In Vivo Characterization of ARN14140, a Memantine/Galantamine-Based Multi-Target Compound for Alzheimer’s Disease

Alzheimer’s disease (AD) is a chronic pathological condition that leads to neurodegeneration, loss of intellectual abilities, including cognition and memory, and ultimately to death. It is widely recognized that AD is a multifactorial disease, where different pathological cascades (mainly amyloid and tau) contribute to neural death and to the clinical outcome related to the disease. The currently available drugs for AD were developed according to the one-target, one-drug paradigm. In recent times, multi-target strategies have begun to play an increasingly central role in the discovery of more efficacious candidates for complex neurological conditions, including AD. In this study, we report on the in vivo pharmacological characterization of ARN14140, a new chemical entity, which was obtained through a multi-target structure-activity relationship campaign, and which showed a balanced inhibiting profile against the acetylcholinesterase enzyme and the NMDA receptor. Based on the initial promising biochemical data, ARN14140 is here studied in mice treated with the amyloidogenic fragment 25–35 of the amyloid-β peptide, a consolidated non-transgenic AD model. Sub-chronically treating animals with ARN14140 leads to a prevention of the cognitive impairment and of biomarker levels connected to neurodegeneration, demonstrating its neuroprotective potential as new AD agent.


Results
Neurodegeneration was induced by i.c.v. injection of Aβ [25][26][27][28][29][30][31][32][33][34][35] . ARN14140 was introduced into the lateral ventricle in mice for 7 days of chronic infusion through an Alzet mini pumps implant. To guarantee full coverage for the entire infusion period, compound doses (2.5 and 7.5 μ g/day for 7 days) were selected based on the in vitro potency of the molecule on both targets. Since the molecular weight of ARN14140 is 506.73, we estimated a constant concentration of the compound of almost 12 μ M. This value was about 6 fold higher than the K i for the NMDA receptor (2.3 μ M) and 17 fold higher than the IC 50 for the AChE enzyme (0.7 μ M). Measurements were taken 7-9 days later.
Behavioral readouts included two complementary procedures to assess memory functions. We measured the following biomarkers of neuronal function and integrity: i) hippocampal lipid peroxidation as a cell integrity marker; ii) hippocampal TNFα as a neuroinflammation marker; iii) hippocampal Synaptophysin (Syn) as a synaptic integrity marker. We also looked at Bax and Bcl-2, two proteins involved in apoptotic pathways. Finally, we examined brain tissue integrity to quantify the damage in hippocampal pyramidal cell layers CA1. We assessed the cholinergic loss by immunolabeling of vesicular ACh transporter (VAChT) in cortex, nucleus basalis magnocellularis (Meynert), and hippocampal formation. Figure 2 reports the effect of ARN14140 on Aβ 25-35 -induced spontaneous alternation deficits in mice during the 7 days. Treatment with Aβ 25-35 induced significant spontaneous alternation deficits when compared to Sc.Aβ (Fig. 2a). ARN14140 dose-dependently reduced the Aβ 25-35 -induced memory deficits with full effect at a dose of 7.5 μ g/day. Conversely, treatments did not affect total arm entries as reported in Fig. 2b. Figure 2 also reports the results of the passive avoidance test. Figure 2c shows that, with respect to untreated mice or those treated with Sc.Aβ , the Aβ 25-35 treatment induced highly significant deficits in mice performance both as step-through latency and as escape latency during the retention session (Fig. 2c,d, respectively). Treatment with ARN14140 before training dose-dependently reduced the Aβ 25-35 -induced deficits, with significant prevention at a dose of 7.5 μ g/day on both parameters ( Fig. 2c,d). The ARN14140 effect was not associated with the typical NMDAR antagonist drawbacks such as increased step-through latency or lower shock sensitivity (data not shown).

Behavioral tests.
In summary, the present findings indicate that chronic infusion of ARN14140 in the lateral ventricles fully protects mice from the development of short-term memory deficits after i.c.v. injection of Aβ [25][26][27][28][29][30][31][32][33][34][35] . Since the Y-maze performance mimics spatial learning and is driven at the hippocampal level, it can be concluded that ARN14140 provides a functional neuroprotective effect for this type of memory. ARN14140 was also effective in the passive avoidance test, which is a fear-motivated test classically used to assess short or long-term memory. Therefore, the data also suggest ARN14140 has an effect on this type of short-term memory. Both memory types are typically affected in AD patients 18 . Biochemical analyses. We then analyzed the effect of ARN14140 on specific markers related to neurodegeneration to assess whether the functional neuroprotection seen behaviorally could be associated with a cellular neuroprotection.
