neurotoxic effects of dichlorophenyl methylsulphones related...

Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 858

Neurotoxic Effects ofDichlorophenyl Methylsulphones

Related to Olfactory MucosalLesions

BY

CARINA CARLSSON

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2003

Contents

INTRODUCTION ..........................................................................................1 Background ................................................................................................1 Olfaction an overview ...........................................................................2 Metabolic capacity of the olfactory mucosa and olfactory bulb ................6 Olfactory toxicants .....................................................................................6 Multivariate data analysis in QSAR-modelling .........................................9

OBJECTIVES...............................................................................................10

MATERIALS AND METHODS..................................................................11 Metabolic activation and covalent binding in vitro..................................11 Tissue-localisation of bound substance....................................................12 Histopathology .........................................................................................12 Localisation and quantification of GFAP.................................................13 Behavioural measurements ......................................................................13 QSAR-modelling......................................................................................15

RESULTS AND DISCUSSION...................................................................16 Tissue-binding..........................................................................................16 Lesions in the olfactory mucosa...............................................................21 Neurobehavioural changes .......................................................................25 Effects in the olfactory bulb (GFAP quantification and localisation)......30 Screening for OM toxicity and QSAR .....................................................32

SUMMARY AND CONCLUSIONS ...........................................................36

ACKNOWLEDGEMENTS..........................................................................38

REFERENCES .............................................................................................40

ABBREVIATIONS

2,5-diClPh-MeSO2 2,5-dichlorophenyl methylsulphonea 2,6-diClPh-MeSO2 2,6-dichlorophenyl methylsulphoneb 3-MI 3-methylindole BC basal cell BG Bowman’s gland BSA bovine serum albumine CYP cytochrome P450 DM dorsal meatus EPL external plexiform layer GL glomerular layer GFAP glial fibrillary acidic protein GSH glutathione H&E haematoxylin and eosin IDPN β,β´-iminodiproprionitrile ip intraperitoneal IPL internal plexiform layer LOAEL lowest observed adverse effect level MCL mitral cell layer MWM Morris water maze OB olfactory bulb OE olfactory neuroepithelium OM olfactory mucosa ON olfactory neuron ONL olfactory nerve layer OR odorant receptor PAS periodic acid Schiff PCA principal component analysis PLS partial least squares projection to latent

structures PVL periventricular layer SC sustentacular cell QSAR quantitative structure-activity

relationships a methylsulphonyl-2,6-dichlorobenzene (2,6-(diCl-MeSO2-B) in Paper II. b methylsulphonyl-2,5-dichlorobenzene (2,5-(diCl-MeSO2-B) in Paper II.

1

INTRODUCTION

Background This thesis deals with the highly potent olfactory mucosa (OM) toxicant 2,6-dichlorophenyl methylsulphone (2,6-diClPh-MeSO2) and its non-toxic 2,5-chlorinated isomer. They are derived from chlorinated benzenes, which are widespread environmental pollutants used in the synthesis of pesticides, phenols and dyes, and formed in waste incinerators (1, 2). Both substances bind firmly in the OM and the anterior parts of the olfactory bulb (OB) following a single intraperiotoneal (ip) dose in mice (3, 4). The 2,6-isomer induces OM necrosis with permanent loss of olfactory neuroepithelium (OE) and subepithelial olfactory nerves (ONs) together with a long-lasting increase in glial fibrillary acidic protein (GFAP) in the OB (5). Following shedding of the OE the basement membrane is repopulated by an atypical respiratory-like epithelium. Within a few weeks, bilateral polyps are formed which will partly occlude the nasal lumen. This dramatic metaplasia persists for at least 46 weeks. In contrast, no olfactory toxicity is observed following a high dose of the 2,5-dichlorinated isomer. The cause of this striking difference in olfactory toxicity is currently not understood. The relationships between metabolic activation, covalent binding, and tissue-specific toxicity of these substances were explored in this thesis. Because the OM and bulb are important parts of the sensory system, changes in neurobehavioural performance were also determined. Also other benzene derivatives, sharing the 2,6-positioning of substituents, have been reported to be toxic to the OM (6-8). Therefore the toxicity of some 2,6- or 2,5-substituated benzene derivatives was evaluated. Multivariate QSAR-modelling was applied in order to identify physicochemical characteristics determining olfactory mucosal toxicity and the possibility of predicting olfactory toxicity was investigated.

2

Olfaction an overview A primary function of the sense of smell is to gain information about the environment. Failure to perceive proper olfactory information can be detrimental. Odorants provide cues for the mammalian infant to locate the mother’s nipple, and later in life helps in finding and judging the quality of food. In both predators and prey the ability to locate the other by its smell is essential. Odours are also important in maternal bonding, individual recognition and in social interactions. Sexual behaviour and thus reproductive success is regulated by odours (9).

The nasal cavity is divided by a cartilaginous septum in two equal halves, each containing a dorsal, lateral, and ventral meatus formed by turbinates (Figure 1). The turbinates are lined with respiratory and olfactory epithelium. The olfactory mucosa (OM) is composed of the olfactory neuroepithelium (OE), containing olfactory neurons (ONs), sustentacular cells (SCs), and basal cells (BCs), and the underlying lamina propria (LP) where the mucous producing Bowman’s glands (BGs), blood vessels, and the axon bundles of the ONs are found (10) (Figure 2). The sustentacular, or supporting, cells play a crucial role in regulating the mucus ion composition, and have metabolic capacity (11). The ONs are bipolar cells with dendrites extending to the epithelial surface and with axons collecting in bundles that pass through the cribriform plate on their way to the olfactory bulb (OB). There are a thousand different odorant receptors (ORs) located at cilia on the olfactory dendritic knobs (12). Odorant molecules activate the ORs, which send impulses via the ON axons to the OB where further processing occurs. A topographic representation of the OM exists because each OR is expressed by a subpopulation of ONs located in one of four zones in the OM (13). Moreover, the axons of ONs expressing the same ORs converge at the same glomeruli in the OB (14-16).

3

Figure 1. Structure of the rat nasal cavity, modified from Mery and coworkers (17). Transversal section through the head at level III (18) showing the locations of the meatuses and different epithelia. Abbreviations: et, ethmoturbinate, ns, nasal septum. (1) Dorsal meatus, (2) lateral meatus, (3) ventral meatus. Arrowheads indicate the transition zone between olfactory epithelium and respiratory epithelium.

4

Figure 2. Schematic drawing illustrating the major morphological features of the olfactory mucosa and the olfactory bulb. Abbreviations: ONL, olfactory nerve layer, GL, glomerular layer, EPL, external plexiform layer, MCL, mitral cell layer, IPL, internal plexiform layer, PVL, periventricular layer. (1) Olfactory receptor-containing cilia, (2) olfactory dendritic knob, (3) olfactory neuron cell body, (4) sustentacular cell, (5) basal cell, (6) basement membrane, (7) Bowman’s gland with duct, (8) blood vessel, (9) olfactory nerve bundle, (10) cribriform plate, (11) tufted cell, (12) mitral cell.

5

The OE has a unique feature of constantly replacing old ONs with newly formed ONs. The BCs are considered to give rise to neural progenitor cells that ultimately mature into ONs (reviewed in (19)). This capacity to regenerate new ONs is also retained following chemically induced damage to the OE, axonal nerve transection, and olfactory bulbectomy (20-23). Interestingly, the topographic pattern of the ONs projection onto the OB is restored to normal after direct damage to the OE produced by intranasal irrigation of toxic compounds such as Triton X-100 or inhalation of MeBr, but not after mechanical transection of the ONs. The capacity for neurogenesis, axon growth, and establishment of a functional synaptic connection with the OB is maintained into old age (24).

The OBs are bilateral extensions of the telencephalon and constitute 4% of the brain mass in an adult rat (25). The OB is organised in six layers (Figure 2) with the incoming ONs forming the outermost olfactory nerve layer (ONL) (26). At the glomerular layer (GL) the ONs form tight bundles, glomeruli, where the first synapses occur with the dendrites of mitral and tufted cells (the secondary neurons). The external plexiform layer (EPL) contains the cell bodies of the tufted cells. The mitral cell layer (MCL) contains the cell bodies of the mitral cells, whereas the internal plexiform layer (IPL) contains the collaterals of the tufted and mitral cells. There is also a deep periventricular layer (PVL) containing granule cells. The tufted and mitral cells synapse onto dendrites of pyramidal cells in the olfactory cortex. The olfactory cortex projects to a number of other brain structures, i.e. the neocortex, thalamus, hypothalamus, hippocampus, and the deep nuclei of the amygdala (26-28).

There are two types of glia in the OB, ensheathing glia and astrocytes. In the developing brain the astrocytes secrete molecules that will help guiding the growing olfactory axons into formation of the glomeruli (29). Another important function of the astrocytes is to degrade released neurotransmitters (30, 31).

6

Metabolic capacity of the olfactory mucosa and olfactory bulb There is a wide range of xenobiotic metabolising enzymes present in the OM. The overall cytochrome P450 (CYP) concentration in the OM is about seven times higher than in the nearby respiratory mucosa and some CYP-activities are even higher in the OM than the liver (32). This makes the OM a potential target for toxicity mediated by chemicals that are activated by metabolism. The most important sites of expression of a number of xenobiotic metabolising enzymes in the OM are the BGs and the SCs (33). Several CYPs have been identified in the OM of mammals with the major forms belonging to the CYP2 family. One of the major CYPs in the rat OM is CYP2A3 (orthologous to mouse CYP2A5 and human CYP2A6) (34, 35). It is expressed at much higher levels in the OM than in the liver. In addition, CYP2G1 is OM specific (36). In rat nasal tissue many CYP-enzymes seem to resist induction by xenobiotics, whereas CYP1A1/1A2 and CYP2E1 could be slightly induced (2-fold or 6-fold, respectively) compared to the liver (37). Other enzymes, such as carboxyl esterase and the Phase II enzymes epoxide hydrolase, glutathione S-transferases, and an olfactory-specific UDP-glucuronosyltransferase have been identified in the OM (33, 37).

The OB is via the ONs in direct contact with the external environment, allowing organic compounds, metals and endogenous substances to be transported via the ON axons into the OB (38-40) and thus circumvent the blood-brain-barrier. The presence of xenobiotic metabolising enzymes in the OB could protect secondary and tertiary neurons from toxic chemicals. Indeed, some CYPs have been identified also in the OB. Among them CYP2D4 is inducible (10-fold) in the granular neurons by the neuroleptic drug clozapine and the neurotoxic solvent toluene (41), whereas CYP4A is constitutively expressed. In addition, CYP 2B1/2 have been localised to neurons and astrocytes in the glomerular and granular cell layers (42). Moreover, cytochrome P450 oxido-reductase immunoreactivity is localised to neurons in the EPL, GL, and the MCL (43).

