Natural mineral fibers: conducting inhalation toxicology studies – part A: Libby Amphibole aerosol generation and characterization method development

Abstract

Background

Asbestos has been classified as a human carcinogen, and exposure may increase the risk of diseases associated with impaired respiratory function. As the range of health effects and airborne concentrations that result in health effects across asbestos-related natural mineral fiber types are not fully understood, the National Institute of Environmental Health Sciences has established a series of research studies to characterize hazards of natural mineral fibers after inhalation exposure. This paper presents the method development work of this research project.

Results

A prototype nose-only exposure system was fabricated to explore the feasibility of generating natural mineral fiber aerosol for in vivo inhalation toxicity studies. The prototype system consisted of a slide bar aerosol generator, a distribution/delivery system and an exposure carousel. Characterization tests conducted using Libby Amphibole 2007 (LA 2007) demonstrated the prototype system delivered stable and controllable aerosol concentration to the exposure carousel. Transmission electron microscopy (TEM) analysis of aerosol samples collected at the exposure port showed the average fiber length and width were comparable to the bulk LA 2007. TEM coupled with energy dispersive spectrometry (EDS) and selected area electron diffraction (SAED) analysis further confirmed fibers from the aerosol samples were consistent with the bulk LA 2007 chemically and physically.

Conclusions

Characterization of the prototype system demonstrated feasibility of generating LA 2007 fiber aerosols appropriate for in vivo inhalation toxicity studies. The methods developed in this study are suitable to apply to a multiple-carousel exposure system for a rat inhalation toxicity testing using LA 2007.

Keywords:

• Asbestos
• Libby amphibole
• fiber
• aerosol generation
• inhalation toxicology

Background

Asbestos is a group of minerals that occur naturally in the environment and is composed of soft and flexible fibers resistant to heat, electricity, and corrosion. Asbestos has been classified as a human carcinogen, and asbestos exposure may increase the risk of diseases associated with inflammation and scarring in the lung and pleural lesions, leading to an impaired respiratory function (Mossman and Gee 1989; Noonan 2017; Solbes and Harper 2018). To date, the asbestiform varieties of six minerals regulated as asbestos include chrysotile, actinolite, tremolite, anthophyllite, crocidolite, and grunerite (amosite). However, there are related natural mineral fibers (e.g. erionite and winchite/richterite) that are not currently regulated as ‘asbestos’ but have similar chemical and physical properties and are associated with disease in humans (NIOSH 2011).

In some cases, humans are exposed to, or have the potential for exposure to, naturally occurring asbestos and related fibrous minerals inadvertently, due to contamination of other materials with fibers, or due to natural outcroppings of fibrous material in the environment. One such exposure was via contaminated vermiculite, a hydrous phyllosilicate mineral which undergoes significant expansion when heated. Vermiculite has a number of industrial and consumer uses, including insulation, soilless growth medium/soil conditioning, and packing/shipping. The vermiculite mine near Libby, Montana (MT) was the major source (nearly 80%) of the world’s supply of vermiculite from 1920 to 1990 for products used in commercial and residential settings (Winters et al. 2011). Mixed minerals, primarily winchite, richterite, with a small amount of tremolite with a range of crystal morphologies, were found in the Libby vermiculite ore, and referred to as Libby amphibole asbestos. An early study in 1986 revealed an increased mortality rate among Libby vermiculite mine workers (McDonald et al. 1986). A growing body of evidence further suggested that human exposure to Libby amphibole asbestos increased risk of lung cancer, mesothelioma, and nonmalignant respiratory disease in both occupational and nonoccupational populations (Benson et al. 2015; Szeinuk et al. 2017; Pfau et al. 2019; Dodson et al. 2020).

