ZnT3KO mice are reported to initially have normal cognition that becomes impaired with age, indicating that the lack of synaptic zinc is not the direct cause of impaired cognition, but induces neurodegenerative effects leading to impairment. Consequently, we sought to investigate the effects of disruption of Zn²⁺ modulation of neurotransmission in cognition, protein expression and activation, and neuronal excitability in ZnT3KO mice. Basal biochemical and activity-dependent alterations in neurotrophic NMDAR signaling and BDNF mRNA expression in ZnT3KO mice before the onset of cognitive deficits were assessed in acute hippocampal slice cultures from juvenile WT and ZnT3KO mice. Neuronal hyperactivity before the onset of cognitive deficits in ZnT3KO was assessed using EEG recordings in young WT and ZnT3KO. Progressive age-dependent cognitive decline, neurodegenerative protein expression, and neurodegenerative alterations in both biochemistry and brain cytoarchitecture associated with epileptiform activity were assessed in cohorts of aged WT and ZnT3KO at 3, 6, 12, and 15 months of age. All animal experiments were approved by the University of California, Irvine Institutional Animal Care and Use Committee and conformed to all University and USDA animal care regulations. Homozygous male and female 129Sv wild-type (WT) and ZnT3KO mice (from Richard Palmiter, University of Washington) were housed in the vivarium at the University of California, Irvine, maintained on a 12-h day/night cycle in separate WT and ZnT3KO colonies, and provided food and water ad libitum. After the eighth generation all of the adult mice in the colonies were genotyped, using the Transnetyx genotyping service, to check for genetic drift. For experiments using adult mice, dams and pups from litters born within 5 days of each other were combined in one cage to a maximum of 3 dams and 14 pups until weaning at 21 days of age. The weaned mice were sexed and housed 5 per cage by sex, with equal numbers of males and females. Equal numbers of males and females were used for experiments unless health precluded participation of an individual. For experiments using hippocampal slices from 3 to 4-week-old mice, hippocampal tissue from litters consisting of at least 4 pups were combined as one sample. GraphPad Prism 7 software was used for statistical analysis. Data is presented as mean ± SEM, with raw data from two compared conditions analyzed by two-tailed Student's t-test, reporting t-test statistics (t), degrees of freedom (df), and p-value (p) in the results.

OLM, also referred to as “novel place learning” or “novel object location,” is a memory task that assesses rodents' preference for novelty; in this case, the novelty of the object is placement in a different location (44). Mice were handled for 2 min for 5 consecutive days and habituated to the testing arena for 5 min for 6 consecutive days, with days 4 and 5 of handling overlapping with days 1 and 2 of habituation. On training day, two identical objects were introduced into the training area, spaced equally apart. Animals were allowed to explore the objects for 10 min to become familiarized with the objects and their location. After 24 h, the mice were tested for long-term object location memory by returning the mice to the testing arena with one of the previous objects placed in the same location and the other object placed in a novel location, and allowing the mice to explore the objects. All extra-maze cues in the room were kept constant throughout training and testing. Habituation, training, and testing were all filmed and analyzed in EthoVision software. Exploration of the objects was scored by two investigators blind to subjects' condition. Time of exploration (t) was determined by the subject being within 1 cm of the object and the elongated line from eyes to nose would intersect the object. The discrimination index (DI) was determined by:

Three- to four-week-old WT or ZnT3KO mice were used to prepare acute hippocampal slices. Animals were anesthetized with 5% isoflurane and decapitated and their brains were rapidly removed and rinsed with ice-cold artificial cerebrospinal fluid (ACSF). The hippocampus was dissected out and sliced into 350 μm-thick sections with a chilled Stoelting tissue chopper and transferred to individual wells in a 48-well plate containing cold ACSF. The ACSF was replaced with 500 μL of 37°C Neurobasal medium and stabilized at 37°C and 5% CO₂ for 1 hour before experimental procedures. Slices from two subjects were pooled for each experiment to result in an equal number of slices for each condition in the experiment. Conditions were no treatment, KCl, and KCl with NMDAR subunit-specific antagonists.

