Investigation of blast induced neurotrauma in a rodent models

Primary blast injury is the direct result of the explosive generated supersonic blast wave's interaction on the body, by inducing rapid changes in atmospheric peak pressure negative pha

CHAPTER 1 INTRODUCTION

1.1 Overview of Traumatic Brain Injury

Traumatic brain injury (TBI) is one of the foremost causes of disability and death

in both civilian settings and theatres of war TBI can be simply defined as damage to the brain tissue resulting from external mechanical forces (Center for Disease Control)(CDC)) In terms of gross clinical classification, TBI can be further sub-classified

as penetrating or closed head injury TBI severity ranges from mild to moderate to severe TBI severity is clinically determined using the Glasgow Coma Scale (GCS), which assesses the conscious state of a TBI patient in terms of eye, verbal and motor responses The severity of TBI using the GCS is categorised as mild (GCS >13), moderate (GCS 9 - 12) and severe (GCS <9) The Department of Veteran Affairs (VA) and Department of Defense (DoD) have also consolidated a classification (Table 1) for the diagnosis of concussion/TBI in the military (Table 1) This classification table utilises

an array of criteria that includes structural imaging, loss of consciousness (LOC), altered level of consciousness (ALOC), post-traumatic amnesia (PTA) as well as the GCS In this study, the effects of explosive blast waves on the central nervous system (CNS) and functions of the brain were investigated

Table 1: VA/DoD Classification of TBI Severity

Structural Imaging Normal Normal or abnormal Normal or abnormal Loss of Consciousness 0-30min >30min and <24 hrs >24hrs

Glasgow Coma Scale (Best

available score in the first

24 hrs)

The time-course effects of TBI can be defined as a two- stage process: acute primary brain injury followed by secondary brain injury Primary injury cannot be prevented by pharmacological or surgical intervention and is due to the direct and initial insult to the brain caused by the application of external mechanical forces These mechanical forces include (i) direct and blunt force impact causing rapid acceleration

penetrating forces that cause localised brain tissue damage These mechanical forces may cause immediate necrosis of torn and overstretched cells (give reference)

For closed head TBI, the most common form is the coup-contra coup injury which can be distinguished by either focal or diffused axonal injuries (DAI) (Anderson and McLean, 2005) Coup-contra coup injury happens when an impact or violent motion brings the head or skull to a sudden deceleration while the brain is still accelerating, causing the brain to slam into the inner skull and bounced to in the opposite direction, thereby resulting in focal injuries at both sides of the brain These acceleration/deceleration forces can also cause stretching and shearing of the brain tissue when rotational forces are involved In addition, relative differences in the tissue densities moving at different speeds can cause diffused axonal injury (Anderson and McLean, 2005)

Secondary brain injury is the resultant neurochemical and neuro-inflammatory complications from arising from the primary mechanical insult (Zhang et al., 2004) Physiologically, these manifest as tissue ischemia, hypoxic damage, edema, and increased intracranial pressure (ICP) Secondary brain injury is usually characterized by resultant neuronal cell death and functional deficits after injury, including slightly compromised cells and nearby health cells (Borgens RB and P, 2012) Following primary brain insult and necrosis of injured tissue, a delayed cascade of cytotoxic apoptosis commonly attributed to NMDA excitotoxicity of the injured neurons occur (Yi

JH and AS., 2006, Mehta A et al., 2012 ), with generation of free radicals (Pun PB et al.,

2009, O'Connell, 2012), mitochondrial dysfunction and disruption (Robertson CL et al.,

2006, Cheng et al., 2012), and irreversible metabolic disturbances (Tsutsui S and PK.,

2012, Xu F et al., 2012) Importantly, secondary brain injury presents a window that may be responsive to potential therapeutic interventions to improve neurological outcome after TBI

1.2 Overview of Blast-induced Neurotrauma

Exposure to explosive blast devices like improvised explosive devices (IEDs), land mines and rocket propelled grenades (RPG) accounts for almost two thirds of the casualties sustained by the US military in Iraqi and Afghanistan wars (Ling G et al.,

2009, Cernak and Noble-Haeusslein, 2010, Belmont PJ Jr, 2012 ) In the recent wars however, advancement in protective armour may reduce penetrating and blunt impact TBIs caused by the blast and protecting from blast lung injury Nevertheless, with advanced far-forward medical care, more injured soldiers are being diagnosed with some form of blast related TBI or blast-induced neurotrauma (BINT) (Courtney and Courtney, 2011, Ling and Ecklund, 2011a) A recent study suggests that up to 89% of studied brain trauma patients from these wars might have sustained some form of BINT (MacGregor AJ et al., 2011) BINT has become a major area of concern in current military medicine and has been termed the “signature wound” of the recent conflicts (Warden., 2006, Ling G et al., 2009, Ling and Ecklund, 2011b) In deployed US military, from year 2000 to the first quarter of 2012, around 244 217 combat personals suffered some form of TBI of which mild TBI (mTBI) alone accounts for 76.8% of the total

(www.DVBIC.org)

1.21 Physics of Explosive Blast

A blast wave generated by an explosive detonation begins with a rapid and instantaneous rise in air pressure that lasts from less than a millisecond (i.e IEDs), to less than a hundred millisecond (i.e air/fuel bombs, nuclear) (Courtney and Courtney, 2011) In an ideal Friedlander blast wave, which is generated by an explosion in an open field, there is a rapid rise in blast overpressure followed by an exponential decay leading to a negative pressure phase (IG, 2001 , Cernak and Noble-Haeusslein, 2010) (Fig 1) The blast wave progresses from the source of the explosion as a sphere of compressed air (shock front) that is followed by an area of rapidly expanding gases travelling at supersonic speed (Rossle, 1950) It is the interaction of the blast wave with

results in blast injury (Bowen et al., 1968, IG, 2001 ) Traditionally, these injuries were thought to occur primarily to the gas-filled organs (auditory, pulmonary and gastrointestinal systems) Hence, the Bowen’s curves were generated as an estimation

of tolerance of the lung to peak overpressure and the positive phase duration of blast exposure The Bowen’s curves (Bowen et al., 1968) and updated estimates by Axelsson (Axelsson H and JT., 1996), were based on studies of thirteen mammalian species which allowed for an extrapolated estimate on human blast-induced lung injury criteria and mortality rate However, the Bowen’s curves were predominately geared towards lung injury and do not serve as a good estimate for CNS tolerance or BINT (Geoffrey Ling et al., 2009)