Extensive characterization of chemical and physical properties, as well as the collection and processing of samples representative of Libby, MT air (referred to as Libby amphibole 2007, or LA 2007 as used in this article), were performed by United States Geological Survey (USGS) and The Agency for Toxic Substances and Disease Registry (ATSDR) (Lowers et al. 2012). The LA 2007 bulk material was collected, processed, and characterized by the USGS on behalf of the US Environmental Protection Agency (EPA). This material was used in conducting in vivo toxicity studies (Gavett et al. 2016) and is representative of the Libby amphibole asbestos mineral compositions found in the air and the vermiculite mine upon its collection in 2007 (Lowers et al. 2012). Initial animal studies conducted by the US EPA further showed that inhalation of relatively short Libby amphibole fibers generated from the bulk material caused inflammatory, fibrogenic, and tumorigenic responses in rats recapitulating asbestos-related disease in exposed humans (Gavett et al. 2016).

As the range of health effects and air concentrations that result in health effects across natural mineral fiber types are not fully understood, the National Institute of Environmental Health Sciences (NIEHS) has established a series of research studies to evaluate the effects of inhalation exposure to these fibers, including Libby amphibole. However, fundamental challenges regarding the technical feasibility of this work were identified and required solutions prior to testing the biological effects, even in short-term studies. Notably, it was imperative to generate test atmospheres with chemical and physical properties as close to the test material as feasible, particularly with regard to fiber size. In addition, nose-only exposure is required for study of asbestos and related natural mineral fibers, for human health and safety purposes, and to conserve the limited test material. However, in order to generate a nose-only atmosphere, the requisite air flows are low relative to whole-body exposure, which present engineering challenges for generation of atmospheres with the desired accuracy, precision, and homogeneity. This article presents the method development work conducted to generate and characterize LA 2007 fiber aerosol for nose-only inhalation toxicity testing. The results from this work are expected to better support and facilitate the design and conduct of in vivo fiber toxicity/carcinogenicity studies using a nose-only exposure system.

Methods

Test material

The test material, LA 2007, was originally provided to the US EPA after collection, processing, and analysis by the USGS (Lowers et al. 2012). Research Triangle Institute (RTI, Research Triangle Park, NC) received the material from the US EPA on behalf of the NIEHS. Battelle received the material from RTI on behalf of the NIEHS. The digital image of the bulk LA 2007 is presented in Figure 1.

Figure 1. Digital image of bulk LA 2007.

Aerosol generation and exposure system

A prototype aerosol generation and exposure system was built by Battelle (for noncommercial purposes) to explore the feasibility to generate and deliver LA 2007 aerosol for nose-only inhalation toxicity studies. The prototype system contained a single 24-port exposure carousel and was designed to be expandable to a multi-carousel exposure system to facilitate larger-scale exposures for conduct of inhalation toxicity studies. The generation and distribution system were designed to operate at a future capacity required to supply up to five 96-port carousels at concentrations ranging from 0.1 to 10 mg/m³. The extra aerosol flow generated in the prototype system was vented through an exhaust system equipped with a HEPA filter.

The prototype system, as shown in Figure 2, consisted of an aerosol generator, a distribution/delivery system, and an exposure carousel. The aerosol generator was housed in a glovebox and the exposure carousel was housed in a biosafety cabinet.

Figure 2. Schematic of the prototype asbestos aerosol nose-only exposure system.

The aerosol generation was conducted using a slide bar aerosol generator (Figure 3) developed by Battelle. The generator used a pneumatically driven slide bar with an incorporated metering port to accurately meter a small volume per minute of dry test material into a disperser air. Test material was gravity fed into the metering port from an incorporated bulk test material hopper. The slide bar was then moved by a pneumatic actuator to the deliver position (as shown at the bottom of Figure3), and the test material was blown from the metering port by a timed puff of compressed air (air shot) and suspended to form aerosol. The aerosol then passed through an air stream of a single-jet disperser to be further dispersed and deagglomerated. The rate of delivery of test material was controlled by the size of the metering port and the forward-and-backward cycling rate of the slide bar controlled by an electric control unit.

Figure 3. Schematic diagram of the slide bar aerosol generator. (Top schematic: slide bar in fill position; Bottom schematic: slide bar in deliver position).