Hippocampal slices were incubated with KCl (20 mM, Sigma). The following drugs were used to modulate NMDA receptor activity: NR2A antagonist [(R)-[(S)-1-(4-bromo-phenyl)-ethylamino]-(2,3-dioxo-1,2,3,4-tetrahydroquinoxalin-5-yl)-methyl]-phosphonic acid (PEAQX or NVP-AAM077), Sigma; NR2B antagonist 4-[2-[4-(cyclohexylmethyl)-1-piperidinyl]-1-hydroxypropyl]phenol (ifenprodil), Sigma. Drugs were reconstituted with sterile H₂O and diluted in Neurobasal media.

Hippocampal slices were transferred to RIPA buffer containing 10 μL/mL Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific), homogenized for 20 s with an IKA Ultra-Turrax T8 homogenizer, and centrifuged at 14,000 x g for 1 h. The supernatant was removed and stored at −80°C until immunoblotting. Protein concentration was assayed using the Thermo Fisher Scientific Pierce BCA protein assay kit.

Proteins were prepared for SDS-PAGE using the BioRad Criterion system. Twenty micrograms of proteins were equalized with RIPA buffer, diluted 1:1 with Laemmli sample buffer (BioRad)/β-Mercaptoethanol prepared at a ratio of 95:5, heated to 100°C for 5 min, loaded into wells of Criterion 4-20% Tris/HCl 1.0 mm precast gels, and separated by gel electrophoresis for 2 h at 125 mV in BioRad tris/glycine/SDS running buffer. Proteins were then transferred onto PVDF membranes (Immobilon) by Western blot for 3 h at 4°C in Towbin buffer (BioRad) with Precision Plus Protein All Blue Standards (BioRad) to identify molecular weights. Membranes were blocked overnight at 4°C in 5% non-fat milk or 5% BSA with 0.1% Tween 20 in TBS and probed with primary antibodies overnight at 4°C in 2.5% non-fat milk or BSA with 0.05% Tween 20 in TBS. Membranes were then washed three times for 5 min each with 0.1% Tween 20 in TBS and incubated in goat anti-mouse and goat anti-rabbit IgG (H+L) HRP-conjugated secondary antibodies (Thermo Fisher Scientific) in 2.5% non-fat milk or BSA with 0.05% Tween 20 in TBS for 2 h at room temperature. Membranes were washed three times for 5 min each with 0.1% Tween 20 in TBS and incubated for 5 min in SuperSignal West Femto, Dura, or Pico chemiluminescent substrate and imaged on a BioRad ChemiDoc XRS+ imaging system and analyzed for quantification with ImageLab software (BioRad). All data were normalized to GAPDH (glyceraldehyde 3-phosphate dehydrogenase). Primary antibodies used were mouse anti-GAPDH (Abcam 125247), rabbit anti-GAPDH (Abcam 37168), rabbit anti-BDNF (Abcam 6201), rabbit anti-MAPK (ERK1/2; Cell Signaling Technology 4695), rabbit anti-phospho-p44/42 MAPK (ERK1/2; Cell Signaling Technology 9106), rabbit anti-Akt (pan; Cell Signaling Technology 4691), rabbit anti-phospho-Akt (Cell Signaling Technology 2965), and mouse anti-calbindin (Swant 300). Secondary antibodies were Alexa Fluor 488 goat anti-rabbit, Alexa Fluor 488 goat anti-mouse, Alexa Fluor 594 goat anti-rabbit, and Alexa Fluor 594 goat anti-mouse (Molecular Probes).

