Slight respiratory irritation but not inflammation in mice exposed to (1-->3)-beta-D-glucan aerosols.

Airway irritation effects after single and repeated inhalation exposures to aerosols of beta-glucan (grifolan) were investigated in mice. In addition, the effects on serum total immunoglobulin E (IgE) production and histopathological inflammation in the respiratory tract were studied. The beta-glucan aerosols provoked slight sensory irritation in the airways, but the response was not concentration dependent at the levels studied. Slight pulmonary irritation was observed after repeated exposures. No effect was found on the serum total IgE levels, and no signs of inflammation were seen in the airways 6 h after the final exposure. The results suggest that, irrespective of previous fungal sensitization of the animals, inhaled beta-glucan may cause symptoms of respiratory tract irritation but without apparent inflammation. Respiratory tract irritation reported after inhalation of fungi may not be entirely attributed to beta-glucan.


Introduction
Occupants in buildings with moisture or mold damage often complain of eye and upper respiratory tract irritation. Based on epidemiological studies, fungi (molds) have been suggested to be responsible for many of the irritative symptoms. Fungi contain several specific agents such as microbially produced volatile organic compounds, mycotoxins, and the polyglucose cell-wall component (1 0/3)-b-Dglucan. 1 Á 4 Airborne b-glucan levels in indoor air have been used as indicators of total fungal cell mass and related to symptoms of irritation, cough, tiredness and headache. 5,6 Using the pure substance, a slightly increased prevalence of throat irritation and cough has been demonstrated in healthy, nonsensitive persons after a 4 h inhalation exposure to b-glucan at 210 ng/m 3 . 7 The sensation of irritation experienced by persons exposed to molds or to other agents in their homes may originate either from an inflammatory reaction in the tissue or from a direct stimulation of the nerve endings; the two processes do not have to be related. A standardized mouse bioassay [American Society for Testing and Materials (ASTM) 8 ] is used to assess the latter by determining changes in the breathing pattern of mice during inhalation of aerosols, gases or vapors. Thus, effects on various parts of the respiratory system (i.e. upper and lower respiratory tract and conducting airways) can be recorded. With the head-out body plethysmography, we have previously shown that inhalation of Stachybotrys chartarum extract aerosols that contained b-glucan provoked sensory irritation in the airways of mice, whereas aerosols of ovalbumin or phosphate-buffered saline (PBS) did not cause this effect. 9 In view of this, we undertook experiments to investigate the potency of b-glucan to cause upper respiratory tract irritation and inflammation. Mice were exposed to an aerosol of b-glucan at different concentrations, and the effects on the respiratory pattern were recorded. Furthermore, the reaction in mice pre-exposed to a fungal extract was evaluated during repeated exposures to b-glucan.

Animals
A total of 30 female, 6-week-old (13 weeks old at the end) BALB/cJBom mice (M&B A/S, Ry, Denmark) and 24 male 5-week-old Ico:OF1 mice (IOPS Caw; Iffa Credo, Saint-Germain sur L'Arbresle, France) were used in the experiments (Table 1). To study the potency of b-glucan to provoke sensory irritation, the OF1 strain was used as this strain has been widely used in the ASTM method. 10 For another purpose (i.e. to illustrate immunological changes due to repeated exposures in mice pre-sensitized to a fungal extract), we chose the BALB/c mice. The animals were housed in a carefully controlled environment in an animal room that operated on a 12-h dark/light cycle at 19.5 Á/20.58C and at 60 Á/65% relative humidity. Groups of four mice were housed in stainless steel cages (42 cm deep )/24 cm wide )/13 cm high) bedded with non-autoclaved cellulose wadding (Katrin Care Wadding; Metsä Tissue, Mänttä, Finland). Food (Labfor R36; Lactamin AB, Stockholm, Sweden) and tap water were provided to the animals ad libitum .