Prior to aerosol generation, the bulk test material was sieved by gently brushing the material through a stainless-steel screen with a pore size of approximately 750 µm to break up large agglomerates without damaging the fibers. No test material was left on the screen (by visual inspection) after the pre-sieving process to eliminate any potential shifting on fiber size distribution. The sieved test material was then fed into the slide bar generator. Downstream of the single-jet disperser, humidified dilution air was added to increase the volumetric flow rate in the distribution line. The slide bar aerosol generator, single-jet disperser and dilution air assembly were all housed within a glovebox located in the exposure room.

The diluted aerosol stream was conveyed from the glovebox via a high-velocity distribution line to a nose-only inhalation exposure carousel housed in a bio-safety cabinet. The stainless-steel distribution line was grounded and bonded to prevent electrostatic charge. To reduce concentration and pressure fluctuations produced by the slide bar aerosol generator, a 5.5 L stainless-steel buffer chamber was installed in the distribution line and downstream of the aerosol generator.

A Battelle fabricated stainless-steel air ejector pump (as illustrated in Figure 2) was used to deliver a pre-determined quantity of aerosol flow from the distribution line to a delivery line. In the air ejector pump, compressed air flowed through a jet nozzle generating a vacuum downstream the nozzle that induced aerosol flowing from the distribution line to the delivery line. The aerosol was further diluted with dilution air to achieve the target exposure concentration in the delivery line, which was then introduced into the exposure carousel.

The nose-only inhalation exposure carousel, developed by Battelle (Cannon et al. 1983, 1988), was designed to provide aerosol for inhalation toxicity test to multiple rodents. The exposure carousel (Figure 4), made of stainless-steel, consisted of three stackable tiers with eight ports per tier, which provided a total of 24 ports for animal exposure and test atmosphere sampling. Each port received approximately 500 mL/min of exposure atmosphere, for a total carousel inlet flow of approximately 12 L/min.

Figure 4. Three-tier 24-port nose-only exposure carousel.

Aerosol entered the exposure carousel from the top then flowed radially to each of the 24 evenly spaced exposure ports. Mounted to the applicable number of ports were rodent nose-only inhalation exposure tubes. The exposure port design minimized the impact the animals may have on the exposure atmosphere as the exhaled air from each rat is unable to reach the breathing zone of other rats.

During exposure test, aerosol concentration in the exposure carousel was controlled manually by adjusting the cycling rate of the aerosol generator, adjusting the pressure of the ejector pump, and/or changing the dilution air flow. The aerosol concentration at exposure port was determined by gravimetric analysis of filter samples. Filter samples were collected at exposure ports using 25-mm EMfab^TM air monitoring filters (Pallflex^® TX40HI20-WW, Pall Laboratory, Port Washington, NY). The filter media were weighed prior to and after sampling using a six-digit calibrated microbalance, and the net gain of the filter mass was divided by the total volume of sampled air to determine the aerosol mass concentration. The sampling flow rates were approximately 0.4–0.5 L/min with the sampling duration varying depending on the test performed.

A MicroDust Pro real-time aerosol monitor (RAM) (Casella CEL Ltd, Bedford, UK) was used to evaluate the temporal stability of the aerosol concentration within the exposure carousel. The MicroDust Pro RAM is a forward light-scattering photometer that can provide real-time aerosol mass concentration measurement. For a given aerosol material and particle size distribution, the RAM photometer voltage response is directly related to the aerosol mass concentration. The temporal stability of the aerosol concentration was reported as the relative standard deviation (RSD) of the RAM voltage measurement.