Twenty micrograms of proteins were equalized with RIPA buffer and loaded into adjacent wells of an S&S manifold-I dot-blot apparatus housing a nitrocellulose membrane. The samples were drawn through the apparatus and into the membrane by vacuum and blocked overnight at 4°C in 5% non-fat milk with 0.1% Tween 20 in TBS and probed with primary antibodies overnight at 4°C in 2.5% non-fat milk with 0.1% Tween 20 in TBS. Antibodies used were rabbit anti-NPY (Abcam 10980) and mouse anti-MAP2 (Millipore MAB4318). Membranes were washed three times for 5 min each with 0.1% Tween 20 in TBS and incubated in goat anti-mouse and goat anti-rabbit IgG (H+L) HRP-conjugated secondary antibodies (Thermo Fisher Scientific) for 2 h at room temperature. Membranes were washed three times for 5 min each with 0.1% Tween 20 in TBS and incubated for 5 min in SuperSignal West Dura chemiluminescent substrate and imaged on a BioRad ChemiDoc XRS+ imaging system and analyzed for quantification with QuantityOne 1-D analysis software (BioRad). All data were normalized to MAP2 adjacent samples.

Hippocampal slices from 3- to 4-week-old WT or ZnT3KO mice were prepared as previously described for chemical modulation of neuronal activity and incubated in 20 mM KCl or vehicle for 3.5 h at 37°C and 5% CO₂, transferred to 1 mL of Trizol (Invitrogen 15596-026) and homogenized for 20 s with an IKA Ultra-Turrax T8 homogenizer. RNA was isolated using the Trizol method (Invitrogen) and stored at −80°C until purification using a Qiagen RNeasy mini-kit (Qiagen 74104) with on-column DNase I digestion according to the manufacturer's protocol. RNA concentrations and purity were measured using a NanoDrop 2000 spectrophotometer and analyzed using NanoDrop software (Thermo Fisher Scientific).

cDNA was prepared from up to 1.0 μg RNA using the BioRad iScript Reverse Transcription (RT) Supermix system (BioRad 170-8840) and RT-PCR was performed with a PCR Sprint thermal cycler using the BioRad iScript RT protocol. The resulting cDNA was diluted 1:5 in DNAse/RNAse-free H₂O and used for qPCR by the SYBR green method (BioRad 170-8880) with a BioRad DNAengine Peltier thermal cycler using the iQ SYBR green protocol. Primer sequences (Eurofins MWG Operon) used were:

TCGTTCCTTTCGAGTTAGCC (mBDNFexS)

TTGGTAAACGGCACAAAAC (mBDNFexAS)

AGGTATCCTGACCCTGAAG (mActinS)

GCTCATTGTAGAAGGTGTGG (mActinAS).

The resulting data was analyzed using MJ Opticon Monitor (BioRad) and mRNA expression was measured with satisfactory reproducibility between triplicates and fold change relative to average ΔCt was calculated as 2^−ΔΔCt and normalized to actin.

Free-floating 40 μm-thick sections of 4% paraformaldehyde (PFA)-fixed mouse brain sections were permeabilized and blocked with 5% BSA/0.3% Triton X-100/PBS for one-to-two hours and incubated overnight at 4°C with primary antibodies in 2.5% BSA/0.015% Triton X-100/PBS. Primary antibodies used were mouse anti-calbindin 1:1,000 (Swant 300), rabbit anti-NPY 1:500 (Abcam 10980), mouse anti-GAP-43 1:1,000 (Calbiochem CP09L), rabbit anti-synaptoporin 1:750 (Synaptic Systems 102002), mouse anti-synaptophysin 1:1,000 (Cell Signaling 7H12) and mouse anti-synaptophysin 1:1000 (Abcam ab8049). The sections were washed three times for 5 min each with PBS and incubated for 1 hour at room temperature with secondary antibodies in 2.5% BSA/0.015% Triton X-100/PBS. Secondary antibodies used were goat anti-rabbit and goat anti-mouse fluorescent conjugated Alexa 488 and Alexa 594 1:5,000 (Invitrogen). The sections were washed again three times for 5 min each with PBS and incubated in Hoescht 33342 solution 1:10,000 (Invitrogen) in PBS for 5 min, washed again as before and mounted on a microscope slide with Vectashield mounting media (Vector Laboratories) with a microscope cover glass. Slides were dried for 1 h at 37°C and stored at 4°C.