Olfactory toxicants In addition to 2,6-diClPh-MeSO2, 2,6-dichlorobenzonitrile (dichlobenil), methimazole, β,β´-iminodiproprionitrile (IDPN), and 3-methylindole (3-MI) are examples of substances that cause olfactory toxicity after non-inhalation routes of administration.

7

The herbicide dichlobenil induces a regio-specific toxicity, preferentially in the dorsomedial region of the OM in mice and rats (44). The mechanism of dichlobenil-induced toxicity has been extensively studied. Dichlobenil is rapidly metabolised to a reactive intermediate by CYP2A5 and CYP2G1 in the mouse OM (45). The metabolite is covalently bound to the BGs (44), which are also the site of the initial necrosis. Ultrastructural changes in the BGs are observed as early as 1 h following a single ip dose (46). The antithyroid drug methimazole is another compound that causes OM degeneration (47, 48). At lower doses the lesions are preferentially localised to the dorsal meatus (DM), whereas at higher doses nearly complete destruction of the OM can occur. Similarly to dichlobenil, methimazole is firmly bound in the BGs (49), where necrosis is first observed. 2,6-diClPh-MeSO2 and 2,5-dichlorophenyl methylsulphone, (2,5-diClPh-MeSO2) (Figure 3) accumulate in a similar way as dichlobenil in the OM, but also in the OB (3). 2,6-diClPh-MeSO2 binds firmly to the BGs whereas there are no reports on the cellular sites of binding in the OM for 2,5-diClPh-MeSO2. The 2,6-chlorinated isomer induces necrosis in the BGs following a single ip dose as low as 4 mg/kg. In addition, tissue binding and toxicity increases by glutathione (GSH) depletion, and decreases by metyrapone pre-treatment (4). These findings support a metabolism-catalysed mode of action. Secondary to the BG necrosis, degeneration and detachment of the OE occurs leaving a denuded basement membrane. Within a week the basement membrane is repopulated by a ciliated respiratory-like epithelium (5). Subepithelial fibrotic masses will develop and rupture the epithelium at the roof of the DM and fuse with the nearby ethmoturbinates. Fibrous polyps will be formed, followed by cartilage and bone metaplasia in the fibrotic masses. This severe tissue remodelling is most pronounced in the DM of the OM and persists for at least 46 weeks after a single dose. In addition to the initial mucosal lesion the 2,6-chlorinated isomer increases GFAP levels in the OB (5). In contrast, the 2,5-chlorinated isomer remains non-toxic following a single ip dose as high as 130 mg/kg, and there is no increase of GFAP. The cause of this striking difference in toxicity and the nature of the firm binding of 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2 in the mouse OM are currently not known.

Other compounds containing a 2,6-substituted benzene structure have been reported to induce tumours following metabolic activation in rodent OM, i.e. the herbicides alachlor, (2-chloro-2’-6’-diethyl-N-(2-methoxymethyl)-acetanilide) (50, 51), acetochlor, (2-chloro-N-(ethylmethyl)-N-(2-ethyl-6-methyl-phenyl)-acetamid) (52), and the drug metabolite xylidine (2,6-dimethylaniline) (53-55). As proposed by Bahrami et al (3), these findings suggest that the 2,6-positioning of substituents and

8

an electron-withdrawing group in the primary position of the benzene ring may be an arrangement that facilitates olfactory mucosal toxicity.

S OOCl Cl

Cl

ClO OS

I II

CH3 CH3

Figure 3. Chemical structures of (I) 2,6-dichlorophenyl methylsulphone (2,6-diClPh-MeSO2) and (II) 2,5-dichlorophenyl methylsulphone (2,5-diClPh-MeSO2).

As reported in the literature, OM degeneration can be associated with deficits in olfactory dependent behaviour. 3-MI treated rats failed to learn an olfactory conditioning test but were capable of learning a passive avoidance test demonstrating that the learning ability remained intact. This behavioural impairment was ascribed to loss of olfaction (56). Similarly, dichlobenil, IDPN, and methimazole treated rats showed a transient olfactory deficit assessed in the buried pellet test that was not associated with deficits in the hippocampus-dependent Morris water maze (MWM) (57). In the same study, the degree of mucosal damage was correlated with the functional impairment in the buried pellet test.

9

Multivariate data analysis in QSAR-modelling Projection methods such as principal component analysis (PCA) and partial least squares projection to latent structures (PLS) are useful tools in several fields of research were multivariate data needs to be analysed simultaneously. Quantitative structure-activity relationship (QSAR) is one such application (58). QSAR analysis aims to identify physicochemical properties which determine the biological or pharmacological activity of a class of chemicals and to calculate a mathematical model to forecast such activity of other chemicals before they have been tested experimentally. A PLS model is calculated from physical and chemical descriptors (X-matrice) and the observed biological responses (Y-matrice) for a training set of compounds. The PLS model can then be used to predict Y from X for new observations. A fundamental assumption in QSAR modelling is that the observed biological response is generated by the same mechanism for all test-chemicals.

10

OBJECTIVES

The overall objective of this study was to clarify the mechanism behind the differential toxicity of 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2 in the OM in mice. Another goal was to investigate whether the two isomers can give rise to behavioural impairments in mice and rats. Finally, I wanted to identify which chemical properties of the molecule determine OM toxicity. In order to meet these objectives, the following studies were performed.

I The ability of rat olfactory and liver microsomes to metabolise 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2 to covalently bound protein adducts was compared. The cellular sites of covalent binding of 2,6-diClPh-MeSO2 in the nasal region of rats were investigated by light microscopy autoradiography in Paper I. The ability of the two isomers to induce olfactory toxicity at a short post-treatment time was also confirmed in rats.

II Neurobehavioural toxicity of the test compounds was

assessed as spontaneous behaviour (SB), radial arm maze learning (RAM) and Morris water maze learning (MWM). In Paper II, SB and RAM were evaluated in mice dosed with 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2. In Paper III the behavioural impairments induced by 2,6-diClPh-MeSO2 were further evaluated in rats and mice of both sexes using the SB, RAM and MWM.

III Histopathology was used to identify novel OM toxicants

and a preliminary quantitative structure-activity relationship (QSAR) model for 2,6-dichlorinated benzene derivatives was developed in Paper IV.

11

MATERIALS AND METHODS

Metabolic activation and covalent binding in vitro The covalent protein binding was determined according to the method of Wallin et al. (59) with a few modifications (Paper I).

Incubations were made in capped tubes containing rat OM or liver microsomes (80 µg protein), a NADPH-generating system, and 2,6-diClPh-14C-MeSO2 (8.8 µM) or 2,5-diClPh-14C-MeSO2 (9.3 µM). Effects of some CYP substrates / inhibitors and GSH on the covalent binding of 2,6-diClPh-14C-MeSO2 to microsomal protein were also investigated. The substrates / inhibitors were pre-incubated for 10 min in 37 °C before 2,6-diClPh-14C-MeSO2 was added. Vials kept on ice were used as negative controls. The incubations were stopped by transferring 80 µl of the incubation mixture to a glass microfiber filter paper (Whatman GF/C, diameter 25 mm). Each filter paper was extracted in a series of solvents with occasional agitations, once in 95 % ethanol, twice in methanol, twice in acetone and once in n-heptane. This was done to remove unbound parent substance and metabolites formed during the incubations. The filters were then dried at room temperature. Hionic-Fluor™ (Packard instrument Co, USA) was added and the samples were counted in a liquid scintillator analyser (Tri-Carb 1900CA, Packard instrument Co.). The covalent binding was calculated after subtraction of the negative control.

For determination of Km and Vmax the amount of labelled and unlabelled substance was varied. Apparent Km and Vmax values were estimated by fitting the observed velocity (pmol/min/mg protein) versus the substrate concentrations to the Michaelis-Menten equation.

V= Vmax * [S] / Km + [S]

12

Tissue-localisation of bound substance Rats were injected with 2,6-diClPh-14C-MeSO2 and after 24 h the localisation of bound metabolites was investigated using either tape-section autoradiography for whole body distribution, or light microscopy autoradiography for identifying target cells in the OM.

For light microscopy autoradiography transversal sections (5 µm) were taken through levels II, III and IV, according to the system of Young (18). The sections were extracted in a graded series of ethanol and xylene, for removal of unbound substance, and dipped in liquid film emulsion (NTB-2; Kodak, Rochester, NY, USA). After 4-8 weeks of exposure, the film was developed and the sections were stained with toluidine blue or haematoxylin/eosin (H&E) for light microscopy examination.

For tape-section autoradiography whole animals were embedded in aqueous carboxymethyl cellulose, frozen in hexane cooled with solid carbon dioxide and sectioned sagitally (20 µm) according to (60). Following freeze-drying the sections were pressed on to X-ray film (Structurix, AGFA, Belgium). Following 8-12 weeks of exposure (–20 °C), the autoradiograms were developed.

Histopathology The entire nasal region was dissected, gently flushed through the nasopharyngeal duct and fixed in 4 % phosphate-buffered formaldehyde (pH = 7.4). The fixed specimens were decalcified in a formaldehyde solution (4%) containing Na2-EDTA (5.5%) before being embedded in paraffin or Historesin (Kulzer Histo-Technik, Germany). Transversal sections were cut through levels II, III and IV, according to the system of Young (18) and stained with toluidine blue, H&E or periodic acid Schiff (PAS). The sections were examined using a Leica (DM RXE) light microscope and photographs were taken with a digital camera (Hamamatsu, Japan) and processed in Adobe Photoshop 6.0.

13

Localisation and quantification of GFAP The localisation of GFAP was investigated in longitudinal sections (5µm) of the OB using the Sigma IMMH-6 kit (Biotinylated Goat anti-Rabbit IgG) (Paper II). Endogenous peroxidase activity was removed by incubating the sections in 3% H2O2. Non-specific background staining was blocked with 5% bovine serum albumine (BSA). The sections were incubated with the undiluted Sigma IMMH-6 kit primary antibody (rabbit anti-GFAP), for 1 hour and were then incubated with the undiluted secondary antibody (biotinylated goat anti-rabbit immunoglobulin) for 20 min. The antigen-antibody complex was conjugated with ExtrAvidin-conjugated peroxidase for 20 min and was visualised with 3-amino-9-ethylcarbazole (AEC).