Fig 1: Schematic of an ideal Friedlander wave form describing the relationship of pressure wave versus temporal changes for high explosive detonated in a free field with no surfaces nearby that can interfere There is an initial rapid rise in positive blast overpressure, hitting a peak overpressure, followed by an exponential decay leading to a negative pressure phase Diagram taken from Wikipedia

Generally, blast injuries can be described into four different injury categories in relation to their mechanisms of injury: primary, secondary, tertiary and quaternary

i Primary blast injury is the direct result of the explosive generated supersonic blast wave's interaction on the body, by inducing rapid changes in atmospheric

peak pressure

negative phase

positive phase

ii Secondary blast injury is due to the impact of bomb fragments, debris or objects put in motion after being accelerated by the blast wind which can lead to penetrating ballistic or blunt force injuries

iii Tertiary blast injury occurs as a result of victims being thrown by the blast pressure wave For example, victims may be thrown onto the ground or propelled through the air, and striking solid objects resulting in blunt trauma

iv Quaternary blast injury is defined as any explosion-related injury or illness not due to any of the above, such as burns and inhalational injuries High temperatures generated from the explosive gases can also render fatal burns to victims close to the detonation

1.22 Experimental Blast Injury Models

With the understanding that the brain is a vulnerable target for blast injury (Kaur

C et al., 1995 , Cernak et al., 2001, Ling G et al., 2009, Cernak and Noble-Haeusslein, 2010), there has been a flurry of research geared towards the understanding of BINT However, BINT remains a difficult injury to predict and diagnose (Hoge et al., 2009) and the mechanisms by which blast exposure results in BINT are still poorly understood A wide spectrum of blast injury models have been investigated ranging from non human

primates, large porcine models, rodents shock tubes to in vitro cell cultures (Kane MJ et

al., 2012 ) and even computational modeling, in a bid to better understand blast effects

Current research groups investigating BINT utilize both large and small-animal models, and with different methods of generating the blast pressure Majority of BINT research however, makes use of small animal blast exposure models that rely on compressed gas-driven shock tubes to generate blast (Elsayed, 1997, Cernak et al.,

2011, Chavko et al., 2011, Garman et al., 2011b, Leonardi et al., 2011, Reneer et al.,

special membranes rupturing at predetermined pressure thresholds These shock tubes generate blast over pressure (BOP) with significantly lower peak threshold pressure but longer positive overpressure durations Some groups utilize traditional chemical explosives as the pressure generator (Kaur C et al., 1995 , Moochhala SM et al., 2004, Annette Säljö et al., 2009, Risling et al., 2011, Rubovitch V et al., 2011 ) For this study,

we used an explosive model in a bunker using the same blast device as in Risling et al.,

2011 In addition, majority of the blast studies conducted exposed the experimental animals to full body blasts without any protection (Cernak et al., 2011, Chavko et al., 2011) Few make use of selective protection to the head or body using Kevlar materials (Bauman et al., 2009b, Garman et al., 2011a) or custom made acrylic holders (Cernak

et al., 2011, Chavko et al., 2011) Blast studies employing large animals models were limited Most use porcine models (Säljö A et al., 2008, Bauman et al., 2009a, Shridharani et al., 2012) and only one study employed non-human primates (Lu J et al., 2012) Besides investigating single blast exposure, BINT resulting from multiple blast exposures have also been recently investigated (Ahlers ST et al., 2012, Balakathiresan

et al., 2012)

Similar to Bowen’s studies which provides an estimate of blast mortality tolerance based on the peak overpressure and duration of blast exposure, most of the BINT studies also describe their studies as a result of varying exposures of peak overpressures However, the blast overpressures used to induce BINT span over a wide range For example, studies using rodents have experimental overpressures ranging from very low level blast exposure 20 – 60KPa (Moochhala SM et al., 2004, Annette Säljö et al., 2009, Pun et al., 2011, Rubovitch V et al., 2011 ), to 90 to 170 KPa (VandeVord et al., 2011, Ahlers ST et al., 2012, Bir et al., 2012, Valiyaveettil et al., 2012) and the highest at 339 KPa (Cernak et al., 2001, Cheng et al., 2010, Cernak et al., 2011, Koliatsos VE et al., 2011) Annex A provides a list of experimental BINT studies conducted and the pressure profiles reported Unfortunately, most research groups only indicate the peak pressure and positive pressure duration, without specifying the impulse data

