Nasal Dosimetry of Inhaled Gases and Particles: Where Do Inhaled Agents Go in the Nose?

Abstract

The anatomical structure of the nasal passages differs significantly among species, affecting airflow and the transport of inhaled gases and particles throughout the respiratory tract. Since direct measurement of local nasal dose is often difficult, 3-dimensional, anatomically accurate, computational models of the rat, monkey, and human nasal passages were developed to estimate regional transport and dosimetry of inhaled material. The computational models predicted that during resting breathing, a larger portion of inspired air passed through olfactory-lined regions in the rat than in the monkey or human. The models also predicted that maximum wall mass flux (mass per surface area per time) of inhaled formaldehyde in the nonsquamous epithelium was highest in monkeys (anterior middle turbinate) and similar in rats and humans (dorsal medial meatus in the rat and mid-septum in the human, near the squamous/nonsquamous epithelial boundary in both species). For particles that are 5 μm in aerodynamic diameter, preliminary simulations at minute volume flow rates predicted nasal deposition efficiencies of 92%, 11% and 25% in the rat, monkey, and human, respectively, with more vestibular deposition in the rat than in the monkey or human. Estimates such as these can be used to test hypotheses about mechanisms of toxicity and supply species-specific information for risk assessment, thus reducing uncertainty in extrapolating animal data to humans.

Introduction

Many inhaled compounds, including formaldehyde and chlorine, induce nasal tissue damage in laboratory animals that varies by location within the nose and is distributed in locations that vary by species (Morgan et al., 1986; Klonne et al., 1987; Monticello et al., 1989; Morgan, 1994; Wolf et al., 1995; Monticello et al., 1996). These site- and species-specific differences in lesion formation are probably due to both regional tissue susceptibility and dose (Morgan and Monticello, 1990). Understanding the relative roles that tissue susceptibility and dose play among species for various toxicants helps reduce uncertainty in regulatory guidelines that are based on nasal lesions in laboratory animals.

Nasal dose depends on the amount of inhaled material delivered by inspired air and the absorption and other transport characteristics of the nasal lining. As inspired air flows through the nasal passages, the folds, grooves, and protrusions of the nasal walls and turbinates divide the airflow into streams of various sizes (Morgan et al., 1991; Hahn et al., 1993; Kimbell et al., 1997a; Kepler et al., 1998; Subranamiam et al., 1998). Nasal turbinates enhance the uptake and deposition of inhaled material by increasing surface area and forcing changes in the direction of passing air. Since nasal anatomy is species-specific, airflow and thus uptake and deposition in the nose follow species-specific patterns as well.

Within a species, nasal uptake patterns for inhaled gases are compound-dependent, determined by the compound’s chemical properties such as diffusivity in air and liquid, solubility, and reactivity with nasal lining components. Nasal particle deposition patterns are size-dependent, influenced by the particles’ inertia, diffusivity, and sedimentation. Uptake and deposition patterns are measures of the amounts delivered to specific sites within the nose and are therefore important components of nasal tissue dose.

Nasal doses of inhaled gases and particles can be difficult to measure experimentally due to rapid metabolic and physical clearance mechanisms and what are often substantial limitations on the ability to make localized measurements. One way to complement experimental measurements is to calculate uptake and deposition patterns throughout the nose and across species using 3-dimensional (3-D), anatomically accurate computer models. Model predictions of where inhaled agents go in the nose can be used to test hypotheses about transport mechanisms and to help extrapolate animal data to people for human health risk assessment.

To illustrate 3-D nasal computational modeling and its uses, this paper briefly describes three examples of inter-species comparisons based on model predictions: (1) olfactory airflow, (2) air-phase transport of inhaled formaldehyde, and (3) nasal deposition of particles that are 5 μm in aerodynamic diameter in the F344 rat, rhesus monkey, and human nasal passages.