Immunization of mice with a fungal extract
For fungal immunizations of two mouse groups (groups Sc and Sc/glu), the fungal extract was prepared from a non-toxic strain of S. chartarum (Sc) (American Type Culture Collection 208877, previously stored as University of Kuopio strain 10) according to Korpi et al. 9 This extract has been shown to contain b-glucan, endotoxin and ergosterol for 6.7 ng/ml, 0.02 ng/ml, and 6.4 mg/ml, respectively. 9 These concentrations were determined per milliliter of the dialyzed and sterile-filtered solution before lyophilization.
Mice were either immunized intraperitoneally with 50 mg S. chartarum extract in adjuvant twice with a 2week interval, or treated with equal amount of PBS (200 ml), or not injected at all (see Table 1). Commercial Imject Alum (Pierce, Rockford, IL, USA) was used as an adjuvant according to the manufacturer's instructions. The experiments were performed in agreement with the European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes (Strasbourg, 18 March 1986, adopted in Finland 31 May 1990). The study was approved by the Animal Care and Use Committee of the University of Kuopio, and by the provincial government.

Generation of glucan aerosols
We used one of the available forms of fungal bglucans, grifolan, which is a gel-forming 1,6- branched (1 0/3)-b-D-glucan isolated from liquidcultured mycelium of Grifola frondosa (kindly provided by Prof. Ohno, Tokyo University of Pharmacy and Life Science, Tokyo, Japan). A water-soluble form of b-glucan was chosen because the S. chartarum extract studied earlier (with its b-glucan moiety) was readily soluble in water and PBS. The resulting grifolan has been analyzed to contain 91% sugar (glucose as a component sugar) and 0.3% protein. 11,12 Two b-glucan batches were used in this experiment. At first, the glucan was dissolved in 0.5 N NaOH to give 0.5% (w/v) solution, and the consecutive working solutions were diluted in sterile water (Medipolar, Oulu, Finland). From the glucan solution (or from sterile water), polydisperse aerosols were generated with an atomizer as described previously. 9 For safety precautions, all the exposure apparatuses were held in a safety hood and the exhaust air was filtered prior to leading it into exhaust air duct.
To determine the glucan concentration of aerosols in the exposure chamber and to ascertain absence of bacterial contamination by determining endotoxin concentration, air samples (2 l/min) were collected at the outlet of the exposure chamber on a polycarbonate filter (ATTP03700, 0.8 mm, 37 mm; Millipore, Molsheim, France). The b-glucan and endotoxin concentrations were determined from a selection of the filters using specific Limulus lysates. 13 A sidestream from the outlet of the chamber was led via a dilutor to the optical particle counter (Hiac/Royco model 5000; Pacific Scientific, Silver Springs, MD, USA) to register the number and size distribution of the generated aerosols continuously during the experiments. This device separates particles into six different classes: from 0.3 to 0.5 mm, from 0.5 to 1.0 mm, from 1.0 to 3.0 mm, from 3.0 to 5.0 mm, from 5.0 to 10 mm, and particles over 10 mm. This generator produces also particles B/0.3 mm. 9