Recent reviews had proposed three possible mechanisms that may lead to BINT (Cernak and Noble-Haeusslein, 2010, Courtney and Courtney, 2010, Bolander et al., 2011) The three mechanisms via which BINT may occur are (i) through direct blast pressure propagation through the skull, (ii) via indirect transmission through the vascular system and (iii) through acceleration and/or rotation of the head due to skull flexure Although the direct transcranial propagation was deemed most significant for BINT, (Cernak et al., 2011), selective shielding of the thorax and abdomen showed reduced mortality and brain injury This suggests that the blast wave may transfer kinetic energy through the vasculature and trigger pressure oscillations in blood vessels leading to brain injury (Cernak et al., 2011) However, many other studies did not provide evidence to reinforce this mechanism Both Bauman (Bauman et al., 2009a) and Garman (Garman et al., 2011a) reported that even with thoracic shielding on respectively, swine and rodent models, they still recorded high ICP similar to the applied blast pressure Hence this suggesting that blast pressure was unlikely to be transferred through the vascular systems Other reported ICP by (Chavko et al., 2007) and (Säljö A

et al., 2008); showed similar or even higher ICP recordings than the applied blast pressure, with no protection, suggesting a mechanism of direct transfer of energy through a skull/brain interface Interestingly, multiple studies (Bolander et al., 2011, Leonardi et al., 2011, VandeVord et al., 2011) indicate an alternative skull flexure hypotheses, especially the superior rat skull, which presents a very high stain rate to the brain through the skull/dura interface However, a swine study by (Shridharani et al., 2012) demonstrated a lower attenuation ratio (ratio of air pressure versus ICP) compared to the rodent studies of (Leonardi et al., 2011) and (Säljö A et al., 2008), and questions the degree of energy transfer to the brain tissue Interestingly, a study using human subjects and low level of acoustics wave pressure (80Hz) showed that though the skull acts as an attenuator of higher frequencies, internal cerebral membranes such

as the falx cerebri can reflect and focus shear waves within the brain (Clayton et al., 2012)

Due to the complex nature of the blast pressure transmission and the ambiguity

attempting to identify the mechanism causing TBI and the added complexities associated with mTBI and its associated neurological presentations

1.3 Study Objectives

The mechanisms resulting in BINT has not been elucidated This study aims to investigate the pathophysiological effects primary blast overpressure induced injury in the brain using two different rodent blast models: a shocktube and modified open blast model Histopathology examinations of the brain after blast injury will be correlated to systemic biomarker changes and behavioral outcomes, to better understand the relationship between pressure profiles varying blast overpressure and resultant brain

injury

CHAPTER 2 MATERIALS AND METHODS

2.1 Establishment of a Blast Tube Model

2.11 Animal subjects

Male Sprague-Dawley rats (Taconic, Denmark) weighing 300-350 grams were used in these experiments After 1 week quarantine, animals were food restricted to 85% of their starting weight to ensure motivation to do the behavioral task The animals were allowed a weekly weight increase of 5% and had full access to water All rats were housed in pairs, separated by an acrylic divider, under conditions of room temperature

on a 12 hrs regular light/dark cycle All animal experimentation protocols in this research project were approved by Karolinska Institutet Institutional Animal Care and Use

Committee (permit number N143/09) Table 2 showed the breakdown of animals usage

Table 2: Number of animals used for each injury and behavioural test and the time of sacrifice

2.12 Blast tube model and explosive protocols

The blast tube (Fig 2) used for this study was designed at the Swedish Defence Research Agency (FOI) and described by Clemedson and Criborn (Clemedson CJ and

Criborn CO, 1955) The tube was initially used for rabbits and in vitro experiments on

muscular tissues (Clemedson CJ et al., 1956, Clemedson CJ and Pettersson H, 1956) Subsequently, the tube was modified by Suneson for experiments with rats and used by Saljo with an electric igniter for her thesis work (Säljö A et al., 2000) After further modification, the blast tube was used with a non‐electric NONEL interval igniter (Risling

M et al., 2002a, Risling M et al., 2002b)

Figure 2: Schematic diagram and picture of blast tube at Södersjukhuset, Sweden

The Swedish army explosive M/46 SPRANGDEG46 PTR-NSP 71 PETN (pentaerythritol tetranitrate) was used as the blast source The shocktube had previously been certified for up to 6 g of PETN The PETN was mounted with tape on a non-steel NONEL branch tube igniter at the closed end of the blast tube (Fig 2) The NONEL tubing was connected to a control box in an adjacent room Male adult rats, weighing 300 to 380g, were deeply anaesthetized by an intra-peritoneal injection of 2.4 ml/kg of a mixture of 1 ml Dormicum® (5 mg/ml Midazolam, Roche), 1 ml Hypnorm® (Janssen) and 2 ml of distilled water The anesthetized rat was mounted at a distance of

1 m from the charge on either a wire mesh holder or with a specially fabricated metal body armour (BA) that protected the animal’s body along the horizontal length of the tube (Fig 3)

Figure 3a) The anesthetized animal was mounted in a meshed metallic net to reduce acceleration movement but not BOP exposure; b) Anesthetized animal was mounted with full body armour protection

For safety, all personnel were counted before the experiment and were held in an adjacent room separated by a steel door during the blast After the steel door was closed and secured, accelerated ventilation in the blast tube lab was initiated and the experiment supervisor ignited the charge In the event of ignition malfunction, the charge would be removed and destroyed after 15 min After ignition of the charge, the animal was rapidly removed from the blast tube setup and examined for airway obstruction or bleeding Body weight was monitoredbefore and after injury (up to 3 days) for any changes

2.13 Optimizing explosives based on mortality

Pilot studies were carried out initially to determine the level of BOP that will result

in brain injury The amount of explosives used ranged from 2 g to 5 g PETN, without

BA The studies showed that without armour BA protection, all animals did not survive more than 3 g PETN (~350 KPa) There was 50% mortality at 2.5 g explosives (2 of 4 animals died) and 100% survival at 2 g explosive (~240kPa) Post-mortem examination

of animals that did not survive 2.5 g and 3 g explosives were determined to have died predominantly due to pulmonary haemorrhage Behavioural assessment was conducted

on animals exposed to 2 g explosive without BA for up to 2 weeks to determine if there was any degree of cognitive injury Our results showed that 2 g explosive BOP (around 240kPa) was not enough to induce any significant behavioural changes in the test animals and gross post mortem observations did not detect any obvious lung injury

