Life extension factor klotho enhances cognition

Abstract

Aging is the primary risk factor for cognitive decline, an emerging health threat to aging societies worldwide. Whether anti-aging factors such as klotho can counteract cognitive decline is unknown. We show that a lifespan-extending variant of the human KLOTHO gene, KL-VS, is associated with enhanced cognition in heterozygous carriers. Because this allele increased klotho levels in serum, we analyzed transgenic mice with systemic overexpression of klotho. They performed better than controls in multiple tests of learning and memory. Elevating klotho in mice also enhanced long-term potentiation, a form of synaptic plasticity, and enriched synaptic GluN2B, an N-methyl-D-aspartate receptor (NMDAR) subunit with key functions in learning and memory. Blockade of GluN2B abolished klotho-mediated effects. Surprisingly, klotho effects were evident also in young mice and did not correlate with age in humans, suggesting independence from the aging process. Augmenting klotho or its effects may enhance cognition and counteract cognitive deficits at different life stages.

Figure 1. The KL-VS allele is associated with better cognitive performance in three independent aging populations without dementia and in a meta-analysis of the populations

(A–D) Neuropsychological scores from tests spanning multiple cognitive domains (Table S4; Figure S1). Global composite Z-scores of 718 aging individuals (52–85 years of age) that were non-carriers (black; n=530) or carriers (purple; n=188) of a single KL-VS allele were obtained from three independent cohorts without cognitive impairments. In each cohort, an individual composite score was standardized and scaled to reflect performance as a measure of the number of standard deviations from the global average of that cohort (global composite Z-score). Higher scores indicate better cognitive performance. (A–C) Global composite Z-scores in (A) Cohort 1 (179 non-carriers, 41 carriers), (B) Cohort 2 (331 non-carriers, 135 carriers), (C) Cohort 3 (20 non-carriers, 12 carriers), and (D) Meta-analysis of the cohorts. All subjects had a MMSE score of 28 or greater and no dementia. Data were analyzed by linear models, accounting for effects of age, sex, and education and testing for effects due to KL-VS genotype. APOEε4 carrier status had no significant effects (Tables S5–S6). °p=0.06, *p<0.05, **p<0.01, ***p<0.001 vs Non-Carrier (linear regression t-test). See also Tables S2-S6 and Figure S1. Data are means ± SEM.

Figure 2. KL-VS-associated cognitive enhancement is independent of age

(A–C) Global composite Z-scores decreased as a function of age in non-carriers (empty squares) and carriers (purple squares) of the KL-VS allele in (A) Cohort 1, (B) Cohort 2, and (C) Cohort 3. (D) In meta-analysis of all cohorts, KL-VS-associated cognitive enhancement tended to decrease with advancing age (ANOVA, KL-VS:age interaction, p=0.10). Data were analyzed by linear models, accounting for effects of age, sex, and education and testing for effects due to an age by KL-VS interaction. APOEε4 carrier status had no significant effects on these measures. See also Tables S5 and S6. Data are means ± SEM.

Figure 3. Elevation of systemic klotho levels occurs in human KL-VS carriers and enhances mouse survival, learning, and memory independent of age

(A) Fasting morning serum klotho levels in individuals from Cohort 1 (55–85 years of age) that were non-carriers (n=118) or carriers (n=38) of a single KL-VS allele. Data were analyzed by a linear model, accounting for effects of age, sex, and education and testing for effects due to KL-VS genotype. APOEε4 carrier status had no effect (Table S7). (B) Hippocampal levels of klotho in NTG and KL mice (n=13–14 mice per genotype, age 3 months). Representative western blots for klotho and actin are shown above; images for each protein were from the same gel. (C) Kaplan-Meier curves show increased survival of heterozygous KL mice from line 46 (Kuro-o et al., 1997; Kurosu et al., 2005) compared to NTG littermates (n=22–29 mice per group, p<0.01 by log rank test). Proportional hazard testing revealed the KL effect was independent of age (p=0.76). (D-E) KL and NTG mice (n=8–9 mice per genotype) were tested in the Morris water maze at 10–12 months of age. (D) Spatial learning curves (platform hidden). Data represent the daily average of total distance traveled to the platform. Mixed model ANOVA: KL effect p<0.05. (E) Results of a probe trial (platform removed) 1 h after completion of hidden-platform training showing latency to reach the original platform location. (F–G) An independent cohort of mice (n=17–19 mice per genotype) was tested in the water maze at 4–7 months of age. (F) Spatial learning curves (platform hidden). Mixed model ANOVA: KL effect p<0.05. (G) Probe trial results. *p<0.05, **p<0.01, ***p<0.001 (t-test). See also Tables S7, S8 and Figure S2. Data are means ± SEM.

Figure 4. Klotho elevation also improves working and context memory without altering other behaviors in young mice

(A) Percent alternations among arms by 3–4-month-old mice during exploration of a Y-maze (n=8–10 mice per genotype). (B) Percent time 6-month-old mice spent freezing at baseline and 24 h after context training in a fear conditioning task (n=6–7 male mice per genotype). (C) Movements during exploration of an open field (n=13–14 mice per genotype; p=0.30 by two-tailed t-test). (D) Percent time spent exploring the open arms of an elevated plus maze (n=14–15 mice per genotype; p=0.60 by two-tailed t-test). *p<0.05, ***p<0.001 vs chance or as indicated by bracket (t-test). See also Figure S3. Data are means ± SEM.