Asbestos fiber analysis methods

Fiber morphology, structure, and mineralogical composition are important discriminating factors in classifying types of asbestos. With respect to biopersistence, previous studies reported fiber dimensions could be the most important determinant of pathogenicity and cancer development (Davis et al. 1986; McConnell et al. 1999; Oberdorster 2000; Boulanger et al. 2014). Microscopy is the preferred means to characterize fiber materials, specifically when analyzing asbestos fibers (US EPA 2004). Asbestos fiber lengths greater than 5 µm are well-suited for optical microscopy methods. Further, light polarization techniques, such as polarized light microscopy (PLM), can be employed to determine the characteristic optical properties of different types of asbestos (EPA 1993). The PLM technique, however, cannot distinguish fibers smaller than 0.3 µm in diameter and does not allow for the ability to perform chemical analysis at the same time. When fibers are smaller or the amount of material available is limited, more sensitive microscopy techniques are necessary (Virta 2002). Electron microscopy techniques are particularly useful for amphibole materials, providing resolution for fibers with 0.1 µm diameters using scanning electron microscopy (SEM) or less than 0.01 µm diameters using transmission electron microscopy (TEM).

Due to apparent visual similarities, particularly among the amphiboles, characterization by morphology alone is not considered a reliable methodology to identify asbestos fibers, thus microscopy is typically combined with other analytical approaches (Boulanger et al. 2014). Elemental analysis of asbestos can be performed using spectroscopic methods such as x-ray fluorescence (XRF) or x-ray photoelectron spectroscopy (XPS). As electron microscopy is typically coupled with energy dispersive spectrometry (EDS), this technique is commonly used to determine the elemental composition of fibers in combination with microscopic morphology assessment to help to identify asbestos (EPA. 2014). In addition to elemental composition, structural characterization can be used to identify asbestos materials. Selected area electron diffraction (SAED) is also often coupled with electron microscopy, and thus provides crystal structure information for asbestos fibers (EPA. 2014).

To standardize the methodology needed for asbestos analysis and reporting, several environmental monitoring methods have been established by various regulatory agencies. The US EPA has created a method for the analysis of asbestos in bulk building materials (EPA 1993). Methods were also established for TEM analysis of asbestos in air samples by the National Institute for Occupational Safety and Health (NIOSH) (NIOSH 1994) and by the International Organization for Standardization (ISO) (ISO 1995). These methods allow for the direct determination of fiber dimensions and chemical compositions of asbestos materials and have been used as reference methods for fiber counting and analysis.

On the basis of the strengths and weaknesses of different analytical techniques, the TEM with EDS and SAED analytical techniques and the ISO Method 10312 were used in this study for fiber counting, analysis, and identification. These methods are consistent with the methodology used and reported by the EPA (Meeker et al. 2003) for characterization of samples obtained from the Libby Superfund Site in Libby, MT.

Aerosol samples for analysis were collected at an exposure port of the exposure carousel using EMSL Zefon TEM Asbestos Cassettes (EMSL Analytical, Inc., Cinnaminson, NJ). Bulk test material samples were collected from the bulk material containers for analysis. The samples were submitted for TEM with SAED and EDS analysis at EMSL Analytical Inc. (Cinnaminson, NJ).

Transmission electron microscopy analysis of the aerosol samples was performed using a JEOL 1200 EX II (Akishima, Tokyo, Japan) scanning transmission electron microscope equipped with a PGT energy dispersive X-ray analyzer and side mount digital camera. Asbestos structures were identified by a combination of morphology, elemental chemistry via EDS, and SAED. Any structures greater than 0.5 µm in length with an aspect ratio (the ratio of a fiber length to its width) of at least 3:1 were counted following the ISO 10312 method counting rules to document length, width, aspect ratio, and structure type. For each sample, approximately one hundred (100) fibrous structures were analyzed following the OSHA 100-fiber/sample stopping rule (OSHA 1994). Chemical analysis of the counted fibers from the samples was performed using EDS. For each TEM sample, representative images (approximately one image per aerosol sample and 20 images per bulk material sample) of fiber morphology, EDS spectra, and SAED patterns were collected to assess chemical composition and crystalline structure. The system (microscope/EDS detector) generated Cliff-Lorimer factors were used to calculate the chemical composition within the PGT Spirit software.