After understanding the lower limits of blast pressure that did not induce injury, animals were exposed to a higher amount of explosive but with protective BA subsequently (Fig 3b) This BA used was designed to mimic the body armour worn by the soldiers in the deployed scenario The fabricated metal BA protects the animals from possible lung or bodily injury but still allows full blast pressure transmission to the cranial For the main study, the effects of BOP to cause CNS injury was examined using 5 g explosive on animals with full BA in comparison to 2 g explosive on animals

2.2 Modified Open Field Primary Blast Overpressure Model

To further understand the CNS injury profiles at different BOPs, an open field blast injury model was conducted This model allows us to examine a more specific blast overpressure spectrum by varying distance from blast source and pressure characteristics that can result in blast induced brain injury This model was established

at ATREC Pte Ltd, Singapore Explosives tests were conducted in a 10 by 10m concrete and steel reinforced room

2.21 Animal subjects

Male Sprague-Dawley rats (NUS CARE (Singapore) and ARC (Australia) weighing 300-350 grams were used in these experiments For this set of studies, the animals used were housed in two different locations to explore the possibilities of carrying out non-invasive magnetic resonance imaging at Biopolis, A*STAR Hence, for the pilot testing and first two sets of blast experiments, animals used were housed in DSO and thereafter, at A*STAR, Singapore

Experimental procedures were separated into 2 sections; the biomarker group and the behavioral group For the behavioral group, the animals were trained in the selected tasks (except for BWT) till a stable baseline before subjected to the injury procedure Briefly, after 1 week quarantine after arrival at housing facilities, animals were food restricted to 85% of their starting weight to ensure motivation to do the behavioral task The animals were allowed a weekly weight increase of 5% and had full access to water All rats were housed in pairs, separated by a perforated metal divider, under conditions of room temperature on a 12 hrs regular light/dark cycle Animals were trained daily and rested for the weekend until reaching stable baseline Baseline performance was established prior to injury exposure and behavioral testing was carried out after blast exposure until sacrifice at up to 1 month after injury For the biomarker

the housing for 1 week before injury exposure All animal experimentation protocols in this research project were approved by the DSO Institutional Animal Care and Use Committee (protocol number DSO/IACUC/09/74) A schematic of the experiment

procedure is shown below (Fig 4)

Fig 4: Schematic of Open Field experimental blast model for Biomarker group (left) and Behavioural group (right)

2.22 Experimental Setup

Computer stimulations of the different amount of explosives and the resultant pressure intensity was carried out to estimate and achieve the BOP required as below and the amount of explosive to be used

Injury Distance (m) Pressure (KPa) Positive duration (ms)

The required blast intensity and duration was an extension of the BOPs tested in the shocktube model The modified open field blast model was finalized using 5kg of TNT (with PETN core) at a fixed distance of 2m and 3m away from the explosives The layout of the blast model is shown in Fig 5 and Fig 6 The explosive was placed in the center of the room and was elevated one meter above the ground using a specially fabricated wooden frame (Fig 5) This was done to reduce possible blast pressure reflections from the ground and to permit equal radiation of pressure outwards to the animals Animals were placed in a metal rectangular wire mashed cages (1.5m by 0.8m

by 0.5m LWH) (Fig 6) at predetermined distances of either 2m or 3m from the explosive The wire mesh cages were placed in four quadrants relative to the explosive and diagonally opposite to each other for each distance The four cages were elevated similarly at 1 m above the ground by metal rack stands and were connected in series with a connecting beam secured to the rack stand This arrangement not only prevented the cages from moving during the blast and also ensured consistency between each blast test

Fig 5: Photograph of the experimental setup The 4 animal cages were placed in the four quadrants facing the 5kg TNT (with PETN booster) explosive at the epicentre The cages were secured to a 1 m high supporting rack and the 4 racks were linked up by a connecting beam to improve stability and structural integrity

Fig 6: A close-up view of the wire mesh animal cage The pressure sensor was placed at the same height and distance with the animals for accurate blast pressure measurements

Fig 6 Supporting Rack

2.23 Pilot test

A pilot study was conducted to establish the experimental protocols of rodent blast injury model and the resultant mortality of the animals when exposed to the predetermined blast intensity Animals with body armour will survive the blast exposure

at both mild and moderate pressure Animals at mild blast pressure had a 90% survival rate and none of animals suffered any adverse effects or injury or burns With the establishment of the mortality rate and calibration of the required blast distances, subsequent blast experiments were then conducted to investigate the degree of BINT in animals model with and without body protection

2.24 Injury induction - Blast exposure

Under continuing anesthesia (ketamine/xylazine 75mg/10mg/kg at 0.2ml/100g

given ip), rats were subjected to explosive BOP with/without full body protection with

just head exposure Full body protection consists of a customized acrylic holder that covers the entire thoracic and lower body of the rodent whilst leaving the head exposed (Fig 7a-b) The customized acrylic body armour consists mainly of a cylindrical holder to contain the body (75mm in diameter, separated in two pieces) with a cone shape head opening (with a 40mm diameter opening, 10mm length) An additional piece of cone shape acrylic prevents head rotation The anaesthetized animal was allowed to lie comfortably inside the holder and the head opening is wide enough for the neck region

to prevent any suffocation Soft sponges were added around the neck region prevent any contact injury to the neck during blast exposure The ears of the animals were taped down with surgical micropore tape to prevent direct blast pressure to the tympanic membrane In addition, aqueous gel (Aquasonic® 100 Ultrasound Gel) was spread onto the facial region, including the whiskers and eye and any other parts of the body that was exposed to prevent any instances of burn injury When the animal were prepared and secured inside the armour, cable ties were used to fully secure the armour casing

as well as for the animal cages (Fig 8) A total of 5 blast experiments were carried out

Fig 7a-b: Custom-made acrylic “body armour” to shield and protect thoracic region of the animals

