Impairment of cognitive function by chemotherapy: association with the disruption of phase-locking and synchronization in anterior cingulate cortex

Animals

A total of 59 male Sprague–Dawley rats were used in these experiments; they were housed
in standard laboratory conditions (temperature at 23?±?1 °C and 12/12 h light–dark
cycles with lights on at 7:30 a.m.) The rats were allowed to adapt to their environments
for at least 5–7 days before the first injection of cisplatin and were handled for
3–5 min daily for 2–3 days before beginning the behavioral experiments. Rats had free
access to water, but had moderate food deprivation during the decision-making testing.
All surgical and experimental procedures were conducted according to the guidelines
laid down by the NIH in the US regarding the care and use of animals for experimental
procedures and were approved by the Committee on Use and Care of Animals at City University
of Hong Kong and the licensing authority to conduct experiment from Department of
Health of Hong Kong (No. 13–89 in DH/HAP/8/2/5 Pt.2).

Drug treatments

The cisplatin dosage and dosing schedule used in the present study were based on the
clinical usage of cisplatin treatment into the abdominal cavity in cancer patients,
slightly modified according to previous description in rodents 4]. Cisplatin (International Laboratory, USA) was dissolved in 0.9 % NaCl (0.4 mg/ml),
warmed to 45 °C and injected with 1.0 ml/100 g rat. The rats weighing 250–300 g at
the beginning of the experiment received intraperitoneal injections of cisplatin or
0.9 % NaCl once every 7–8 days for a total of 6 times in recent memory testing group
and 8 time for remote memory group (Table 2).

Table 2. Experimental design

Behavioral assessments

Behavioral assessments began after 3 cisplatin administrations. The animals were tested
sequentially in a series of behavioral tests in the following order: (1) open field
test (OFT), (2) elevated plus-maze (EPM), (3) Morris water maze (MWM), (4) rat Iowa
gambling task (RGT). The rats were given a recovery phase of 5 days before, and 5 days
after MWM to maintain a fair general condition during chronic cisplatin treatment.
All behavioral procedures were conducted during the light phase of the day. To minimize
possible circadian influences on the rats, cisplatin and control rats were observed
alternate basis.

Open field test (OFT)

The OFT is often used to assess exploration in a novel environment and offers a preliminary
screen for anxiety-related behavior in rats 19]. Before the actual testing, animals were habituated to the behavioral testing room
for 1–2 h sessions on 2 consecutive days. During a 5-min testing, rats were placed
individually in the center of a 40 × 40 cm square drawn in the middle of a black square
arena (80 × 80 × 40 cm) and then allowed to explore the field freely. The apparatus
was cleaned between rats using 70 % ethanol. Spontaneous exploration behavior in the
OFT were observed and recorded by ANY-maze (Stoelting Co., Wood Dale, IL, US), as
indicated by total horizontal distance traveled (m), number of rearings (times), time
spent (s) and number of entries (times) in the center and peripheral areas, and freezing
time (s).

Elevated plus-maze (EPM)

The EPM is a popular behavioral test for anxiety-like behavior and is thought to result
from natural aversion of rats to explore elevated and open areas 20]. The apparatus was made of brown acrylic plastic with two sets of opposing arms (open
arms: 50 × 10 cm; closed arms: 50 × 10 × 40 cm), 50 cm from the ground. Rats were
placed individually in an open field for 5 min and then allowed to start exploring
the maze freely from the junction of the two sets of arms (10 × 10 cm) facing one
open arm in a 5-min test. The maze was cleaned between rats using 70 % ethanol. Time
spent in each arm was recorded using the ANY-maze with entries being defined as 85 %
of the area of the animal being present in the area entered. Time spent in open arms,
especially the percentage of open arm time versus closed arm time was evaluated to
assess anxiety.

Morris water maze (MWM)

The MWM was used to assess and compare spatial learning and memory in cisplatin-treated
and control rats as described previously 26], 67]. Experiments were performed in a black circular tank (a diameter of 150 cm with 60 cm
depth), which was filled with the opaque water (22–24 °C) by adding black nontoxic
paint. A circular platform (diameter of 10 cm and 25 cm high) submerged 2.0 cm below
water level in the center of one of the four quadrants (target quadrant). Behavioral
data of the animal in the water maze were acquired using the ANY-maze by a digital
video camera located above the center of the tank.