Results

Test system characterization and optimization

In addition to having the capability of supplying aerosol for multi-carousel exposure testing, the prototype system was designed and characterized to meet the operation goals of (a) delivering stable LA 2007 aerosol concentration at 10 mg/m³ and (b) maintaining fiber size at the exposure ports equivalent to that of the bulk test material. The test system presented in Figure 2 was the final configuration deemed to meet the criteria specification. Modifications made during method development are summarized below.

Achieving stable aerosol concentration

Generating and delivering a stable asbestos aerosol concentration to a nose-only exposure system was a challenge, due to the relatively small volume and low flow rate of nose-only exposure carousels compared to whole-body exposure chambers. Fluctuations in aerosol concentrations were noticed during initial characterization tests, which were attributed to the slide bar aerosol generator intermittently delivering test material to the air stream. To smooth out periodic concentration fluctuations of the exposure atmosphere and to achieve the requisite stable aerosol concentration, a stainless-steel buffer chamber was added to the distribution line. Selection of the buffer chamber volume considered the need to balance the chamber size with a sufficient volume to smooth out the periodic fluctuations, yet not too large of a volume that would result in long aerosol fibers settling out of the test atmosphere during passage though the chamber.

In addition to adding the buffer chamber, aerosol concentration stability in the exposure carousel was also improved by pre-sieving the test material by brushing LA 2007 through a stainless-steel screen to break up large agglomerates as described below.

Minimizing long fiber loss

Another engineering challenge was to generate and deliver test atmospheres that have fiber size distribution as close to the bulk LA 2007 material as feasible. TEM results for aerosol samples collected from the exposure ports during initial characterization testing revealed the average fiber length was approximately 45% shorter than that measured for the bulk LA 2007 material. The observation indicated a significant loss of longer fibers during aerosol generation and delivery to the exposure carousel.

To reduce the loss of the longer fibers during aerosolization and transport, the buffer chamber was reduced to half of the original size (from a volume of 11 L to 5.5 L) and the airflow through the chamber was redirected from the original horizontal flow to vertical flow (from top to bottom, as shown in Figure 2).

In addition, a 100 µm pore size stainless-steel sieve used to break up the large agglomerates in the bulk test material resulted in an average fiber length shorter than the bulk material. As such, the original sieve was replaced by a 750 µm pore size sieve. Moreover, to further limit potential breakage/shearing of LA 2007 fibers during the sieving procedure and ensure longer fibers passed through the sieve, the test material was gently brushed through the sieve rather than applying pressure to force the material through the screen.

Following the modification to the buffer chamber and sieving process, 3-h aerosol generation tests were conducted using unsieved test material and sieved test material (using the 750 µm pore size sieve). The mean aerosol concentration was measured by collecting 180-min filter samples during generation. Results are summarized in Table 1.

Table 1. Aerosol concentration and temporal stabilities.

Test material	Measured aerosol concentration (mg/m^3)	RAM RSD (%)
Sieved test material LA 2007	11.6	12
Unsieved test material LA 2007	10.0	19

The mean aerosol concentrations were within 20% of target (10 mg/m³). For aerosol generation using sieved LA 2007, stable aerosol concentration was observed, with RAM RSD of 12%. Higher RAM RSD was observed when generating aerosol using the unsieved LA 2007, indicating the necessity for sieving the test material prior to aerosolization.

Duplicate aerosol samples for TEM analysis were collected at the exposure ports during aerosol characterization to evaluate fiber size distributions. Samples of bulk LA 2007 (unsieved and sieved) were also collected and analyzed together with the aerosol samples. The results are presented in Table 2. In general, the average fiber length was comparable across the aerosol samples, sieved bulk samples and un-sieved bulk samples, indicating fiber length was not altered during aerosolization and delivery to the exposure carousel. While the average length result of the sieved bulk and aerosol samples appear to be different, the length ranges overlap, and the overall aspect ratios are the same. The length difference could be due to natural variability in LA 2007. Higher average fiber width was measured in aerosols generated with unsieved LA 2007, implying bundles of fiber were present in the aerosol. On the basis of aerosol concentration, temporal stability, and TEM data, the sieved LA 2007 was selected for subsequent aerosol generation and exposure testing.