Fig 8: Picture showing the animals secured inside the body armour and tied down to the wire mesh cages The exposed heads were covered with adequate amount of aqueous gel to prevent burn injury

2.3 General Experimental Procedures

2.31 Behavioural Testing

A series of neurobehavioral tests that were reported to be sensitive behavioral measures for TBI were carried out to assess the recovery of the animals from the injury exposure in the aspects of memory, motor coordination and sustained attention The neurobehavioral tasks, RAM (test memory), beam-walking test (test motor coordination) and 5CSRTT (test sustained attention) were carried out as detailed below Baseline performance was established prior to injury exposure and behavioral testing was carried out after blast exposure until sacrifice at up to 1 month after injury

2.311 Radial Arm Maze

The radial arm maze (RAM, Fig 9) is a common tool used to investigate and measure specific aspects of spatial working and reference memory of rodents This task

is based upon the principle that the animals have evolved an optimal strategy to explore their environment using their spatial memory abilities and obtain food with the minimum amount of effort Notably, damage to temporal lobe structures, particularly the hippocampus, normal aging, and a variety of pharmacological agents would cause impairment in spatial memory that could lead to decreased performance on the radial arm maze

Central Arena Reward dish

The RAM (Panlab, Spain) consists of an octagonal central platform with eight automated sliding guillotine doors giving access to eight radiating arms of equal lengths (Wx1690mm, Lx1250mm, Hx1450mm)(Fig 9) The apparatus is made of black plexiglas and mounted on a tripod with adjustable height Each arm has lateral walls higher on the proximal side of the arm than on the distal side In the distal extremity of each arm contains a detachable recessed cup for baiting with food pellet (45mg, raspberry flavored, Test diet) Each arm contains 2 sets of photoelectrical cell mounted

on proximal and distal end of the arm to differentiate between arm entries and visits Automation of the RAM by the MazeSoft software (Panlab, Spain) allowed; (i) the location of the rat detected by the photoelectrical cells and visualized on the computer screen, (ii) controlling the opening and closing of the guillotine doors to each radial arms Prominent extra-maze visual cues are present to allow spatial recognition of arm position

The experimental protocol of RAM was conducted in three main phases; namely: (i) habituation, (ii) training, (iii) actual trials Three days before habituation, animals were deprived of food until their body weight was reduced to 85% of the initial free-feeding weight During habituation phase (span of 2-3 days) the food restricted rats were familiarized to the maze All the maze doors were kept opened and food rewards were scattered around the maze to entice the rats to explore After the rats were able to freely explore the maze and consumed the food rewards at the food disc at the distal arm, the protocol continued into the training phase (span of 10 or more days) The rats were trained to retrieve a food pellet from each selected baited arm only once Here, the rats were placed in the center arena with all doors closed and after 30 s, all eight doors to the radial arms will open and the rats allowed to freely explore until the entrance into one arm was detected Then all doors would close, except corresponding to the arm being visited After the animal turned back to the central area, the open door would then

be closed; all doors would remain closed until the confinement time (5s) had elapsed and then reopened again for a new choice This cycle of events was repeated

number of visits to each arm were automatically recorded by the MazeSoft software After each run, the maze was cleaned with absorbing paper to prevent a bias due to olfactory intra-maze cues Over time, the rats would also learn that certain arms were not baited (i.e reference memory) and avoid them accordingly Once the rats achieved

a stable baseline performance (>85% accuracy) over a three-day period, they would be induced with injury or sham-operated, then tested again at 72 h after injury to evaluate spatial memory changes

The confinement time was a critical feature of the maze in that they restricted the rat to the central platform area between choices for a short period, hence preventing the animal from developing a biased response habit For example, without temporary confinement between each arm choice, the rat could successfully solve this task by simply always turning right/left after each choice and entering the first arm away from the previously chosen one This simple strategy does not require an accurate knowledge of the spatial environment or spatial memory, and would give a biased response pattern Unless the investigator is primarily interested in studying response patterns, it is best to have the rat confined to the central platform prior to making each arm choice The current protocol defined a five seconds delay during confinement of the rats to central arena This would increase the difficulty of the rats to remember which previous arms it had visit (hence working memory errors) and increase the duration required for training

From the experimental trials, the following data may be obtained for each rat for data analysis and interpretation:

i Number of revisit to baited arms: A working memory error used to indicate spatial working memory performance

ii Number of visit to non-baited arms: A reference memory error used to indicate spatial reference memory performance

iii Latency to retrieve all baited arms: A measure of the level of motivation of the rat

on the RAM task

2.312 Beam-walking Test (BWT)

The beam-walking task (BWT) is a commonly used test procedure for the assessment of balance and coordination The traversing of the narrow beam by the rodent subject involves both central and peripheral neural processes for proper integration of sensory inputs (i.e proprioception & balance) and the subsequent elicitation of motor adjustment (i.e muscle tone & limb movement) As compared to the rotarod (also used for measurement of motor coordination & balance), the BWT may be potentially less stressful for the rodent subject since negative reinforcement (i.e foot shock) is not actively used during task training Subjects are also generally able to master the beam walking task in a shorter time (i.e within a day), as compared to the long training period required to achieve baseline performance on the rotarod

Fig 10: Laboratory set-up of Beam Walk Apparatus

The apparatus used in the beam-walking test is very basic It consists mainly of a narrow flat beam (2.5cm wide) leading to a brightly decorated goal-box Foam cushions are placed below the beam to cushion any accidental fall The conduct of the beam walking test consists of three phases; namely: 1) Habituation, 2) Progressive beam training, and 3) Actual trials In the habituation phase, the food-restricted rat is first placed into the goal-box to feed on the rat chow provided Subsequently, progressive beam training is initiated, starting from the beam-end closest to the goal-box (Point A) and progressing gradually to the opposite end (Point C) (Fig 10) A neurologically intact