GFAP levels in the OB were quantified using a sandwich enzyme-linked

immunosorbent assay (ELISA) method adapted from (61) (Paper III). Microtiter plates (Costar, Cambridge, MA) were coated with a polyclonal rabbit anti-cow GFAP antibody (5 µg/ml). Unspecific binding was blocked with 1 % BSA in phosphate buffered saline. The wells were emptied but not rinsed and the homogenates and the GFAP-standards (0-300 ng/ml in phosphate buffered saline containing 0.05% Tween-20) were incubated at room temperature for 1 h. The wells were washed and incubated for 1 h at room temperature with a monoclonal mouse anti-GFAP antibody (40 ng/ml), washed, and incubated for 30 min at 37 °C with a horseradish peroxidase-conjugated goat anti-mouse antibody (150 ng/ml). The plates were washed and incubated with 1,2-phenylenediamine dihydrochloride (0.4 mg/ml in 0.1 M citrate/phosphate buffer, pH 5.0) for 15 min in the dark. The reaction was stopped with 1 M H2SO4, and the absorbance was determined at 490 nm in a spectrophotometer (Bio-Rad microplate reader, Model 3550). The linear part of the calibration curve was used for titer calculations (2.5-120 ng GFAP/ml). Total protein in the homogenates was quantified using a fluorescence-based assay (62) with BSA as the standard.

Behavioural measurements Detailed descriptions of each test are found in Paper II and Paper III.

The spontaneous behaviour (SB) test is a motor activity test that measures

the ability to habituate to a new environment. Habituation is defined as a decrease in the behavioural variables in response to the decreased novelty of the test cage environment. According to the procedures described in (63-65), spontaneous motor activity was measured by an automated device,

14

consisting of macrolon rodent test cages (40 x 25 x 15 cm) each placed within two series of infrared beams (high and low levels) (RAT-O-MATIC, ADEA Elektronic AB, Uppsala, Sweden). The variables locomotion, rearing, and total activity, were measured over three consecutive 20-min periods. The motor activity test room, in which all 12 ADEA activity test chambers were placed (each identical to the home cage), was well secluded, and used only for this purpose. Each test chamber was placed in a soundproof wooden box with 12 cm thick walls, front panels, and day lighting.

The radial arm maze (RAM) test measures hippocampus-dependent spatial learning ability and working memory, which is influenced by the ability of the animal to collect preferentially visuo-spatial information. However, in the absence of external visual cues or sufficient lightning, accurate arm choice can be guided by information provided by intramaze olfactory cues (66-74). The RAM- procedure was modified and adapted to evaluate maze-learning performance in mice (75). The 8 arms of the maze extended radially from a central hub. The maze was placed on a table in the centre of the test room. At testing, all mice/rats were placed on a total food deprivation for 24 hours. For the learning trials, a regular food pellet was placed at the extremity of each arm. At the start of each test trial, the mouse/rat was placed in the central hub and then monitored for its instrumental learning performance, i.e., the latency until all 8 pellets were collected and the number of revisiting errors per animals. Within each arm, photocell beams registered every arm entrance. An upper time limit of 10 min was applied throughout. The maze-learning test room was secluded, without any explicitly arranged extra-maze cues. Each mouse/rat was tested for one trial only, on each of three consecutive days. Each animal was observed carefully to ensure that each pellet was consumed.

As opposing to the RAM-test, the Morris water maze (MWM) task (76) is a visual guided hippocampus-dependent spatial working memory task that is not considered to be dependent on olfactory information (77, 78). The water maze was a circular, thermostatically controlled (25°C) bath. A platform was placed at a constant position, at 1 cm below the water level. Before testing, all rats were allowed to swim for 60 s with the platform removed. After the platform had been replaced at a set position in the pool, each animal was given one test session on each of 4 consecutive days. At each test session, five trials were presented to the rat. For all trials, the rat was placed at the same point in the pool and allowed to swim around to find the submerged platform and escape from the water. The location of the platform remained

15

constant throughout the first 3 days but was on the fourth day moved to the opposite quadrant of the pool for reversal trials. The spatial acquisition was measured as the decreased latency to find the platform over training sessions.

QSAR-modelling A QSAR model for olfactory toxicity induced by 2,6-dichlorobenzene derivatives was built by using the partial least squares projections to latent structures (PLS) method (79). The commercially available software Simca v9.0 (Umetrics Inc., Umeå, Sweden) was used for the calculations. A total of 31 physicochemical descriptors for each compound were calculated using the HyperChem v5.11 molecular modelling software with the ChemPlus add-on (Hypercube Inc., Gainesville, Florida, USA) or derived from the literature (80, 81) and were used as X-data in the PLS. We needed to establish specific threshold doses for the olfactory toxicity of all 2,6-substited compounds. The lowest dose, for each compound, causing BG degeneration/necrosis either observed in this study or obtained from the literature (3, 7, 44) was defined as the lowest observed adverse effect level (LOAEL). The non-toxic compounds were assigned the highest dose tested (1.16 mmol/kg) as the LOAEL. The LOAEL was converted to the negative logarithm of the molar dose (mol/kg) and was used as Y-data in the PLS. To investigate the validity of the calculated PLS model, internal validation was used (79). The cross-validated predicted variance Q2 gives a measure of the model’s prediction quality, and the quantitative measure of the explained variation or the “goodness of fit” of the modelled Y matrix is expressed as R2Y. Additionally, response permutation was used to check the degree of over fit for the model and the statistical significance of the PLS modelled R2Y and Q2 values (82).

16

RESULTS AND DISCUSSION

Tissue-binding Previous studies have revealed a differential toxicity of the 2,6- and 2,5-chlorinated isomers of dichlorophenyl methylsulphone in the OM of mice, despite their similar tissue binding recorded by tape-section autoradiography. To explore whether a differential metabolic activation could explain this isomer-specific toxicity, the ability of OM microsomes to metabolise and covalently bind the two isomers was investigated. Because a larger amount of microsomes can be obtained from the rat OM, we used rats in this study. For comparison with previously reported results in mice we performed a tape-section (Paper III) and a light-microscopy (Paper I) autoradiography study on the 2,6-isomer in rats. Phorone pre-treatment was used in order to reduce the GSH concentration in the OM, thus reducing GSH-conjugation of the putative reactive intermediate. Phorone reduces the concentration of GSH in several organs, in mice and rats, by forming adducts with GSH (83). Furthermore, phorone does not interact with the microsomal mixed-function oxidase system or with GSH-dependent enzymes (84).

As shown in Figure 4, there was a marked dose-dependent covalent binding of 2,6-diClPh-14C-MeSO2 to microsomal proteins of the rat OM (Paper I). The apparent Km and Vmax values were calculated to 4.7 mM and 38.5 pmol/min/mg protein, respectively. In contrast, the covalent binding of 2,5-diClPh-14C- MeSO2 was low. No significant covalent binding to liver microsomes was observed with either substance. Addition of the CYP inhibitors metyrapone and tranylcypromine abolished the covalent binding to olfactory microsomes, confirming a CYP-catalysed metabolic activation.

17

Figure 4. Concentration-dependent covalent binding of 2,6-diClPh-14C-MeSO2 and 2,5-diClPh-14C-MeSO2 to rat olfactory microsomes. The incubations were performed with a NADPH-generating system at 37°C for 30 min.

Addition of dichlobenil reduced the covalent binding of 2,6-diClPh-MeSO2 to olfactory microsomes by 68 %, suggesting that common CYP forms were involved in the bioactivation of these toxicants. CYPs known to metabolise dichlobenil are CYP2A5 and the olfactory specific CYP2G1(45).

The tape-section autoradiography results in rats were in concordance with previously results in female mice (3, 4). There were no differences in the gross distribution pattern in male and female rats (Paper III). Twenty-four hours after an ip injection of 2,6-diClPh-14C-MeSO2 there was a high uptake of radioactivity in the OM and the lateral nasal glands (Figure 5). A distinct uptake of radioactivity was also observed in the peripheral region of the OB, considerably exceeding all other regions of the brain.

18

Figure 5. Tape-section autoradiogram showing the head of a rat injected with 2,6-diClPh-14C-MeSO2. There is a high uptake of radioactivity in the olfactory mucosa (OM) and the lateral nasal glands (LG). The peripheral regions of the olfactory bulb (OB) also show a distinct labelling.

Light microscopy autoradiography (Figure 6) of the OM revealed a high

level of silver grains confined to the BGs in the LP, whereas no silver grains were observed in the OE after an ip injection of a non-toxic dose (1 mg/kg) of 2,6-diClPh-MeSO2. The highest level of binding occurred in BGs in the dorso-medial part of the olfactory region and at the tips of the ethmoturbinates, whereas the level of binding was lower in BGs in the lateral meatus. Since the tissue sections were stepwise extracted with organic solvents the distribution of silver grains was considered to show the localisation of firmly bound adducts. The results show that the activating CYP enzyme was residing in the BGs. In addition to the BGs, the SCs in the OE are a major site of expression for various CYP enzymes (37). The lack of covalent binding in the SCs could be due to a lack of the activating enzyme in these cells. However, one of the most abundant OM CYPs, CYP2A5, has been localised to both SCs and BGs in the mouse OM by immunostaining (Piras E. and A. Franzén, personal communication). This CYP is involved in the metabolism of dichlobenil and is probably also involved in the metabolic activation of 2,6-diClPh-MeSO2. Another possible explanation is that the bioavailability of 2,6-diClPh-MeSO2 to SCs following blood-borne exposure is low. A low subtoxic dose, as in this case, could be efficiently metabolised in the large mass of BGs in the LP and therefore not reach the SCs in the avascular epithelium (first pass metabolism). In a previous study, a distinct

19

covalent binding of another olfactory mucosal toxicant, phenacetin, was found in both SCs and in BGs in cultured explants of the nasal septum. In contrast, there was a very low binding in the SCs but a high binding in the BGs following injection of phenacetin in vivo (85). The SCs could consequently be a more sensitive target for inhaled than for blood-borne olfactory toxicants.