First, we studied lipid peroxidation, a well-known marker of oxidative stress, which is associated with neurodegenerative diseases 19,20 . As shown in Fig. 3a, treatment with Aβ 25-35 induced a highly significant increase in  10 .
lipid peroxidation in mouse hippocampus compared to Sc.Aβ /V-injected mice. ARN14140 infusion significantly prevented the Aβ 25-35 -induced increase of lipid peroxidation. The effect of ARN14140 was complete at 7.5 μ g/day, suggesting that, at this dose, the compound fully reverses the Aβ 25-35 -induced neuronal oxidative damage.
AD has been associated with activation of inflammatory pathways including the hyper secretion of pro-inflammatory, neurotoxic cytokines such as TNFα by reactive microglia and monocytes 21 .
As shown in Fig. 3b, after treatment with Aβ [25][26][27][28][29][30][31][32][33][34][35] , the hippocampal content of TNFα increased compared to ScAβ -injected mice (+ 89% increase, p < 0.05 using a two-column test, red bar vs. empty bar). As reported in the same figure, continuous i.c.v. infusion of ARN14140 for 7 days tended to prevent the increase in TNFα levels, although the ANOVA failed to reach significance, a similar attenuation was noted with both regimens, 2.5 and 7.5 μ g/day. Since TNFα is thought to be part of the initiation mechanism of immune-mediated neurotoxic inflammation 22 , the data suggest that ARN14140 can possibly provide protection against AD-related inflammatory events. We then investigated the Synaptophysin expression in hippocampus. Syn is a 38 kDa protein, present in virtually all neurons, which participates in synaptic transmission. Reduction of Syn is an index of synaptic function and many reports suggest that Syn levels are lowered by neurodegeneration 23,24 .
In our study, treatment with Aβ 25-35 induced a highly significant decrease of hippocampal Synaptophysin content (− 34%) as compared to Sc.Aβ -injected mice (cfr white bar with red bar in Fig. 3c). At a dose of 7.5 μ g/day, ARN14140 significantly prevented the Aβ 25-35 -induced decrease in Syn in hippocampus, indicating that, at the top dose, ARN14140 can fully protect synaptic function.
Lastly, we focused on the Bcl-2 protein family. An increase of apoptotic cell death has been reported in the brain tissue of AD patients 25,26 . This event has been associated with a dysregulation of the Bcl-2 proteins balance. The Bcl-2 proteins are a family of evolutionarily related proteins mainly involved in regulating programmed cell death (apoptosis). Bax is the most studied member of the family and has pro-apoptotic activity, while Bcl-2 itself is the most investigated family member with anti-apoptotic activity.
The continuous infusion of ARN14140 for 7 days dose-dependently reduced the increase in Bax content in mouse hippocampus and rebalanced the altered Bax/Bcl-2 ratio (see Fig. 4a,b). In this study, both doses led to a significant effect although the max was observed with the top dose.
In summary, it has been reported that Aβ peptide affect Bax and Bcl-2 expression levels in AD human neurons 27 . Our findings indicate that ARN14140 can rebalance these mechanisms through which can possibly reduce AD-associated apoptosis.
Morphological alterations. We studied the neuroprotective potential of ARN14140 at a cellular level by evaluating its effect on Aβ 25-35 induced CA1 mouse hippocampus cell loss. Cresyl violet was used as a vital stain and a significant decrease in cell count was observed in Aβ 25-35 treated mice 9 days after i.c.v. injection of the peptides (Fig. 5a,b). The continuous infusion of ARN14140 treatment dose-dependently prevented the effect in viable cells, with a significant difference when compared to the Aβ 25-35 /V-treated group at a dose of 7.5 μ g/day (Fig. 5c).
We then analyzed the acetylcholine marker in different samples. VAChT was stained. VAChT is responsible for loading into neuronal endings ready for secretion. A decrease in VAChT indicates less acetylcholine, most likely due to loss of cholinergic neurons. The qualitative analysis of immunolabeling of VAChT showed large immunoreactive fibers in the hippocampus especially in the CA3 area (see Fig. 6a) and a more punctae labeling in the different cortex layers of Sc.Aβ /V-treated animals (Fig. 6b). In Aβ 25-35 -treated mice, a net decrease in immunolabeling was observed (Fig. 6c,d), which could be attributed to the small loss of neurons measured in the hippocampus.
Poor brain integrity due to the chronic implantation of the pump meant that drug analysis could only be done on qualitative bases (marked glial injury, with high variability among mice). Nonetheless a clear trend of increased density of cholinergic terminals was observed in those mice receiving ARN14140. The effect was detectable at the high dose with an intense labeling of cholinergic terminals in the hippocampus and cortex, as pointed out in Fig. 6g,h.