2.313 Five-Choice Serial Reaction Time Task

The Five-Choice Serial Reaction Time Task (5CSRTT) was originally developed for use with rats by Trevor Robbins’ group in Cambridge to specify the psychology and underlying neurobiology of attentional processes More recently, it has also been used

to study impulse control The task requires the detection of brief flashes of light presented pseudo-randomly across five spatial locations Detection is signaled by a nose-poke response from the rat Stimuli are presented rapidly in multiple trials, and hence the task does have some degree of analogy with continuous performance tests used in humans (Leonard, 1959; Wilkinson, 1963) As has been discussed in many reviews (Bushnell, 1998; Robbins, 2002), manipulations of the basic task parameters provide relatively independent behavioral indices of dissociable aspects of attention, impulse control, and even, compulsive behavior

The apparatus used for the 5CSRTT consists of an operant chamber placed within a sound-attenuating box Within the chamber, five niches are found on one side and each of these niches contains a light emitting diode (i.e for stimulus presentation) and an infrared motion sensor On the opposite side of the chamber, a food magazine linked to a food pellet dispenser is installed for implementation of positive reinforcement

to correct responses (Fig 11)

Fig 11: Interior view of the five-choice chamber

The conduct of the 5CSRTT for assessment of sustained attention consists of three main phases; namely: 1) Response Training, 2) Progressive 5CSRTT Training, and 3) Actual trials A food restriction protocol is also in place to ensure motivation on the task and appropriate nutrition of the rats Briefly, in the “Response Training” phase, the food-restricted subjects are placed into the chamber for habituation to the context and food pellets, and, subsequently, the shaping of the nose-poke and pellet retrieval behavior The subjects are then moved on to the “Progressive 5CSRTT Training” in which a progressive training protocol trains the subjects to detect and nose-poke the appropriate niche for food reward with shortening time window over sessions Upon satisfaction of prescribed performance criteria (>60% correct; less than 20% missed; over 3 consecutive days) at stipulated test parameters (1s stimulus duration; 5s limited hold), the subjects may be used to evaluate the test substances Retest reliability is superb for this test but residual effect (i.e crossover effect) of agent used should be considered

5 niches with LED & infrared sensors Food Magazine (ie for reward)

From the automated data acquisition system, a comprehensive set of data from each test session is derived as follows:

i Correct trials

ii Incorrect trials

iii Missed trials

iv Total trials

v Premature responses

vi Latencies

2.32 Histopathology

2.321 Paraffin fixed samples

Animals were sacrificed by an overdose of pentobarbital (KI) or ketamine+xylazine mixture (DSO) and transcardially perfused with Hartman's solution, followed by fixative containing 4% formaldehyde and in phosphate buffer (pH 7.0) The brain, including the olfactory blub, spinal cord, cerebellum and left lung were dissected out, and further post fixed for 2-3 hrs in 10% buffered formalin The brains were then dehydrated in an ascending series of alcohol, cleared with xylene, and then embedded

in paraffin wax Paraffin sections of 4µm thickness were then cut and microwaved in citrate buffer for antigen retrieval and blocked with peroxidise blocking reagent (S2023, DAKO UK Ltd, UK)

2.322 Histopathological examination

A systematic histopathological examination of the various brain structures was conducted For apoptosis staining, sections were stained according to the protocol

provided in the ApopTag® Peroxidase In Situ Apoptosis Detection Kit (S7100,

Chemicon International, Inc, MA USA) For immunohistochemistry, sections were also incubated with mouse monoclonal anti-neuronal nuclei (NeuN) (MAB377, Chemicon International, Inc, MA USA) diluted 1:600 in PBS; rabbit polyclonal anti-glial fibrillary acidic protein (GFAP) (AB5804, Chemicon International, Inc, MA USA) diluted 1:1500 in PBS; ionized calcium binding adaptor molecule (IBA-1) (019-19741, Wako Pure

(AB3594, Chemicon International, Inc, MA USA) diluted 1:100 in PBS; biotinylated Ricinus Communis Agglutinin I (RCA 120) (B-1085, Vector Laboratories) diluted 1:1000

in PBS; and rabbit polyclonal Amyloid β precursor protein (APP) (AB17467, Abcam) diluted 1:100 in PBS; for detection of NeuN, GFAP, IBA-1, AQP-4, lectin and APP, respectively Subsequent antibody detection was carried out using either anti-mouse (rat adsorbed) or anti-rabbit IgG (ImmPRESS Ig reagent kit, Vector Laboratories) except for lectin which was carried out using horseradish peroxidase streptavidin (SA-

5004, Vector Laboratories) Staining and their specific uses were stated in table 4 All samples were then visualised using 3,3’-diaminobenzidine (DAB) and examined under

a light microscope (Olympus, Japan) For fluorescence staining, FITC-conjugated and Cy3-conjugated secondary antibodies were used instead Fluorescence images representing at least one brain each from three rats at different time points were captured under a confocal microscope (Olympus Fluoview TM1000)

Table 4

ApopTag® Peroxidase Detects apoptotic cells in situ by labeling and detecting

DNA strand breaks by the TUNEL method

types including cerebellum, cerebral cortex, hippocampus, thalamus, spinal cord and neurons in the PNS

fluorescein, and is used to label degenerating neurons in

ex vivo tissue of the central nervous system

Glial fibrillary acidic protein

(GFAP)

GFAP is the main constituent of intermediate filaments in astrocytes and serves as a cell specific marker that distinguishes differentiated astrocytes from other glial cells

Ionized calcium-binding

adapter molecule 1 (IBA-1)

A calcium-binding protein that play a role in macrophage activation and function and is a marker of microglia

mediated water flux represents an integral element of brain volume and ion homeostasis

Amyloid precursor protein

(APP)