Four consecutive days of probe trainings were performed to assess special learning.
Rats were released from 4 randomized release points facing the tank wall to learn
to locate a hidden platform in 60 s using visual orientation cues on the walls of
the tank and on the curtains around the tank. Rats either found the platform by themselves
or were guided by the experimenter, and remained on the platform for 10 s. Four trials
were conducted each day with intervals of 30 min between each. Twenty-four hours or
thirty days after the last training session, a single probe test was performed to
assess recent or remote spatial memory, respectively. The hidden platform was removed
to assess memory retention for the submerged platform location and rats were released
from the farthest position of original platform and allowed to search the target quadrant
that previously contained the escape platform. Swim paths for all trails were recorded
and latency to reach platform (sec), distance traveled (m), swimming speed (m/s),
and the amount of time spent in the target quadrant (% of total time) were calculated
to assess spatial recent or remote memory acquisition and retention/retrieval.

Rat Iowa gambling task (RGT)

The rat Iowa gambling task was used to detect decision-making. The rats made choices
from options associated with different amounts of reward (food pellets) but also different
amounts and likelihoods of penalties (time-outs) 18]. The training and test procedures for the rat were identical to the ones previously
described by Rivalan et al. 49].

The task was performed in four polyvalent conditioning boxes (28 × 30 × 34 cm) adapted
from five-choice serial reaction time chambers (Imetronic, Pessac, France). In each
box, there were 4 nose-poke holes illuminated with white LED on the front curved wall,
a food dispenser providing food pellets (45 mg, TestDiet, USA) at the back wall and
a transparent central opening partition (7 × 7 cm) dividing the box into two chambers
at the middle. Infra-red detectors were equipped in holes to detect the nose poke
and connected to the food dispenser.

Prior to the actual training, food was restricted over a 3-day period following a
1-day fasting period. Simultaneously, 50 reward pellets per rat were put inside the
cage every day to make sure the rats became habituated to their taste. During the
training phase, daily food was restricted for each rat to about 5 g of food per 100 g
body weight and rats usually took 5–7 days to make the association between nose-pokes
in illuminated holes and food rewards in the food dispenser. In order to guarantee
that the selection of the nose poke was a conscious choice, the rats were trained
to associate a single nose-poke with one food pellet delivery according to a criterion
of at least 100 pellets obtained within a 30-min session, followed by two consecutive
nose-pokes with one food pellet delivery with the same criterion. Two final 5-min
training sessions were conducted to habituate the rats with the quantity of pellets
that could be obtained during the test. The first session was set by two nose-pokes
with two pellets at a time (maximum 30 pellets) after making a choice and the other
with one pellet (maximum 15 pellets).

The test procedure was performed the following day and lasted 60 min or was cut off
by 250 pellets obtainment. Rats were free to make choices among the four holes (A-D)
as during the training phase; however, different choices were associated with different
outcomes set as follows: choices A or B related to two pellets each time as immediate
reward, but had separately 1/2 probability to trigger a long penalty time-out (222 s)
or 1/4 probability for a very long penalty time-out (444 s), during which period no
pellet can be obtained; choices C or D associated to smaller immediate reward (one
pellets each time), but also smaller penalty (1/4 chance for 12 s time-out, or 1/2
chance for 6 s time-out). Although the immediate reward of choice A and B was twofold
that of choice C and D, in the long run the choice C or D would allow for the rats
to obtain more food pellets compared to choice A or B. Thus the choices C and D were
advantageous choices, while the choices A and B were disadvantageous choices. The
percentage of advantageous choices (number of nose-pokes (C?+?D) / number of nose-pokes
(A?+?B?+?C?+?D) * 100 %) was used as a criterion to distinguish the good (70 % preference
of advantageous choices during the last 20 min of RGT test), undecided (30 % – 70 %
preference) and poor (30 % preference) decision-makers. Among the good decision makers,
delayed-good decision-making was determined when the percentage of advantageous choices
was below 70 % at 30 min and above 70 % in the last 20 min.

Electrophysiological recordings

The electrophysiological recordings in both the control and cisplatin-treated rats
were performed after the series of behavioral tests.