Images were captured with an Axiovert 200 inverted microscope and processed using AxioVision software (Zeiss) for visualization of calbindin and NPY immunolabeling. An Apotome device (Zeiss) was used for optical “Z” sectioning of multifluorescence signals, then a maximum intensity projection of all z planes was created for analysis of synaptoporin and GAP-43 immunolabeling. This allowed visualization of individual processes through multiple planes of focus.

Calbindin and NPY immunolabeling in specific hippocampal regions were quantified as percent area in two fields from two samples for each condition. The mean fluorescence was measured in each sample in cell-free areas to determine background autofluorescence and the images adjusted to normalize the level of background across green and red channels. Synaptoporin puncta were counted at 63X magnification with three fields imaged for each subject in the molecular layer of the dentate gyrus of ZnT3KO mice and values were expressed as ratio relative to WT puncta for each subject. Puncta were counted if the size and intensity were above a threshold that was determined to be background fluorescence. GAP-43 levels were analyzed in ten fields from each subject imaged at 63X magnification and the percent area of GAP-43 immunofluorescent labeling in each field was quantified using the AxioVision (Zeiss) automated measurement program and the average percent area was calculated.

40 μm-thick PFA-fixed mouse brain sections were mounted on microscope slides with 0.5% gelatin (Sigma) mounting solution and dried for 5 min at 45°C. The slides were then immersed in 100% ethanol for 5 min, 70% ethanol for 2 min, and rinsed in deionized water for 2 min. Slides were then immersed in a solution of 0.06% KMnO₄ (Sigma) for 10 min with gentle shaking and rinsed in deionized water for 2 min. Slides were then immersed in a solution of 0.0004% Fluorojade B (Millipore) in 0.1% acetic acid for 20 min with gentle shaking in the dark, rinsed in deionized water three times for 1 min each and dried at 45°C for 10 min. Slides were then immersed in xylene for three changes of 1 min each and coverslipped with DPX mountant (Sigma).

Two-to-four images of the CA3 region of the hippocampus were captured at 20X with an Axiovert 200 inverted microscope and processed using AxioVision software (Zeiss).

Background intensity was normalized between samples and Fluorojade puncta were marked if the size and intensity were above a threshold that was determined to be background fluorescence and counted as positive labeling if the Fluorojade puncta was co-localized with Hoescht staining to eliminate non-specific Fluorojade label. The number of Fluorojade- and Hoescht-positive cells counted in the multiple images was averaged to quantify degenerating neurons for each animal.

EEG activity was recorded in 2-month-old freely moving mice using a Neurologger recording device (TSE Systems) and processed with Spike2 v8.07 (Cambridge Electronic Design) analysis software. Surgeries were performed as previously described in Vogler et al. (45). Briefly, mice were anesthetized with 3% isoflurane, the scalp removed, then 0.255 mm diameter electrodes were implanted 1 mm deep into 0.26 mm holes drilled into the skull and fastened with cyanoacrylate glue. A mounting pedestal was fashioned with dental cement over the electrode wires and exposed skull, and the wound sealed with cyanoacrylate glue. Two recording electrodes were placed at −1.34 mm A/P, ± 1.5 mm M/L and two at −5.34 mm A/P, ± 3.00 mm M/L relative to bregma and the reference electrode was placed at −7.00 mm A/P, 0.50 mm L relative to bregma. Mice were intraperitoneally injected with 3 mg/kg ketoprofen prior to surgery, provided with 1.25 mg/ml acetaminophen in their drinking water for 3 days post-surgery, and housed individually in a home cage for the duration of the experiment. Mice were briefly anesthetized with isoflurane at least 4 days post-surgery and the Neurologger was affixed to the pedestal. The mice were then returned to their home cage, which was then placed in a grounded Faraday cage for a recording period of 20–26 h. Recordings with poor signal from any electrode were rejected. EEG data during movement detected by the accelerometer function of the Neurologger device was removed manually and the raw EEG data was bandpass filtered from 1 to 80 Hz. Spike threshold was a peak amplitude 2.5X greater than the average baseline to determine spikes/hour.