Table 2. LA 2007 fiber size distribution by TEM.

Property		Bulk LA 2007 (unsieved)	Bulk LA 2007 (sieved)	Aerosol sampling at exposure port
				Unsieved	Sieved
Length (µm)	Average^a	3.98 ± 1.20	3.82 ± 1.11	5.23 ± 0.37	4.79 ± 0.86
	Range	0.60–16.90	0.60–20.00	1.00–11.50	1.10–15.40
Width (µm)	Average^a	0.24 ± 0.07	0.26 ± 0.07	0.61 ± 0.04	0.32 ± 0.05
	Range	0.05–1.10	0.05–1.00	0.10–1.32	0.08–1.00
Aspect Ratio		17	15	8.6	15
Total fibers counted		103	103	16	44

^aData shown are average ± standard deviation.

Characterization testing demonstrated the aerosol generation and delivery system was able to deliver stable LA 2007 aerosol to the exposure carousel with fiber dimensions equivalent to the bulk test material chemically and physically.

Test system performance

Following initial characterization, 5-day aerosol generation tests were conducted to further characterize the aerosol generation and exposure system. Sieved LA 2007 aerosol was generated over five consecutive days for a duration of 3 h per day at a target concentration of 10 mg/m³.

Aerosol concentration

Aerosol concentration was measured daily by collecting duplicate filter samples at the exposure port for each 3-h generation test. The daily mean aerosol concentrations and concentration temporal stability measured by a RAM are presented in Table 3. Daily mean aerosol concentrations were within 20% of target for each exposure day. Aerosol concentration was stable, with a temporal variability (RAM RSD) less than 15% over the 5-day testing period. Higher RAM RSD (13%) measured on Day 3 was due to a concentration adjustment to reach the target concentration.

Table 3. Aerosol concentration at the exposure port during the 5-day generation tests.

Test Day	Target aerosol concentration (mg/m^3)	Mean aerosol concentration (mg/m^3)^a	Percentage of target (%)	RAM RSD (%)
0	10	10.5 ± 0.1	105	8
1	10	10.8 ± 0.3	108	5
2	10	9.7 ± 0.2	97	5
3	10	8.2 ± 0.1	82	13
4	10	10.0 ± 0.2	100	4

^aMean concentration based on duplicate samples; data shown are mean ± difference.

TEM with EDS and SAED analysis

Over the five-day testing period, samples for TEM analysis were collected at an exposure port on Days 0, 2, and 4. The results are presented and discussed below.

Fiber size distributions

A summary of fiber dimension distribution determined by TEM for aerosol samples is presented in Table 4. For comparison, the average results of bulk LA 2007 samples measured during the study are presented in Table 4. The average fiber lengths and widths collected at the exposure ports ranged from 4.44 to 4.79 µm for fiber length and from 0.28 to 0.36 µm for fiber width. These dimensions were comparable to the bulk test material (approximately 5.76 µm for fiber length and 0.33 µm for fiber width), considering the length/width ranges overlap and the inherent fiber size variability in the bulk LA 2007 test material as reported in other studies (Lowers et al. 2012; Wang et al. 2023). The determined fiber dimensions (including length, width, and aspect ratio) of aerosol samples were consistent over the 5 days.

Table 4. LA 2007 fiber dimensions characterization during the 5-day evaluation testing.

Properties	Bulk LA 2007 (Unsieved)	Aerosol sample at exposure port
		Test day 0	Test day 2	Test day 4
Number of TEM samples (Average number of structures counted/Sample)	4 (106)	4 (87)	4 (62)	4 (53)
Average length^a (µm)	5.76 ± 2.55	4.65 ± 1.81	4.44 ± 1.23	4.79 ± 1.30
Range of length (µm)	0.25–28.30	0.56–60.40	0.50–34.20	0.62–29.20
Average width^a (µm)	0.33 ± 0.04	0.36 ± 0.11	0.28 ± 0.09	0.34 ± 0.09
Range of width (µm)	0.06–2.00	0.10–2.20	0.05–1.30	0.05–1.30
Average aspect ratio	17	13	16	14

^aData shown are average ± standard deviation.

