Pulmonary Toxicity of Perfluorinated Silane-Based Nanofilm Spray Products: Solvent Dependency
Asger W. Nørgaard,* Jitka S. Hansen,* Jorid B. Sørli,* Marcus Levin,*,† Peder Wolkoff,* Gunnar D. Nielsen,* and Søren T. Larsen*,1
*Danish NanoSafety Centre, National Research Centre for Working Environment (NRCWE), DK-2100 Copenhagen, Denmark; and †Department of Micro- and Nanotechnology, Technical University of Denmark, DK-2800 Lyngby, Denmark
1To whom correspondence should be addressed at National Research Centre for the Working Environment, Lersø Parkallé 105, DK-2100 Copenhagen East, Denmark. Fax: +45 39165201. E-mail: [email protected].
Received July 3, 2013; accepted September 16, 2013
A number of cases of pulmonary injury by use of aerosolized surface coating products have been reported worldwide. The aerosol from a commercial alcohol-based nanofilm product (NFP) for coating of nonabsorbing surfaces was found to induce severe lung damage in a recent mouse bioassay. The NFP contained a 1H,1H,2H,2H-perfluorooctyl trialkoxysilane (POTS) and the effects were associated with the hydrolyzed forms of the silane; increase in hydrolyzation resulted in faster induction of compro-mised breathing and induction of lung damage. In this study, the impact of the solvent on the toxicity of POTS has been investi-gated. BALB/cA mice were exposed to aerosolized water-based NFPs containing POTS, and solutions of hydrolyzed POTS in methanol, ethanol, and 2-propanol, respectively. No acute respira-tory effect was observed at exposure concentrations up to 110 mg/m3 with an aqueous solution of POTS. However, exposure to POTS in methanol resulted in a decrease of the tidal volume—an effect that did not resolve within the recovery period. After 27 min of exposure, the tidal volume had decreased by 25%, indicating par-tial alveolar collapse. For POTS in ethanol and 2-propanol, a 25% reduction of the tidal volume was observed after 13 and 9 min, respectively; thus, the tidal volume was affected by increase of the chain length. This was confirmed in vitro by investigating lung sur-factant function after addition of POTS in different solvents. The addition of vaporized methanol, 2-propanol, or acetone to aero-solized POTS in methanol further exacerbated the tidal volume reduction, demonstrating that the concentration of vaporized sol-vent participated in the toxicity of POTS.
Commercial impregnation spray products, including nano-film products (NFPs), consist of one or more surface active compounds, solvent, and in some cases a propellant. The sol-vent may be water or composed of one or more organic sol-vents, eg, alcohols, butyl acetate, and naphtha. Several cases of
pulmonary injury have been associated with use of impregna-tion spray products (Daubert et al., 2009; Lazor-Blanchet et al., 2004; Pauluhn et al., 2008; Vernez et al., 2006), and recent lit-erature reveals that the products that cause respiratory problems among users often are based on an organic solvent vehicle (eg, Nørgaard et al., 2010a; Pauluhn et al., 2008). Although there are several water-based impregnation products on the market—usually for absorbing surfaces such as concrete, leather, textiles, and wood—these rarely give rise to respiratory cases among users. This is striking, because the active components, eg, acryl polymers or organosilanes, in products based on organic sol-vent may be similar to those based on water. Furthermore, the incidence of cases related to exposure of impregnation sprays has often been observed when a product formulation has been altered; this also includes changes in the solvent composition (Hubbs et al., 1997). This prompted us to investigate whether the properties of the solvent may affect the acute airway toxic-ity of the inhaled aerosols from a well-characterized group of common impregnation products, NFPs.
NFPs sprayed onto a hard surface form a ca. 20-nm “nano-film” by self-organization during evaporation of the solvent, inducing nonstick properties (Sepeur, 2008). The film arises from the sol-gel process (Hench and West, 1990; Schmidt, 2006) that involves series of hydrolysis and condensation reac-tions between organo-functionalized (eg, fluorinated) silanes. The result is an interconnected network of functionalized silox-anes. The process is similar to the preparation of silica sols by polymerization of silicon alkoxides by addition of water and an acid catalyst (Hench and West, 1990). The NFPs are available for a wide range of surfaces such as bathroom tiles, concrete, floors, and textiles. NFPs for coating of nonabsorbing materi-als (eg, floor materials and tiles/ceramics) have been charac-terized previously by various mass spectrometric techniques (Nørgaard et al., 2009, 2010b,c, 2011). They were identified as solutions of 1H,1H,2H,2H-perfluorooctyl trialkoxysilane
toxicological sciences 137(1), 179–188 2014doi:10.1093/toxsci/kft225Advance Access publication October 4, 2013
(POTS) in 2-propanol and a hexadecyl trialkoxysilane in etha-nol, and their hydrolysates and condensates (Fig. 1). No solid material was found in the products. The airway effects of the aerosolized form of these NFPs have been investigated in a mouse bioassay (Nørgaard et al., 2010a). The NFP contain-ing the perfluorinated silane induced an irreversible and severe reduction of the tidal volume, progressing to pulmonary edema and tissue damage. The difference between the no-observed adverse effect level and the exposure concentration leading to lethal lung damage was found to be approximately 10%. Both the perfluorination and the degree of hydrolysis influenced the toxicity of POTS.