Exposure and mouse bioassay
The six groups of naïve/glu mice were each exposed once to different glucan concentrations (from 37 to 1189 mg/m 3 ) or to sterile water (control experiment) ( Table 1). In the immunized mice, the first aerosol exposure to glucan was performed 12 days after the second immunization of the PBS/glu and Sc/glu mice. The immunized mice were exposed to glucan aerosol on days 0, 1, 7, 8, 14, 15, and 21 (Table 1).
Before each aerosol exposure, there was initially a 15 min acclimation period followed by a 15 min control (baseline) period when baseline respiratory parameter values were obtained for each mouse. The exposures lasted for 15 Á/20 min and were followed by a 15 min recovery period. During the acclimatiza-tion, control and recovery periods, filtered compressed air was fed into the exposure chamber.
The mouse bioassay followed the ASTM E 981-84 standard, 8 as described by Korpi et al. 9 Briefly, the respiratory function of the four simultaneously exposed animals was monitored. The baseline values of different respiratory parameters collected during the control period are presented in Table 2. Based on the baseline values [mean and standard deviation (SD)], each breath was then classified either as normal or as indicating sensory irritation, pulmonary irritation, pulmonary irritation in phase 1, airflow limitation within the conducting airways of the lungs (i.e. bronchoconstriction), or their combinations with each animal acting as its own control. 14 Á 16 Due to the high number of individual breaths analyzed and the low coefficients of variation, using four mice per group is sufficient to establish limits of detection. The smoothing polynomial spline program 14 was used to determine changes in each breath variable or category, and mean responses of four simultaneously exposed mice were calculated. This time series analysis was used to define time Á/response relationships for each group of mice and each exposure regimen. The 95% confidence intervals were also calculated for each smoothed curve.
The animals were observed daily. The body weights of the animals were also recorded on every exposure day. Six hours after the final aerosol exposure, mice were sacrificed with pentobarbital (Mebunat † ), and serum samples as well as tissue samples from the respiratory tract were collected from the PBS/glu and Sc/glu mice that were exposed repeatedly to b-glucan. Serum samples were collected also from the groups PBS and Sc to measure antibody levels without inhalation exposure. Tissue samples were collected also from two naïve/glu mice 6 h after their exposure to 1189 mg/m 3 b-glucan (see also Table 1).

Measurement of total IgE concentrations and histopathology of nasal cavity and lungs
The serum total IgE levels were determined with an enzyme-linked immunosorbent assay as described in Korpi et al. 9 The mouse skull with nasal cavity and lungs was collected and fixed with 4% formaldehyde solution. The mouse skull was decalcified for 6 h in 5% nitric acid solution (Riedel-de Haën KG, Seelze, Germany). After that, multiple sections were cut to reveal the distribution of epithelial types and submucosal glands in the nasal cavity (modified after reference 17). The resulting three or four blocks of the tissue were embedded in paraffin, sectioned at a thickness of 5 mm, and stained with hematoxylin and eosin (HE). The left and right lung (all lobes) including samples from hilar lymph nodes of the lungs were Respiratory irritation by b-glucan routinely processed, embedded in paraffin, sectioned at 5 mm, and stained with HE. An experienced pathologist examined the stained sections of nasal cavity and lungs with light microscopy.

Statistical analysis
The serum total immunoglobulin E (IgE) levels were compared between the mouse groups by the Mann Á/ Whitney U-test (SPSS for MS Windows, release 9.0) to detect the immunological changes induced by bglucan aerosol inhalation. The body weights of the mice at the first and last exposure day were compared with the Wilcoxon signed Rank Test to detect changes due to repeated exposures. Differences with pB/ 0.05 were considered significant.

Glucan aerosols
During repeated exposures, the number of produced particles (of diameter from 0.3 to /10 mm) was approximately 2)/10 9 particles/m 3 except for the fifth and the sixth repeated exposure experiments, when the counts were approximately 7)/10 8 parti-cles/m 3 due to the change of the nozzle plate. The diameter of the majority of aerosols produced from solutions of b-glucan was in the range from 0.3 to 3.0 mm when analyzed in the scale from 0.3 to /10 mm with the optical particle counter. The proportion of particles of 0.5 Á/3.0 mm and the total number of generated aerosols became higher when the amount of b-glucan in the solution was increased (i.e. 425 or 1163 mg/ml). Endotoxin was not detected on any of the nine analyzed filters (detection limit is 10 pg/ml 13 ).

