NPTX2 and cognitive dysfunction in Alzheimer's Disease

Abstract

Memory loss in Alzheimer's disease (AD) is attributed to pervasive weakening and loss of synapses. Here, we present findings supporting a special role for excitatory synapses connecting pyramidal neurons of the hippocampus and cortex with fast-spiking parvalbumin (PV) interneurons that control network excitability and rhythmicity. Excitatory synapses on PV interneurons are dependent on the AMPA receptor subunit GluA4, which is regulated by presynaptic expression of the synaptogenic immediate early gene NPTX2 by pyramidal neurons. In a mouse model of AD amyloidosis, Nptx2^-/- results in reduced GluA4 expression, disrupted rhythmicity, and increased pyramidal neuron excitability. Postmortem human AD cortex shows profound reductions of NPTX2 and coordinate reductions of GluA4. NPTX2 in human CSF is reduced in subjects with AD and shows robust correlations with cognitive performance and hippocampal volume. These findings implicate failure of adaptive control of pyramidal neuron-PV circuits as a pathophysiological mechanism contributing to cognitive failure in AD.

Keywords: Alzheimer's disease; alzheimer’s disease; dementia; immediate early gene; inhibitory neuron; mmediate early gene; neuroscience.

Figure 1.. NPTX2 levels are reduced in human postmortem AD brain and DS brain, but not in ASYMAD brain.

(A,B,D and E) Representative western blot images (A) and quantification of NPTX2 (B), NPTX1 (D) and NPTXR (E) normalized to PSD95 in the frontopolar cortex (FPC), precuneus (PCU), occipital gyrus (OCC), middle frontal gyrus (MFG), middle temporal gyrus (MTG) and parietal gyrus (PAR) from controls and AD subjects. NPTX2 is down-regulated in all assayed brain regions of AD individuals. FPC: control, n = 7; AD, n = 8. PCU: control, n = 15; AD, n = 19. OCC: control, n = 7; AD, n = 7. MFG: control, n = 9; AD, n = 16. MTG: control, n = 10; AD, n = 18. PAR: control, n = 5; AD, n = 5. (C) Nptx2 mRNA is reduced in AD brain. FPC: control, n = 9; AD, n = 16. PCU: control, n = 7; AD, n = 6. (F,G) Western blot assays reveal no significant change of NPTX2 expression in MFG from subjects with asymptomatic AD (ASYMAD). Control, n = 8; ASYMAD, n = 10. (H, I) Western blot assays show significant reduction of NPTX2 in MFG of individuals with Down syndrome (DS). Control, n = 6; DS, n = 6. (J) Nptx2 mRNA is reduced in MFG of individuals with DS. Control, n = 5; DS, n = 5. *p<0.05, **p<0.01, ***p<0.001 by two-tailed t test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.002

Figure 1—figure supplement 1.. Immediate early gene expression in human postmortem AD brain.

(A) NPTX2 levels are reduced in human postmortem AD brain. Quantification of NPTX2, NPTX1 and NPTXR levels in Figure 1A with actin as reference protein. NPTX2 is down-regulated in all assayed brain regions of AD individuals when referenced to actin. Frontopolar cortex (FPC): control, n = 7; AD, n = 8. Precuneus (PCU): control, n = 15; AD, n = 19. Occipital gyrus (OCC): control, n = 7; AD, n = 7. Middle frontal gyrus (MFG): control, n = 9; AD, n = 16. Middle temporal gyrus (MTG): control, n = 10; AD, n = 18. Parietal gyrus (PAR): control, n = 5; AD, n = 5. (B) Egr1 and Arc expression in different brain regions of control and AD subjects. Western blots show that expression of Egr1 and Arc are not significantly altered in most tested brain areas from AD individuals. FPC: control, n = 7; AD, n = 8. PCU: control, n = 15; AD, n = 19. OCC: control, n = 7; AD, n = 7. MFG: control, n = 9; AD, n = 16. MTG: control, n = 10; AD, n = 18. PAR: control, n = 5; AD, n = 5. *p<0.05, **p<0.01, ***p<0.001 by two-tailed t test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.006

Figure 1—figure supplement 2.. NPTX2 levels are reduced in human postmortem DS-AD brain.