In this study, a number of commercially available NFP prod-ucts, based on either water or an organic solvent in addition to synthesized solutions of POTS in various solvents, have been tested. To mimic relevant human exposure, we used an acute inhalation model in mice to evaluate the impact of the sol-vent on the airway toxicity of POTS in NFPs. The NFPs were selected due to their simplicity, (ie, one active compound solu-ble in a range of solvents), thus allowing the study of effects possibly associated with the properties of the solvent.
MATerIALS ANd MeTHOdS
Animals. Inbred BALB/cA male mice aged 5–7 weeks and weight 25.5 ± 2.3 g were purchased from Taconic M&B (Ry, Denmark) and were housed in polypropylene cages (380 × 220 × 150 mm) with pinewood sawdust bedding (Lignocel S8, Brogaarden, Denmark). The cages, housing up to 10 mice each, were furnished with bedding materials, gnaw sticks, and plastic nesting houses. Cycles of 12-h light (from 6 am to 6 pm) followed by 12-h dark were applied. Cages were sanitized twice weekly. Food (Altromin no. 1324, Altromin, Lage, Germany) and municipal tap water were available ad libi-tum. Treatment of the animals followed procedures approved by The Animal Experiment Inspectorate, Denmark (No. 2006/561-1123-C3).
(98%) were obtained from Sigma Aldrich (Brøndby, Denmark). The lyo-philized pulmonary surfactant formulation Alveofact was a kind gift from Lyomark Pharma (Oberhaching, Germany).
Nanofilm products. Seven commercially available water-based NFPs were included in this study. The following were obtained from NanoCover (Aalborg, Denmark): “Absorbing Floor Materials,” “Textiles and Leather,” a concentrated version of “Textiles and Leather,” and “Wood and Stone.” The products “Floors, Joints, and Concrete” and “Leather and Textiles” were obtained from Nanonordisk (Aalborg, Denmark), whereas the products “Special Textile Coating” and Wood and Stone were obtained from NanoLotus (Odense, Denmark) and Percenta AG (Glücksburg, Germany), respectively. According to the product manuals, 20–25 ml/m2 is required for coating of tex-tiles, whereas 25–50 ml/m2 is required for coating of floor materials, concrete, stone, and wood. Analysis of all products by electrospray ionization mass spec-trometry (ESI-MS) revealed that they all contained hydrolyzed forms of POTS. Thus, they were considered identical (see below). The product Special Textile Coating was selected as representative for the whole group of water-based NFPs and used for the mouse bioassay exposures. In addition to the commer-cially available products, a water-based concentrate of the Nanocover Textiles and Leather product was included for mouse exposure at high concentrations.
The concentration of nonvolatiles in Special Textile Coating (termed NFP) and Textiles and Leather concentrate (termed NFP concentrate) was estimated in the following way: 1 ml of the product/concentrate was transferred to a 2 ml glass vial and purged to dryness by a gentle stream of nitrogen. The mass of the nonvolatile fraction was determined gravimetrically in triplicate and found to be 0.33 ± 0.02% and 14.83 ± 0.06% for the product and concentrate, respec-tively. One NFP for nonabsorbing floor materials from NanoCover, which contained POTS in 2-propanol (cf. Nørgaard et al., 2010c), was used for the experiments in the capillary surfactometer (CS, see below).
Preparation of POTS solutions. Solutions of hydrolyzed POTS were prepared in methanol, ethanol, and 2-propanol, respectively. Water (3.0 mmol) and formic acid (0.3 mmol) were added to 1.0 mmol of 1H,1H,2H,2H-perfluorooctyl triethoxysilane in a 2.5-ml polypropylene vial. Further, 100 µl of 2-propanol (1.3 mmol) was added before the mixture was shaken gently for 2 h resulting in clear, homogenous solutions. The mixtures were then trans-ferred to Teflon flasks and diluted with methanol, ethanol, or 2-propanol until 1.2% wt/wt solutions of the POTS were obtained. All solutions were analyzed by ESI-MS prior to their use in mouse exposure experiments (Nørgaard et al., 2010c).
Mass spectrometric analysis. NFP (1 µl) or hydrolyzed POTS solution was diluted 102–103 times with methanol in a 2 ml glass vial and either acidi-fied with 10 µl formic acid or alkalized with 20 µl of a 25% aqueous solution of ammonium hydroxide. The diluted samples were then infused directly into the electrospray ion source of a Bruker micrOTOF-Q mass spectrometer (Bruker Daltonik, Bremen, Germany) by the means of a syringe pump (KD Scientific, Holliston, MA) at a flow rate of 3.3 µl/min. The capillary voltage was 4.5 kV (positive mode) or 3.2 kV (negative mode), the nebulizing gas pressure was 1 bar, and the drying gas flow and temperature were 4 l/min and 190°C, respec-tively. For collision-induced dissociation (CID), nitrogen was used as collision gas, and collisions were carried out at energies ranging from 12 to 25 eV. An ESI calibrant solution containing fluoroalkyl phosphazines (Agilent Tune Mix) was used for external mass calibration resulting in errors of ±5 mDa for most mass measurements.