Our morphological findings confirm at a cellular level that ARN14140 infusion had a beneficial effect on development of neurotoxicity. The effect on CA1 hippocampal cell loss was clear and quantitative, allowing in turn a quantitative analysis of the trend of cholinergic markers. Discussion AD is a complex and multifactorial disease whose causes are still partially unknown. It is widely accepted that AD is associated with a series of concomitant biochemical events that contribute to the pathophysiology of the disease. In the search for more effective therapeutic approaches to AD, an emerging option is to design multi-target molecules. The concept is that multi-target molecules can act at several levels of the AD-related neurodegenerative cascade/s 28 at the same time. This should provide a profile with greater therapeutic benefits and reduced side effects, compared to single-target compounds.
Pursuing this line of research, we recently designed a novel class of multi-target molecule: dual NMDAR/ AChE inhibitors. As starting scaffolds, we used two well-known currently available drugs, galantamine and memantine. We "linked" them into a single compound to form a new class of molecules acting on both targets at the same time. Galantamine was chosen because of its intrinsic dual mechanism of action as an anticholinesterase inhibitor (potentiation of endogenous ACh content) and allosteric modulator of neuronal nicotinic receptors (increase of glutamate levels) 29,30 . Memantine was chosen because of its reduced NMDA-related side effects, which may be due to its selective interaction with extrasynaptic NMDAR (selective blockade of extrasynaptic NMDAR does not affect the postsynaptic NMDAR, which is believed to mediate the beneficial glutamate effects in the CNS) 31 .
ARN14140 was the best representative of this new chemical class, which has been the subject of an international PCT patent application (WO2013160728A1).
Based on the initial biochemical characterization, which has already been published 10 , ARN14140 was chosen because it is a fairly balanced molecule. It interacts equally with the AChE enzyme and the NMDA receptor. In this study, the molecule was tested for its ability to prevent the biochemical and behavioral detrimental effects of the i.c.v. injection of Aβ [25][26][27][28][29][30][31][32][33][34][35] . The model is thought to mimic some key phenotypes of AD (mainly cognitive/ memory impairment and neurodegeneration), and has already been used to determine neuroprotective effect of galantamine and memantine alone 32,33 .
In the present study, ARN14140 was given i.c.v. to minimize all possible issues of brain penetration, since preliminary data showed a suboptimal pharmacokinetic (PK) profile. To maintain a constant inhibition of the AChE and NMDA targets for the required length of time, the compound was given by chronic infusion with osmotic mini-pumps, providing a continuous flow to interact with the targets. For these reasons, ARN14140 is a prototype suitable to explore a concept but insufficient for becoming a medicine. We are now seeking to chemically optimize the PK profile and to explore new delivery strategies, such as transdermal patches, which could minimize the unfavorable plasma profile and maximize ARN14140 therapeutic efficacy.
The Aβ 25-35 -treated animals show a statistically relevant cognitive impairment as well as alteration of the major markers of neurodegeneration and cell death. ARN14140 was able to rescue the behavioral impairment in two different tasks and to balance the levels of biomarkers of neurodegeneration, synaptic plasticity, and apoptosis, which were modified by treatment with Aβ [25][26][27][28][29][30][31][32][33][34][35] . We should recall here that ARN14140 was simultaneously administered to Aβ [25][26][27][28][29][30][31][32][33][34][35] , showing that the molecule could be able to tackle neurodegeneration in the very early stages of the disease. However, based on our data, we could not argue whether ARN14140 is also able to reverse the neuronal damage already present in the brain of AD patients quite in advance of the symptom onset. When administered to mice for 7 days at a dose of 7.5 μ g/day, ARN14140 completely reverted the Aβ 25-35 -induced neurotoxicity in terms of behavioral tasks and biomarkers. We believe this highly positive feature was due to the synergistic potentiation arising from the simultaneous engagement of NMDAR and AChE. To support this hypothesis is the increasing evidence that the glutamatergic and cholinergic neuronal systems influence each other, and that their joint dysfunction is crucial to the effects produced by AD 34 . In particular, based on our recent findings on memantine and galantamine in combination 35 , we may hypothesize that the galantamine moiety could determine AChE inhibition, tackling memory loss and cognitive impairment. In addition, the galantamine function is expected to contrast NMDA-mediated neurotoxicity through the allosteric modulation of nicotinic receptors, thus potentiating memantine's efficacy in promoting neuronal survival. This interpretation is, however, rather speculative and cannot rule out the involvement of additional, as yet unknown, mechanisms. Our data suggests that ARN14140 could serve as the initial prototype for a pharmacological advancement with respect to the current AD therapies, since it merges complementary activities that could otherwise only be achieved by administering a cocktail of molecules 36 .