APP is the precursor molecule whose proteolysis generates beta amyloid (Aβ), that is the primary component of amyloid plaques found in the brains of Alzheimer's disease patients

2.34 Blood gas analysis – iSTAT

Blood was withdrawn as above and blood parameters including pO2, pH, pCO2, TCO2, lactate and metabolic status were measured by i-STAT Portable Clinical Analyzer (Abbott Laboratories Inc., New Jersey, USA) Blood was collected in heperainized syringe (BD A-Line 3ml ABG ST) and transferred to iSTAT cartages CG4+ and Chem 8+

2.35 Serum Cytokines

Serum cytokines that include Interleukin (IL)-1α, IL-β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12p70, IL-13, granulocyte-macrophage colony-stimulating factor (GM-CSF), and interferon-gamma (INF-γ) and necrosis factor-alpha (TNF-α) levels were measured using

a Bio-Plex system (-Plex Pro™ rat cytokine Th1/Th2 12-plex immunoassay kit

#171-K1002M; Bio-Rad Laboratories, Inc., USA) 4 to 6 samples per group were used

2.36 Serum NR2 antibody

Serum NR2 antibody level was detected using Gold Dot NR2 from CIS Biotech (US) The Gold Dot NR2 Antibody assay is a serological enzyme-linked immunosorbent assay (ELISA) for the quantitative determination of antibodies to the NR2 subunit of human NMDA glutamate receptor in serum High levels of NR2 antibodies will indicate occurrences of brain stroke or ischemic events

2.37 Fresh frozen tissue samples

Animals were sacrificed with an overdose of ketamine+xylazine mixture A midline incision from the abdomen region to the sternum was made and than a further

“V” shape cut was made from the sternum to expose the cardiac cavity, exposing the heart A cardiac puncture was carry out to withdraw blood into vacutainers and processed for serum or for blood gas analysis The brain including spinal cord and olfactory bulb were collected The brain tissue was further divided into different regions, including frontal cortex, mid cortex, hippocampus, and cerebellum Samples were immediately quick frozen in liquid nitrogen Frozen samples were stored in -80oC until use for protein extraction for western blot and ELISA

2.38 Protein extraction

Brain tissues including mid-cortex and hippocampus were homogenized in tissue protein extraction reagent (T-PER) (Thermo Scientific) for total proteins extraction Tissues were cut into average weight of 300mg and placed into 2ml microcentrifuge tube (Qiagen) with 0.7ml of T-PER solution and a metal bead The tissues tubes were then homogenized using Tissue Lyser II (Qiagen) for 2mins @ 20

Hz Lyses samples were then centrifuge at 10 000xG for 5 min in 4 oC The supernatant were than collected and protein concentrations were determined using Bradford method and measured at 595 nm Further dilution was done with T-PER to

obtain final concentration of 3µg/ml Samples were then stored in -80oCuntil use for ELISA and multiplex analysis

2.39 Tissue Specific Inflammatory cytokines assay

The relative concentrations of 12 pro-inflammatory cytokines to control samples

in the protein supernatant from the cortex and cerebellum tissue lysate of rats at designated time-points after blast injury were determined with a Rat Inflammatory Cytokines Multi-Analyte ELISArray kit (Mer004A, Qiagen, CA, USA) The tissue homogenates for the ELISArray measurements were prepared as for western blotting and ELISArray measurements were performed according to the manufacturer’s protocol Separately, due to the small volume of hippocampal tissue available, the hippocampal protein supernatant was measure for tumor necrosis factor-alpha (TNF-α), interleukin (IL)-1α, IL-β, IL-2, IL-4, IL-6, IL-10, granulocyte-macrophage colony-stimulating factor (GM-CSF), and interferon-gamma (IFN-γ) levels using the Bio-Plex system (Rat Cytokine TH1/2 12-plex Panel #171-K1002M; Bio-Rad Laboratories, Inc., USA) at 24 h, 72 h, 2 weeks and 1 month post-blast

2.4 Statistics

One-way analysis of variance was carried out to determine changes in systemic

cytokine expression after blast Unpaired Student’s t-test was used to compare

differences between injury groups and sham at various time-points post-blast Data are

expressed as mean ± SEM, where appropriate Significance was accepted at P-value of

less than 0.05

CHAPTER 3 RESULTS

Blast Tube model

3.1 Behavioural Assessment

3.11 Radial Arm Maze– Memory retention

Animals were trained to baseline and before injury did not commit any working memory errors After blast exposure, 5g BA animals showed significantly more working memory errors at 7 days post-blast (Fig 12) In contrast, 2g blast animals showed similar minimal working memory errors compared to sham animals However, 5g blast animals did not show any deficits at 2 weeks post-blast and were similar to 2g blast and

sham animals

Figure 12: Radial Arm Maze: Working memory error of sham (red), 2 g blast (blue) and 5 g explosive with body armour (green) (5 g BA) animals before and up to 2 weeks post-blast Deficit in short term memory can be observed at 1 week post blast for 5g BA animals compared to sham and 2g blast groups

At baseline and before blast exposure, all animals committed not more than one reference memory errors At 5th day after blast exposure, 5g blast animals showed persistent more reference memory errors up to 2 weeks At 5th day and 2 weeks after blast injury, 5g blast animals made significantly more reference memory errors compared to sham and 2 g blast animals respectively (Fig 13) In contrast, both 2g

blast and sham animals made averagely not more than 1 reference memory error after blast.

