A multifunctional therapeutic approach to disease modification in multiple familial mouse models and a novel sporadic model of Alzheimer’s disease

NMZ retained the neuroprotective actions of CMZ [15], providing efficacy against OGD (Fig. 2a) and against exposure to oAβ (250nM or 5 μM). Cell viability was reduced by ~ 15 % and 60 % after vehicle treatment and exposure to oAβ, 250nM and 5 μM, respectively (Fig. 2b, c). NMZ demonstrated superior protection compared to CMZ in mitigating oxidative and low and high dose oAβ insults.

NMZ was first tested in APP/PS1 double transgenic mice that manifest aberrant accumulation of human Aβ driven by FAD mutations in APP₆₉₅ ^K670N/M671Land PS1^M146L. We have previously measured NMZ brain bioavailability and correlated brain levels with efficacy in cognition enhancement in scopolamine-induced amnestic mice, demonstrating that the brain concentration of NMZ required for memory consolidation after amnestic insult is approximately 0.5–1.0 nM [19]. Our previous PK/PD studies [19] guided dose selection herein, since NMZ was effective by both i.p. and p.o. drug administration at doses of 1 mg/kg and 20 mg/kg/day, respectively. Therefore, NMZ was administered at 1 mg/kg/day i.p. and 20 mg/kg/day in drinking water, treating randomized groups of both transgenic and age- and sex-matched wildtype (WT) littermates from 2.5 months of age for 12 weeks. In the last 2 weeks of treatment, animals were tested in the radial arm water maze (RAWM). NMZ treated transgenic animals performed significantly better than APP/PS1 control animals in trial blocks 2 and 4 (p < 0.05) (Fig. 2d).

In addition to improvement of cognition, whole brain homogenates were analyzed for various biomarkers. In whole brain homogenates, a significant reduction of oAβ was observed for NMZ treated animals for Aβ_1–42 species (Fig. 2e). Classic Aβ pathology was studied using quantitative Thioflavin S stereological analysis of serial coronal sections. NMZ treatment demonstrated almost a 3-fold reduction in Thioflavin S-positive Aβ deposit volume in both the cortex and hippocampus (Fig. 2f and h). Lower levels of inflammatory markers in human blood and in rat models of hypoxia-ischemia in vivo and ex vivo, including the pro-inflammatory cytokine tumor necrosis factor alpha (TNF-α), have been reported previously for CMZ [12]. Total brain homogenates from APP/PS1 mice and WT littermates analyzed for TNF-α showed significant NMZ-induced reduction (Fig. 2e). This was mirrored by Iba-1 staining of coronal slices that showed reductions in activated microglia in the hippocampi and cortex of NMZ treated mice (Fig. 2g). Thus, in the one FAD model, in which WT mice are littermates, NMZ treatment significantly reduced elevation of TNF-α; a result compatible with the reported anti-inflammatory actions of CMZ.

The increased age of LaFerla’s 3 × Tg (APP₆₉₅ ^K670N/M671L, PS1^M146L, tau^P301L) transgenic mouse model at treatment onset and the more complex pathology represents an increased therapeutic challenge. The effect of NMZ was first studied in LTP in the CA3-CA1 pathway in hippocampal slices from 16-month-old male mice [37]. Since the 3 × Tg is a homozygous line, the 129/SvJ x C57BL/6 background was used as an outbred line to provide control slices to give baseline performance. NMZ treatment resulted in a significant increase in fEPSP after LTP induction compared to untreated transgenics, to levels similar to WT control (Fig. 3a). To investigate the involvement of the NO/cGMP pathway in the action of NMZ, ODQ (1H-[1, 2, 4] oxadiazolo-[4, 3-a] quinoxalin-1-one), a selective inhibitor of soluble guanylyl cyclase (sGC) was used to block cGMP production, and added 30 min prior to induction. LTP induced in the presence of both NMZ and ODQ was not significantly different to that obtained in untreated 3 × Tg slices. Examination of the field responses during theta-burst stimulation (TBS) indicated that NMZ facilitated LTP in 3 × Tg mice by enhancing the depolarization induced by the high frequency bursts (Fig. 3b); this effect was also attenuated by inhibition of cGMP production by ODQ.

In light of the findings in APP/PS1 mice, we chose to use a similar treatment protocol in 3 × Tg mice. Drug treatment started at 10 months of age for females and 12 months of age for males (to account for early pathology in females). In the last week of the trial, the animals were tested in the step through passive avoidance (STPA) task. The performance of NMZ treated 3 × Tg mice was comparable to untreated WT and significantly better than vehicle treated 3xTg mice (Fig. 3c). The procognitive activity of NMZ is proposed to be mediated via re-activation of CREB, leading to increased expression of gene products, notably brain-derived neurotrophic factor (BDNF). In cerebral homogenates, both pCREB and BDNF levels were significantly increased in NMZ treated 3 × Tg mice (Fig. 3d and e). Aβ_1–42 measured in the insoluble guanidine extracted fraction and oAβ in the soluble fraction was lowered with NMZ treatment (Fig. 3f). Staining using the Aβ antibody 6E10 revealed fewer β-amyloid plaques in the hippocampal formation, particularly in the subiculum (Fig. 3h), although the cortex did not show a significant difference in plaque pathology. Immunohistochemical analysis using AT8 an antibody directed against phosphorylated tau showed a significant decrease in staining in the CA1 region of NMZ treated mouse brains (Fig. 3i). In contrast, total tau was unchanged, measured by western blot of total brain extract using antibody HT7 (a human-specific monoclonal antibody recognizing amino acids 159–163) (Fig. 3g).