The ESI-MS spectrum of the product Special Textile Coating (NanoLotus) recorded in the negative mode is shown in Figure 2a. Two major clusters of peaks can be observed around m/z 485 and m/z 861, which correspond to the silanes and disiloxanes of POTS with a varying number of methoxy groups (see structures in Fig. 1). Assignments of sum formulae for the individual peaks are summarized in Table 1. Two types of ions were observed: deprotonated molecules (M−H)− and their formic acid adducts (M+HCOO)−; the products contain formic acid in concentrations up to 1%. Similar adduct ions were observed previously during the monitoring of the curing processes of the NFPs (Nørgaard et al., 2010b). The structural assignment of the observed ions was
FIg. 1. Silanes and siloxanes of a 1H,1H,2H,2H-perfluorooctyl trialkox-ysilane (POTS). a, Fully hydrolyzed silane (silane triol); b, mono-methoxylated silane; and c, dimethoxylated disiloxane. See assigned ions 1–6 in the mass spectrum shown in Figure 2a.
made on the basis of their CID spectra. For instance, the CID spectrum of the mono-methoxylated silane (m/z 485) recorded at 25 eV is shown in Figure 2b. Aside from the precursor ion, the spectrum shows several neutral losses, eg, CH
4O (m/z 453), CH
2O
2 (m/z 439), and C
2H
4O
2 (m/z 425). In addition, series
of less dominant peaks occur with internal differences of 18 Da (H2O), 20 Da
(HF), and 100 Da (C2F
4). The ions of m/z 287 and 267 are the results of a partial
breakup of the perfluorinated chain, whereas its complete removal from silicon can be observed at m/z 111 and 97. Analysis of Special Textile Coating in the positive mode showed that the majority of the observed ions could be assigned as polyethylene glycol and poly siloxane (Supplementary Figure S1).
In general, ESI-MS analysis of the NFPs and the prepared solutions of POTS all yielded similar spectra (Supplementary Figures S3 and S4).
Generation of test atmospheres and particle characterization. The water-based NFPs, the water-based NFP concentrate, POTS solutions, and the corre-sponding solvent controls were aerosolized at a flow rate of 0.4–0.6 ml/min by infusion from a glass syringe into a Pitt no. 1 jet nebulizer (Wong and Alarie, 1982) by means of an infusion pump (New England Medical Instruments, Medway, MA). The exposure air stream (25 l/min, < 2% relative humidity medical grade air) was subsequently led through a Vigreux column to ensure mixing and then directed into a 20 l exposure chamber of glass and stainless steel resulting in an approximate air exchange rate of 1/min (Clausen et al., 2003). Outlet air was passed through a series of particle and active coal filters before the exhaust to the atmosphere.
Vapors of methanol, 2-propanol, and acetone were generated by sparging the solvent in a U-shaped glass tube (length: 300 mm, i.d.: 15 mm) serving as an evaporation unit.
Prior to sparging, the evaporation unit was placed in an oil bath and thermo-stated to 160°C to ensure complete evaporation of the solvent. A dry air stream (1.5 l/min, medical grade) was led through the evaporation unit and the solvent vapor was mixed with the NFP aerosol. The flow of organic solvent into the evaporating unit was 0.5 ml/min. Experiments are listed in Table 2.
Particle size distributions were measured by a Fast Mobility Particle Sizer (FMPS, TSI Model 3091, Shoreview, MN), which measures the electrical mobility equivalent particle size (D
m) in 32 channels with midpoints ranging
from 6.04 to 523.3 nm. Sampling was conducted through the standard FMPS cyclone Model 1031083 (d
50 = 1 µm). Particles were measured at positions
in the center of the exposure chamber by use of 1/8 in Teflon tubing (length 10–15 cm) and 1/4 in conductive flexible tubing (approximately 30 cm). The instrument was operated without animals at a sampling frequency of 1 s and a sample flow of 10 l/min. The total flow through the chamber was 25 l/min. Size distributions were calculated as a 30-s average after aerosol size and concentra-tion had stabilized within the chamber.
The mass-concentration of the particle exposure was measured by filter sampling using Millipore Glass fiber filters (Millipore Co, Billerica, MA) placed in closed-face 25-mm cassettes. The air flow through the filters was 2 l/min for 2–36 min. Differential weighing of the filters was carried out on a Sartorius Microscale Model M3P (Sartorius GmbH, Göttingen, Germany) after 24 h of acclimatization (humidity 50% ± 2.5%, 20°C ± 0.2) before and after filter exposure. No filter data were obtained for the NFP. Thus, the exposure concentration was estimated from the results of POTS 1A–C, 2, and 3 (Table 2 and Supplementary material).