Effects of glucan aerosol inhalation on the murine respiratory pattern
When naïve/glu mice were exposed to various bglucan concentrations, a distinct increase in the time of braking (TB) was observed (Table 3). However, the proportion of the sensory irritation (SI) breaths did not clearly reflect the increase in TB values (Table  3). This is mainly due to the filtering of the data and the prerequisite of three consecutive breaths in the same abnormal category. The TB (and SI) response did not show a clear dependence on b-glucan concentration, and the respiratory response was of the same magnitude at the concentration range from 37 to 883 mg/m 3 (Table 3). In the repeated exposures,  Groups were naïve/glu, four mice in each exposure. Diagnosis of SI requires an increase in time of breaking (TB) greater than 2.0 times the SD of the mean baseline value for that mouse. An increase in TB of 20% compared with the value during control period represent the limit of detection (LOD) in this bioassay, and the LOD of the breath classification for SI 2% 18,19 . n.a., not analyzed. a slight SI response was also observed (Table 4). TimeÁ/response plots of increases in TB and SI breaths due to exposure to b-glucan are illustrated in Figs. 1 and 2 Slight pulmonary irritation was noticed occasionally in the repeated exposure experiments at a bglucan level of 1189/53 mg/m 3 (mean9/SD), as evidenced by an increase in breaths classified as The dilution of b-glucan solution for the aerosol generator was 125 mg/ml and the mean b-glucan concentration9/SD was 1189/53 mg/m 3 . The respiratory parameter data from the fifth and seventh exposures of the group PBS/glu could not be processed due to computer program failure. Diagnoses of SI or PI require an increase in TB or time of pause (TP), respectively, greater than 2.0 times the SD of the mean baseline value for that mouse. An increase in TB of 20%, and an increase in TP of 12% compared with the values during control period represent the limit of detection (LOD) in this bioassay, and the LODs of the breath classifications for SI and PI are 2% for both 18,19 . n.a., not analyzed. pulmonary irritation (PI) ( Table 4). This PI effect was not observable during the single exposures. When the maximum SI effect occurred within 3 min after the onset of the exposure, the maximum PI effect occurred later (i.e. 6 Á/20 min after the onset of the exposure). Slight b-glucan-induced pulmonary irritation in phase 1 (rapid shallow breathing) (B/12% of breaths) was observed on few exposure occasions of the groups Sc/glu and naïve/glu. b-glucan aerosols did not cause bronchoconstriction in any of the experiments.

Serum total IgE levels and tissue samples
Immunization with the fungal extract increased the serum total IgE levels in mice significantly (p0/0.002) (mean9/SD, serum total IgE level in the PBS group was 3739/123 ng/ml and in the group Sc was 11,8839/7234 ng/ml). However, no statistically significant differences (p /0.4) in the serum total IgE levels were caused by seven b-glucan aerosol exposures (group PBS/glu, 3589/71 ng/ml and group Sc/glu, 98489/6305 ng/ml) compared with the mean serum total IgE levels of the groups PBS and Sc, respectively. The bodyweights of the mice did not change significantly after repeated exposures in either of the groups PBS/glu or Sc/glu (p /0.1) (data not shown).
The lung specimens of the seven PBS/glu and seven Sc/glu mice were normal (i.e. no signs of inflammation or fibrosis) at histopathological examination (data not shown). In one PBS/glu and one Sc/ glu mouse, there was a slightly increased number of perivascular/peribronchial lymphocyte aggregates in the lungs. The lung specimens of the two heavilyexposed naïve/glu mice were normal.
The histopathological analysis of the nasal cavity did not reveal any signs of inflammation or tissue damage in any of the Sc/glu mice, the PBS/glu mice, or either of the two naïve/glu mice.

