In vitro chemical and cellular tests applied to uranium trioxide with different hydration states.

A simple and rapid in vitro chemical solubility test applicable to industrial uranium trioxide (UO3) was developed together with two in vitro cellular tests using rat alveolar macrophages maintained either in gas phase or in alginate beads at 37 degrees C. Industrial UO3 was characterized by particle size, X-ray, and IR spectra, and chemical transformation (e.g., aging and hydration of the dust) was also studied. Solvents used for the in vitro chemical solubility study included carbonates, citrates, phosphates, water, Eagle's basal medium, and Gamble's solution (simulated lung fluid), alone, with oxygen, or with superoxide ions. Results, expressed in terms of the half-time of dissolution, according to International Commission on Radiological Protection (ICRP) classification (D,W,Y), varied for different hydration states of UO3, showing a lower solubility of hydrated UO3 in solvents compared to basic UO3 or UO3 heated at 450 degrees C. Two in vitro cellular tests on cultured rat alveolar macrophages (cells maintained in gas phase and cells immobilized in alginate beads) were used on the same UO3 samples and generally showed a lower solution transfer rate in the presence of macrophages than in the culture medium alone. The results of in vitro chemical and cellular tests were compared, with four main conclusions: a good reproducibility of the three tests in Eagle's basal medium the effect of hydration state on solubility, the classification of UO3 in terms of ICRP solubility criteria, and the ability of macrophages to decrease uranium solubility in medium.


Introduction
Uranium trioxide (U03 * xH20) is an important intermediate compound in the uranium ore treatment cycle. At the Comurhex factory in Narbonne, France, uranium trioxide is formed by calcining ammonium diuranate (ADU) between 3500 and 400°C. Industrial U03 obtained in this way is more or less hydrated but may still contain traces ofADU or U308 (uranium sesquioxide). In dust form, this compound may be inhaled by the worker, and to protect individuals it is essential to know its physicochemical properties and solubility in vivo.
To determine the rate ofdissolution ofcompound in the lungs, various in vitro chemical and cellular solubility tests have been carried out with alveolar macrophages to categorize U03 within the D, W, or Y transferability classes as defined by the International Commission on Radiological Protection (ICRP) (1). Ifthe materials are well characterized, in vitr tests will enable the problems ofdissolution to be investigated and the mechanisms concerned, such as oxidation or complexation, to be elucidated. However, these rapid in vitro tests, which use simulated biological liquids, as well as macrophage cultures, must be validated by in vivo tests following administration by inhalation.
The results ofin vivo experiments agree that the transferability ofU03 is ofthe D type, as Morrow et al. (2) discovered in dogs and Stradlingetal. (3) in rats. However, the results ofin vitro tests using different solvents are much more variable. For example, Cooke and Holt (4) obtained a class Wdissolution in a simulated lung liquid, Pasquier and Bourguignon (5) a class D dissolution with a simulated serum, Kalkwarf(6)50% class D, 50% classY with a Gamble's solution, and Eidson (7) a class D behavior. Finally, Stuart et al. (8) studied the hydration ofU03 and its dissolution mechanisms in saline solution and showed that the solubility ofU03 increased with its hydration state. However, the solubility ofU03 has yet to be measured using in vitro cellular tests.
The aim of the present research is to suggest an in vitro methodology that can be used to characterize the physicochemical properties of any sample of U03 under investigation. Three in vitro tests, one chemical (9) and two cellular, were developed to determine the solubility ofU03 in various solvents or biological environments and to shed light on the dissolution mechanisms. The two in vitro cellular tests were carried out using alveolar macrophages from rats, employing two different protocols. Inone case, the macrophages were maintained in a gas phase using a method adapted from that described by Voisin et al. (10). In the other case, the macrophages were incorporated into alginate beads, using the method ofLirsac et al. (11). The U03 hydration problem, which results from aging ofthe dust with the natural humidity of air, is discussed and probably explains the variable results found by the different investigators.