E4FAD mice carry targeted replacement of mouse APOE with human APOE4. The 5 × FAD background causes deposition of human Aβ, which is more severe in apoE4 carrier mice than in mice carrying the apoE2 or apoE3 isoforms [30]. The E4FAD mouse is therefore an important AD model for drug testing of predictive efficacy in the human apoE4 carrier population. Male E4FAD mice at 3.5 months of age were, as in the previous mouse models, administered NMZ i.p. (1 mg/kg/day) and p.o. in hydrogel (20 mg/kg/day) for 12 weeks before sacrifice, allowing comparison with published studies on male EFAD mice at the same age [38]. A significant cognitive deficit was not observed using the STPA task, nor nesting behavior assessment up to 6 months of age, compared to C57BL/6 WT mice. However, after tissue extraction, Aβ_1–42 showed a significant reduction in the soluble, insoluble and oligomeric forms (Fig. 4a, b). Levels of pCREB and total CREB, measured in hippocampal homogenates, showed a significant increase in pCREB and no significant differences in total CREB in the NMZ treatment group (Fig. 4c). Activation of CREB is intrinsically linked to synaptic function; and the scaffold protein, postsynaptic density protein 95 (PSD-95), plays a critical role in synaptic plasticity and synaptic dysfunction in AD [39, 40]. Immunoassay of homogenates from E4FAD mice showed a significant increase of PSD-95 in the hippocampus of E4FAD mic3 after treatment with NMZ (Fig. 4d).

Aldh2 ^−/− mice, compared to age and sex-matched littermates, show a progressive decrease in performance, beginning at 2.5–3 months of age, in both the novel object recognition (NOR) and Y-maze tasks, measures of non-spatial and spatial working memory [36]. Given the positive outcomes of NMZ treatment in the three familial AD transgenic mouse models, we decided to forego i.p. delivery in Aldh2 ^−/− mice, administering drug solely by the oral route (20 mg/kg/day) for 12 weeks from 3 months of age. The Aldh2 ^−/− mouse represents a model for accelerated aging manifested by accumulation of lipid peroxidation products that fail to be cleared by Aldh2 enzyme action and pathology reflected by significantly increased levels of HNE-adducted proteins. We hypothesized that NMZ would counteract the downstream effects of this pathology on neuronal and synaptic function without impacting upstream HNE-adduct formation. Marked increases in HNE adduct formation occur early in Aldh2 ^−/− mouse brains, characterized by increases in both the density and number of immunoreactive bands in immunoblots, which was not significantly perturbed by administration of NMZ (Fig. 5a). Aldh2 ^−/− mice demonstrated age-related increases in soluble Aβ and phospho-tau that were significant at 6 months of age [36] and attenuated by NMZ treatment (Fig. 5b, c), without change in total tau. Age-related changes in synaptic and neuronal markers were also significantly attenuated by NMZ treatment (Fig. 5d), as were age-related increases in activated caspases 3 and 6 (Fig. 5e). These progressive biochemical changes linked to synaptic and neuronal dysfunction are intrinsically linked with AD in humans.

Aldh2 ^−/− and WT mice were tested in the Y-maze and NOR before randomization into treatment groups. In both tasks, untreated Aldh2 ^−/− mice showed a significant cognitive deficit that was more distinct with age; NMZ treatment having no effect on WT mice. A significant deficit was seen in Aldh2 ^−/− mice prior to randomization; whereas, NMZ treated mice up to 6 months of age showed no cognitive deficit in the Y-maze and NOR (Fig. 6a, b).

A final ex vivo experiment was carried out using the acetylcholine receptor agonist, carbachol, applied to hippocampal slices from NMZ-treated Aldh2 ^−/− mice, WT littermates, and their vehicle controls. Carbachol stimulation of hippocampal slices from 6 month old WT mice that had been administered NMZ or vehicle control for 12 weeks caused significant and similar elevation of pCREB (Fig. 6c). Slices from vehicle treated Aldh2 ^−/− mice showed significantly attenuated basal pCREB response relative to WT and no significant effect of carbachol stimulation, whereas hippocampi from Aldh2 ^−/− mice that had undergone NMZ treatment showed basal response and stimulation in response to carbachol comparable to WT littermates. NMZ treatment of Aldh2 ^−/− mice was able to restore normal basal and stimulated pCREB signaling.

All animal care and procedures were conducted with approved institutional animal care protocols at UIC (OACIB), Columbia University (IACUC) and Queen’s University (UACC), in accordance with the NIH Guide for the Care and use of Laboratory Animals.

