Re-engineering a neuroprotective, clinical drug as a procognitive agent with high in vivo potency and with GABA A potentiating activity for use in dementia

NMZ is a GABA
_A
potentiator and shows a sedative effect at higher doses than CMZ

The core anticonvulsant, sedative, and anxiolytic properties of CMZ derive, at least
in part, from GABA potentiation. Measurements using the Xenopus oocyte model expressing ?
₁
?
₂
?
₂
GABA
_A
receptors confirmed that NMZ retained GABA
_A
potentiating activity in vitro (Fig. 1a). As a measure of sedation, motor impairment was assessed using latency to fall
on a rotating rod for C57BL/6 mice (Fig. 1b). Preliminary experiments showed a significant effect of CMZ (50 mg/kg i.p.), therefore,
this and an equimolar dose of NMZ were selected. All animals were trained in the task
for 3–5 days prior to experiment, until latency to fall reached 100 ± 10 s. At 10 min
post injection, CMZ and NMZ groups had a latency to fall of 14.2 ± 9.8 and 38.7 ± 5.4 s
respectively (p  0.001), showing motor impairment over vehicle treated animals (108.1 ± 6.2 s).
At 30 min, only CMZ showed a significant deficit (35.0 ± 12.8 s; p  0.001), versus
vehicle (115.8 ± 6.3 s; NMZ 93.5 ± 5.4 s), and at 60 min, CMZ retained significant
sedative actions (73.0 ± 6.5 s; p  0.05; vehicle 122.1 ± 10.7 s; NMZ 109.1 ± 7.0 s).
To test more profound sedation, loss of righting reflex (LORR) was assessed in male
C57BL/6 mice (Fig. 1c). The duration of LORR, defined as failure to place four paws on the ground within
30 s after placing the mouse on its back, was measured every 2 min for 2 h post drug
administration. All animals recovered fully after treatment. CMZ significantly induced
LORR at doses lower than NMZ. At the 125 mg/kg dose, NMZ induced LORR for 4.8 ± 1.7 min,
which was not statistically significant; whereas CMZ at a molar equivalent dose induced
a significant LORR (23.8 ± 2.7 min). Only at the highest dose (175 mg/kg) did LORR
reach significance for NMZ (LORR = 90.5 ± 20.9 min, compared to a molar equivalent
dose of CMZ: LORR = 66.2 ± 7.7). It is important to note that the dose of NMZ needed
to induce a transient sedative effect, reflected by LORR, was almost 200-fold higher
than the procognitive dose reported below.

Fig. 1. Retention of GABA
_A
potentiating, and attenuated sedative activity in NMZ relative to CMZ. a Oocytes expressing the ?
₁
?
₂
?
₂
GABA
_A
receptor (n = 6) showed a dose response to increasing concentrations of NMZ in the
presence of GABA (6 µM). Addition of picrotoxin (200 µM) caused 96 ± 2 % inhibition
of the potentiated GABA response (n = 4) (O). Data show mean ± SD normalized to the
saturated 200 µM GABA response. b Male C57Bl/6 mice (n = 5–16) were injected i.p. with CMZ (45 mg/kg) or an equimolar
dose of NMZ before testing for their latency to fall on a rotating rod (RR). NMZ showed
less sedation than CMZ at various time points. Data show mean ± SEM. Statistical significance
relative to vehicle is indicated by *p  0.05, **p  0.01, ***p  0.001, using one-way ANOVA with Dunnett’s post hoc test. c Male C57BL/6 mice (n = 4–5) were injected with escalating doses of CMZ and NMZ and
loss of righting reflex (LORR) was measured over 2 h. NMZ showed less sedation than
CMZ, and no LORR was observed for NMZ until 125 mg/kg. Data show mean ± SEM. Non-zero
statistical significance by one-sample t test is indicated by *p  0.05, ***p  0.001