Discussion
The primary objective was to study whether b-glucan aerosols were capable of stimulating the nerve endings directly and whether the response was dependent on previous fungal sensitization. The results showed that aerosolized form of water-soluble bglucan at concentrations from 37 to 1189 mg/m 3 induced a minor SI response, and that the effect was not affected by preceding fungal exposure. SI response is not a general response to any type of inhaled material. For instance, saline solution of approximately 55 mg/m 3 or ovalbumin solution of approximately 109 mg/m 3 does not induce the SI effect. 9 However, inhalation of the fungal (S. char- tarum ) extract showed a distinct SI response. Based on the b-glucan content of the extract, working dilutions used, the air volume and the volume of the working solution aerosolized during an exposure, we can estimate that the maximum b-glucan concentration in the air was 10 mg/m 3 in those experiments. 9 In the present experiments, b-glucan concentrations from 0 to 1189 mg/m 3 were tested for their SI potency, and these concentrations were found to exert lower increases in time of braking and SI breaths than the S. chartarum extract. Thus, bglucan does not contribute significantly to the SI response provoked by aerosols of fungal extract.
The reason for the fact that the SI response was not observed at every exposure level or at every occasion of repeated exposure remains unexplained. The generated particles reached the murine airways, as the majority of the particles were */at each concentration level */in the range from 0.3 to 3 mm. 20 Since the biological effect probably depends on the mass rather than the particle count number, we transformed the results from the optical particle counter from the count number into mass fraction, assuming the same density as that of NaCl for all particles. Even for the 6 )/10 9 particles/m 3 generated from the most concentrated b-glucan solution (1163 mg/ml), the mass median diameter and geometric SD were low (2.5 mm and 1.9). This suggests that the particles were respirable. 21 In indoor environments, microbial growth on moist building materials has also been related to the lower respiratory tract symptoms, such as cough, wheezing and dyspnoea. 22 Á 26 When the amount of bglucan was used as an indicator of mold growth in buildings, increased levels of b-glucan (i.e. /5Á/10 ng/m 3 ) in the air were reported to be associated with cough and general symptoms. 27 The ability of bglucan to increase the prevalence of cough was also verified experimentally, when healthy non-sensitive persons were exposed for 4 h but to much higher levels of b-glucan (210 ng/m 3 ) than normally present in the environment. 7 In the present study, after repeated exposures to soluble b-glucan only slight signs of pulmonary irritation were noticed in mice. Thus, the results from this study suggest that these sensory and pulmonary irritation responses reported after inhalation of fungi are probably not attributed to b-glucan.
Inhalation of aerosols of different mold species for 5 weeks results in a massive cell influx in the lungs of guinea pigs, akin to the reaction after inhalation of bacterial endotoxin. 28,29 In this study, repeated inhalation exposures to b-glucan within 22 days did not induce notable inflammatory changes in the lungs of the mice. Histological changes have both been detected and been absent after acute or chronic inhalation exposures to b-glucan. 29 Á 34 The reasons for different observations regarding the degree of cell infiltration may be due to different glucan delivery protocols as well as different b-glucan preparations. It should be noted, however, that any particular form of b-glucan has not been systematically proven the most potent when assessing both inflammatory cell influx 31,34 and other inflammatory markers (e.g. cytokine and nitric oxide production in in vitro studies). 11,35 Á 38 Taken together, these results suggest that the inflammation caused by inhalation of fungi is not caused by the b-glucan component in the fungal cell wall.
So far, no evidence of b-glucan being an allergen by the inhalation route has been presented. This is supported by the results from this study, since repeated b-glucan exposure did not increase the serum total IgE levels in mice, irrespective of whether the mice were pre-exposed to a fungus. This corroborates the previous suggestions according to which b-glucan may act as an adjuvant exerting additive or synergistic effects only when combined with other agents, such as allergens (ovalbumin) either via inhalation exposure or via injection. 39 Á 41 However, various gel-forming b-glucans have different effects on balancing helper T-cell activity (i.e. on the production of Th1 or Th2 antibody subclasses and cytokines). 42 In conclusion, a single inhalation of b-glucan (grifolan) aerosol provoked a slight sensory irritation response in mice without a dose Á/response relationship at b-glucan levels from 37 to 883 mg/m 3 . Mice repeatedly exposed to an average b-glucan level of 1189/53 mg/m 3 developed a minor pulmonary irritation response in addition to sensory irritation. A previous sensitization to a fungal extract did not affect the response. No inflammatory changes in the nasal cavity or in the lungs could be demonstrated after the exposures. Inhalation of b-glucan did not increase the production of serum total IgE.