(A) Western blot assays show NPTX2 down-regulation in superior frontal gyrus (SFG) of individuals with Down syndrome and Alzheimer’s disease (DS-AD). Control, n = 8; DS-AD, n = 12. (B) Nptx2 mRNA is reduced in SFG of individuals with DS-AD. Control, n = 8; DS-AD, n = 12. *p<0.05, by two-tailed t test. Data represent mean ± SEM. See patients information in Figure 1—source data 3. DOI: http://dx.doi.org/10.7554/eLife.23798.007

Figure 2.. miRNAs dysregulation and NPTX2 down-regulation in AD brains.

(A) Nptx2 pre-mRNA levels are identical in frontopolar cortex (FPC) of AD subjects and control. Control, n = 9; AD, n = 16. (B) microRNAs predicted to bind with Nptx2 3’UTR by TargetScan. (C) Taqman assays show miR-182 and miR-1271 are increased in FPC of AD individuals. Control, n = 9; AD, n = 16. *p<0.05, **p<0.01 by two-tailed t test. (D) miR-96, miR-152 and miR-182 are up-regulated in individuals with Down syndrome (DS). Triploid miR-155 served as positive control. Control, n = 14; DS, n = 18. *p<0.05, **p<0.01, ***p<0.001 by two-tailed t test. (E) miR-96, miR-182 and miR-1271 target the same sequence in Nptx2 3’UTR. (F–H) Cultured mouse cortical neurons are transduced with lentivirus expressing nontargeting miRNA (LV-NT) or miR-96, miR-1271 and miR-182. (F) Expression of miR-96, miR-1271 and miR-182 reduce NPTX2 protein level. (G) miR-96 reduces Nptx2 mRNA. (H) Nptx2 pre-mRNA is preserved by miR-96 and miR-1271 expression. n = 5 wells from three independent culture except n = 4 wells for LV-NT group in Figure 2G. *p<0.05, ***p<0.001 by nonparametric one way ANOVA with Tukey post hoc test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.008

Figure 2—figure supplement 1.. Methylation of Nptx2 promoter in human brain.

(A) Representative pyrosequencing traces show high methylation of Nptx2 promoter in pancreatic cell line AsPC1 cells, and low methylation in human brain. (B) Nptx2 promoter methylation in brains is not different between control and AD subjects. n = 8 per group. Two-tailed t test was performed. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.009

Figure 2—figure supplement 2.. Nptx2 mRNA is targeted by miRNAs that are upregulated in AD brain.

(A) miR-1271 up-regulation correlates with reduced Nptx2 mRNA. Control, n = 9; AD, n = 16. Pearson correlation coefficient analysis was performed. (B,C) Assays of Nptx2 mRNA and miRNA in AD precuneus (PCU) region. (B) Nptx2 mRNA is reduced in PCU area of AD subjects with preserved Nptx2 pre-mRNA. Control, n = 7; AD, n = 6. (C) miR-152 and miR-182 are increased in AD PCU region compared with control. n = 7 per group. *p<0.05, **p<0.01, ***p<0.001 by two-tailed t test. (D) Nptx2 pre-mRNA is reduced in superior frontal gyrus (SFG) of subjects with Down syndrome. Control, n = 14; DS, n = 18. **p<0.01 by two-tailed t test. (E,F) miRNAs directly target Nptx2 3’UTR. (E) Wild-type (WT) or miR-binding site mutated (Mut) Nptx2 3’UTR is inserted downstream of a luciferase reporter. Mutated nucleotides are in red. (F) miRNA mimics are able to reduce the luciferase activity in HEK293 cells, and this effect is partially abolished by mutation of miRNA binding site on Nptx2 3’UTR. n = 9–16 wells. * miRNA mimic vs control RNA; ^# WT Nptx2 vs mutant Nptx2. Two-tailed t test. (G) Western blot assays show no significant changes of Gephyrin and Homer1, which are predicted targets of miR-1271, in human postmortem brain. Frontopolar cortex (FPC): control, n = 7; AD, n = 8. Precuneus gyrus (PCU): control, n = 15; AD, n = 18. (H) mRNA levels of Gephyrin and Homer1 in brain are not significantly altered in AD subjects. FPC: control, n = 9; AD, n = 16. PCU: control, n = 7; AD, n = 6. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.010