Samples of U03
The U03 industrial dust from the Comurhex factory in Narbonne (France) takes the form of small agglomerates ofvarious colors measuring between 1 and 10 mm in geometric diameter. The different-colored agglomerates were separated manually and powdered to give a grain size between 1 and 10 ytm. Four basic samples were obtained and characterized by X-ray or infrared spectrometry (IR). These samples included the original mixture or industrial U03 (A), the orange U03 extract (B), the yellow ADU (C) and an oxidized compound U03 and U308 mixture (D).
Examination of aging on samples A and B with air contact showed the production ofyellow spots on the orangebackground ofthe dust. This led us to prepare additional samples to study the effect ofhydration state on solubility: the cx and f3 samples were prepared by heating in a furnace at 37°C at 100% humidity to give a quicker aging or hydration of samples A and B and the -y sample, which corresponds to the B compound dehydrated in an oven at 450°C. The five samples used in the present investigation were A, B, ca, ,B, and zy, details of which are shown in Table 1.

Physicochemical Properties
The characterization ofthe main physicochemical properties ofthe five samples, important for understanding in vitro tests, involves the following tests.
Paile-Size Distributon. Dust samples, ground to produce particles between 1 and 10 ptm in diameter, were aerosolized and the airborne particles sampled with a Andersen Mark II cascade impactor device. The fourth stage ofthis apparatus corresponded to an activity mass aerodynamic diameter (AMAD) of 3.3 ,umm. X-Ray Diffraction. This technique allows the crystalline form of the constituent compounds to be determined. The apparatus used was a Philips PW 1730 spectrometer. Solid InfraredAnalysis. This technique, similar to X-ray diffraction, gives a characteristic absorption spectrum ofthe vibration bands ofthe constituent compounds atom groups. Analyses were carried out with a Fourier IR converting spectrometer (1760 X, Perkin Elmer).
Each test was conducted over a 2(-day period, and the collected samples were analyzed by fluorimetry with a FDTU1 fluorimeter (CEA, France). The results are given as a nondissolved uranium fraction (F) with a time functionF = 2fi exp-(0.693t/hi), wherefi is the initial fraction ofthe i compound and 71 the dissolution half-time. The sum i = 1 or 2 was computed using a nonlinear regression program.
In Vitro Cellular Tests Alveolar Macrophages. The alveolarmacrophages (AM) were harvested by bronchoalveolar lavage from rats (OFA strain), anesthetized with pentobarbital, and sacrificied by exsanguination.
Phagocytosis. A suspension ofU03 (AMAD = 3.3 jAm, ag = 1.7, concentration = 110 mg/L) was prepared in the cell culture medium and sonicated for 10 min before contact with the cells. The particles were incubated with the macrophages for 1 hr in a culture flask. The adherent cells were washed with a saline solution and either transferred for the assays in gas phase or added to the alginate solution to form beads.
CeUl Survival. CELLS IN GAS PHASE. After phagocytosis, the macrophages (1 x 106 cells per plate) were deposited on a Gelman membrane (0.2 itm pore size) and applied to the surface ofa reservoir filled with a nutrient medium (10). The cells were in direct contact with air that had been saturated with water at 37°C and enriched with 5% CO2. For sample A, the nutrient medium was medium 199 (Gibco) and for sample B, the nutrient medium was BME (Biomerieux). Both media were supplemented with 10% fetal bovine serum, 2 mM L-glutamine, kanamycin, penicillin, and streptomycin. The concentration of uranium was 2 ,ug/106 cells. Control plates with particles alone (i.e., without cells) were prepared to determine the dissolution rate of U03 in the nutrient medium.
CELLS IMMOBRIZED IN BEADs. A suspension ofalveolar macrophages that had engulfed the U03 (2 jAg/106 cells) was added to an alginate solution (2%) to obtain a final concentration of 1.5% . This suspension was then extruded through a succession of catheters with decreasing internal diameters, according to the method described by Lirsac et al. (11). The droplets were allowed to fall into Gibco culture medium supplemented with 10% fetal bovine serum and 10 mM calcium, thus forming beads made up ofan alginate network. Control beads without macrophages were prepared with the U03 suspension (2 ,ug/106 cells) to test the dissolution rate in the nutrient medium.
CeUl Viabii. The viability ofcells maintained inthe gas phase was estimated by measuring the adenosine triphosphate (ATP) content using the method ofMcEbroy, as modified by Voisin et al. (10). The viability of the cells immobilized in beads was estimated by measuring chemiluminescence at 37°C with luminal according to the method of Dyer and Wesleid (14).
Twenty-five beads containing a total of 0.5 x 106 macrophages were added to 200 uL of 10 3 M luminol. The cells were stimulated by 200 AL ofzymosan opsonized with rat serum, and the chemiluminescence was measured during 30 min.
Measurementof Uranium Dissolution. Each day, the nutrient medium contained in the reservoir (AM in gas phase) or in the flask (AM in beads) was replaced, and the amount ofsolubilized uranium in the medium was assayed. At the end of the experiment, the uranium concentration remaining on the membranes or in the beads was assayed, making it possible to calculate the total amount of uranium present initially.