The Morris water maze task (MWM) has been used to assess memory in ZnT3KO mice by several research groups. Those previous findings indicate that 6-10-week-old ZnT3KO mice are unimpaired (46), while 3-month-old ZnT3KO gave variable results in different studies (43, 47, 48), and 6-month-old ZnT3KO mice exhibit significant impairment (43). As ZnT3 and synaptic Zn²⁺ are heavily concentrated in the hippocampus, we selected a task requiring hippocampal-dependent spatial memory; specifically, the object location memory (OLM) task (44, 49), which has no requirement to learn a task or to introduce a motivating factor. Instead, OLM relies on rodents' inherent preference for novelty over familiar circumstances, including familiar objects placed in novel locations (50, 51), and requires hippocampal function for encoding, consolidation, and retrieval of memory for the location of the objects (52, 53). We profiled hippocampal-dependent memory in ZnT3KO mice at 6 weeks, 3 months, and 6 months of age in both male and females, finding normal cognition at 6 weeks (DI, WT, 26.31 ± 3.714, n = 13; KO, 30.26 ± 2.859, n = 16), impairment at 3 months (DI, WT, 27.22 ± 4.083, n = 9; KO, 12.07 ± 3.93, n = 14; t = 2.567 df = 21, p = 0.018) and profound impairment at 6 months (DI, WT, 28.63 ± 5.193, n = 8; KO, −2.972 ± 6.136, n = 14; t = 3.487 df = 20, p = 0.0023; Figure 1), with no difference between male and females in performance at each age (data not shown). Thus, our results suggest that the lack of synaptic Zn²⁺ does not directly cause cognitive impairment as young ZnT3KO have intact cognition, but does trigger progressive neurodegenerative processes as ZnT3KO age, leading to impaired cognition.

Because Zn²⁺ activates TrkB receptors and inhibits NMDA receptors (3, 25, 40, 54), we investigated the effects of the removal of synaptic Zn²⁺ in the expression and activation of two proteins found in both signaling pathways: Erk1/2 and AKT (31, 55–60). Acute hippocampal slices from 3-4-week-old ZnT3KO and WT mice were treated for 10 min with vehicle or 20mM KCl to stimulate vesicle release to assess basal expression and activity-induced phosphorylation levels, followed by protein extraction and analysis by Western blot. Densitometry normalized to GAPDH showed reduced levels of AKT and Erk1/2 in ZnT3KO under basal conditions (Figure 2A and Supplementary Figure 1A). The percent increase of p-AKT after treatment with KCl was similar in both genotypes (data not shown). However, the increase in p-Erk1/2 after treatment with KCl was significant in WT mice (Figure 2B and Supplementary Figure 1B), with a reduced percent increase of p-Erk in ZnT3KO mice (p-Erk1/2 % increase, WT, 72.8 ± 3.9%, n = 4; KO, n = 5, 41.1 ± 3.2%; t = 6.341 df = 7, p = 0.0004). Levels of p-Erk1/2 at basal conditions were also reduced in ZnT3KO mice (p-Erk NT, WT, 0.41 ± 0.02, n = 4; KO, 0.26 ± 0.02, n = 5; t = 4.054 df = 7, p = 0.0048). These results invited further investigation of the role of synaptic Zn²⁺ in activation of Erk1/2.

NMDARs have differential effects on the activation of Erk1/2 depending on NMDAR location and subunit composition, but conflicting results have been reported (31, 33, 57). Consequently, we investigated the effects of removal of synaptic Zn²⁺ on differential activation of Erk1/2 by NMDAR subunits using antagonists of NR2A (PEAQX) and NR2B (ifenprodil) NMDAR subunits. Acute hippocampal slices were treated with vehicle, 20 mM KCl, or 20 mM KCl+antagonist and the protein samples were prepared for Western blot. Treatment with KCl+ifenprodil reduced levels of p-Erk1/2 in both WT and ZnT3KO, with similar percent decrease in both genotypes (p-Erk1/2 % increase, WT −35.19 ± 1.41%, n = 4; KO, −33.87 ± 3.13%, n = 6, t = 0.3234 df = 8, p = 0.7546; Figure 3A and Supplementary Figure 2A). This indicates that the lack of synaptic Zn²⁺ does not alter NR2B-mediated activation of Erk1/2. Unexpectedly, treatment with KCl+PEAQX increased levels of p-Erk, with a significantly greater increase in the ZnT3KO (p-Erk1/2 % increase, WT, 10.95 ± 6.21%, n = 5; KO, 78.90 ± 11.55%, n = 5; t = 5.181 df = 8, p = 0.0008; Figure 3B and Supplementary Figure 2B). This observation suggests that synaptic Zn²⁺ inhibits NR2A-mediated Erk1/2 activation as the removal of Zn²⁺ resulted in increased p-Erk1/2.