The GSH content in OE is high in comparison with other extra-hepatic tissues (86) indicating an important role in the detoxification system at this site. Indeed, GSH appeared to play a protective role as a scavenger of the putative reactive 2,6-diClPh-MeSO2 -metabolite because a concentration-dependent reduction of covalent binding to olfactory microsomes was observed following addition of GSH to the incubation medium (Paper I). The observed lack of inhibitory effect of methyl-GSH implies that the thiol-group in GSH is important for the trapping of the 2,6-diClPh-MeSO2 -intermediate. In rats pretreated with the GSH-depleting agent phorone, both tape-section and light microscopy autoradiography revealed a markedly higher level of 2,6-diClPh-MeSO2 binding than in non pre-treated rats (Figure 6). This observation is in agreement with a previous study showing that pretreatment with phorone increases the toxicity of 2,6-diClPh-MeSO2 in the OM of mice (4).

20

Figure 6. Microautoradiograms showing the dorsal meatus of rats killed 24 h after an ip injection of a non-toxic dose of 2,6-diClPh-14C-MeSO2 (1 mg/kg). The binding of 2,6-diClPh-14C-MeSO2 is localised to the Bowman’s glands in the lamina propria, demonstrated by bright field (A and C) and corresponding dark-field (B and D) photomicrographs. In the rat pre-treated with the GSH-depleting agent phorone the binding of 2,6-diClPh-14C-MeSO2 is markedly increased (C and D). (Toluidine blue stained sections, magnification x 200).

21

As opposed to the highly toxic 2,6-dichlorinated isomer, no marked covalent binding of the non-toxic 2,5-diClPh-MeSO2 to rat olfactory microsomes was observed (Figure 4). This finding confirms that this isomer was not metabolised at a high rate to a reactive intermediate by microsomal OM CYP enzymes. Interestingly, previous tape-section autoradiograms revealed a high and selective binding of the 2,5-chlorinated isomer in the OM of mice (3). In that study the uptake of the non-toxic isomer in the OM was comparable to that of the toxic 2,6-chlorinated isomer and did therefore not explain the strikingly different toxicity of the two isomers. The identity of the bound radioactive substance at this site has not been determined. The selective binding of 2,5-diClPh-MeSO2 reported in the OM of mice in vivo could be due to an association of the parent compound with a ligand-binding protein residing in the Bowman's glands (87). This conclusion is supported by our findings that 2,5-diClPh-MeSO2 is not metabolically activated to a significant degree.

Lesions in the olfactory mucosa The toxicity of the two dichlorophenyl methylsulphone isomers was examined in the rat OM at 24h after dosing, representing a short post-treatment time (Paper I). The toxicity of 2,6-diClPh-MeSO2 at two doses (16 mg/kg and 65 mg/kg) was further investigated at 3 weeks, representing a long post-treatment time (Paper III).

22

Where there any differences in 2,6-diClMeSO2-B-derived toxicity in rats and mice? The acute OM toxicity findings in rats injected ip with 2,6-diClPh-MeSO2 (16, 65 mg/kg) were similar to those previously described in mice (3), however less pronounced. Strikingly, at three weeks after dosing there were dramatic differences in the extension and character of the metaplastic changes when comparing mice and rats.

In mice dosed with 2,6-diClPh-MeSO2 (16, 32, 65 mg/kg), severe olfactory mucosal metaplasia and complete disappearance of the BGs in the DM were observed at three and 13 weeks after dosing (Paper II and III). Large fibrous bilateral polyps containing cartilage and bone were filling the dorsal nasal lumen and fusing with nearby ethmoturbinates. These findings confirm previously described metaplastic changes and severe tissue remodelling first observed at one week after dosing and persisting after 46 weeks (5).

In rats, three weeks following dosing with 2,6-diClPh-MeSO2 (16, 65 mg/kg), the OE was disorganised and partly replaced by a respiratory-like epithelium, preferentially in the DM, nasal septal walls, and the tips of the ethmoturbinates in both sexes (Figure 7). Noteworthy, all animals in all dose groups had large areas of intact olfactory epithelium. In the low-dose groups only a few scattered metaplastic cells/animal was observed. The PAS-stained BGs in the LP were reduced in numbers or completely absent at the sites of metaplasia. Moderate fibrosis was observed in the most affected areas, but the axon bundles appeared normal. Cyst-like structures consisting of PAS-positive material growing downward and folding into the LP were frequently observed.

Twenty-four hours following administration of 2,6-diClPh-MeSO2 (16 mg/kg) there was an undulating, necrotic, and partly detached OE in the dorso-medial part of the olfactory region, including the tips of the ethmoturbinates (Paper I). In addition, the BGs in the LP were necrotic and had lost their secretory granules. The axon bundles had a normal morphological appearance. Twenty-four hours after a single high dose (65 mg/kg), the morphological changes were more extensive. The basal membrane was completely denuded or covered with a thin layer of cells in some areas. The BGs were necrotic or had completely disappeared. In contrast, a single ip injection of 2,5-diClPh-MeSO2 (65 mg/kg) did not induce OM lesions, in accordance with previously reported results in mice (3).

23

Were the OM lesions caused by isomer-specific CYP-activation? The results of the covalent binding study imply that the 2,6-diClPh-MeSO2-induced toxicity in the OM indeed resulted from a formation of a CYP-catalysed reactive intermediate, which was covalently bound to the tissue. The immunostaining for CYP2A has been reported to be more extensive in the dorsomedial part of the olfactory region than in the lateral part (37). The cellular sites of mouse CYP2A5 expression in the OM are the BGs and the SCs (Piras, Franzén, personal communication). The colocalisation of CYP2A5 and the 2,6-diClPh-MeSO2-induced acute BG necrosis in the dorsomedial part of the olfactory region suggests that CYP2A5 may be involved in metabolising 2,6-diClPh-MeSO2 to a reactive intermediate (88). The localistion of phase II detoxifying enzymes might also be important. The microsomal epoxide hydrolase is localised in the apical cytoplasm of SCs in all regions of the rat OM except for the DM (89). Glutathione-S-transferases are preferentially localised to the BGs, BCs and SCs in nasal regions covered by OE (90).

There was a discrepancy between the observed localisation of covalent binding in the BGs and the widespread toxic lesion in the OM. As previously suggested for the nasal toxicant dichlobenil (8), this discrepancy implies that a primary necrosis in the BGs is followed by a secondary lesion characterized by degeneration and detachment of the OE.

24

Figure 7. Histological sections showing details of the olfactory mucosa (level III) of rats 3 weeks after injection of corn oil, (A, nasal septum) or 2,6-diClPh-MeSO2, 16 mg/kg (B, nasal septum), 65 mg/kg (C, dorsal meatus, D, ventral ethmoturbinate). A large cyst with PAS-positive contents is seen in the olfactory epithelium in B. Note otherwise largely intact olfactory epithelium in B. In C there is fibrosis and lack of Bowman’s glands in lamina propria. The olfactory epithelium is replaced by a ‘respiratory-like’ epithelium. The ventral ethmoturbinate in D appears intact. OE; olfactory neuroepithelium, RE; respiratory-like epithelium, LP; lamina, BG; Bowman’s gland. Periodic acid Schiff (PAS), magnification x 200.

25

Neurobehavioural changes 2,6-diClPh-MeSO2 is a potent olfactory toxicant capable of inducing permanent loss of OE and subepithelial ONs. Other olfactory toxicants, such as dichlobenil and methimazole, induce neurobehavioural impairments correlated to the degree of OM lesions (57). Therefore we evaluated the SB and the performance in the RAM in mice dosed with 2,6-diClPh-MeSO2 (Paper II). The non-toxic 2,5-dichlorinated isomer was used as a negative control. In Paper III we reinvestigated the 2,6-diClPh-MeSO2-induced behavioural effects comparing rats and mice of both sexes. In order to clarify if 2,6-diClPh-MeSO2 induces a general spatial-learning disability, the learning capacities of the rats were also evaluated in the MWM learning task.

Were there any isomer-, species- or sex-dependent differences in behaviour? In Paper II we found that 2,6-diClPh-MeSO2 gave rise to significant, long-lasting but transient, increases in SB, in all three motor activity variables in mice, at one, 2, and 12 weeks after dosing (32 and 65 mg/kg) (Figure 8). At 12 weeks only the high dose group retained significantly elevated counts in all variables. Unexpectedly, also the non-toxic isomer gave rise to significant increases in SB that lasted for two weeks. Also in the RAM 2,6-diClPh-MeSO2 induced significant, long-lasting but transient impairment of performance. A dose-response relationship was observed at 2 weeks that persisted at 12 weeks (Paper II). 2,5-diClPh-MeSO2 did not induce any impaired performance in the RAM.

As shown in (Paper III) the results of both SB and RAM learning in 2,6-diClPh-MeSO2 dosed (16 and 65 mg/kg) rats were similar as in mice at two weeks after dosing, in both female and male animals. However, in the rats we found a deficit at a lower dose (16 mg/kg) than in mice (Figure 9). The performances of the 2,6-diClPh-MeSO2 treated male and female rats in the MWM, were similar to the performances of their respective control groups at all test days. All groups learned the task of locating the platform within the circular pool from day to day.

Is hyperactivity in concordance with the behaviour of anosmic animals? Anosmia can be induced by intranasal infusion of ZnSO4, resulting in total shedding of the OE (91). In addition, another commonly used method to achieve completely anosmic rats is bilateral removal of the olfactory bulbs

26

(68). Rats rendered anosmic by these methods have been tested for behavioural impairments. ZnSO4 irrigated rats are not hyperactive (91), and one study reports a reduced spontaneous motor activity (66). As a contrast, olfactory bulbectomised rats show hyperactivity among other behavioural impairments. However, this increased motor-activity can be attenuated by traditional tricyclic anti-depressants, as well as by selective serotonin reuptake inhibitors (78). Consequently, a reduced olfactory ability per se is not likely to render the animals hyperactive. Considering this, and the fact that also the “non-toxic” isomer induced hyperactivity (Paper II), it is not likely that loss of smell due to loss of OE was the cause of the observed 2,6-diClPh-MeSO2 induced hyperactivity. Furthermore, mice dosed with 16 mg/kg showed extensive lesions in the OM and no hyperactivity in the spontaneous behaviour test whereas rats dosed with 16 mg/kg were hyperactive although only minor OM lesions were observed (Paper III). These results imply that an unknown direct effect in the brain caused the hyperactivity.