Physicochemical Properties
The results ofthe X-ray diffraction and IR characterization of the U03 samples are given in Table 1. Examples ofthe IR spectra of samples B, (3, and -y (corresponding, respectively, to the U03 extracted on its own, after hydration, and after drying 1 day in an oven), are given in Figure 1. The change ofwavelength of the uranyl peak is due to the hydration rate of U03. In Vitro Chemical Tests In Vitro Cellular Tests Table 3 shows the results for the dissolution tests of samples A, B, ,3, and y in the gas-phase test. The results are shown as percent ofuranium solubilized in 24 hr and highlight the effect of hydration on solubility, especially for compound B.
The results for cumulative percentage dissolution of industrial U03 (sample A), are shown in Table 4. The results obtained with the gas-phase and alginate-beads protocols are compared. The average quantity of uranium linked to the macrophages after phagocytosisandbeforethetests was0.27 ± 0.11 tg (n = 21).
The dissolution half-time for sample A in macrophages was 12.2 days for the gas-phase test and 10.8 days for the beads test. The corresponding test reference values were found to be 4.6 and 5.5 days, respectively.
Measurement ofATP viability in alveolar macrophages with U03 and in control cells maintained in gas phase is shown in    Figure 2. Figure 3 shows the measurements of chemiluminescence for reference control macrophages and the macrophages with phagocytized U03 in the beads test. Results are given for different days (1-4 days).