Acute airway effects. To assess the acute effects on respiration of the NFPs in different solvents, groups of mice (n = 4–10) were exposed head out only for 60 min. The 60-min exposure period was followed by a 15-min recov-ery period in which the mice were exposed to laboratory air. Prior to exposure, a 15-min baseline period was recorded for each mouse. To assess exposure-related effects, the respiratory parameters during exposure were compared with baseline levels, ie, each mouse served as its own control. For each mouse, mean values of each minute during the experiment were calculated. Exposure concentrations were measured concurrently by filter sampling as described above and given in Table 2. The mice were euthanized immediately after each experiment.
Collection of respiratory parameters. The Notocord Hem (Notocord Systems SA, Croissy-sur-Seine, France) data acquisition software was used to collect respiratory parameters. The mice were placed in body plethysmographs and exposed head out only (cf. Vijayaraghavan et al., 1993). The acquisition program calculates inter alia the respiratory frequency (breaths/min); time of inspiration (ms); time of expiration (ms); time from end of inspiration until the beginning of expiration, termed time of brake (TB, ms); time from end of expi-ration until beginning of the next inspiration, termed time of pause (TP, ms); tidal volume (VT, ml); and mid-expiratory flow rate (VD, ml/s). Stimulation of the trigeminal and vagal nerve endings causes respective increases in TB and TP, which are markers of upper/sensory and lower airway irritation,
FIg. 2. A, Electrospray ionization mass spectrometry spectrum of the water-based NFP Special Textile Coating. The insert shows the ions around m/z in more detail. B, CID spectrum of m/z 485 recorded at a collision energy of 25 eV. Analysis of all the other water-based NFPs in addition to the pre-pared 1H,1H,2H,2H-perfluorooctyl trialkoxysilane (POTS) solutions resulted in similar spectra (cf. Supplementary Figures S3 and S4). Thus, these mass spectra are representative of hydrolysates and condensates of POTS analyzed by negative electrospray ionization. The numbers 1–6 are referring to the struc-tures shown in Figure 1.
respectively. Comprehensive descriptions of the breathing parameters and their analysis have been made elsewhere (Alarie et al., 1998; Larsen and Nielsen, 2000; Nielsen et al., 1999). Data acquisition and calculations were performed as described previously (Larsen et al., 2004).
In vitro surfactometry. The pulmonary surfactant Alveofact was dis-solved at a concentration of 5.6 mg/ml in sterile milliQ water over night. The pulmonary surfactant function was assessed using a CS (Calmia Medical Inc, Toronto, Ontario, Canada). This assay resembles the conditions in the human terminal bronchioles (Enhorning, 2001). Surfactant sample (0.5 µl) was intro-duced into the narrow section (0.25 mm in diameter) of a glass capillary and placed in a water bath at 37°C. The CS then gradually increased the air pressure until the liquid blocking the capillary was extruded. A sample with a func-tioning surfactant will prevent the liquid from collapsing back into the narrow section, and in this case, no resistance is measured during the 120 s of airflow. However, if the surfactant is damaged, the liquid will collapse and the airflow
will be blocked until the CS has raised air pressure sufficient to remove the blockage. This cycle continues during the measured time frame. Surfactant function was assessed as the time of unobstructed airflow in the glass capillary (time open, in %).
To assess pulmonary surfactant inhibition by the POTS/solvent mixtures, the Alveofact was incubated with a solvent alone (control group) and a mix-ture of solvent and POTS (exposure group). More specifically, Alveofact was incubated with POTS in 2-propanol with added methanol, ethanol, and 2-pro-panol (equal volumes of POTS solution and solvent), respectively, at 37°C for 30 min. The solvent control samples (water, methanol, ethanol, and 2-pro-panol) were likewise mixed in equal volumes with 2-propanol before incuba-tion with Alveofact. The final concentration of Alveofact was 4 mg/ml in all experiments.
Following incubation, the organic solvents were removed by evaporation. This was done at a pressure of 200 mbar and temperature of 34°C for 30 min (methanol and ethanol) or 60 min (2-propanol) in a drying cabinet (Heraeus
aCalculated from infusion flows and chamber ventilation rate (see the materials and methods section).bSpecial Textile Coating.cAll nonvolatiles are assumed to be POTS.dNo filter data. Exposure concentration estimated from POTS solutions 1A–C, 2, and 3 (see Supplementary material).
TABLe 1elemental Compositions of Selected Ions Observed in the eSI-MS Spectrum of Special Textile Coating (Fig. 2a)
Measured Mass Theoretical Mass Error (mDa) Ion Type Elemental Composition
type RVT220) connected to a vacuum pump (Neuberger Laboport). After evap-oration, the samples were analyzed with the CS. Three to 15 replicates of each sample were analyzed.