NMZ reverses memory deficits in WT animals by consolidating memory

NMZ was designed to add procognitive activity to the neuroprotective actions of CMZ,
therefore, NMZ (1 mg/kg) was tested in male C57BL/6 mice using the step-through passive
avoidance (STPA) behavioral task, in which mice learn to associate a mild electric
shock (0.5 mA) with the dark side of a light–dark box and latency to enter is assessed
24 and 48 h after training. A variety of amnestic agents, administered i.p. 30 min
before training, have been shown to cause loss of memory without hindering the ability
of animals to learn the task. The muscarinic receptor antagonist scopolamine (1 mg/kg)
41], the NMDA receptor antagonist MK-801 (0.1 mg/kg), the benzodiazepine diazepam (0.5 mg/kg)
42], and the nitric oxide synthase (NOS) inhibitor L-NAME (50 mg/kg), were able significantly
to inhibit memory as reflected by decreased latency in STPA (Fig. 2a). Vehicle treated animals reliably show a latency in testing at or close to the
cutoff threshold of 300 s, which was reduced after treatment with amnestic agents.
Administration of NMZ (1 mg/kg i.p.) 20 min prior to training restored memory as shown
by significantly increased latency (scopolamine 98.1 ± 16.6 vs. 240.7 ± 25.3 s treated;
MK-801 94.6 ± 9.3 vs. 255.1 ± 15.4 s treated; diazepam 166.7 ± 16.6 vs. 250.9 ± 22.5 s
treated; L-NAME 22.0 ± 10.4 vs. 211.7 ± 31.4 s treated 50 min prior to training).
To test the effects of oral drug delivery, animals were treated with NMZ (20 mg/kg)
in drinking water for 24 h prior to training, again resulting in reversal of a scopolamine-induced
deficit, tested at 24 h (231.5 ± 19.3 s), or 48 h post training (Fig. 2b).

Fig. 2. Reversal of induced amnestic deficits by multiple agents in STPA. a Male C57BL/6 mice (n = 5–10) treated with vehicle or NMZ (1 mg/kg, single dose i.p.;
or 20 mg/kg/day, oral drinking water) after being administered with diverse amnestic
agents were tested for their latency to enter the dark side of the STPA apparatus at 24 h after training. NMZ reversed memory deficits induced
by scopolamine (1 mg/kg), MK-801 (0.1 mg/kg), diazepam (0.5 mg/kg), and L-NAME (50 mg/kg).
b Male C57BL/6 mice with scopolamine-induced deficits treated with NMZ showed reversal
of latency to enter the dark side of the chamber 24 and 48 h after training. Statistical
significance is indicated by ***p  0.001 compared with the respective saline control for each time point, using unpaired
t test. c In scopolamine induced deficits, NMZ demonstrated reversal of cognitive deficits
when administered i.p. between 40 min prior to training (?40 min) and 90 min after
training (+90 min); but no significant effect was observed when administered 60 or
120 prior to training. Data show mean ± SEM. Statistical significance is indicated
by *p  0.05, **p  0.01, ***p  0.001 compared with the respective saline control for each treatment, using one-way
ANOVA with Dunnett’s post hoc test

In the STPA task, animals are tested at least 24 h after drug administration, therefore,
confounding locomotor drug effects are highly improbable. The task is also suited
to correlate pharmacokinetics with pharmacodynamics (PK/PD). Using scopolamine-induced
amnesia, the time of administration of NMZ (1 mg/kg i.p.) was varied from 120 min
prior to 90 min post training (Fig. 2c). NMZ was procognitive when administered within 40 min prior to the start of training,
while not effective when given earlier (?120 min: 134.0 ± 11.3 s; ?60 min: 125.2 ± 37.1 s;
?40 min: 263.0 ± 17.1 s; ?20 min: 280.0 ± 11.4 s). When administered after training,
the procognitive effect was significant at least up to 90 min (+30 min: 243.0 ± 34.0 s;
+60 min: 300.0 ± 0.0 s; +90 min: 228.1 ± 71.9 s). Task acquisition during training
did not differ significantly between vehicle and NMZ treated animals (data not shown).