Were the 2,6-diClPh-MeSO2-induced RAM impairments due to loss of olfaction? The performance of anosmic animals as well as the role of olfaction in maze learning has been investigated. The ability of rats to learn to follow intramaze cues (odour elicited by the reward or odour tracks elicited by the rat itself) was assessed in a RAM procedure (70). The results showed that animals rewarded only if they followed intramaze (olfactory) cues performed worse than animals rewarded if they followed extramaze (visual) cues. The authors conclude that intramaze cues appear to be overshadowed by extramaze spatial cues in the conventional RAM-task. More recent reports show that in the absence of external visual cues or sufficient lightning, accurate arm choice can be guided by information provided by intramaze olfactory cues (66-69, 71). One interesting behavioural impairment in olfactory bulbectomised rats is impaired RAM-learning (67). Consequently, the impaired acquisition of the RAM in 2,6-diClPh-MeSO2 dosed mice and rats observed in Paper II and III could be related to the metaplasia of the OE, making them less able to use olfactory cues in their searching strategies. Interestingly, however, the histopathological lesions in the OM of rats were dramatically less severe than those previously observed in mice at the same dosing and post-treatment time. In fact, 3 weeks after dosing, histopathological examination revealed that large areas of olfactory epithelium remained intact, particularly in all rats but also in the mice. From these findings one would not expect these animals to be completely anosmic or to have a severely impaired olfactory perception at the OR level. Only as

27

little as 10 % of intact OE is sufficient for retaining olfactory ability (92). However, due to the zonal expression of ORs, the possibility for partial anosmia for the ORs specifically located in the DM exists. The DM is actually defined as being one of four distinct receptor zones (13). Notably, we failed to demonstrate any 2,6-diClPh-MeSO2 induced reduction in olfactory ability in a simple odour discrimination test (unpublished data). In Paper III low dose rats with minor metaplastic changes in the OM were impaired in the RAM, whereas low dose mice with bilateral polyps and extensive metaplasia where not. These findings does not support that the impairments in the RAM were dependent on loss of olfactory ability due to OM lesions.

28

Figure 8. Spontaneous behaviour at 2 weeks after a single injection of 2,6-diClPh-MeSO2 (32 or 65 mg/kg), 2,5-diClPh-MeSO2 (65 mg/kg) or vehicle in female mice. Locomotion, rearing, and total activity were measured over three consecutive 20 min periods. Statistical analysis was performed using ANOVA with a Split-plot design. Pair wise testing between each of the treatment groups was performed with the Tukey HSD test. Bars represent mean ± SD (n = 12). Letters in upper case p < 0.01 and lower case p < 0.05, where Arepresents comparison versus the control treatment group, Brepresents comparison between the two doses of 2,6-diClPh-MeSO2 treatment and Crepresents comparison between 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2 treatment.

29

Figure 9. Radial arm-maze learning at 3 weeks after a single injection of 2,6-diClPh-MeSO2 (16 or 65 mg/kg) or vehicle in female mice and rats. Numbers of arm reentries (errors), and latency to collect 8 pellets were measured over three consecutive days. Statistical analysis was performed using ANOVA with a Split-plot design. Pair wise testing between each of the treatment groups was performed with the Tukey HSD test. Bars represent mean ± SD (mice, n = 9; rats, n = 8). Letters in upper case p < 0.01 and lower case p < 0.05, where Arepresents comparison versus the control treatment group, Brepresents comparison between the two doses of 2,6-diClPh-MeSO2 treatment.

30

Can hyperactivity affect RAM-learning? A similar behavioural profile as reported in 2,6-diClPh-MeSO2-exposed rats in Paper III was recently described in mice lacking the M1 muscarinic acetylcholine receptor (93). These mice showed a pronounced increase in locomotor activity measured in several tests, as well as an increased number of arm revisiting errors in the RAM. The latencies to retrieve all 8 pellets were however unaffected. Moreover, the performance in MWM was unaffected. The increased arm revisiting errors observed was suggested to be caused by the general hyperactivity associated with the lack of the M1 muscarinic acetylcholine receptor. In the present study hyperactivity was not exclusively correlated with RAM impairments, e.g. the low dose male mice were impaired in the RAM, but showed no statistical significant hyperactivity. Nevertheless, the hyperactivity observed in the SB could contribute to the RAM-deficits in the other dose groups.

Effects in the olfactory bulb (GFAP quantification and localisation) Immunohistochemistry for GFAP in the OB was performed in mice dosed with 2,6-diClPh-MeSO2 or 2,5-diClPh-MeSO2 (Paper II). Also, to further explore any possible species and/or gender differences GFAP was quantified in the OB of 2,6-diClPh-MeSO2 dosed rats and mice of both sexes (Paper III).

Was the GFAP-reaction dependent on the initial olfactory mucosal lesion? GFAP is a sensitive and specific marker for rapid astrocytic response to injury and disease in the CNS. Increased GFAP has been found following cryogenic brain lesions, stab wounds, hypothermia, and toxic lesions (94). Increase in GFAP has also been associated with conditions such as scrapie, Alzheimer’s disease and Creutzfeldt-Jakob disease. The normal content of GFAP in adult brains also increases with age. The reactive astrocytes are thought to be involved in the healing process following injury to the CNS. The proliferation of astrocytes observed after experimental brain injury has been proposed to be a result of mitogens released from vascular sources and damaged tissue and may represent an attempt to compensate for the acute loss of astrocytes (95). Three weeks after dosing (16 and 65 mg/kg), there were significant dose-dependent increases in GFAP levels in the OB (Paper

31

III). The increase was 3-fold in the rat high dose groups and 8-fold in the mice high dose groups. Two weeks following a single ip dose of 2,6-diClPh-MeSO2 (65mg/kg), there was a strong GFAP immunoreactivity in the GL in the mouse OB (Paper II). A low and scattered GFAP immunostaining was also observed in the remaining areas of the OB. Control animals showed only few GFAP immuno-positive cells in the OB. There was no difference observed between the 2,5-diClPh-MeSO2 treated mice and control mice.

As proposed in (5) the GFAP-reaction is likely to be a consequence of the degeneration and subsequent regeneration of the OE. In addition, GFAP-induction in the OB has been observed following treatment with other olfactory toxicants such as dichlobenil, methimazole, and IDPN (49, 96). However, a direct effect in the brain causing GFAP induction can not be excluded because rats showing only minor metaplastic changes had increased GFAP levels. Both 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2 accumulates in the OB providing another possible source of biochemical interaction. Several CYP enzymes (e.g. CYP2B1/2, CYP2D4) are expressed in the OB, and the possibility of a CYP-catalysed activation of 2,6-diClPh-MeSO2 also in the OB can not be ruled out.

Was the behaviour affected by the GFAP-rection in the olfactory bulb? There was a 4-fold increase in GFAP in mice dosed with 16 mg/kg 2,6-diClPh-MeSO2 whereas no hyperactivity in the SB-test was observed at this dose (Paper III). This observation does not support the possibility that the increased motor activity was caused by the GFAP-reaction. The 2,5-diClPh-MeSO2-induced increase in motor activity without increased GFAP also supports this conclusion. The time-course of GFAP-induction caused by 2,6-diClPh-MeSO2 (5) correlates well with the time-course of RAM deficits in mice shown in Paper II. Also, in Paper III all groups that displayed RAM deficits had increased GFAP levels in the OB, supporting a possible causal relationship. GFAP-immunoreactive astrocytes are localised mainly in the GL in the rat OB (23, 97). The OB glomerulus is the site of the first synapse in the olfactory pathway (98). There is clearly a possibility that the GFAP induction (or related biochemical disturbances) could affect the ability of the bulb to correctly process information provided by the ONs, thus leading to alterations in behaviour dependent on olfactory learning or olfactory input.

There is strong evidence that processes associated with olfactory learning takes place in the OB. The OB receives direct cholinergic input from the basal forebrain and the presence of cholinergic receptors (both muscarinic and nicotinic) as well as serotonergic receptors have been demonstrated in several layers of the OB (99, 100). It has been shown that cholinergic afferents play a crucial role in olfactory learning, mainly by activating muscarinic receptors (101, 102). As mentioned, 2,6-diClPh-MeSO2 dosed

32

animals display a similar behavioural profile as mice lacking the M1 muscarinic acetylcholine receptor (93). The importance of acetylcholine in OB synaptic plasticity has also been demonstrated (103, 104). In addition, dopamine receptors are localized to the OM and the OB, where they are believed to play a role in olfactory memory formation (105-107).

Screening for OM toxicity and QSAR A number of 2,6-dichlorinated benzene derivatives and some other structurally related compounds were evaluated for OM toxicity. The BGs are the primary target for metabolism-activated olfactory toxicity and BG degeneration was used as the criterion for olfactory toxicity. The doses administered were equimolar to those known to give BG necrosis in dichlobenil-treated mice. If no toxicity was observed the dose was increased up to a dose equimolar to 200 mg/kg dichlobenil. After 24 h the nose was dissected and subjected to histopathological examination. The observed LOAELs obtained, in addition to the known LOAELs for a number of other 2,6-dichlorinated benzene derivatives, were together with 31 physicochemical descriptors for the substances subjected to multivariate QSAR-modelling.

Were there any novel olfactory toxicants? We identified two novel OM toxicants out of the 20 test chemicals, 2,6-dichlorobenzaldehyde oxime, and 2,6-dichloronitrobenzene (Paper IV). The remaining 18 substances caused no visible changes in the OM.

2,6-dichlorobenzaldehyde oxime: At 0.58 mmol/kg and 0.29 mmol/kg the BGs were necrotic with a reduced or absent PAS staining in all mice (Figure 10 B). Vacuoles were also observed in the LP but no morphological changes occurred in the axon bundles. The OE in the dorso-medial part of the olfactory region including the tips of the middle ethmoturbinates was disorganized with vacuoles in both the basal and the apical parts. At 0.15 mmol/kg no changes was observed.

2,6-dichloronitrobenzene: A dose-dependent necrosis in the BGs was

observed. At 0.58 and 0.29 mmol/kg the BGs glands in all mice were necrotic with a negative PAS-reaction and there were large vacuoles in the LP (Figure 10 C). At 0.15 and 0.07 mmol/kg degenerated, vacuolated or necrotic BGs glands were observed in all mice. At 0.58 mmol/kg the OE in all regions of the sections was detaching and sloughing into the nasal lumen in 2/3 mice. At 0.29 mmol/kg, the OE was slightly disorganised with

33

vacuoles in the apical and basal parts. At 0.15 and 0.07 mmol/kg the OE was apparently intact.

None of the 2,5-dichlorinated compounds were toxic to the OM at the doses tested, supporting the structure-activity relationship proposed by Bahrami et al (3). Figure 10 D shows the apparently normal OM in the DM of a mouse dosed with 2,5-dichloronitrobenzene.