Statistics. Effects on respiratory parameters were evaluated by ANOVA when more than 2 groups of mice were compared. In case of significant differ-ences between multiple groups, pairwise comparisons by Tukey’s method were performed. The TB and TP data were log transformed to fulfill the assumptions of the ANOVA. When only 2 groups were compared, this was done by means of a 2-sample t test. The CS values were compared by the Kruskal-Wallis test and group differences assessed by the Mann-Whitney’s U test. For all analyses, a significant difference was accepted at p levels < .05. Calculations were performed using the Minitab Statistical Software version 15 (Minitab Inc, State College, PA).
reSuLTS
Airway Effects
The toxicological effect of the solvent was investigated by exposing groups of mice to (1) a commercial water-based NFP; (2) a concentrated water-based NFP; and (3) solutions of POTS in methanol, ethanol, and 2-propanol (see Table 2). Figure 3 shows effects of upper and lower airway irritation and effect on the tidal volumes for all groups before, during, and after exposure. Some upper and lower airway irritation was observed (ANOVA, p < .001, Figs. 3a and 3b, respectively, Table 3).
FIg. 3. Effects on respiratory parameters during and after exposure to different nanofilm products (NFPs). Exposures are summarized in Table 2 and statis-tical analyses in Table 3. A–C, Mice were exposed to a water-based NFP; a water-based NFP concentrate; and 1.2% solutions of 1H,1H,2H,2H-perfluorooctyl trialkoxysilane (POTS) in methanol, ethanol, or 2-propanol (POTS 1–3, respectively). Effects on time of brake (A), time of pause (B), and tidal volume (C) are shown. D, Mice were exposed to POTS in methanol in addition to POTS in methanol concurrently with vaporized methanol, 2-propanol, or acetone (POTS 1A–C, respectively). A control group was exposed concurrently to aerosolized methanol and vaporized 2-propanol gas. Effects on tidal volume are shown. For each mouse, the mean of a 1-min period was calculated as percentage of the baseline value. Group means are shown.
A short-lasting sensory irritation was observed due to POTS 3 and partly due to POTS 2 exposure (Fig. 3a and Table 3). The TB elongation had completely resolved after 30 min of exposure. After 35 min of exposure, a consistent lower airway irritation developed in the POTS 3 group and the TP elonga-tion persisted during the recovery period (Fig. 3b and Table 3). However, the most prominent effects were observed on the tidal volume (Fig. 3c). After 60 min of exposure, a significant difference in tidal volume was observed between the groups (ANOVA; p < .0001). The commercial water-based product Special Textile Coating and the water-based concentrate did not alter the tidal volume and both were significantly dif-ferent from the 3 alcohol solutions of POTS (1–3) (Tukey; p < .0001), where 40%–60% reductions were observed after 60 min of exposure.
It appeared that increase in chain length of the alcohols increased the initial rate of reduction in the tidal volume (Fig. 3c). A 25% reduction in tidal volume was obtained after 27, 13, and 9 min for POTS dissolved in methanol, ethanol, and 2-propanol, respectively (ANOVA; p < .0001, Table 3). The reduction in tidal volume in the 3 alcohol groups did not differ significantly at the end of the exposure period (Tukey; p > .05). The effect on tidal volume was not reversible within the 15-min recovery period.
To further study whether a solvent may alter the tidal volume reduction of POTS, we exposed a group of mice to aerosolized POTS in methanol (POTS 1) concurrently with admixed vapor-ized alcohols (Fig. 3d). Addition of vaporized methanol sig-nificantly decreased the time to reach a 25% reduction in tidal volume when compared with POTS in methanol only (t test,
p < .0001, Fig. 3d and Table 3). Concurrent exposure to vapor-ized 2-propanol also decreased the time to reach 25% reduc-tion in tidal volume in comparison to POTS in methanol alone (p = .001, Fig. 3d and Table 3). A control group exposed to methanol and vaporized 2-propanol showed a transient reduc-tion in tidal volume, which partly reversed during the recovery period; irreversibility of the response was only seen in groups exposed to POTS (Fig. 3d and Table 3).
Gas phase mixing of vaporized 2-propanol with the aero-sol of POTS dissolved in methanol could theoretically lead to an exchange of POTS alkoxy groups (Fig. 1) and thus change its toxicological properties. In order to eliminate this possibility, we exposed mice to vaporized acetone, which cannot react with POTS, concurrently with aerosolized NFP in methanol. In this experiment, the mice showed a reduction in the tidal volume similar to the mice exposed to POTS in methanol together with vaporized 2-propanol (t test; p > .05, comparing time to a 25% reduction, Fig. 3d and Table 3).
In Vitro Data—Capillary Surfactometry
Incubation of Alveofact with water, methanol, ethanol, and 2-propanol showed that the functioning surfactant was able to maintain the capillary open >98% of the time measured (Fig. 4). Thus, the incubation did not inhibit the surfactant function of Alveofact. POTS in 2-propanol mixed with an equal volume of solvent resulted in a significantly inhibited surfactant as apparent from the reduction in the time of open capillary.