Overall, the present study further confirms that multi-target compounds can play an important role in the treatment of complex neurological disorders, and may represent a novel paradigm in anti-AD drug discovery research.  All experiments and protocols were authorised and approved by the French Ministry of Research, as well as by the Regional Animal Welfare Committee. All efforts were made to minimise the number of animals used.
Spontaneous alternation performances. The spatial working memory was examined by measuring the spontaneous alternation behaviour of the mice in the Y-maze 14,32,[38][39][40][41] . The Y-maze is made of grey polyvinylchloride. Each arm is 40 cm long, 13 cm high, 3 cm wide at the bottom, 10 cm wide at the top, and converging at an equal angle. Seven days after Aβ [25][26][27][28][29][30][31][32][33][34][35] or Sc.Aβ peptide injection, each mouse was placed at the end of one arm and allowed to move freely through the maze during an 8 min session. The series of arm entries, including possible returns into the same arm, was recorded visually. An alternation was defined as entries into all three arms on consecutive occasions. The number of maximum alternations was the total number of arm entries minus two and the percentage of alternation was calculated as actual alternations/maximum alternations × 100. Measured parameters included the percentage of alternation (memory index) and total number of arm entries (exploration index).
Passive avoidance test. The contextual long-term memory of the animals was assessed using a two-session step-through passive avoidance procedure 32,40,41 .
The apparatus consisted of two compartments (15 × 20 × 15 cm high), one illuminated with white polyvinylchloride walls and the other darkened with black polyvinylchloride walls and a grid floor for electrical shocks. A guillotine door separated each compartment. A 60 W lamp positioned 40 cm above the apparatus lights up the white compartment during the experiment. Scrambled foot shocks could be delivered to the grid floor using a shock generator scrambler (Lafayette Instruments, Lafayette, USA). Training: 8 days after Aβ [25][26][27][28][29][30][31][32][33][34][35] or Sc.Aβ peptide injection, each mouse was placed into the white compartment. After 5 s, the door was raised. When the mouse entered the darkened compartment and placed all its paws on the grid floor, the door was closed and the foot shock (0.3 mA) delivered for 3 s. The latency spent to enter the dark compartment and the number of vocalisations was recorded. None of the treatment affected the step-through latency (H = 0. 31 each mouse was placed again into the white compartment. After 5 s, the door was raised and the step-through latency (latency to enter the dark compartment) was recorded up to 300 s. Lipid peroxidation. The quantification of lipid peroxidation in tissue extracts is based on Fe(III)xylenol orange complex formation according to Hermes-Lima et al. 42 . Six mice for each group were sacrificed by decapitation, without anesthesia nor sedation, 9 days after Aβ [25][26][27][28][29][30][31][32][33][34][35] or Sc.Aβ peptide injection. The brains were rapidly removed, weighed and kept in liquid nitrogen until assayed. After thawing, one hippocampus per mouse was homogenized in cold methanol (1/10 w/v), centrifuged at 1,000 g for 5 min and the supernatant placed in an Eppendorf tube. The reaction volume of each homogenate was added to FeSO 4 1 mM, H 2 SO 4 0.25 M, xylenol orange 1 mM and incubated for 30 min at room temperature. After reading the absorbance at 580 nm (A 580 1), 10 μ L of cumene hydroperoxide (CHP) 1 mM was added to the sample and incubated for 30 min at room temperature, to determine the maximal oxidation level. The absorbance was measured at 580 nm (A 580 2).
The level of lipid peroxidation was determined as CHP equivalents according to: CHPE = A 580 1/A 580 2 × [CHP (nmol)] and expressed as CHP equivalents per mg of tissue and as a percentage of the control group data (V-treated Sc.Aβ -administered mice).

Histology/immunohistochemistry. Nine days after Aβ
Examination of the CA1 area was performed using digitalized slices using a Nanozoomer virtual microscopy system (Hamamatsu, Massy, France). CA1 measurement and pyramidal cell counts were processed using a 40 × objective using ImageJ software (NIH). Data were expressed as mean number of viable cells per mm, from at least four slices × 3 CA1 fields × 2 hemispheres for each group, according to the previously reported method 12 .
For immunohistochemical labeling of the vesicular acetylcholine transporter (VAChT), sections were incubated overnight at 4 °C with Rabbit monoclonal anti-VAChT (SOSV00002G, Thermofisher). Then, sections were incubated with Rabbit biotinylated secondary antibody (B8895, Sigma-Aldrich) and incubated for 1 h in avidin-biotin complex (ABC Vector laboratory). Signal was detected with the diaminobenzidine kit (Vector, Laboratories). Examination of the stained area was performed using digitalized slices using Optic microscope (Leica).