trials with the same sample location. Mice were tested in squads of 4–5 to minimize variations in inter-trial intervals, which was 5–10 min for all animals throughout 12-day training period. The method of tail tendon breaking test was used to examine collagen cross-linking, as described41. In brief, a collagen fibre was teased from a mid-tail section of the lateral tail tendon and tied to a 2-g weight. The fibre was suspended into a bath containing 7 M urea at 45 °C. The fibre breaking time was determined in quadruplicate for each mouse. Skeletal muscles (quadriceps) and dorsal skin were dissected from mice, fixed in 10% neutralized formalin at 4 °C overnight, and embedded into paraffin. Paraffin sections were prepared at 5-μm thickness and subjected to haematoxylin and eosin staining. Images were collected using an Axioskop II light microscope (Zeiss) and analysed using Image J. We adopted the bone volume fraction procedure established in the literature42. In brief, the left intact femurs were removed and analysed via LaTheta LCT-100A X-ray microtomography scanner (Aloka) through mouse physiology core facility at Albert Einstein College of Medicine. The distal part of femur encompassing the cancellous bone was analysed. The trabecular and cortical bone regions were outlined for each tomography slice by the software of the scan system. Bone volume fraction was calculated as the trabecular bone volume divided by the total bone volume. A calibration phantom was used for calibration of each scan. Mice under anaesthesia were perfused with 4% paraformaldehyde, and brains were removed, post-fixed in 4% paraformaldehyde, and infiltrated in 20–30% sucrose. Brain sections were made at 20-μm thickness via a cryostat, blocked with serum, penetrated with Triton-X 100, treated overnight at 4 °C with primary antibody, followed by reaction with fluorescence-conjugated secondary antibody (Jackson), and imaged under a confocal microscope. For BrdU staining, sections were pre-treated with 1 M HCl for 30 min at 37 °C, followed by 5-min treatment with 0.1 M sodium borate, pH 8.5. Primary antibodies included rabbit anti-Iba-1 (Wako), rabbit anti-GFAP (Millipore), mouse anti-TNF-α (Abcam), mouse anti-NeuN (Millipore), and goat anti-Cre antibody (Santa Cruz). For Nissl staining, freshly isolated mice brains were fixed in 4% paraformaldehyde in PBS for overnight at 4 °C. The fixed whole brains were then subjected to cryosectioning coronally, and frozen sections were stained to detect Nissl body by using the NovaUltra Nissl stain kit (IHCWORLD) according to the manufacturer’s instruction. Serial brain sections across the MBH were made at 20 μm thickness, and every five sections were represented by one section with staining and cell counting. For western blotting, animal tissues were homogenized, and proteins were dissolved in a lysis buffer, and western blots were conducted as previously described3. Proteins separated by SDS–PAGE were identified by immunoblotting with primary rabbit anti-pRelA, anti-RelA and anti-β-actin antibodies (Cell Signaling) and horseradish peroxidase (HRP)-conjugated anti-rabbit secondary antibody (Pierce). Promoter sequence of the rat Gnrh1 gene was PCR amplified (−1934 to +21) from a rat genomic DNA preparation, and subcloned into the pGL3-basic luciferase reporter vector (Promega) using standard cloning strategies. pcDNA expressing CAIKK-β or DNIκB-α versus control were previously described13, 14, 15, 37, or pcDNA expressing HA-RelA was provided by A. Lin. RelA shRNA and control (GFP) shRNA vectors were obtained from Addgene, as studied in the literature43. RelA shRNA: 5′-GCATGCGATTCCGCTATAA-3′; control shRNA: 5′-ACAGCCACAACGTCTATAT-3′. Expression plasmids for c-Jun or c-Fos were provided by D. Stocco. Vectors expressing c-Jun or c-Fos shRNA or scramble shRNA control were provided by L. Fahana. c-Jun shRNA: 5′-AGTCATGAACCACGTTAAC-3′; c-Fos shRNA: 5′-TCCGAAGAGAACGGAATAA-3′; scramble shRNA: 5′-GTTATTACTGTTCGATCGC-3′. 12-O-tetradecanoylphorbol-13-acetate (also known as phorbol 12-myristate 13-acetate or TPA) and calphostin-C were from Sigma-Aldrich. TPA or calphostin-C was dissolved in dimethylsulphoxide (DMSO) and applied in cell culture medium at a final concentration of 0.2 μM or 0.01 μM, respectively, and DMSO did not exceed 0.1% of cell culture medium. GT1-7 cells were previously described13, and cultured in a standard humidified incubator at 37 °C and 5% CO2 with DMEM cell culture medium supplemented with 10% FBS, 2 mM l-glutamine, and PenStrep (50 U ml−1 penicillin G, 50 μg ml−1 streptomycin). Transfection of cultured cells with luciferase plasmids and expression plasmids was performed through Lipofectamine 2000 (Invitrogen). The dual luciferase reporter assay (Promega) was performed according to the manufacturer’s instruction, and co-transfection of the pRL-TK vector expressing Renilla luciferase was used to control firefly activity internally. Empty plasmids pGL3 and pcDNA3.1 were used as negative controls. RNA was extracted by TRIzol (Invitrogen) and analysed via SYBR green real-time PCR (StepOnePlus real-time PCR system, Invitrogen). Testosterone and oestradiol were measured using testosterone and oestradiol EIA kits (Cayman Chemical). GnRH was measured using the luteinizing hormone-releasing hormone EIA kit (Phoenix Pharmaceuticals). Mice were pre-implanted with intracerebroventricular (i.c.v.) cannula in the hypothalamic third ventricle, and after a ~3-week recovery, they were subjected to neurogenesis assay or survival assays. In the neurogenesis assay, mice were daily pre-injected with GnRH or vehicle at the dose of 1 ng per day through cannula for 3 days, subsequently a single i.c.v. injection of BrdU (Sigma) at the dose of 10 μg (defined as day 0), and continued to receive daily i.c.v. injections of GnRH (1 ng per day) or vehicle for 7 days before they were killed for brain sectioning. In the survival assay, mice pre-received daily i.c.v. injections of GnRH (1 ng per day) or vehicle for 3 days, then daily i.c.v. injections of BrdU (10 μg per day) together with daily i.c.v. injections of GnRH (1 ng per day) or vehicle for 7 days (last day was defined as day 7), and followed by continued daily i.c.v. injections of GnRH (1 ng per day) or vehicle until day 30 when mice were killed for brain sectioning. Kolmogorov–Smirnov test was used to determine parametric distribution of data. Analysis of variance (ANOVA) and appropriate post-hoc analyses were used for comparisons involving more than two groups. Two-tailed Student’s t-tests were used for comparisons involving only two groups. Lifespan analysis was performed using Kaplan–Meier survival analysis; the mutant and control survivorship curves were compared in pairs and P values were obtained with log-rank test. Maximal lifespan of mice were statistically analysed using Chi-squared test according to the literature44. All data were presented as mean ± s.e.m. P < 0.05 was considered significant. Download references We thank other Cai laboratory members for technical assistance, and L. Farhana, D. Stocco, L. Eckhardt, T. Ohshima, A. Lin and D. Tantin for reagents. This study was supported by National Institutes of Health (NIH) grants R01 AG 031774, R01 DK078750, and American Diabetes Association grant 1-12-BS-20 (all to D.C.). D.C. is a recipient of Irma T. Hirschl Scholarship.