Mice were randomly grouped into NMZ or vehicle treatment groups and mice were treated with drug or vehicle for 10–12 weeks; where applicable littermates were sex and age-matched. APP/PS1 mice (6–8 weeks) were administered drug (1 mg/kg qd) or saline i.p. supplemented with drug (20 mg/kg qd in drinking water). 3 × Tg mice (males 12 months; females 10 months) and E4FAD mice (male 4 months) were administered drug using the same protocol, with the exception that drug in hydrogel replaced drinking water [38]. Aldh2 ^−/− mice (3 months) were treated with drug (20 mg/kg qd) in hydrogel. Primary cultures of cortical neurons were prepared from DIV 16 days old rat embryos and subjected to transient OGD or oAβ insults as described [15].

LTP measurements were made on 400 μm 3 × Tg mouse slices from 16-month-old male mice using ten theta bursts. Slices were perfused with ODQ (10 μM) and/or NMZ (50 μM) in aCSF perfusate for 30 min before inducing LTP as measured by fEPSP slope measured for 50 min.

The RAWM task has proven informative in the analysis of short-term spatial working memory in AD models and was performed as previously described [58]. NOR and Y-maze were performed as described: preliminary data, tested using a three-way analysis of variance, showed no sex differences in either task for Aldh2 ^−/− mice, therefore NOR and Y-maze data from male and female mice were combined for analysis [36]. In the NOR, on the test day, animals were allowed to explore the objects until they accumulated a total of 30 s of exploration time, and the discrimination index (difference in time exploring the novel and familiar object, divided by total exploration time) determined. In the Y-maze, spontaneous alternation rate was calculated as the total triads containing entries into each of the three arms without repeated entry into a previously visited arm, divided by the total number of arm entries. STPA has been widely used to test long-term working memory and was performed as previously described [20].

Brains were separated by hemisphere for immunohistochemistry or ELISA/immunoblot studies. Frozen hemi-brains were homogenized and serially extracted for biomarker measurement. For EFAD mice, proteins were extracted from the hippocampus using a three step serial extraction as described [38]. Levels of oAβ (Biosensis) Aβ42 (Invitrogen), TNF-α (Invitrogen), pCREB (Biosource), and BDNF (Promega) were determined for soluble and insoluble fractions using sandwich ELISA kits and reported as weight/unit per milligram of soluble protein. Samples for immunoblot analysis of Aβ were incubated in Laemmli buffer containing 6 M urea for 30 min prior to electrophoresis. The following antibodies were used: mouse monoclonal to Aβ (W0-2, Millipore, recognizes residues 4–10 of Aβ), rabbit polyclonal antibody to 4-hydroxynonenal (HNE11-S antibody, Alpha Diagnostics International), mouse monoclonal antibody to PHF-tau (AT8 antibody, Pierce Thermo Scientific), rabbit polyclonal antibody to caspase-3 (ab90437, Abcam), rabbit polyclonal antibody to caspase-6 (ab52295, Abcam), mouse monoclonal antibody to PSD-95 (Pierce Thermo Scientific), mouse monoclonal antibody to total tau (ab64193, Abcam), rabbit polyclonal antibody to phospho-CREB (Ser 133, 06–519, Upstate Biotechnology), rabbit polyclonal antibody to total CREB (06–863, Millipore), mouse monoclonal antibody to β-actin (Sigma). Serial coronal brain sections (30 μm) of paraformaldehyde-fixed hemi-brains were cut frozen using a Leica cryostat in six adjacent series and immunohistochemically processed using antibodies directed against Aβ/APP, a phospho-specific (Ser202/Thr205) tau antibody, and microglia marker Iba1 (primary, 6E10, Covance, 1:2000; AT8, ThermoFisher, 1:1000; anti-Iba-1, Wako, 1:1000; secondary goat anti-mouse IgG for 6E10/AT8, 1:200; goat anti-mouse IgM for anti-Iba-1, 1:200). Quantitative Thioflavine-S staining was performed on frozen hemi-brain 20 μm serial slides. Aβ deposition stereology in cortex and hippocampus were analyzed using quantitative Metamorph software. Brain slices, were incubated with carbachol or vehicle (Basal) for 30 mins and snap frozen. Immunoblot analysis for pCREB was performed using 30 μg protein of hippocampal homogenate, and immunoreactive bands were quantitated by densitometry. Hippocampal slices from WT and Aldh2 ^−/− mice that had been treated with NMZ or vehicle control for 12 weeks, were incubated with carbachol (50 μM) or vehicle for 30 min and snap frozen, before immunoblot analysis of pCREB in homogenates.

Unless otherwise stated, all data are expressed as mean ± SD or SEM and were analyzed by one-, two- or three-way analysis of variance with either Bonferroni’s or Tukey’s post-hoc test, and/or a Student’s t test for unpaired data. A p-value of 0.05 or less was considered statistically significant. All statistical analysis was done of GraphPad Prism.

The Office of Animal Care and Institutional Biosafety (OACIB) UIC; The Institutional animal care and use committee (IACUC) Colombia University; The University Animal Care Committee (UACC) Queen’s University.