Comparative investigations of the biodurability of mineral fibers in the rat lung.

The biodurability of various glass fibers, rockwool, and ceramic fibers was examined in rat lungs and compared with natural mineral fibers. Experiments were based on studies that have shown that the biodurability of fibers is one of the essential factors of the carcinogenic potency of these materials. Sized fractions of fibers were instilled intratracheally into Wistar rats. The evenness of distribution of fibers in the lung was checked by scanning electron microscopy (SEM) or careful examination of the fiber suspension before treatment. After serial sacrifices up to 24 months after treatment, the fibers were analyzed by SEM following low temperature ashing of the lungs. Parameters measured included number of fibers, diameter, and length distribution at the various sacrifice dates, so that analyses could be made of the elimination kinetics of fibers from the lung in relation to fiber length (FL). Size selective plots of the fiber elimination correlated with fiber diameters enables the mechanism of the fiber elimination (dissolution, fiber breakage, physical clearance) to be interpreted. The half-time of fiber elimination from the lung ranges from about 10 days for wollastonite to more than 300 days for crocidolite. The biodurability of man-made vitreous fibers (MMVF) is between these values and is dependent on the chemical composition of the fibers and the diameter and length distribution. Results indicate that the in vivo durability of glass fibers is considerably longer than expected from extrapolation of published data on their in vitro dissolution rates.


Introduction
The persistence of mineral fibers in the lung or the serosal tissue is thought to be related to the potency of tumor induction (1,2), but it is important to distinguish between persistence and biodurability of a fiber (3). Persistence can be defined as the long-term residence of fibers in the same location. High biodurability is a precondition for the persistence. In most investigations the entire lung was ashed and fibers were analyzed from this sample; therefore for practical reasons persistence and biodurability are used congruently, although for mechanistic reasons of tumor induction there is a difference.
The methodological approaches to evaluate the durability of fibers include in vitro and in vivo testing. In vitro investigations on the dissolution rate of fibers attempt to use a solution that simulates the environment of fibers in the lung. This is not possible with just one solution, because fibers may be located in areas where different conditions exist such as the alveolar space, in the inter-This paper was presented at the Workshop on Biopersistence of Respirable Synthetic Fibers and Minerals held [7][8][9] September 1992 in Lyon, France.
Parts of this study were funded by the German Bundesanstalt fuOr Arbeitsschutz stitium, or in alveolar macrophages, where fibers would be subjected to the acid medium of lysosomes. In vivo investigations start with inhalation or intratracheal instillation of fibers into experimental animals after which the clearance of fibers from the lung is measured by serial sacrifices. The mechanisms responsible for clearance are dissolution, breakage of fibers, and physical dearance ofwhole fibers. Element analysis may also provide important information on durability (4).
Clearly in vivo elimination of fibers may differ considerably from results obtained by in vitro dissolution, not only because of the complexity of the different dissolution conditions in the lung, but also because other important mechanisms are involved. The slower processes of breakage and alveolar clearance can be ignored for very soluble fibers, but often mineral fibers have quite a low in vitro solubility (5) and therefore investigation of in vitro dissolution alone may result in misleading condusions.
Principles and Limitations of the Investigations of Biodurability Application Method: Inhalaton versus

Intratracheal Ins nllaaon
The inhalation route is usually the appropriate method to deliver fibers to the lungs of experimental animals. Under certain cir-cumstances, like the availability of a limited amount of sized fibers, intratracheal instillation may also be used (6,7). An advantage of the intratracheal instillation is that a precise starting point for kinetic analysis can be defined.
The evenness of the distribution of the fibers should be checked by SEM or by careful examination of the fiber suspension before treatment, to avoid the formation of granulomas, which were reported at an intratracheal injection dose of 20 mg (7). Comparison of results of inhalation studies with those where intratracheal instillation has been used should be done for validation.

Physical Mechanisms of Fiber Elimination
The key mechanisms for the physical elimination of fibers from the lung are mucociliary clearance; alveolar clearance, which is mostly mediated by alveolar macrophages; transport via lymphatic channels; and mechanical migration of fibers in lung tissue due to the respiratory tidal movement (8). These forces also may be responsible for the breakage especially of partly disintegrated fibers in the lung.
The elimination of durable fibers from the lung is dependent on physical mechanisms, which are influenced by fiber length, fiber diameter, and fiber mass retained in lungs. Alveolar macrophages can only phagocytize fibers up to 10 pm in length, prior to their removal by the ciliated airways (9,10).

Retained Mass
The phenomenon of "dust overloading" of the lung is observed with isometric particles at a lung burden of about 1 mg/g lung in rats (11). A similar effect was observed with fibers (10). Fibers may possess an "intrinsic" toxicity, which can affect the integrity of alveolar macrophages and thus their ability to eliminate fibers at a lower retained mass compared to isometric "nuisance" particles. As a result, the kinetics of fiber elimination from the lung are influenced not only by diameter and length distribution, but also by the retained fiber mass. The dependence of the physical clearance mechanisms on all these parameters makes it difficult to compare biodurability results obtained for different fiber types.