Representative fiber length distribution plots are presented in Figure 5 and representative fiber width distribution plots are presented in Figure 6 for aerosol sample and the bulk LA 2007 sample. Representative TEM images are shown in Figure 7 for aerosol samples and in Figure 8 for bulk LA 2007 test material.

Figure 5. Representative fiber length distributions. (Top plot: result of one aerosol sample as an example. Bottom plot: result of one bulk LA 2007 material sample as an example.)

Figure 6. Representative fiber width distributions. (Top plot: result of one aerosol sample as an example. Bottom plot: result of one bulk LA 2007 material sample as an example.)

Figure 7. Representative TEM images of LA 2007 fibers – aerosol samples. (Scale bar is 800 nm for all photos.)

Figure 8. Representative TEM images of LA 2007 fibers – bulk test material. (Scale bar is 500 nm for all photos.)

Aerosol fiber identification and chemical composition

Selected area electron diffraction patterns presented by amphibole fibers are uniform rows of closely spaced spots with patterns that are influenced by the orientation of the fiber to the direction of the electron beam (Millette 1987). The presence of concise spots conveys the crystallinity of the fiber. Upon measurement, sets of zone-axis patterns of spots are used to describe types of amphiboles in conjunction with chemical composition measured by EDS (Meeker et al. 2003). The electron diffraction patterns for the aerosol sample and the bulk LA 2007 test material indicated the aerosol and bulk material were highly crystalline and were consistent with the standards of amphibole asbestos from the Libby amphibole, MT region (Meeker et al. 2003). Representative SAED images for aerosol sample and bulk LA 2007 test material are shown in Figure 9.

Figure 9. Representative SAED images. (Top two scans were from aerosol samples. Bottom two scans were from bulk LA 2007. Scale bar is 0.2 1/Å for all scans.)

Table 5 presents the elemental oxide content of the fibers determined by EDS for aerosol sample and bulk LA 2007 test material. The aerosol results were consistent over the five days of exposure testing and were consistent with the characterization of the bulk LA 2007 test material. Representative EDS spectra are shown in Figure 10 for the aerosol sample and bulk LA 2007 test material. Note that all the EDS spectra of asbestos material prepared on copper TEM grids contain background Cu peaks at approximately 8 and 9 keV as shown in Figure 10. These peaks were not included in the determination of elemental composition presented in Table 5. As shown in Figure 9, O peak presents in the EDS spectra of the aerosol samples and the bulk material sample; the O peak is not labeled in the spectra of the aerosol samples due to low intensity.

Figure 10. Representative EDS Spectra. (Top three spectra were from aerosol samples collected on Day 0 (A), Day 2 (B), and Day 4 (C). Bottom spectra were from a bulk LA 2007 sample.)

Table 5. Fibers element oxide content for LA 2007 by TEM via EDS.

Sample ID.	Test Day	Elemental oxide content (Weight %)
		CaO	MgO	SiO2	FeO	K2O	Na2O
Bulk LA 2007^a	NA	11.9 ± 1.6	17.9 ± 1.2	57.5 ± 1.3	10.1 ± 1.8	0.9 ± 0.5	2.0 ± 0.6
Aerosol samples	0^b	9.5	17.4	61.2	10.4	0.4	1.1
	2^b	9.5	17.4	60.4	10.0	0.7	2.0
	4^b	14.7	14.8	56.9	10.6	1.3	1.8
	Overall^c	11.2 ± 2.7	16.5 ± 2.2	59.5 ± 3.3	10.3 ± 0.3	0.8 ± 0.4	1.6 ± 0.9

^aData shown are average ± standard deviation, n = 4 samples, 20 spectra per sample.