Exogenous Zn²⁺ has differential modulation of NMDAR subunits dependent on membrane voltage and Zn²⁺ concentration (24, 25), highlighting the complex interactions of endogenous Zn²⁺ in NMDAR-mediated activation of Erk1/2. To further explore this relationship, we investigated the effects of the removal of synaptic Zn²⁺ on basal and activity-dependent expression of BDNF, which is linked to NMDAR activation of Erk1/2 signaling (34, 61), and is critical for cognitive function and neuronal health. To determine if the removal of synaptic Zn²⁺ alters BDNF expression, we assayed BDNF protein levels in age-matched hippocampus tissue from 6-, 12-, and 15-month-old WT and ZnT3KO mice. Proteins extracted from tissue lysates and analyzed by Western blot normalized to GAPDH showed an age-dependent decrease in BDNF levels becoming significant by 15 months of age (BDNF, WT, 0.93 ± 0.09, n = 6; KO, 0.59 ± 0.07, n = 6; t = 2.932 df = 10, p = 0.015; Figure 4A and Supplementary Figure 3). We also investigated whether the increase of BDNF mRNA expression that occurs after depolarization with KCl (62) is altered in ZnT3KO mice using acute hippocampus slices from 3-4-week-old WT and ZnT3KO mice. The slices were incubated for 3.5 h in vehicle or 20 mM KCl, then assayed through RT-PCR and qPCR. The results showed no significant difference in basal levels of BDNF mRNA and a suppressed increase after treatment with KCl in ZnT3KO mice (Ratio BDNF/actin, NT, WT 1.91 x 10⁻³ ± 5.28 x 10⁻⁷, n = 6, KO, 2.31 x 10⁻³ ± 1.86 x 10⁻⁶, n = 6, p < 0.05; ratio BDNF/actin KCl, WT 2.83 ± 0.72, n = 6; KO 1.22 ± 0.32, n = 6; t = 4.93, df = 10, p < 0.0001; Figure 4B). These results suggest that ZnT3KO mice have impaired activity-dependent increases in transcription of BDNF mRNA but may also have compensatory mechanisms to maintain BDNF homeostasis that become less effective as they age.

ZnT3KO mice exhibit enhanced susceptibility to induced seizures, suggesting that the lack of synaptic Zn²⁺ causes elevated neuronal activity that reaches seizure threshold with less stimulation, or results in reduced capacity to inhibit excessive activity. Consequently, we investigated the possibility of epileptiform activity in ZnT3KO hippocampus using the biochemical and morphological markers of decreased calbindin expression in the dentate gyrus, increased neuropeptide Y (NPY) expression in the CA3, and aberrant mossy fiber sprouting into the molecular layer of the dentate gyrus. Our results show that ZnT3KO hippocampus has an age-dependent reduction in calbindin protein levels (12 mo., WT, 1.04 ± 0.05, KO, 0.85 ± 0.05, n = 8, t = 2.697 df = 14, p = 0.0173; 15 mo., WT 0.94 ± 0.04, n = 7, KO 0.80 ± 0.04, n = 7, t = 2.314 df = 12, p = 0.0392; Figure 5A and Supplementary Figure 4A), with a larger percent decrease of calbindin immunoreactivity in the dentate gyrus (WT, 19.85 ± 0.75, n = 2, KO 2.62 ± 0.15, n = 2, t = 22.36 df = 2, p = 0.002) compared with CA1 (WT, 31.9 ± 1.5, n = 2, KO, 33.98 ± 1.5, n = 2, t = 0.9843, df = 2; Figures 5B, C). We also found an age-dependent increase in NPY protein levels in ZnT3KO hippocampus (12 mo., WT, 0.97 ± 0.04, n = 5, KO 1.16 ± 0.05, n = 5, t = 3.017, df = 8, p = 0.0166; 15 mo., WT, 0.94 ± 0.14, n = 4, KO, 1.64 ± 0.04, n = 4, t = 4.902, df = 6, p = 0.0027; Figure 5D and Supplementary Figure 4B), particularly in CA3 region (12 mo., WT, 1.66 ± 0.06. n = 2, KO, 4.72 ± 0.16, n = 2, t = 17.86, df = 2, p = 0.0031; Figures 5E, F).