Figure 10. The olfactory mucosa in the dorsal meatus of mice 24 h after a single injection of A) corn oil (control), B) 2,6-dichlorobenzaldehyde oxime (0.58 mmol/kg), C) 2,6-dichloronitrobenzene (0.58 mmol/kg) D) 2,5-dichloronitrobenzene (0.58 mmol/kg). The arrows indicate PAS-positive Bowman’s glands. BM: basement membrane, LP: lamina propria, OE: olfactory neuroepithelium.

What physicochemical descriptors influenced BG necrosis? It has been shown for dichlobenil and 2,6-diClPh-MeSO2 (Paper I) that BG necrosis is the result of a CYP-catalysed (probably CYP2A5) metabolic

34

activation. When building a QSAR model for OM toxicity we have to assume that the BG necrosis induced by the other 2,6-dichlorinated derivatives was also the result of a CYP-catalysed activation.

The PLS model gave two significant principal components (PCs), which explained 72 % (R2Y=0.72) of the variation in the response Y-data (threshold doses for BG necrosis). The first PC explained 50 % while the second PC explained another 22 % of the variation. There was an apparent differentiation between the active and non-active 2,6-chlorinated benzene derivatives. The PLS weight-plot (Figure 11) gives information about how the physicochemical variables combine to form the quantitative relationship of the measured responses. OM toxicity increased with an increase in heat of formation (Hf), molecular dipolar momentum (µ), and the value of the substituent electronic descriptors σm and σp. Negatively correlated descriptors were describing the hydrophobicity of the molecule (π and logP), and electronic properties such as electron affinity (Ea), the energy difference in frontier molecular orbitals (∆E) and the electronic charge on the main substituent carbon (C1).

The Ea and the ∆E are measures of susceptibility towards nucleophilic and electrophilic attack and thereby measures of the molecular reactivity in chemical reactions (108). As the ∆E decreases the molecule becomes more susceptible toward molecular reactions. We can thus presume that the toxicity of a 2,6-dichlorobenzene derivative increases with a decrease in ∆E and Ea. Also, the weight plot suggested that a 2,6-dichlorobenzene derivative with a high hydrophilicity and a high dipolar momentum should be toxic. Further, a substituent group that increases the negative charge at C1 increases the toxicity, supporting that an electron withdrawing substituent in this position facilitates toxicity, as previously suggested (3). In summary, these electronic effects determine the OM toxicity of a 2,6-dichlorinated benzene derivative by improving the reactivity for nucleophilic reactions such as a CYP 2A-catalysed arene oxide formation.

35

Figure 11. PLS weights of the histopathological response. The toxicity descriptor is defined as log (1/LOAEL). The chemical descriptors are named as in Table 3, Paper IV. Examples of descriptors positively correlated with olfactory mucosal toxicity are Hf and σm and negatively correlated descriptors are ∆E and Ea.

Can the olfactory mucosal toxicity of a substance be predicted by its physicochemical properties? The Q2 intercept (-0.02) was well below the limit of 0.05, suggesting that the predictability of the model was within appropriate boundaries (109). The model predicted the OM toxicity of the training set consisting of 17 2,6-dichlorinated benzene derivatives, 6 active and 11 inactive compounds, with a root mean square error of prediction (RMSEP) of 0.29 of the LOAEL (Paper IV). This QSAR model should be considered a preliminary model because the only way to verify the validity of a QSAR model is by testing with an external validation set of additional compounds.

36

SUMMARY AND CONCLUSIONS

There was a marked covalent binding of 2,6-diClPh-14C-MeSO2 to rat microsomal proteins, whereas no significant covalent binding was observed with the 2,5-isomer. The binding was confined to the BGs as shown by light microscopy autoradiography. At a short post-treatment time 2,6-diClPh-MeSO2 induced OM lesions in rats similar to those previously reported in mice, however less extensive. At a longer post-treatment time there were dramatic differences. In rats, patches of the OE in the DM were replaced by an atypical, ciliated, respiratory-like epithelium. In the corresponding mice, there was severe metaplasia accompanied by bilateral large polyps filling the nasal lumen. No olfactory toxicity was observed in 2,5-diClPh-MeSO2 treated rats or mice.

2,6-diClPh-MeSO2 induced dose-dependent, long-lasting, hyperactivity in the SB-test, and acquisition deficits in the RAM-test in mice. At 12 weeks, significantly elevated counts remained in the 2,6-diClPh-MeSO2 high dose group (65 mg/kg). In 2,5-diClPh-MeSO2 treated mice, an unexpected hyperactivity was observed at one and two weeks, but not at 12 weeks. No effect was observed in the RAM-test. At 2 weeks, hyperactivity and RAM acquisition deficits were observed also in 2,6-diClPh-MeSO2-dosed rats. There were no major differences between male and female rats or mice. No effects were observed in the MWM.

The GFAP reaction in the OB previously described in mice was confirmed in mice and rats of both sexes. The lowest dose in this study (16 mg/kg) induced GFAP in both species. The GFAP was localised to the glomerular layer in the OB of 2,6-diClPh-MeSO2-dosed mice.

Two novel OM-toxicants were identified, 2,6-dichlorobenzaldehyde oxime and 2,6-dichloronitrobenzene. A preliminary QSAR model was built based on these and previously reported OM toxicities. Physicochemical properties positively correlated to OM toxicity were identified as: heat of formation, molecular dipolar momentum, and the substituent electronic descriptors. Negatively correlated descriptors were variables describing the hydrophobicity of the molecule and electronic properties such as electron affinity, the energy difference in frontier MO, and the electronic charge on the primary carbon.

37

From these findings the following conclusions could be drawn:

• A CYP-catalysed transformation of 2,6-diClPh-MeSO2 to a reactive intermediate and a subsequent covalent metabolite binding (protein adduct formation) in the BGs underlies the high acute toxicity of 2,6-diClPh-MeSO2 in the OM of rats and mice (Paper I).

• Both 2,6-diClPh-MeSO2 and 2,5-diClPh-MeSO2 are neurotoxic. • The observed increased motor activity was not dependent on the

OM lesion or the GFAP-induction in the OB (Paper II and III). I propose that 2,6-diClPh-MeSO2 caused a direct effect in the brain leading to hyperactivity.

• No exclusive causal mechanism behind the RAM deficits was

identified. I propose that the hyperactivity and a biochemical disturbance in the brain contributed to the impairment in RAM acquisition (Paper II and III).

• There was not one specific physicochemical variable that

selectively separated the toxic from the non-toxic 2,6-dichlorinated benzene derivatives. When it comes to predicting biological effects of chemical substances the entire chemical domain has to be considered. However, a 2,6-dichlorinated benzene derivative with a large, polar, and strong electron withdrawing substituent in the primary position has the potential of being an OM toxicant (Paper IV).

38

ACKNOWLEDGEMENTS

Most of this work was performed at the Dept. of Environmental Toxicology, Uppsala University. I wish to express my sincere gratitude to those who have contributed to this thesis, and especially to:

Ingvar Brandt, my supervisor, for giving me the opportunity to complete my PhD, and for guidance through the work. Eva Brittebo, co-author, for knowing all about nose histology and for putting energy into things. Anna Franzén, co-author, for end-less electron microscopy sessions and joyful company at “scientific meetings”. Fariba Bahrami, co-author, for introducing me to “the nose-project”. Anders Fredriksson, co-author, for all your skill and knowledge in behavioural science. Örjan Lindhe, “room-mate” (should have been co-author), for always helping with whatever needed, for stimulating discussions regarding science, recipes, children, and a lot more… Margareta Mattsson, for technical assistance, and without whom this work would never have been completed. Annette Axén, for patient, repeated, technical assistance with the electron microscope. Anette? This thing doesn’t work?? Åke Bergman, Tatiana Cantillana, Christina Larsson, for synthesis of test-substances, co-authoring, and valuable chemical advice. Mikael Harju and Mats Tysklind, co-authors, for QSAR-modelling and discussions about chemical structures and predictability.

To all present and former workers at Ekotox, for contributing to a crazy and enjoyable environment and athmosphere. I especially thank :Katrin for yor help with some manuscripts and for teaching me about multivariate data analysis!! Ulrika and Per for discussing nose & behaviour problems. Peter Pärt, for taking me on as “ex-jobbare” for a hundred years ago and for getting me started as a PhD-student.

39

Last, but above all, Micke, Erika, Jacob, Lina, and Robert, my beloved family, for just existing, for giving me love and support, and for providing a completely different planet to live on.

This work was supported by the Oscar and Lili Lamm’s Foundation. I thank the Royal Swedish Academy, CF Liljewalchs, Knut & Alice Wallenbergs stiftelse and Stiftelsen Facias, for awarding me travel stipends.

40

REFERENCES

1. Weber, R., et al., Formation of PCDF, PCDD, PCB, and

PCN in de novo synthesis from PAH: mechanistic aspects and correlation to fluidized bed incinerators. Chemosphere, 2001. 44(6): p. 1429-38.

2. Bakke, J.E., et al., Metabolism of the mercapturic acid of 2,4',5-trichlorobiphenyl in rats and mice. Xenobiotica, 1983. 13(10): p. 597-605.

3. Bahrami, F., et al., Localization and comparative toxicity of methylsulphonyl-2,5 and 2,6- dichlorobenzene in the olfactory mucosa in mice. Toxicol Sci, 1999. 49: p. 116-123.

4. Bahrami, F., et al., Target cells for methylsulphonyl-2,6-dichlorobenzene in the olfactory mucosa in mice. Chem Biol Interact, 2000. 128(2): p. 97-113.

5. Bahrami, F., et al., Persistent olfactory mucosal metaplasia and increased olfactory bulb glial fibrillary acidic protein levels following a single dose of methylsulphonyldichlorobenzene in mice: Comparison of the 2,5- and 2,6-dichlorinated isomers. Toxicol Appl Pharmacol, 2000. 162: p. 49-59.

6. Eriksson, C. and E.B. Brittebo, Metabolic activation of the herbicide dichlobenil in the olfactory mucosa of mice and rats. Chem Biol Interact, 1991. 79(2): p. 165-77.

7. Brittebo, E.B., et al., Toxicity of 2,6-dichlorothiobenzamide (chlorthiamid) and 2,6- dichlorobenzamide in the olfactory nasal mucosa of mice. Fundam Appl Toxicol, 1991. 17(1): p. 92-102.