^bn = 2 samples, one spectrum per sample.

^cData shown are average ± standard deviation, n = 6 samples, one spectrum per sample.

Discussion

A nose-only exposure system design was selected for conducting in vivo inhalation toxicity studies of naturally occurring asbestos and related natural mineral fibers. A nose-only exposure system can conserve limited test materials and generate less exhaust air and wastewater contaminated with asbestos (resulting from periodical cleaning of the exposure carousel/chamber), due to its relatively low flow rate and small volume compared to a whole-body exposure system. The low flow rate and small volume of a nose-only exposure system, however, make it more challenging to generate a continuous, stable, and controllable aerosol concentration from powder-like dry mineral fiber materials such as asbestos.

In this study, stable and controllable aerosol mass concentration was achieved using a slide bar aerosol generator combined with a buffer chamber installed downstream of the generator. The slide bar aerosol generator metered and aerosolized dry test material to provide controllable aerosol mass concentration, while the buffer chamber dissipated the pressure wave created by the cyclic aerosol generation to maintain concentration stability. Moreover, aerosol concentration stability was further improved by breaking up the large agglomerates present in the bulk LA 2007 material prior to aerosol generation.

Another imperative goal and criterion for in vivo inhalation toxicity studies of asbestos and natural mineral fibers are to generate test atmospheres with chemical and physical properties as close to the bulk test material as feasible, especially for fiber size distributions. This presented another fundamental challenge as natural settling of long fibers during transport from the generator to the exposure carousel could change fiber size distributions. Two approaches developed in this study effectively reduced fibers settling. First, a high velocity grounded stainless steel distribution line was installed to transport fiber aerosol from the generator to the exposure carousel that effectively prevented fiber settling through the transportation line. Second, the volume and flow direction of the buffer chamber were optimized thereby eliminating long fiber loss through the buffer chamber. Besides eliminating long fiber settling, it is important not to alter the fiber size distributions when breaking up large agglomerates in bulk LA 2007 during the sieving process. To prevent this, the bulk material was brushed gently through a large pore (∼750 µm) stainless-steel screen without damaging the fibers or leaving test material on the screen after the process.

Characterization tests on the prototype exposure system using LA 2007 further demonstrated that the aerosol generation, delivery, and exposure methods developed in this study are appropriate to conduct in vivo inhalation toxicity studies with other asbestos and related natural mineral fibers of concern. It is worth noting that the prototype nose-only exposure system has been expanded to a multiple-carousel system suitable for exposure to a range of concentrations for in vivo inhalation toxicity studies and the details will be reported in a subsequent article.

Conclusions

The prototype nose-only exposure system successfully delivered a stable aerosol concentration of LA 2007 to the exposure carousel at the target concentration of 10 mg/m³ and was further validated in five-day aerosol generation tests. TEM analysis of aerosol samples collected at the exposure ports demonstrated average fiber length and width comparable to the characterization of the bulk LA 2007 material. TEM coupled with EDS and SAED analysis further confirmed fibers from the aerosol samples were consistent with the bulk LA 2007 chemically and physically. In summary, the prototype nose-only exposure system developed and optimized in the current study demonstrates the feasibility to generate LA 2007 aerosols appropriate for in vivo inhalation toxicity studies.

Authors contributions

AW and AG designed and supervised the experimental work, and MG and DP executed the experiments. BE and RR conducted the TEM with SAED and EDS analysis. JR and JS coordinated the TEM samples collection and interpreted the data. AW drafted the manuscript and PLY wrote the Background section. BS, KE, GR, PLY, and MS supervised the work, and critically reviewed and revised the manuscript. The manuscript was written through contributions of all authors. All authors read and approved the final manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The dataset supporting the conclusions of this article is included within the article.

Additional information

Funding

This research was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (NIEHS), Intramural Research projects ES103316-05 and ES103374, and NIEHS Contract HHSN273201400015C (Battelle Memorial Research Institute, Columbus, OH).