Fig 13: Radial Arm Maze: Reference memory error of sham (red), 2 g blast (blue) and 5 g BA blast (green) animals before and up to 2 weeks post-blast 5 g BA animals had significantly increased reference memory errors compared to 2 g blast animals at 5 d and 2 weeks post-blast compared to 2g

blast * = p < 0.05 for btwn sham and 2g Blast and # =p<0.05 for versus 2g Blast

5 g BA blast animals were observed to have higher visit response latencies on 4thday after injury as seen in the longer duration to visit the correct arms in the RAM test However, the animals recovered subsequently (Fig 14) In contrast, 2g blast and sham animals showed similar visit response latencies timing

Fig 14: Radial Arm Maze: Visit response latencies of sham (red), 2 g blast (blue) and 5 g BA Blast (green) animals before and up to 2 weeks post-blast Time taken to visit the correct arms were increased at 4

days after blast for 5g BA groups compared to Sham and 2g Blast * = p <0.05

3.12 Beam-Walking Test – Motor Coordination

Trained animals will transverse the beam within the duration of two seconds Motor coordination of the animals was not affected in both 2g and 5g BA animals (Fig 15) after blast exposure compared to sham animals and their baseline

Fig 15: Beam-Walking Test: Time (s) taken to complete beam walk test for sham (green), 2 g blast (blue) and 5 g explosive with body armour (red) animals before and up to 2 weeks post-blast No significant difference was observed between groups

3.13 Five Choice Serial Reaction Time Task– Sustained Attention

Animals for the test were pre-trained to be above 80% for the correct choice made at baseline before blast exposure Hence, any deviation from the baseline values can be attributed to injury Sustained attention for the maximum 2 seconds stimulus was not affected in blast exposed animals as all groups continued to perform above the 80% criteria point up to 2 weeks post-blast (Fig 16) However, due to time limitation, animals exposed to 2 g blast were only tested up to 7 days after injury There was no observable difference in correct choice response latencies between the groups (Fig 17) Interestingly, 2 g blast animals showed increased latency when making a wrong decision (Fig 18), which did not lead to observable changes in the percentages of correct trials There was no observable differences between blast groups and sham animals (n=3) in the premature responses (Fig 19) and preservatives responses (Fig 20)

Fig 16: 5CSRTT: Correct scores (%) at 2 s stimulus for sham (red), 2g Blast (green) and 5g BA (blue) Animals were trained to attain ≥ 80% baseline accuracy No significant difference was found for both 2 g and 5 g BA from baseline scores and sham animals, indicating no change in the attention process

Fig 17: 5CSRTT: Comparison between sham, 2 g blast and 5 g BA No significant difference was found for both 2 g and 5 g BA from baseline scores

Fig 18: 5CSRTT: Comparison between sham, 2 g blast and 5 g BA 2g Blast group showed increased

time taken to make the incorrect decision at 4th and 5th day after injury # = p <0.05 2g blast versus

sham group

Fig 19: 5CSRTT: Premature responses comparison between sham, 2 g blast and 5 g BA No significant difference was found for both 2 g and 5 g BA animals compared to baseline and sham animals up to 2 weeks post-injury

Fig 20: 5SCRTT: Preservative responses comparison between sham, 2 g blast and 5 g BA No significant difference was found for both 2 g and 5 g BA animals compared to their respective baseline

before injury However, sham animals showed transient increase in response at 3 days post-injury * = p

<0.05

3.14 Elevated Plus Maze - Anxiety test

5 g BA animals did not show increased time spent in the closed arms compared

to sham animals 3 days after blast Interestingly, 2 g Blast animals showed increased time spent in the opened arms and vice versa lesser time spent in the closed arms compared to sham animals (Fig 21)

Fig 21: Comparison between sham, 2g blast and 5g BA anxiety levels 2g Blast group showed reduced anxiety level compared to sham and 5g BA groups No difference between 5g BA and sham group

3.2 Blast-induced cytokine changes

In order to determine whether systemic cytokine expression was related to blast severity, multiplex cytokine assay was carried out for IL-1α, IL-2, IL-4, IL-6, IL-10, IFN-γ and TNF-α whilst a single cytokine assay was carried out for CINC-1 The cytokines IL-1β and IL-4 were below the ELISA kits detection pre- and post-blast and were excluded from further analysis Serum test was not carried out for 2 g blast at 3, 24 and 72h time-points as these group were previously intended for behavioral analysis only and logistic

limitation did not allowed for repeating animals test Sham control animals had serum

levels of IL-1α, IL-2, IL-6, IL-10, IFN-γ, TNF-α and CINC-1 of 0.23±0.24, 102.63±20.28, 167.95±27.03, 32.03±6.65, 15.22±3.64, 2.59±0.66, 52.05±19.1 and 93.13±11.00 pg/ml respectively (Fig 22a-h) Systemic IL-1α, IL-6, GM-CSF, IFN-γ and TNF-α were significantly greater than sham levels at 2 wk after 5g blast injury (Fig 22a,c, d and 22e respectively) CINC-1 was significantly elevated at 3 h after 5 g blast injury (Fig 22g) Additionally, systemic TNF-α and GM-CSF levels for 5 g blast animals were significantly increased compared to 2 g blast animals at 1 and 2 weeks post-blast, respectively

(p<0.05; Student’s t-test) (Fig 22f and 22d) Despite apparent increases in IL-1α at 2

week after 2 g blast injury (Fig 22a), these elevated levels were not statistically

after 2 and 5 g BA blast injury (Fig 22h) at 2 weeks, but the expression level was not significantly different from sham

Pro-inflammatory Cytokines

Fig 22a: Serum cytokine IL-1 α levels up to 2 weeks after exposure to varying blast severity; sham (red), 2

g blast (blue), 5 g BA (green) * = p<0.05 5 g BA vs sham, 2 g 1w and 5 g BA 3h, 24h and 72h.# = p<0.05, for 2 g Blast vs sham and 2 g blast 1wk and 5 g BA (3, 24 and 72) Animals after blast showed increased delayed systemic expression of IL-1 α at 2 weeks after blast exposure For 5g BA animals showed gradual increase in IL-1 α expression from 72hrs after blast and peak at 2 weeks