8. Brittebo, E.B., C. Eriksson, and I. Brandt, Effects of glutathione-modulating agents on the covalent binding and toxicity of dichlobenil in the mouse olfactory mucosa. Toxicol Appl Pharmacol, 1992. 114(1): p. 31-40.

9. Doty, R.L., Odor-guided behavior in mammals. Experientia, 1986. 42(3): p. 257-71.

41

10. Graziadei and P.P.C., The olfactory mucosa of vertebrates, in Handbook of Sensory Physiology, Beidler and L.M., Editors. 1971, Springer-Verlag: Berlin. p. 27-58.

11. Menco, B., Ultrastructural studies on membrane, cytoskeletal, mucous, and protective compartments in olfaction. Microsc Res Tech, 1992. 22(3): p. 215-24.

12. Zhang, X. and S. Firestein, The olfactory receptor gene superfamily of the mouse. Nat Neurosci, 2002. 5(2): p. 124-33.

13. Mori, K., H. von Campenhause, and Y. Yoshihara, Zonal organization of the mammalian main and accessory olfactory systems. Philos Trans R Soc Lond B Biol Sci, 2000. 355(1404): p. 1801-12.

14. Vassar, R., et al., Topographic organization of sensory projections to the olfactory bulb. Cell, 1994. 79(6): p. 981-91.

15. Ressler, K.J., S.L. Sullivan, and L.B. Buck, Information coding in the olfactory system: evidence for a stereotyped and highly organized epitope map in the olfactory bulb. Cell, 1994. 79(7): p. 1245-55.

16. Astic, L. and D. Saucier, Neuronal plasticity and regeneration in the olfactory system of mammals: morphological and functional recovery following olfactory bulb deafferentation. Cell Mol Life Sci, 2001. 58(4): p. 538-45.

17. Mery, S., et al., Nasal diagrams: a tool for recording the distribution of nasal lesions in rats and mice. Toxicol Pathol, 1994. 22(4): p. 353-72.

18. Young, J.T., Histopathological examination of the rat nasal cavity. Fundam Appl Toxicol, 1981. 1: p. 309-312.

19. Schwob, J.E., Neural regeneration and the peripheral olfactory system. Anat Rec, 2002. 269(1): p. 33-49.

20. Cummings, D.M., et al., Pattern of olfactory bulb innervation returns after recovery from reversible peripheral deafferentation. J Comp Neurol, 2000. 421(3): p. 362-73.

21. Christensen, M.D., et al., Rhinotopy is disrupted during the re-innervation of the olfactory bulb that follows transection of the olfactory nerve. Chem Senses, 2001. 26(4): p. 359-69.

22. Schwob, J.E., S.L. Youngentob, and R.C. Mezza, Reconstitution of the rat olfactory epithelium after methyl bromide- induced lesion. J Comp Neurol, 1995. 359(1): p. 15-37.

42

23. Huard, J.M., et al., Adult olfactory epithelium contains multipotent progenitors that give rise to neurons and non-neural cells. J Comp Neurol, 1998. 400(4): p. 469-86.

24. Morrison, E.E. and R.M. Costanzo, Regeneration of olfactory sensory neurons and reconnection in the aging hamster central nervous system. Neurosci Lett, 1995. 198(3): p. 213-7.

25. Cain, D.P., The role of the olfactory bulb in limbic mechanisms. Psychol Bull, 1974. 81(10): p. 654-71.

26. Allison, A.C., The morphology of the olfactory system in the vertebrates. Biol Rev, 1953. 28: p. 195-244.

27. Price, J.L., B.M. Slotnick, and M.F. Revial, Olfactory projections to the hypothalamus. J Comp Neurol, 1991. 306(3): p. 447-61.

28. Price, J.L., Beyond the primary olfactory cortex: olfactory related areas in the neocortex, thalamus and hypothalamus. Chem Senses, 1985. 10(2): p. 239-258.

29. Gonzalez Mde, L., C.J. Malemud, and J. Silver, Role of astroglial extracellular matrix in the formation of rat olfactory bulb glomeruli. Exp Neurol, 1993. 123(1): p. 91-105.

30. Norenberg, M.D. and A. Martinez-Hernandez, Fine structural localization of glutamine synthetase in astrocytes of rat brain. Brain Res, 1979. 161(2): p. 303-10.

31. Norenberg, M.D., Distribution of glutamine synthetase in the rat central nervous system. J Histochem Cytochem, 1979. 27(3): p. 756-62.

32. Dahl, A.R. and W.M. Hadley, Nasal cavity enzymes involved in xenobiotic metabolism: effects on the toxicity of inhalants. Toxicology, 1991. 21(5): p. 345-372.

33. Bogdanffy, M.S., Biotransformation enzymes in the rodent nasal mucosa: the value of a histochemical approach. Environ Health Perspect, 1990. 85: p. 177-86.

34. Bereziat, J.C., et al., Cytochrome P450 2A of nasal epithelium: regulation and role in carcinogen metabolism. Mol Carcinog, 1995. 14(2): p. 130-9.

35. Zhang, J. and X. Ding, Identification and characterization of a novel tissue-specific transcriptional activating element in the 5'-flanking region of the CYP2A3 gene predominantly expressed in rat olfactory mucosa. J Biol Chem, 1998. 273(36): p. 23454-62.

43

36. Su, T., et al., Expression of CYP2A genes in rodent and human nasal mucosa. Drug Metab Dispos, 1996. 24(8): p. 884-90.

37. Thornton Manning, J.R. and A.R. Dahl, Metabolic capacity of nasal tissue interspecies comparisons of xenobiotic-metabolizing enzymes. Mut Res, 1997. 380(1-2): p. 43-59.

38. Persson, E., P. Larsson, and H. Tjalve, Cellular activation and neuronal transport of intranasally instilled benzo(a)pyrene in the olfactory system of rats. Toxicol Lett, 2002. 133(2-3): p. 211-9.

39. Tjälve, H., et al., Uptake of Manganese and Cadmium from the Nasal Mucosa into the Central Nervous System via Olfactory Pathways. Pharmacol Toxicol, 1996. 79: p. 347-356.

40. Dahlin, M., et al., Transfer of dopamine in the olfactory pathway following nasal administration in mice. Pharm Res, 2000. 17(6): p. 737-42.

41. Hedlund, E., et al., Cytochrome P4502D4 in the brain: specific neuronal regulation by clozapine and toluene. Mol Pharmacol, 1996. 50(2): p. 342-50.

42. Rosenbrock, H., et al., Expression and localization of the CYP2B subfamily predominantly in neurones of rat brain. J Neurochem, 2001. 76(2): p. 332-40.

43. Norris, P.J., J.P. Hardwick, and P.C. Emson, Localization of NADPH cytochrome P450 oxidoreductase in rat brain by immunohistochemistry and in situ hybridization and a comparison with the distribution of neuronal NADPH-diaphorase staining. Neuroscience, 1994. 61(2): p. 331-50.

44. Brandt, I., et al., Irreversible binding and toxicity of the herbicide dichlobenil (2,6-dichlorobenzonitrile) in the olfactory mucosa of mice. Toxicol Appl Pharmacol, 1990. 103(3): p. 491-501.

45. Gu, J., et al., Purification and characterization of heterologously expressed mouse CYP2A5 and CYP2G1: role in metabolic activation of acetaminophen and 2,6-dichlorobenzonitrile in mouse olfactory mucosal microsomes. J Pharmacol Exp Ther, 1998. 285(3): p. 1287-1295.

46. Mancuso, M., A. Giovanetti, and E.B. Brittebo, Effects of dichlobenil on ultrastructural morphology and cell replication in the mouse olfactory mucosa. Toxicol Pathol, 1997. 25(2): p. 186-94.

44

47. Brittebo, E.B., Metabolism-dependent toxicity of methimazole in the olfactory nasal mucosa. Pharmacol Toxicol, 1995. 76: p. 76-79.

48. Genter, M.B., et al., Olfactory toxicity of methimazole: dose-response and structure-activity studies and characterization of flavin-containing monooxygenase activity in the Long-Evans rat olfactory mucosa. Toxicol Pathol, 1995. 23(4): p. 477-86.

49. Bergman, U. and E.B. Brittebo, Methimazole toxicity in rodents: covalent binding in the olfactory mucosa and detection of glial fibrillary acidic protein in the olfactory bulb. Toxicol Appl Pharmacol, 1999. 155(2): p. 190-200.

50. Feng, P.C., et al., Metabolism of alachlor by rat and mouse liver and nasal turbinate tissues. Drug Metab Dispos, 1990. 18(3): p. 373-7.

51. Li, A.A., et al., Metabolism of alachlor by rat and monkey liver and nasal turbinate tissue. Drug Metab Dispos, 1992. 20(4): p. 616-8.

52. Ashby, J., et al., Evaluation of the potential carcinogenicity and genetic toxicity to humans of the herbicide acetochlor. Hum Exp Toxicol, 1996. 15(9): p. 702-35.

53. Holland, D.C., et al., Simultaneous determination of xylazine and its major metabolite, 2,6- dimethylaniline, in bovine and swine kidney by liquid chromatography. J AOAC Int, 1993. 76(4): p. 720-4.

54. Bryant, M.S., et al., 2,6-Dimethylaniline--hemoglobin adducts from lidocaine in humans. Carcinogenesis, 1994. 15(10): p. 2287-90.

55. Parker, R.J., J.M. Collins, and J.M. Strong, Identification of 2,6-xylidine as a major lidocaine metabolite in human liver slices. Drug Metab Dispos, 1996. 24(11): p. 1167-73.

56. Peele, D.B., et al., Functional deficits produced by 3-methylindole-induced olfactory mucosal damage revealed by a simple olfactory learning task. Toxicol Appl Pharmacol, 1991. 107(2): p. 191-202.

57. Genter, M.B., et al., Characteristics of olfactory deficits in the rat following administration of 2,6-dichlorobenzonitrile (dichlobenil), 3,3´-iminidipropionitrile, or methimazole. Fundam Appl Toxicol, 1996. 29: p. 71-77.

58. Eriksson, L. and E. Johansson, Multivariate design and modelling in QSAR. Chemometr Intell Lab Sys, 1996. 34: p. 1-19.

45

59. Wallin, H., et al., A rapid and sensitive method for determination of covalent binding of benzo[a]pyrene to proteins. Chem Biol Interact, 1981. 38(1): p. 109-18.

60. Ullberg, S., The technique of whole body autoradiography. Science Tools, the LKB Instrument Journal, special issue, 1977.