Figure 5. Klotho overexpression enhances synaptic GluN2B levels in the hippocampus and cortex

(A-G) Synaptic membrane Fraction 1 (PSD-enriched) and Fraction 2 (non-PSD enriched) isolated from hippocampus or cortex of NTG and KL mice (n=15-18 mice per genotype, age 3-4 months). (A) Representative western blots of hippocampal fractions. (B-E) Quantitation of (B) GluN2A, GluN2B, and GluN2C levels in hippocampal fraction 1, (C) GluN2B in hippocampal fractions 1 (left) and 2 (right), (D) PSD-95 and GluN2B and (E) synpatophysin levels. (F) Representative western blots of cortical fractions isolated from frontal, motor and somatosensory regions (Bregma 0-3.5 mm). (G) Quantitation of GluN2B levels in cortical Fractions 1 (left) and 2 (right). For each fraction, protein levels are relative to NTG levels, arbitrarily defined as 1.0. *p<0.05 vs NTG (t-test). See also Figure S4. Data are means ± SEM.

Figure 6. Klotho overexpression enhances NMDAR-, but not AMPAR-, dependent functions

(A, B) FOS expression in the dentate gyrus of NTG and KL mice immediately following a probe trial in the water maze. (A) Staining with antibodies to FOS revealed more immunoreactive granule cells in the dentate gyrus of KL (bottom) than NTG (top) mice. Scale bar: 100 μm. (B) Quantitation of FOS-positive granule cells (n=7 mice per genotype, age 3-4 months). The mean level in NTG controls was arbitrarily defined as 1.0. (C) Field excitatory postsynaptic potential (fEPSP) recordings from acute hippocampal slices of 3.5-4.5-month-old NTG and KL mice. LTP induction and decay in the dentate gyrus were monitored for 45-50 minutes following theta burst stimulation of the medial perforant pathway. Mixed model ANOVA: KL vs NTG genotype by time effect p<0.01. Number of slices/mice: NTG 4/4, KL 8/6. (D) AMPAR-mediated basal synaptic transmission in acute hippocampal slices of 3.5-4.5-month-old mice at the medial perforant path to dentate granule cell synapse. Number of slices/mice: NTG 5/3, KL 9/5. (E-G) Isolated NMDAR and AMPAR EPSCs measured by whole-cell patch-clamp recordings from dentate granule cells in acute hippocampal slices of 3–4-month-old mice. (E) Representative traces of evoked EPSCs at +40 mV or −70 mV. NMDAR–mediated EPSCs were quantitated between 80–100 ms after stimulation (blue shading) in the top traces and AMPAR–mediated EPSCs at the nadir of the bottom traces. Scale: 100 pA, 50 ms. (F) Quantitation of AMPAR/NMDAR EPSC ratios. Number of slices/mice: NTG 10/3, KL 10/3. p=0.32 (t-test). (G) Decay constant (τ) of isolated NMDAR EPSCs in the presence of NBQX (10 μM). Number of slices/mice: NTG 7/3, KL 5/3. *p<0.05, **p<0.01 (t-test). See also Figure S5. Data are means ± SEM.

Figure 7. Treatment with GluN2B selective antagonists blocks klotho effects on NMDAR currents and cognition

(A,B) Representative traces (A) and quantitation (B) of isolated NMDAR EPSCs in the presence of NBQX (10 μm) at baseline (black) and following perfusion with ifenprodil (Ifen., 3 μM) (red) in the same slices. Number of slices/mice: NTG 6/3, KL 5/3. Two-way repeated measures ANOVA: ifenprodil effect p<0.0001, ifenprodil by KL interaction p<0.05. (C) Time course of NMDAR EPSC decay constant following ifenprodil perfusion in each genotype. Mixed model ANOVA: p<0.0001 for KL vs NTG genotype by time effect. (D) Change in decay constant (τ) between 0-10 min after initiation of ifenprodil treatment in NTG and KL slices. Number of slices/number of mice: NTG 7/3, KL 5/3. (E-F) Mice (n=8-19 per group) received a single i.p. injection of vehicle (−) or ifenprodil (5 mg/kg) 10 min before training in a fear conditioning paradigm or testing in the Y-maze. (E) Percent time mice (age 5-7 months) spent freezing at baseline and 24 h after context training in a fear conditioning task. Two-way ANOVA: ifenprodil by KL interaction p<0.05. (F) Percent alternations among Y-maze arms that mice (age 10-12 months) showed during 3 min exploration. Two-way ANOVA: ifenprodil by KL interaction p<0.09. (G-H) Mice (n=13-19 per group, age 3-5 months) received a single i.p. injection of vehicle (−) or Ro 25-6981 (Ro 25; 5 mg/kg) 30 min before testing in Y-maze. (G) Percent alternations among Y-maze arms. Two-way ANOVA: Ro 25 by KL interaction p<0.05. (H) Percent decrease in alternations following Ro 25 treatment in NTG and KL mice. *p<0.05, **p<0.01, ***p<0.001 vs. NTG or as indicated by brackets by t-test (B, D, H) or Bonferroni-Holm test (E, F, G). See also Figure S6. Data are means ± SEM.