Mossy fiber sprouting is commonly identified in Zn²⁺-rich mossy fiber tracts using Timm's staining to label Zn²⁺. However, ZnT3KO animals have negligible Zn²⁺ levels in mossy fibers. Thus, we used two different markers to visualize aberrant mossy fiber sprouting: synaptoporin, a synaptic protein preferentially expressed in mossy fibers, and GAP-43, a growth-associated protein found in growth cones of sprouting axons. We found an age-dependent increase in synaptoporin puncta counts expressed as a ratio of ZnT3KO relative to WT (6 mo., WT, 1.00 ± 0.09, n = 8, KO, 1.89 ± 0.29, n = 8, t = 2.975, df = 14, p = 0.01; 13 mo., WT, 1.00 ± 0.11, n = 8, KO, 1.69 ± 0.29, n = 7, t = 2.366, df = 13, p = 0.0342; Figures 6A, B) and GAP-43 percent area (6 mo., WT, 1.19 ± 0.09, n = 5, KO, 2.79 ± 0.23, n = 5, t = 6.441, df = 8, p = 0.0002; 12 mo., WT, 1.17 ± 0.27, n = 5, KO, 3.00 ± 0.56, n = 5, t = 2.956, df = 8, p = 0.0183; Figures 6C, D), in the molecular layer of the dentate gyrus in ZnT3KO relative to WT mice. Together, these results provide biochemical and morphological evidence of age-dependent epileptiform activity which becomes evident beginning at 6 months of age in ZnT3KO mice.

Evidence of age-dependent increases in markers of epileptiform activity invited investigation to determine the age of onset of aberrant EEG in ZnT3KO mice. Using a wireless EEG recording device, we analyzed epileptiform spiking activity in freely moving 2-month-old mice in their home cages during ≈24-h recording sessions, finding a 147% increase in the rate of spikes per hour (WT, 0.40 ± 0.06, n = 3; KO, 0.99 ± 0.19, n = 3; t = 2.864, df = 4, p = 0.0457) in ZnT3KO mice (Figures 7A, B). The presence of increased epileptiform spiking in 2-month-old ZnT3KO is possibly the genesis or precursor of more extensive neuronal hyperactivity resulting in markers of epileptiform activity by 6 months of age.

Altered expression and activation of neurotrophic proteins and epileptiform activity may result in synaptic loss and neurodegeneration. We quantified synaptic density using synaptophysin immunolabeling and degenerating neurons through Fluorojade labeling. Image analysis showed a significant reduction in synaptic density in the CA3 region of 13-month-old ZnT3KO mice determined by percent area of synaptophysin immunolabeling (WT, 0.167 ± 0.034, n = 6; KO, 0.070 ± 0.011; n = 6; t = 2.687, df = 10, p = 0.0228; Figure 7C) and frequent degenerating neurons in the CA3 region of the hippocampus in 13-month old ZnT3KO mice (WT, 2.08 ± 0.46, n = 11; KO, 3.70 ± 0.54; n = 11; t = 2.87, df = 20, p = 0.0095; Figure 7D), with no significant difference found in 2- or 3-month-old ZnT3KO mice, indicating that the lack of synaptic Zn²⁺ leads to age-dependent synaptic and neuronal loss.