61. O'Callaghan, J.P., Quantification of glial fibrillary acidic protein: Comparison of slot-immunobinding assays with a novel sandwich ELISA. Neurotoxicol Teratol, 1991. 13(3): p. 275-282.

62. Lorenzen, A. and S.W. Kennedy, A fluorescence-based protein assay for use with a microplate reader. Anal Biochem, 1993. 214(1): p. 346-8.

63. Archer, T., et al., Central noradrenaline depletion antagonizes aspects of D-amphetamine-induced hyperactivity in the rat. Psychopharmacology, 1986. 88: p. 141-146.

64. Archer, T., et al., Marble burying and spontaneous motor activity in mice: Interactions over days and the effect of diazepam. Scand J Psychol, 1987. 28: p. 242-249.

65. Fredriksson, A. and T. Archer, MPTP-induced behavioural and biochemical deficits: a parametric analysis. J Neural Transm Park Dis Dement Sect, 1994. 7(2): p. 123-32.

66. Alessandri, B., M.H. Vonau, and W. Classen, The use of an unbaited radial maze in neurotoxicology: II. Sensory inputs, general malaise and locomotor activity. Neurotoxicology, 1994. 15(2): p. 359-70.

67. Hall, R.D. and F. Macrides, Olfactory bulbectomy impairs the rat's radial-maze behavior. Physiol Behav, 1983. 30: p. 797-803.

68. Kelly, J.P., A.S. Wrynn, and B.E. Leonard, The olfactory bulbectomized rat as a model of depression: an update. Pharmacol Ther, 1997. 74: p. 299-316.

69. Lavenex, P. and F. Schenk, Integration of olfactory information in a spatial representation enabling accurate arm choice in the radial arm maze. Learn Mem, 1996. 2: p. 299-319.

70. Olton, D.S.C., C., Intramaze cues and "odor trails" fail to direct choice behavior on an elevated maze. Anim Learn Behav, 1979. 7(2): p. 221-223.

71. Staubli, U., et al., Olfaction and the "data" memory system in rats. Behav Neurosci, 1987. 101: p. 757-65.

46

72. Zoladek, L.R., W.A., The sensory basis of spatial memory in the rat. Anim Learn Behav, 1978. 6: p. 77-81.

73. Olton, D.S., J.A. Walker, and F.H. Gage, Hippocampal connections and spatial discrimination. Brain Res, 1978. 139: p. 295-308.

74. Olton, D.S. and M.A. Werz, Hippocampal function and behaviour: spatial discrimination and response inhibition. Physiol Behav, 1978. 20: p. 597-605.

75. Fredriksson, A., et al., Prenatal co-exposure to metallic mercury vapour and methylmercury produce interactive behavioural changes in adult rats. Neurotoxicol Teratol, 1996. 18: p. 129-134.

76. Morris R, Spatial localization does not require the presence of local cues. Learn Motivat, 1981. 12(2): p. 239-260.

77. Hodges, H., Maze procedures: the radial-arm and water maze compared. Brain Res Cogn Brain Res, 1996. 3(3-4): p. 167-81.

78. van Rijzingen, I.M., W.H. Gispen, and B.M. Spruijt, Olfactory bulbectomy temporarily impairs Morris maze performance: an ACTH(4-9) analog accellerates return of function. Physiol Behav, 1995. 58(1): p. 147-52.

79. Wold, S., Validation of QSAR's. Quant Struct-Act Relat, 1991. 10: p. 191-193.

80. Skagerberg, B., Principal properties in design and structural description in QSAR.. 1989, Umeå University: Umeå.

81. Hansch, C., A. Leo, and W. Taft, A survey of Hammett substituent constants and resonance field parameters. Chem Rev, 1991. 91: p. 165-195.

82. van der Voet, H., Comparing the predictive accuracy of models using a simple randomization test. Chemometr Intell Lab Sys, 1994. 25(2): p. 313-23.

83. Younes, M., S.C. Sharma, and C.P. Siegers, Glutathione depletion by phorone organ specificity and effect on hepatic microsomal mixed-function oxidase system. Drug Chem Toxicol, 1986. 9(1): p. 67-73.

84. Siegers, C.P., W. Hubscher, and M. Younes, Glutathione-S-transferase and GSH-peroxidase activities during the state of GSH-depletion leading to lipid peroxidation in rat liver. Res Commun Chem Pathol Pharmacol, 1982. 37(2): p. 163-9.

85. Brittebo, E.B., Metabolic activation of phenacetin in rat nasal mucosa: dose-dependent binding to the glands of Bowman. Cancer Res, 1987. 47(5): p. 1449-56.

47

86. Potter, D.W., L. Finch, and J.R. Udinsky, Glutathione content and turnover in rat nasal epithelia. Toxicol Appl Pharmacol, 1995. 135(2): p. 185-91.

87. Dear, T.N., et al., Novel genes for potential ligand-binding proteins in subregions of the olfactory mucosa. Embo J, 1991. 10(10): p. 2813-9.

88. Kudo, H., et al., Dietary zinc deficiency decreases glutathione S-transferase expression in the rat olfactory epithelium. J Nutr, 2000. 130(1): p. 38-44.

89. Genter, M.B., D.M. Owens, and N.J. Deamer, Distribution of microsomal epoxide hydrolase and glutathione S- transferase in the rat olfactory mucosa: relevance to distribution of lesions caused by systemically-administered olfactory toxicants. Chem Senses, 1995. 20(4): p. 385-92.

90. Banger, K.K., et al., Immunohistochemical localisation of six glutathione S-transferases within the nasal cavity of the rat. Arch Toxicol, 1994. 69(2): p. 91-8.

91. van Riezen, H., H. Schnieden, and A.F. Wren, Olfactory bulb ablation in the rat: behavioural changes and their reversal by antidepressant drugs. Br J Pharmacol, 1977. 60(4): p. 521-8.

92. Harding, J.W. and J.W. Wright, Reversible effects of olfactory nerve section on behavior and biochemistry in mice. Brain Res Bull, 1978. 4: p. 17-22.

93. Miyakawa, T., et al., Hyperactivity and intact hippocampus-dependent learning in mice lacking the M1 muscarinic acetylcholine receptor. J Neurosci, 2001. 21(14): p. 5239-50.

94. Eng, L.F. and R.S. Ghirnikar, GFAP and astrogliosis. Brain Pathol, 1994. 4(3): p. 229-37.

95. Hill-Felberg, S.J., et al., Concurrent loss and proliferation of astrocytes following lateral fluid percussion brain injury in the adult rat. J Neurosci Res, 1999. 57(2): p. 271-9.

96. Deamer, N.J., J.P. O'Callaghan, and M.B. Genter, Olfactory toxicity resulting from dermal application of 2,6-dichlorobenzonitrile (Dichlobenil) in the C57Bl mouse. Neurotoxicology, 1994. 15(2): p. 287-293.

97. Bailey, M.S. and M.T. Shipley, Astrocyte subtypes in the rat olfactory bulb: Morphological heterogeneity and differential laminar distribution. J Comp Neurol, 1993. 328: p. 501-526.

98. Kasowski, H.J., H. Kim, and C.A. Greer, Compartmental organization of the olfactory bulb glomerulus. J Comp Neurol, 1999. 407: p. 261-274.

48

99. Hamada, S., et al., Localization of 5-HT2A receptor in rat cerebral cortex and olfactory system revealed by immunohistochemistry using two antibodies raised in rabbit and chicken. Brain Res Mol Brain Res, 1998. 54(2): p. 199-211.

100. Marichal, P., et al., Origin of differences in susceptibility of Candida krusei to azole antifungal agents. Mycoses, 1995. 38(3-4): p. 111-7.

101. Durand, M., et al., Developmental and aging aspects of the cholinergic innervation of the olfactory bulb. Int J Dev Neurosci, 1998. 16(7-8): p. 777-85.

102. Ravel, N., A. Elaagouby, and R. Gervais, Scopolamine injection into the olfactory bulb impairs short-term olfactory memory in rats. Behav Neurosci, 1994. 108(2): p. 317-24.

103. Elaagouby, A. and R. Gervais, ACh-induced long-lasting enhancement in excitability of the olfactory bulb. Neuroreport, 1992. 3(1): p. 10-2.

104. Ravel, N., et al., Scopolamine impairs delayed matching in an olfactory task in rats. Psychopharmacol, 1992. 109(4): p. 439-43.

105. Coronas, V., et al., Dopamine receptor coupling to adenylyl cyclase in rat olfactory pathway: a combined pharmacological-radioautographic approach. Neuroscience, 1999. 90(1): p. 69-78.

106. Coronas, V., et al., Identification and localization of dopamine receptor subtypes in rat olfactory mucosa and bulb: a combined in situ hybridization and ligand binding radioautographic approach. J Chem Neuroanat, 1997. 12(4): p. 243-57.

107. Weldon, D.A., M.L. Travis, and D.A. Kennedy, Posttraining D1 receptor blockade impairs odor conditioning in neonatal rats. Behav Neurosci, 1991. 105(3): p. 450-8.

108. Pearson, R.G., Absolute electronegativity and hardness correlated with molecular orbital theory. Proc Natl Acad Sci, 1986. 83: p. 8440-8441.

109. Eriksson, L., et al., Introduction to multi- and megavariate data analysis using projection methods (PCA & PLS). 1999, Umeå, Sweden: Umetrics inc.

Abstract ContentsABBREVIATIONSINTRODUCTIONBackgroundOlfaction ( an overviewMetabolic capacity of the olfactory mucosa and olfactory bulbOlfactory toxicantsMultivariate data analysis in QSAR-modelling

OBJECTIVESMATERIALS AND METHODSMetabolic activation and covalent binding in vitroTissue-localisation of bound substanceHistopathologyLocalisation and quantification of GFAPBehavioural measurementsQSAR-modelling

RESULTS AND DISCUSSIONTissue-bindingLesions in the olfactory mucosaWhere there any differences in 2,6-diClMeSO2-B-derived toxicity in rats and mice?Were the OM lesions caused by isomer-specific CYP-activation?

Neurobehavioural changesWere there any isomer-, species- or sex-dependent differences in behaviour?Is hyperactivity in concordance with the behaviour of anosmic animals?Were the 2,6-diClPh-MeSO2-induced RAM impairments due to loss of olfaction?Can hyperactivity affect RAM-learning?

Effects in the olfactory bulb (GFAP quantification and localisation)Was the GFAP-reaction de

neurotoxic effects of dichlorophenyl methylsulphones related...

Documents