Methods ofCharacterizing the Fiber Elimination Process
To obtain an exact characterization of the fiber elimination process requires the total count of fibers retained in the lung and their length and diameter distribution for each date of serial sacrifices. Fiber length and diameter are usually log-normal distributions, and can be described on a log-normal plot (Figure 1), which can be plotted for each sacrifice date, for comparison with the original stock material. It is possible to calculate half-times as a characteristic number, if, in a semilog plot, an approximation to a straight line is obtained. In this method, a regression line is calculated, based on the logarithm of the fiber mass or fiber number at each sacrifice date. The clearance rate k, and the 95% confidence limit are given by t112 ± t112(CL)= 1n2 k± ksE*t where ksE represents the standard error of the regression coefficient and t is derived from the t distribution. Although this method leads to a considerable loss of information, it does make it possible to compare the elimination kinetics of different fiber types.
There are certain problems with this method, however. First, the dissolution kinetics is dependent on the fiber diameter, making it difficult to compare monodisperse fibers with log-normal distributed fibers, which in addition may have different breakage patterns. Second, with an increasing lung burden of particles, the fraction deposited in lung compartments from which there is low physical clearance will increase ("sequestered areas" [12]). This may be due to encapsulation of fibers by fibrotic tissue, leading to a considerable retardation of the elimination kinetics

Materials and Methods
The sized glass fiber X607 (Manville Technical Center) consists of CaO: 38.3%, and SiO2: 59.6%. The methods and results for this fiber are presented here as an example for other fibers. The source of the other materials and the methods used are described in detail elsewhere (4,(14)(15)(16).
To characterize the samples, the fibers were suspended in water, briefly sonicated (<1 min), and filtered onto a Nuclepore filter. Fiber number and size distribution were determined from transmission electron microscopy (TEM) or SEM photos. For the majority of the samples, two or three magnifications were chosen so that both the longest and the thinnest fibers could be measured with sufficient precision. To avoid double counting, different fiber length limits were set for the counts at each magnification. The size distribution was approximated to log-normal for all fiber types and was classified by the limits for length and diameter of 10, 50, and 90% of the fiber number.
Two milligrams of fibers, suspended in 0.3 ml of 0.9% NaCl solution were instilled intratracheally into 200 to 220 g female Wistar rats (Central Institute for Laboratory Animal Breeding, Hannover). Three to six rats of each group were sacrificed. After drying, the lungs were subjected to low-temperature ashing; the ash was briefly suspended in water and filtered on a Nuclepore filter (pore size, 0.2 pm, 0.4 pm for glass fibers). The lungs of each animal were analyzed as separate samples. For fibers with short residence times, only samples of the earlier sacrifice dates were analyzed, whereas for persistent fibers samples up to two years after treatment were investigated. Usually 200 fibers per animal were analyzed for size distribution. Further details are presented elsewhere (17).

Results
We analyzed fibers in the ashed lung of rats that had received the glass fiber X607 by intratracheal instillation 2 and 14 days, 1, 2, 3, and 6 months before sacrifice; the elimination kinetics were plotted logarithmically ( Figure 2). Calculated half-times, which are short compared with other glass fibers, are given in Table 1  Clearance kinetics for five fractions of increasing fiber length were also plotted ( Figure 3) and the corresponding halftimes were calculated ( Table 2). The halftime for fibers >20 pm in length was short, indicating that breakage and dissolution were responsible for the elin fraction. For particles < 2.' the half-time calculated frc particles was 68 days, corres value of the alveolar clearar isometric particles, which was approximately 60 days (11). The data from a series of fibers that had been studied earlier (4,14) were recalculated in similar fashion (Tables 3, 4).
The results from recalculation were markedly different for glass fiber 104/E, since in the earlier calculations only the values after 1 and 180 days were considered. As the fiber count had increased after 365 days, these values were incorrectly omitted. The original data set is given in Table 5. Another methodological limitation in the earlier experiments was due to the low fiber number remaining after 1 and 2 years, necessitating very long and tedious examination of the SEM and TEM photographs of the filtered fibers. Parts of the earlier study (4) are being repeated, this time adopting current methodology; five animals are sacrificed at each time interval, and in each lung 200 fibers will be analyzed, making a total of 1000 fibers, which should be sufficient for rigorous statistical analysis.

Discussion
Intratracheal instillation must be performed very carefully to avoid agglomerations of particles and to ensure an even distribution of the particles in the lung. It has been shown that instillation of 2 mg of fiber repeated 10 times resulted in a distribution in the lung very similar to that resulting from inhalation (7). In the present studies a single intratracheal instillation of 2 mg was used, and SEM examination indicated a quite even distribution. Recent results with long fibers show, however, that even 2 mg of fibers may result in agglomer-200 ations. The use of multiple instillations was suggested to achieve optimum dust distribution with a limitation of the doses administered to a maximum of 0.6 mg/g fresh ie after intratracheal weight of lung (18).
In another study of the glass fiber x 607, a half-time of 77 days was calculated nination of this (19), compared with 46 days reported here.

pm in length
A reason for this difference may be the )m the mass of low statistical power, due to the limited sponding to the number of animals used in the inhalation ice of insoluble study, and because intervals examined were only one and two years after exposure (19). It is recommended that in chronic inhalaing.
tion studies larger numbers of animals are necessary in retention and clearance experiments. A further difference between the two studies is that the fibers were adminis-Mass of particles tered to the animals during one year in the lean (95% CL) inhalation experiment (19),  (Table 4). One When the dissolution rates of 30 glass into slow clearance compartments in glass microfiber (104/475), however, fibersofdifferentcompositionwerestudied the lungs. appears to have a very long half-time of at pH 7.4, a 1000-fold difference was Calculation of clearance half-times 1794 days for fibers >5 pm in length, but observed between the most durable fiber, from results of another inhalation experi-the 95% confidence limit of 301 days-oo with a rate of 0.9 ng/cm2/hr and the least ment (20) using the same methods as those would suggest that the number of experi-durable glass fiber type, with a rate of described here are given in Table 3, and mental animals and the number of fibers 869.9 ng/cm2/hr (5). Where the dissoluthese are in relatively good agreement with analyzed may have been too low for adetion rate was measured in vitro and the biothe half-times derived from intratracheal quate statistical analyses. durability was determined in vivo for the