Discussion
Observation ofthe hydration state ofthe extracted U03 (sample B) during chemical and cellular tests (Table 3) led to a closer examination ofthe physicochemical properties ofthe A, B, a, , and y compounds as well as their in vitro reactions.
X-ray comparison ofcompounds A and B (Table 1) shows that these compounds are quite similar, with traces ofU308 and ADU associated with the industrial U03. On the other hand, Figure 1 shows the development ofa single uranium peak at 940 cm-' for ,B, whereas B gives two distinct peaks at 914 and 798 cm-'. The IR spectra are the same for compounds A and B. Addition of water molecules, which occurs with aging of dust, induces a chemical transformation. The -y spectrum ofcompound B heated at 450°C shows the disappearance of the 3430 and 1620 cm-' bands in water and a movement ofthe two uranium peaks at 903 and 782 cm-'. Thus, hydration and heating change the chemical composition of U03.
This physicochemical study demonstrates that the processes involved with U03 dust hydration (i.e., industrial A and extracted B) change the crystalline structure by preferential bonding with H20.
The results ofin vitro chemical assays in various media ( Table  2) lead to several interesting conclusions: a) The hydrated dust, i. e., samples ca and ,B, are the least soluble in all media, the exception being NaHCO3, in which solubility is very rapid with a half-time of only about 1 day. b) Dehydrated U03 -y is the most soluble ofall the compounds. c) Carbonates and citrates easily solubilize U03 in all its various forms with a class D behavior of between 1 and 3 days. Solubilized uranium in the uranyl forms gives carbonate and citrate soluble complexes as described by Hodge et al. (15), which have the compositions UO2(C03) 3 and U02C6H507, respectively. d) All compounds of U03 are rendered insoluble by phosphates, giving half-times longer than 500 days, i.e., a class Y behavior. Insoluble complexes created, such as (UO2)3(PO4)2 or UO2HPO4, with stability constants (Kj) of -49.1 and -10.7, are also described by Hodge et al. (15). e) In Gamble's solution alone, compounds generally have a mixed behavior, between class W and Y, due to the competition between phosphates and carbonates. The phosphatic complexes prevail, however, resulting in a class Y tendency. The addition of 02 to Gamble's solution gives a class D-type high solubility to the U03 due to the accelerated oxidizing of the U03* xH2O dust in a UO2 solution. The action of the superoxide ions in Gamble's solution, due to the addition of pyrogallol with oxygen in the presence of SOD, is more moderate than the action of 02f) The BME or cellular culture medium is richer in proteins than Gamble's solution and gives a reaction of the class-D type with the compounds A, B, and 'y, and W type with the hydrated a and (3 compounds. g) The natural dust hydration in air atmospheres is different from hydration in water or solvent. Samples A and B, which are orange at the start, quickly turn yellow in water (hydration) and dissolve very quickly. In contrast, samples a and (3 hydrated in air dissolve more slowly. With both in vitro cellular tests applied to the industrial compound A (Table 4), the effect of the macrophages is to reduce solubility. A similar effect can be seen in Table 3 when com-pounds A, B, (3, and -y were tested in gas phase. Variable results with U03 compound B were obtained over a 9-month period and have revealed the problem of transformation of the dust due to hydration. Sample B was not stable in time, which explains its decrease in solubility from 38.3 % to 12.8%. The observed reduction in solubility may be due to interactions between the dust and constituents of the cellular membrane or to a compound insolubilization by lysosomes. The intracellular precipitation of uranium in the form ofuranyl phosphate flakes has already been demonstrated in other cell types by Galle (16) and is confirmed by the appearance of insoluble compounds (t,h < 500 days) when phosphates are used in the in vitro test ( Table 2). In the lungs, compounds and precipitates are removed by a mucocilary action and eliminated via the gastrointestinal tract.
Decreasing solubility with increasing hydration has been observed (Table 3), confirming the in vitro chemical tests results ( Table 2). The two types oftest gave results that can be compared between themselves and also with the BME in vitro chemical test. The gas-phase test, which is much easier to implement, seems suited to research on rapidly soluble compounds. It is difficult to maintain cells in a satisfactory state for more than 5 days, but, with the more complicated test using alginate beads, the macrophages survive longer, providing a useful method for studying less-soluble compounds. When the macrophages are maintained in gas phase, an increasing intracellular ATP concentration (Fig. 2) is sometimes observed after survival for 24 hr; this is probably due to a change in cellular activity related to phagocytosis. For the three tests, the A has a class D solubility with an average half-time of4-6 days without macrophages and a limit behavior between D and W with macrophages, the average half-time being 11.5 days. Good agrement between the three tests shows that BME is a representative medium for in vitro chemical testing.

Conclusion
By using in vitro tests on a U03 uranium compound with variable hydration levels, we could compare three different techniques for measuring dissolution. Use of these rapid and easily implemented techniques improves the understanding ofthe mechanisms that underlie the dissolution of a dust in the lungs and enables materials to be classified using ICRP criteria by taking into account their physicochemical characteristics.
The three tests give entirely comparable results and complement one another. The in vitro chemical test has the advantage ofbeing simple to use and may assist in understanding solubilization or complexation phenomena in various solvent systems (i.e., carbonates, phosphates, citrates) and may apply to any class D, W, or Y compound. The addition of a macrophage in vitro cellular test is an essential complement and reveals insolubilization phenomena due to macrophage action. The gas-phase test is fairly easy to implement and applies more to class D (rapidly soluble) compounds, with which the test may be completed in 4-5 days. The macrophage beads test is more difficult to use but applies to class D and W compounds when the effects ofmacrophages can be investigated that period of up to 2 weeks. The main results revealed by this research are a) U03 compounds are able to hydrate with time, which might explain the discrepancies in the results ofmeasurements of solubility by different authors; b) the solubility of U03 decreases with increasing hydration in most solvents, as well as the BME for the three tests; c) the results for the industrial compound A fbr the in vitro techniques in BME medium (average half-time of4-6 days) compare well with those in Gamble's synthetic medium +02 (halftime of 5 days); and d)the distinct insolubilization phenomenon observed in both cellular tests is due to the macrophage action and is reproducible for any hydration level. All these in vitro results should be compared to in vivo results for each ofthe given compounds after inhalation by rats.