Particle Characterization
Particle size distributions of aerosolized products are shown in Figures 5a and 5b. The aerosolized water-based NFP Special Textile Coating had a main wide mode around 100 nm and a smaller plateau around 10–20 nm (Fig. 5a). The NFP concen-trate produced higher concentration within the larger mode but also showed a clearer mode around 10 nm that is not visible with NFP. POTS 1–3 all showed similar modal distributions with a broad mode around 100 nm and the only difference between them was a slightly higher concentration over all sizes for POTS 1 (Fig. 5b). The addition of vaporized solvent, POTS 1A–C (Fig. 5b), did not to change the distribution of particles. In general, the POTS solutions produced similar size distributions indicating that pulmonary particle deposition is comparable.
dISCuSSION
The acute airway toxicity of water-based spray products and POTS dissolved in different solvents was studied. Our data showed that the solvent markedly influenced the toxicity of the product. The most prominent effect was irreversible reduction in the tidal volume due to atelectasis, ie, alveolar collapse, as shown in a previous study (Nørgaard et al., 2010a). Solvents, eg, methanol with added 2-propanol vapor, caused only a transient reduction in the tidal volume, ie, the mice recovered during the 15-min recovery period. This observation is in accordance with other toxicity studies of waterproofing sprays, which indicates
TABLe 3Overview of Statistical results of Mouse exposure experiments
Note. An ANOVA with Tukeys post hoc test was used to compare several groups. A t test was used to compare the rate of VT reduction of POTS 1A–C with POTS 1. Mean and SD are given.
aEffect on tidal volume (VT). Time to reach a 25% reduction in VT.bThe mean value of the last minute of the period is calculated.cEffect on time of brake (TB) during first 5 min of exposure. Mean of period is calculated.dEffect on time of pause (TP) during the 15-min recovery period. Mean of period is calculated.eSpecial Textile Coating.fSignificantly different from POTS 1.gSignificantly different from POTS 2.hSignificantly different from POTS 3.iSignificantly different from solvent control.jNot applicable because some mice did not reduce VT with 25%.
that the toxicity is not due to the solvents per se (Tashiro et al., 1998; Yamashita and Tanaka, 1995).
Short-lasting sensory irritation was observed after exposure to POTS 2 and 3 and was most likely due to the high concen-trations of alcohol (solvent) than to POTS itself. Alcohols are known to induce temporary sensory irritation (Kristiansen et al., 1986, 1988). As shown for formaldehyde, a strong sen-sory irritative effect may only cause a minor change in VT (Nielsen et al., 1999) without any effect on survival of the ani-mals even after long-term exposure (Monticello et al., 1996). Lower airway irritation (TP elongation) developed at the end of the exposure period for POTS 3. Lower airway irritation arises due to vagal nerve stimulation. For alcohols, the effect is revers-ible after end of exposure (Kristiansen et al., 1986, 1988). TP elongation may also occur during inflammatory and edematous responses and may therefore persist after cessation of exposure (Vijayaraghavan et al., 1993). However, the change in VT (and thus the severe clinical effect of atelectasis) did not reflect the pattern of TP. Therefore, the mechanism behind the TP elonga-tion is not expected to cause the VT reduction.
It has previously been suggested that the solvent may alter the toxicity of a product simply by altering the droplet size of the aerosol (Yamashita et al., 1997a,b). However, in the pre-sent study, the particle size distributions of perfluorosilanes were unaltered in different solvents (Fig. 5b); this suggests that the particle size is not responsible for the mechanism behind the change in toxicity. Furthermore, addition of vaporized solvent clearly increased the toxicity of the aerosol without
changing the droplet size distribution. More specifically, using the alcohols as the model solvent group, the toxicity of POTS increased with increasing carbon chain length of the alcohol. The water solubility of the alcohol decreases and the lipophi-licity increases with increasing chain length from log k
o/w −0.74
for methanol to 0.25 for 2-propanol (Sangster, 1989). The frac-tion of alcohol trapped from the inhaled air due to a scrubber effect, ie, dissolved in the upper airway mucosa, may decrease with increasing carbon chain length. Thus, a solvent change from methanol to 2-propanol may result in a larger propor-tion of the alcohol reaching the terminal bronchioles and the alveoli as suggested by the experiments in Figure 3c. However, our in vitro experiments (Fig. 4) clearly demonstrated that it was the carbon chain length/lipophilicity of the solvent that determined the toxicity of POTS on surfactant function. In the lung, alcohols may dissolve in the lung lining fluid containing
FIg. 4. Effect of 1H,1H,2H,2H-perfluorooctyl trialkoxysilane (POTS) 3 on the pulmonary surfactant properties. Pulmonary surfactant was incubated with water, POTS 3, and 2-propanol. Methanol, ethanol, and 2-propanol were added, and the surfactant function was assessed by capillary surfactometry. Bars indicate mean values with SEM of 3 (for surfactant mixed with solvent) and 15 (for surfactant mixed with solvent and POTS 3) analyses per sample. Horizontal bars indicate groups that are significantly different from each other (Mann-Whitney U test, p < .008).