Evoked local field potential (LFP) recording

LFP in the anterior cingulate cortex (ACC) elicited by the electrical stimulation
of the basolateral amygdala (BLA) was used as a quantitative measure of synaptic potency
in BLA-ACC pathway. Eleven male Sprague–Dawley rats (5 control and 6 cisplatin-treated
rats) were anesthetized (with 1.5 g/kg urethane, i.p.) and placed in a stereotaxic
frame with body temperature maintained at 36.5?±?0.5 °C. A glass recording microelectrode
filled with 2.0 M NaCl was slowly lowered into the ACC (AP 2.0 – 3.8 mm, ML 0.5 –
1.0 mm, depth 1.5 – 3.5 mm). A bipolar tungsten electrode was placed in the ipsilateral
BLA (AP ?3.0 to ?3.3 mm, ML 4.8 – 5.3 mm, depth 6.7 – 7.5 mm). To evaluate synaptic
potency test stimuli (500 ?A, square wave pulse, duration: 0.2 ms) were delivered
at 0.033 Hz. After responses stabilized, variations in the stimulus current (from
50–1000 ?A, increasing 50 ?A intervals) were delivered in order to generate input–output
curves (I/O). The details of the procedure were similar to those of previous publications
33], 68].

Long-term potentiation (LTP) induction

To examine the synaptic plasticity, LTP was induced by applying theta burst stimulation
(TBS, three sets of 10 trains with 10 s intervals; each train consisted of 10 bursts
at 5 Hz containing 5 pulses at 100 Hz) to the BLA. The test stimulate intensity was
chosen to evoke 50 % of the maximum amplitude of evoked LFP. Details of the procedure
have been described previously 33].

Multiple electrodes recording

In brief, rats were anesthetized and placed in a stereotaxic frame. Two small holes
(1–2 mm wide) were drilled above the BLA and ACC to insert 16-channel silicon-based
electrodes (Plexon, Dallas, TX). The silver grounding wires from the electrodes were
wrapped around the mounting screws. Electrodes were slowly advanced using a micropositioner
until clear neuronal action potentials in most recording channels were observed on-line
(OmniPlex^® D system, Plexon, Dallas, TX). LFPs and spike firings were recorded with a 64-channel
electrophysiological data acquisition system. LFPs were amplified (×20,000), band-pass
filtered (0.05 – 200 Hz, 4-pole Bessel) and sampled at 1 kHz. Spikes were filtered
(0.3 – 5 kHz, 4-pole Bessel) and sampled at 40 kHz. Signals from the BLA and ACC were
recorded simultaneously during spontaneous activity and visceral stimulation (colorectal
distention, CRD, 60 mmHg) 33].

Histological identification

After completion of the electrophysiological recordings, a DC current (100–500 ?A)
was passed between the recording electrodes and ground to lesion the brain. The rats
were perfused with saline followed by 4 % paraformaldehyde and then the brains were
sliced at 50 ?m and stained with Cresyl violet. Drawings were made of sections showing
electrode tracks related to the structure of the ACC and BLA. A standard rat atlas
was used as reference for reconstruction of the stimulating and recording sites 69].

Data analysis

Statistical analyses

Data from the behavioral tests were analyzed by ANY-maze, POLY File (Imetronic, Pessac,
France), SPSS (IBM SPSS Statistics 20) and GraphPad (Prism 5.0). Data were compared
using student’s unpaired t-test and one-way and two-way analysis of variance (ANOVA)
followed by the Bonferroni post hoc test. Comparisons of proportions of individuals
in the gambling task were conducted using the non-parametric Mann–Whitney test. The
neural data were processed off-line using NeuroExplorer 5 (Plexon, Dallas, TX) and
exported to custom written MATLAB (MathWorks, Natick, MA) programs for additional
analysis. Data sets were compared by one-way repeated-measures ANOVA followed by multiple
comparisons of Bonferroni’s test using Prism 5.0 (GraphPad, La Jolla, CA). Results
were expressed as mean?±?SE. P? 0.05 was considered statistically significant.