NMZ is orally bioavailable in the brain

The loss of procognitive activity when NMZ is administered 60 min before training
is indicative of drug clearance and a T
_1/2
  60 min in mice. The plasma and brain concentrations of NMZ and its denitrated metabolite,
5-(2-hydroxyethyl)-4-methylthiazole (HMZ), were measured by LC–MS/MS (Fig. 3a), since HMZ is itself known to have bioactivity in vivo 43]. At early time points after bolus injections of NMZ at sedating (50 mg/kg i.p.) or
procognitive doses (1 mg/kg i.p.), NMZ brain concentrations of 10.7 ?M and 81.3 nM
were measured, respectively. In STPA, NMZ is procognitive when administered 20 min
prior to training, but loses activity administered 60 min prior to training: corresponding
to brain levels of NMZ of 10.2 and 1.01 nM respectively. However, we also observed
that NMZ administration prior to training is not required to restore memory, but that
NMZ functions by consolidating memory when administered after training. The brain
concentrations of NMZ and HMZ 5 h after injection were only 0.32 and 5.7 nM, respectively.
In place of drug delivery in drinking water, drug was delivered in hydrogel that mice
readily consume for hydration when drinking water is made unavailable. This delivery
method achieves a relatively constant drug concentration throughout the awake-period,
simulating an extended release clinical formulation 44]. Oral administration of NMZ (20 mg/kg) over 24 h, representing a procognitive dose
in STPA, resulted in brain concentrations of NMZ and HMZ of 0.73 and 3.41 nM, respectively.
Under these conditions, brain and plasma concentrations were not significantly different.
Taken together, these measurements indicate that the brain concentration of NMZ required
for memory consolidation after amnestic insult is approximately 0.5–1.0 nM. The measured
concentration of HMZ could be used as a surrogate for the maximum theoretical concentration
of NO released from NMZ, which would be 3–10 nM.

Fig. 3. Pharmacokinetics study on male C57BL/6 mice treated with NMZ, single dose (50 or 1 mg/kg,
i.p.) or supplied in hydrogel (20 mg/kg/day), showing bioavailability in brain and
plasma at various time points after initiation of treatment. The estimated t
_½
for i.p. administration is 10 min

NMZ, but not CMZ, restores LTP in hippocampal slices from APP/PS1 mice

In hippocampal slices from 3-month-old male mice, the effect of NMZ and CMZ on long-term
potentiation (LTP) was measured in the CA3-CA1 pathway. After 15 min of baseline collection,
drug (100 ?M) was added to the bath solution for 5 min prior to induction of LTP using
three trains of ten theta bursts, and the resulting fEPSP were recorded in the CA1
area for 120 min. CMZ had no significant effect, whereas NMZ induced a significant
increase in fEPSP slope after LTP induction compared to the untreated transgenic,
to levels indistinguishable from WT control (Fig. 4). NMZ perfusion of WT hippocampal slices had no significant effect (data not shown).
It is important to note that NMZ treatment had no effect on LTP induced in hippocampal
slices from WT mice, but restored LTP in slices from APP/PS1 mice. This observation
is compatible with the ability of NMZ to activate NO/cGMP/pCREB signaling, when signaling
is impaired.

Fig. 4. Beneficial effects seen in LTP from NMZ treatment in APP/PS1 mice. a, b LTP was measured in the CA1 region of hippocampal sections in 4 month old male APP/PS1
mice or littermate controls (n = 5–8) treated with CMZ (a) or NMZ (b). NMZ showed restoration of LTP in APP/PS1 mice to WT levels, whereas the effects
of CMZ were not significant. Statistical significance was analyzed by two-way ANOVA
with repeated measures: WT veh (n = 6) vs. WT NMZ (n = 6): F(1,10) = 1.106 p  0.05
No Sig.; WT veh vs. APP/PS1 veh: F(1,12) = 18.86 p  0.05 Sig; APP/PS1 veh (n = 8)
vs. APP/PS1 NMZ (n = 7): F(1,13) = 17.71 p  0.05 Sig; WT NMZ vs. APP/PS1 NMZ: F(1,11) = 0.02351
p  0.05 No Sig