10 10010
0
101
102
103
104
105
106
107
108
Mobility equivalent diameter [nm]
dN/d
logD
p [cm
−3 ]
NFP
A
B
NFP conc
10 100
101
102
103
104
105
106
107
108
Mobility equivalent diameter [nm]
dN/d
logD
p [cm
−3 ]
POTS 1POTS 1APOTS 1BPOTS 1CPOTS 2POTS 3
FIg. 5. Particle number size distributions in aerosolized solutions of POTS in water (A) and organic solvent (B) as measured with Fast Mobility Particle Sizer. Vertical bars indicate 1 SD.
pulmonary surfactant and alter its physicochemical properties due to cosolvent properties of the alcohols (Ladaa et al., 2001). Further, it may be speculated that the solubility of perflouro-silane in the lung lining fluid increases with the quantity and lipopholicity of the solvent alcohol, and thereby facilitates contact between perfluorosilane and lung surfactant compo-nents. This hypothesis is supported by the in vitro surfactant experiments. The rate of reduction in tidal volume caused by perfluorsilane inhalation increased in parallel with the lipo-philicity of the cosolvent. In addition to the observation that the toxicity of POTS increased with the lipophilicity of the alcohol, the concentration of alcohol in the inhaled air also affected the toxicity of POTS. During the experiments with alcohol vapor (Fig. 3d), alcohol concentrations in the mouse chamber exceeded 10 000 ppm (Table 2), so that the loading of the upper airway mucosa probably minimized the impact of the scrubber effect and allowed more solvent to reach the lower airways. In these experiments, the toxicity of POTS was aggravated and a further reduction in tidal volume to 30% of baseline was observed.
In conclusion, our study highlights that the respiratory tox-icity of an impregnation spray product does not only depend on the toxicity of the active agents but also on the amount and properties of the used solvents. This study suggests that water as a solvent may reduce the toxicity of an impregnation product. The use of appropriate solvents should, therefore, be incorporated in safe-by-design considerations of impregnation products. However, appropriate risk management has to be introduced dependent on the type of product.
SuPPLeMeNTAry dATA
Supplementary data are available online at http://toxsci. oxfordjournals.org/.
FuNdINg
Danish Environmental Protection Agency; Danish Working Environment Research Fund.
ACkNOwLedgMeNTS
The authors thank Maria Hammer and Michael Guldbrandsen for excellent technical assistance.
reFereNCeS
Alarie, Y., Schaper, M., Nielsen, G. D., and Abraham, M. H. (1998). Structure-activity relationships of volatile organic chemicals as sensory irritants. Arch. Toxicol. 72, 125–140.
Clausen, S. K., Bergqvist, M., Poulsen, L. K., Poulsen, O. M., and Nielsen, G. D. (2003). Development of sensitisation or tolerance following repeated
OVA inhalation in BALB/cJ mice. Dose-dependency and modulation by the Al(OH)3 adjuvant. Toxicology 184, 51–68.
Daubert, G. P., Spiller, H., Crouch, B. I., Seifert, S., Simone, K., and Smolinske, S. (2009). Pulmonary toxicity following exposure to waterproofing grout sealer. J. Med. Toxicol. 5, 125–129.
Enhorning, G. (2001). Pulmonary surfactant function studied with the pulsat-ing bubble surfactometer (PBS) and the capillary surfactometer (CS). Comp. Biochem. Physiol. A Mol. Integr. Physiol. 129, 221–226.
Hench, L. L., and West, J. K. (1990). The sol-gel process. Chem. Rev. 90, 33–72.
Hubbs, A. F., Castranova, V., Ma, J. Y., Frazer, D. G., Siegel, P. D., Ducatman, B. S., Grote, A., Schwegler-Berry, D., Robinson, V. A., Van Dyke, C., et al. (1997). Acute lung injury induced by a commercial leather conditioner. Toxicol. Appl. Pharmacol. 143, 37–46.
Kristiansen, U., Hansen, L., Nielsen, G. D., and Holst, E. (1986). Sensory irri-tation and pulmonary irritation of cumene and n-propanol: Mechanisms of receptor activation and desensitization. Acta Pharmacol. Toxicol. 59, 60–72.
Kristiansen, U., Vinggaard, A. M., and Nielsen, G. D. (1988). The effects of n-butanol vapour on respiratory rate and tidal volume. Arch. Toxicol. 61, 229–236.
Ladaa, T. I., Lee, C. M., Coates, J. T., and Falta, R. W., Jr. (2001). Cosolvent effects of alcohols on the Henry’s law constant and aqueous solubility of tetrachloroethylene (PCE). Chemosphere 44, 1137–1143.
Larsen, S. T., Hansen, J. S., Hammer, M., Alarie, Y., and Nielsen, G. D. (2004). Effects of mono-2-ethylhexyl phthalate on the respiratory tract in BALB/c mice. Hum. Exp. Toxicol. 23, 537–545.
Larsen, S. T., and Nielsen, G. D. (2000). Effects of methacrolein on the respira-tory tract in mice. Toxicol. Lett. 114, 197–202.