Materials and Methods

The computational models needed to be anatomically accurate since they were to be used for species-specific estimates of regional nasal airflow, inhaled gas uptake and particle deposition. Detailed descriptions of the 3-D grids, simulations, and experimental confirmation of model results are given elsewhere (Kimbell et al., 1993, 1997a, 2001a, 2001b; Kepler et al., 1998; Subramaniam et al., 1998) and are described very briefly here. Each computer model was constructed from sequential cross-sections of the nasal passages, obtained from rat and monkey tissue specimens and from human magnetic resonance imaging (MRI) scans. The perimeter of the nasal airway in each cross-section was digitized and an initial grid was constructed for each section using in-house software (Godo et al., 1995). The cross section grids were then connected to each other to form 3-D, gridded (meshed) reconstructions of the nasal passages (Figure 1) using in-house software (Godo et al., 1995) and the commercial mesh generator, Gambit (Fluent, Inc., Lebanon, NH).

The 3-D grids were used to compute the flow of inspiratory air and the transport of inhaled formaldehyde and of 5-μm particles. Particles were assumed to be aerodynamically equivalent to smooth, spherical, unit-density (in g/cm³) particles whose physical diameter is 5 μm. The airflow and gas and particle transport computations were performed by approximating solutions at the grid points to the equations that govern flow and gas and particle transport, representing the conservation laws of mass and momentum (Bird et al., 1960). The approximate solutions to flow and gas transport equations were calculated using the commercial CFD software FIDAP (Fluent, Inc., Lebanon, NH) and particle transport was calculated using in-house software (Asgharian and Anjilvel, 1994). The solutions provided estimates of the direction and speed of airflow and the concentrations of inhaled formaldehyde at every cross-hair in each grid, as well as the deposition sites of particles released into the inspired air from the nostrils.

Results

The computer models predicted that during resting breathing, the fraction of air entering the nostrils that passed through regions lined by olfactory tissue was about 20% in the rat (Kimbell et al., 1997a), about 9% in the monkey (Kepler et al., 1998), and about 3% in the human (Subramaniam et al., 1998). For formaldehyde, the models predicted that maximum uptake rates (mass per surface area per time) in nonsquamous epithelium were approximately 2-fold higher in monkeys than in rats or humans (Kimbell et al., 2001a). The maximum-uptake regions were located near the anterior margin of the middle turbinate in the monkey, in the dorsal medial meatus in the rat, and near the mid-septum in the human. In both the rat and human, the maximum-uptake regions were near the squamous/nonsquamous epithelial boundary (Figure 2; Kimbell et al., 2001a).

For 5-μm particles, preliminary simulations at minute volume flow rates predicted nasal deposition efficiencies of 92%, 11%, and 25% in the rat, monkey, and human, respectively. More vestibular deposition was predicted in the rat than in the monkey or human (Figure 3).

Discussion

Three-dimensional, anatomically accurate models can be used in a number of different applications, including hypothesis testing, risk assessment, and medicine. The computational models described here have provided support for the hypotheses that the distribution of formaldehyde-induced nasal lesions is related to nasal uptake patterns in the rat (Kimbell et al., 1997b) and that the distribution of hydrogen-sulfide-induced olfactory lesions is related to nasal uptake patterns in rat olfactory epithelium (Moulin et al., 2000). Ozone-induced cell proliferation in the rat did not correlate with predicted hot spots of nasal uptake (Hotchkiss et al., 1994) unless tissue susceptibility was taken into account (Cohen Hubal et al., 1997), suggesting that factors other than airflow-dominated delivery play important roles in ozone toxicity.

These models have also been used to compare nasal gas uptake among species (Kimbell et al., 2001a) and to provide quantitative estimates of species-specific uptake for use in human health risk assessment (Kimbell et al., 2001b; Conolly et al., 2003; Schlosser et al., 2003). Medical applications of the computer models described here include predicting the regional deposition of therapeutic drugs delivered via the nasal passages (Kimbell et al., 2003) and studying the effects of surgical interventions such as partial turbinectomy on nasal airflow (Wexler et al., 2003). We have recently extended the models to study how interindividual differences in human nasal anatomy affect airflow and the regional uptake of inhaled reactive gases (Segal et al., 2004). Together with an understanding of tissue susceptibility, these models provide a means for the identification, quantification, and reduction of sources of uncertainty in risk estimates.