Figure 3.. Circuit Rhythmicity and GluA4 expression are disrupted in hAPP/Nptx2^-/- mice.

(A–D) Example extracellular field potentials (i) with hatched area shown on an expanded time base (ii) for WT (A), Nptx2^-/- (B), hAPP (C), and hAPP/Nptx2^-/- (D) mice. For hAPP/Nptx2^-/- trace, grey trace in (ii) shows trace on expanded time base and the black trace shows expanded voltage axis to show gamma-like oscillations nested within population spikes. (E) Power spectra for WT, Nptx2^-/-, hAPP, and hAPP/Nptx2^-/- mice, taken from 400 s of recording, normalised to the power between 3 and 300 Hz. (F) The normalised power of gamma-like oscillations was significantly reduced in hAPP/Nptx2^-/- mice (z-score; WT, 5.1 ± 0.73; Nptx2^-/-, 3.8 ± 0.63; hAPP, 3.9 ± 0.79; hAPP/Nptx2^-/-, 1.7 ± 0.68; p=0.0199, One-way ANOVA). **, p<0.01 vs WT, post-hoc multiple comparisons test. (G) GluA4 expression is not altered in 6 month-old Nptx2^-/- mouse cortex. n = 4 for WT and n = 3 for Nptx2^-/-. (H) Representative western blot images and quantification of GluA4 in forebrains of 6 month-old WT, hAPP and hAPP/Nptx2^-/- mice. hAPP/Nptx2^-/- mice show reduced GluA4 in brain. WT, n = 4; hAPP, n = 6; hAPP/Nptx2^-/-, n = 4. **p<0.01 by nonparametric one way ANOVA with Tukey post hoc test. DOI: http://dx.doi.org/10.7554/eLife.23798.011

Figure 3—figure supplement 1.. NPTX2 expression in amyloidosis mouse model.

Western blot assays show no significant change of NPTX2 expression in 6 month-old male APPswe/PS1∆E9 (hAPP) mouse brain when compared with wildtype (WT). n = 3. Two-tailed t test was performed. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.012

Figure 3—figure supplement 2.. Spontaneous sharp-wave ripples (SWRs) occur more frequently in hAPP/Nptx2^-/- mice.

(A–D) Example extracellular field recordings showing SWRs in unfiltered traces (i), with the boxed area filtered (100 to 250 Hz) to show an individual SWR (ii) and the corresponding wavelet transform of the unfiltered trace (iii) for WT (A), Nptx2^-/- (B), hAPP (C), and hAPP/Nptx2^-/- (D) mice. (E) The incidence of SWRs was significantly greater in hAPP/Nptx2^-/- mice (WT, 0.55 ± 0.13 Hz, n = 12; Nptx2^-/-, 0.28 ± 0.10 Hz, n = 10; hAPP, 0.59 ± 0.14 Hz, n = 17; hAPP/Nptx2^-/-, 1.46 ± 0.32 Hz, n = 14; p=0.0010, One-way ANOVA; **, p<0.01 vs WT in post hoc Dunnet’s multiple comparisons test). (F) The peak frequency if SWRs was significantly reduced in hAPP/Nptx2^-/- mice (WT, 198.7 ± 2.7 Hz; Nptx2^-/-, 198.7 ± 3.4; hAPP, 192.6 ± 2.2 Hz; hAPP/Nptx2^-/-, 188.2 ± 3.4 Hz; p=0.0384, One-way ANOVA; *, p<0.05 vs WT in post hoc Dunnets multiple comparisons test). (G,H) No significant differences were observed in either the SWR peak amplitude (G) (WT, 17.9 ± 2.3 µV; Nptx2^-/-, 18.4 ± 4.3 µV; hAPP, 20.3 ± 4.4 µV; hAPP/Nptx2^-/-, 24.6 ± 4.6 µV; p=0.6773, One-way ANOVA), or duration (H) (WT, 26.6 ± 1.4 ms; Nptx2^-/-, 27.0 ± 1.3 ms; hAPP, 27.4 ± 1.1 ms; hAPP/Nptx2^-/-, 29.1 ± 1.4 ms; p=0.5609, One-way ANOVA). DOI: http://dx.doi.org/10.7554/eLife.23798.013