Spike sorting

The single-unit spike sorting was performed with Off-Line Spike Sorter V3 software
(OFSS, Plexon Inc., Dallas, TX) using combined manual and automatic sorting techniques.
Spikes were identified when 3 SDs higher than the noise amplitude. All waveforms recorded
from each channel were isolated as distinct clusters in 3D space based on the characteristics
of spike waveforms using principal component analysis (PCA). Automatic techniques
were used to generate separation of waveforms into individual clusters. Manual checking
was performed to ensure that the spike waveforms were consistent and cluster boundaries
were clearly separated. All isolated single-units exhibited recognizable refractory
periods (?1 ms) in the inter-spike interval (ISI) histograms 70].

Spectral analysis

Oscillatory brain activity in the theta (4–10 Hz) frequency range is believed to temporally
linked single neuron into assemblies cooperatively facilitate synaptic plasticity
40], and plays a role in many different cognitive functions, including memory and decision
making 45]. To clarify alterations in the theta power spectra following cisplatin treatment,
we calculated the ACC theta power, peak frequency of the theta oscillations, as well
as the theta ratio (theta/(theta?+?delta). In order to achieve this, the raw ACC LFPs
were filtered between 1 and 100 Hz using non-causal zero-phase-shift filter (fourth-order
Butterworth). The power spectral densities (PSD) were characterized using multi-taper
estimates with 7 tapers and 2¹³ frequency bins (NeuroExplorer 5, Plexon, Dallas, TX). The overlapping percentage
of each window is 50 %. The power spectra were normalized so that the sum of all the
spectrum values equals the mean squared value of the signal. The PSD curve was smoothed
with a Gaussian filter (15 bins running average), with the band power being defined
as the area under the curve (AUC) of the corresponding frequency domain. The power
from each animal was averaged over the 16 channels and to compute the theta/delta
ratio, the AUC of the theta band power to the theta plus delta band (1–10 Hz) was
calculated. To expose the time-frequency features of the theta oscillations in both
BLA and ACC, the time-varying power spectra were calculated by FFT with Hann window
function using NeuroExplorer 5. The spectrum units were normalized by raw PSD, so
that the sum of all the spectrum values equals to the mean squared value of the signal.

The phase-locking of single neuron activity to the theta oscillations of local field
potential

To investigate the angular distributions of spikes in relation to the theta oscillation,
and clarify the significance of phase-locking of spikes with the theta oscillation,
the phase distribution and Rayleigh test were performed using in house MATLAB scripts
modified based on the study in 17]. To ensure the validity of the statistical results, only neurons that had at least
50 spikes were used for phase-locking estimation. Before performing the analysis,
22 different frequencies ranging from 1.6 to 64 Hz were selected, such that, f?=?2^x with x?=?{6/8, 8/8, 10/8, 12/8, …, 48/8}. The LFP was then convolved with a series
of Morlet wavelets centered about each of the selected frequencies and each with a
length of 4 cycles. The result of these wavelet transforms, is a matrix of vectors
whose absolute values (or length) and arguments (or angles) represent the amplitude
and phase, respectively, of the LFP at frequency f and time t. At each of the selected frequencies, the phase angle of each spike was estimated
from the inverse tangent of the real and imaginary components of the wavelet transform.
The circular mean of the spike phases was calculated by taking the weighted sum of
the cosine and sine of the angles. The mean angle was then the angle of the results
and the mean vector length (R) the absolute value over the number of spikes. To check
for phase-locking, the Rayleigh test was used to compare against uniformity by calculating
the test statistic and a p value. A neuron was considered phase locked in theta range if the P value is below the threshold of 0.0023 which is 0.05 Bonferroni corrected for multiple
comparisons (0.05/22, twenty-two frequencies were tested) 17].

Synchronized theta oscillations between BLA and ACC by cross-correlation analyses

The synchronized theta activities between the BLA LFP_? and ACC LFP_? were evaluated by cross-correlograms using NeuroExplorer 5 software. We aligned the
2 LFP_?s and chose the BLA LFP_? as reference. The Pearson correlation values were calculated with a lag time ranging
from ?0.5 to 0.5 s with small bins (2 ms). The cross-correlation curves were smoothed
with a Gaussian filter (5 bins running average). We averaged the cross-correlograms
from valid electrode channels in the BLA and ACC and took the second positive peak
as a quantitative measure because it locates at about 0.2 s lag time which represents
theta activity at about 5 Hz 71].