Lazor-Blanchet, C., Rusca, S., Vernez, D., Berry, R., Albrecht, E., Droz, P. O., and Boillat, M. A. (2004). Acute pulmonary toxicity following occupational exposure to a floor stain protector in the building industry in Switzerland. Int. Arch. Occup. Environ. Health 77, 244–248.
Monticello, T. M., Swenberg, J. A., Gross, E. A., Leininger, J. R., Kimbell, J. S., Seilkop, S., Starr, T. B., Gibson, J. E., and Morgan, K. T. (1996). Correlation of regional and nonlinear formaldehyde-induced nasal cancer with prolifer-ating populations of cells. Cancer Res. 56, 1012–1022.
Nielsen, G. D., Hougaard, K. S., Larsen, S. T., Hammer, M., Wolkoff, P., Clausen, P. A., Wilkins, C. K., and Alarie, Y. (1999). Acute airway effects of formaldehyde and ozone in BALB/c mice. Hum. Exp. Toxicol. 18, 400–409.
Nørgaard, A. W., Janfelt, C., Benassi, M., Wolkoff, P., and Lauritsen, F. R. (2011). Nebulization ionization and desorption ionization analysis of reac-tive organofunctionalized silanes in nanofilm products. J. Mass Spectrom. 46, 402–410.
Nørgaard, A. W., Jensen, K. A., Janfelt, C., Lauritsen, F. R., Clausen, P. A., and Wolkoff, P. (2009). Release of VOCs and particles during use of nanofilm spray products. Environ. Sci. Technol. 43, 7824–7830.
Nørgaard, A. W., Larsen, S. T., Hammer, M., Poulsen, S. S., Jensen, K. A., Nielsen, G. D., and Wolkoff, P. (2010a). Lung damage in mice after inha-lation of nanofilm spray products: The role of perfluorination and free hydroxyl groups. Toxicol. Sci. 116, 216–224.
Nørgaard, A. W., Vaz, B. G., Lauritsen, F. R., and Eberlin, M. N. (2010b). Real-time monitoring of the progress of polymerization reactions directly on sur-faces at open atmosphere by ambient mass spectrometry. Rapid Commun. Mass Spectrom. 24, 3441–3446.
Nørgaard, A. W., Wolkoff, P., and Lauritsen, F. R. (2010c). Characterization of nanofilm spray products by mass spectrometry. Chemosphere 80, 1377–1386.
Pauluhn, J., Hahn, A., and Spielmann, H. (2008). Assessment of early acute lung injury in rats exposed to aerosols of consumer products: Attempt to disentangle the “Magic Nano” conundrum. Inhal. Toxicol. 20, 1245–1262.
Sangster, J. (1989). Octanol-water partition coefficients of simple organic com-pounds. J. Phys. Chem. Ref. Data 18, 1111–1127.
Schmidt, H. (2006). Considerations about the sol-gel process: From the classical sol-gel route to advanced chemical nanotechnologies. J. Sol-Gel Sci. Technol. 40, 115–130.
Sepeur, S. (2008). Silane-Technology as the Key to Chemical Nanotechnology. Vincentz Network, Hannover, Germany.
Tashiro, K., Matsumoto, Y., Nishizuka, K., Shibata, K., Yamamoto, K., Yamashita, M., and Kobayashi, T. (1998). Mechanism of acute lung injury caused by inhalation of fabric protector and the effect of surfactant replace-ment. Intensive Care Med. 24, 55–60.
Vernez, D., Bruzzi, R., Kupferschmidt, H., De-Batz, A., Droz, P., and Lazor, R. (2006). Acute respiratory syndrome after inhalation of waterproofing sprays: A posteriori exposure-response assessment in 102 cases. J. Occup. Environ. Hyg. 3, 250–261.
Vijayaraghavan, R., Schaper, M., Thompson, R., Stock, M. F., and Alarie, Y. (1993). Characteristic modifications of the breathing pattern of mice to
evaluate the effects of airborne chemicals on the respiratory tract. Arch. Toxicol. 67, 478–490.
Wong, K. L., and Alarie, Y. (1982). A method for repeated evaluation of pul-monary performance in unanesthetized, unrestrained guinea pigs and its application to detect effects of sulfuric acid mist inhalation. Toxicol. Appl. Pharmacol. 63, 72–90.
Yamashita, M., and Tanaka, J. (1995). Pulmonary collapse and pneumonia due to inhalation of a waterproofing aerosol in female CD-1 mice. J. Toxicol. Clin. Toxicol. 33, 631–637.
Yamashita, M., Tanaka, J., Yamashita, M., Hirai, H., Suzuki, M., and Kajigaya, H. (1997a). Mist particle diameters are related to the toxicity of waterproof-ing sprays: Comparison between toxic and non-toxic products. Vet. Hum. Toxicol. 39, 71–74.
Yamashita, M., Yamashita, M., Tanaka, J., Hirai, H., Suzuki, M., and Kajigaya, H. (1997b). Toxicity of waterproofing spray is influenced by the mist particle size. Vet. Hum. Toxicol. 39, 332–334.