Figure 4.. GluA4 levels are reduced in human postmortem AD brain.

(A) Immunostaining of GluA4 demonstrates GluA4 is enriched on PV-IN in human cortex. Data were collected from four cases including occipital gyrus and parietal gyrus. (B and C) Representative western blot images (B) and quantification of GluA4 (C) in the frontopolar cortex (FPC), precuneus (PCU), middle frontal gyrus (MFG) and middle temporal gyrus (MTG) from controls and AD subjects. GluA4 is significantly down-regulated in PCU and MFG of AD individuals. FPC: control, n = 7; AD, n = 8. PCU: control, n = 15; AD, n = 19. MFG: control, n = 9; AD, n = 16. MTG: control, n = 10; AD, n = 18. *p<0.05 by two-tailed t test. Data represent mean ± SEM. (D) GluA4 levels correlate with NPTX2 in both control and AD group. Pearson correlation coefficient analysis was performed. (E and F) GluA4 expression did not correlate with NPTX2 in young adult brain. n = 10. Pearson correlation coefficient analysis was performed. (G and H) NPTX2 expression in young adult brain was higher than in aged controls. Young, n = 12; Aged, n = 15. **p<0.01 by two-tailed t test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.014

Figure 5.. NPTX levels are reduced in CSF from individuals with clinically diagnosed AD.

(A,B) Western blot assays of NPTX2, NPTX1 and NPTXR in lumbar cerebrospinal fluid (CSF) from age-matched controls and patients with clinically diagnosed AD. AD patients show reduced NPTX2, NPTX1 and NPTXR levels in CSF. Control, n = 36; AD, n = 30. (C) ELISA was developed to quantitate NPTX2 protein in CSF. NPTX2 assayed by ELISA corresponds closely with levels defined by WB. n = 64. Pearson correlation coefficient analysis was performed. (D) ELISA shows significant reduction of NPTX2 in CSF from patients with clinically diagnosed AD. Control, n = 36; AD, n = 28. (E) ELISA assay confirmed the reduction of NPTX2 in AD in second set of CSF sample. n = 36 for control, n = 30 for AD. (F) CSF NPTX2 levels are significantly reduced in individuals with mild cognitive impairment (MCI) compared with healthy controls. Control, n = 72; MCI, n = 17. (G–L) Receiver operating characteristic (ROC) curve analysis of CSF Aβ42 (G), tau (H), p-tau181 (I), NPTX2 (J), tau/NPTX2 (K) and p-tau/NPTX2 (L) as AD diagnostic biomarkers for distinguishing AD from control. Cut off values were determined by maximizing Youden index value. ROC analysis of tau/NPTX2 indicates its diagnostic power is superior to NPTX2 alone, Aß42, tau or p-tau. AUC: area under ROC curve. Control, n = 61–72; AD, n = 50–58. **p<0.01, ***p<0.001 by two-tailed t test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.016

Figure 5—figure supplement 1.. Reduction of NPTXR levels in lumbar cerebrospinal fluid (CSF) from individuals with AD.

(A) Representative western blot images show three major bands with human CSF samples by NPTXR antibody. (B) Patients with clinically diagnosed AD have reduced NPTXR levels in CSF compared with healthy controls. Control, n = 31; AD, n = 28. ***p<0.001 by two-tailed t test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.019

Figure 5—figure supplement 2.. Detection of NPTX in human CSF.

(A) NPTX2 and NPTX1 are detected in lumbar CSF of human subjects as a high molecular weight complex that is resolved into individual NPTXs with reducing reagent on SDS-PAGE. Arrows indicate monomer NPTX2 and NPTX1. BME: β-mercaptoethanol. (B) NPTX2 expression in CSF correlate with levels of CSF NPTX1 and NPTXR in individual samples. n = 66. Pearson correlation coefficient analysis was performed. DOI: http://dx.doi.org/10.7554/eLife.23798.020

Figure 5—figure supplement 3.. Development of NPTX2 ELISA assay.

(A,B) Generation of mouse monoclonal NPTX2 antibody (A) and purification of NPTX2 protein (B) for ELISA assay. Western blots using mouse NPTX2 monoclonal antibody show a 50 kDa band in WT brain lysate that is absent in Nptx2^-/- and Nptx1^-/-; Nptx2^-/-; Nptxr^-/- (triple knockout, TKO) (A). DOI: http://dx.doi.org/10.7554/eLife.23798.021

Figure 5—figure supplement 4.. Reduction of NPTX levels in second set of CSF from individuals with AD.

Representative western blot images and quantification of NPTX2, NPTX1 and NPTXR in second set of lumbar CSF from patients with clinically diagnosed AD. AD patients show reduced NPTX2, NPTX1 and NPTXR levels in CSF compared with healthy controls. Control, n = 36; AD, n = 30. *p<0.05, **p<0.01 by two-tailed t test. Data represent mean ± SEM. DOI: http://dx.doi.org/10.7554/eLife.23798.022

Figure 5—figure supplement 5.. Positive correlation of NPTXs with CSF tau and p-tau.

(A–C) CSF NPTX2 levels correlate with CSF Tau (B) and p-Tau (C), but not with CSF Aβ42 (A) in AD patients. n = 49–51. Pearson correlation coefficient analysis was performed. (D–F) CSF NPTX1 levels correlate with CSF Tau (E) and p-Tau (F), but not with CSF Aβ42 (D) in AD patients. n = 49–51. Pearson correlation coefficient analysis was performed. (G–I) CSF NPTXR levels correlate with CSF Tau (H) and p-Tau (I), but not with CSF Aβ42 (G) in AD patients. n = 49–51. Pearson correlation coefficient analysis was performed. DOI: http://dx.doi.org/10.7554/eLife.23798.023

Figure 6.. NPTX2 expression correlates with cognitive performance and measures of hippocampal volume.

(A) NPTX2 expression in CSF correlates with cognitive function assayed by DRS in AD group. n = 30. p=0.0093 by Pearson correlation coefficient analysis. DRS: dementia rating scale. (B–D) No correlation was observed between CSF Aβ42 (B), tau (C) or p-tau181 (D) with DRS in AD group. n = 28. (E–H) NPTX2 expression in CSF correlates with cognitive function assayed by WCST test (E), semantic verbal fluency test (F), VRT test (G) and CVLT test (H). n = 20–28. WCST: Wisconsin Card Sorting Task; AFV: Semantic Verbal Fluency Test (‘Animals’, ‘Fruits’, ‘Vegetables’); VRT: Visual Reproduction Test; CVLT: California Verbal Learning Test. Pearson correlation coefficient analysis was performed. (I) Representative images of human brain MRI imaging. (J) NPTX2 levels in CSF correlate with hippocampal occupancy. n = 25 AD subjects. HOC: hippocampal occupancy. DOI: http://dx.doi.org/10.7554/eLife.23798.024