New approaches to immunotoxicity testing.

New approaches to immunotoxicity testing are reviewed and discussed. A method of activating T-cells in vivo is presented which circumvents artifacts dur to viability effects encountered with in vitro mitogen assays. The use of adoptive transfer approaches to combine the advantages of in vitro manipulation with in vivo function assays is discussed relative to natural killer cells. The need for an in vitro metabolic activation step coupled to other in vitro immunologic assays is discussed.


Introduction
Previous papers have demonstrated the importance of the immune system with regard to the well-being of all animals. It has also been clearly demonstrated that chemicals can modify the immune response, some for the benefit of the animal, e.g., immunopharmaceuticals, and some in potentially harmful ways, e.g., chemical wastes and environmental pollutants. Adequate coverage of the new approaches to immunotoxicity testing would require a separate symposium. The volume of new methods being applied to all sorts of immunologic investigations is enormous, and the list of literature reports increases monthly. Thousands of compounds and their effects on the immune system require examination, and rapid, economical screening procedures would certainly benefit industry as well as health and regulatory agencies.
General Approaches to Testing for Immunotoxicity Table 1 shows the basic ways in which immune response tests may be carried out. In the first *Department of Health and Human Services, Public Health Service, Food and Drug Administration, Division of Microbiology, 1090 Tusculum Avenue, Cincinnati, Ohio 45226.
February 1982 category of tests the cells involved in a particular immune response are exposed directly to the chemical in vitro and the test for immune function is also carried out in vitro. Although this approach is the most economical in terms of time and animals, it has some problems, most notably: metabolic alteration of the compound may not be accounted for; distribution of the -chemical into the immune compartment is assumed; and cells are more susceptible to damage in the relatively dilute in vitro microenvironment. A compromise approach in terms of realism and economy calls for in vivo exposure to the chemical followed by in vitro testing of immune function of various lymphoid tissue. This is clearly the most common approach currently used in immunotoxicity studies. The major advantages of this approach are that distribution of the chemical to the immune compartment is realistic and metabolism of the compound may occur naturally. This approach will be discussed further. Lastly, animals may be exposed to the chemical in vivo and immune functions tested in vivo. Although this is the most Returning to the in vivo exposure/in vitro analysis scheme, some disadvantages with this approach have recently been noted.

Lymphoproliferative Assays
Figure la depicts the standard lymphoproliferative assay (or mitogen assay) as currently performed. Although chemical contact occurs in vivo, it was thought that preparation of cells for in vitro culture might remove the chemical; thus chemical and activator (or mitogen) might not be present simultaneously. Conversely, cells might be metabolically or physically damaged by the chemical, making them appear nonfunctional in the more hostile in vitro situation, but compensatory mechanisms in vivo might permit normal functions to occur. To test these hypotheses, we used staphylococcal enterotoxin A (SEA), a T-cell specific mitogen (1) with low toxicity to the mouse and with the added advantage of a broad plateau of activation rather a a -X than the sharp dose-related peaks commonly seen with the lectin mitogens concanavalin A or phytohaemmagglutinin (2). SEA was shown to be an in vivo T-cell activator (3). The only in vitro culture required was a 4-hr period for 3H-thymidine uptake (Fig. lb). Thus, chemical and activator could be present simultaneously in vivo. During these studies, we noticed that certain agents administered in vivo, such as antithymocyte serum (ATS), which disrupts T-cell membrane integrity, showed a greater suppressive effect on in vitro mitogen responsiveness at far lower doses than were required to suppress in vivo activation of T cells by SEA. Table 2 shows the effects of ATS on in vitro survival of cells, as determined by vital dye staining. Spleen cells from ATS-treated animals were 97-98% viable at the time of harvest. Four hours of culture had only a slight negative effect on in vitro survival of ATS-treated cells. The rate of cell death of 48-hr cultured cells (the routinely used LP assay time) was greatly accelerated by ATS pretreatment compared with the in vitro death rate of controls. Magnification of the apparent damage in the in vitro culture is advantageous because greater sensitivity is attained; however, further work has shown that realism may be compromised. Experiments have indicated that the shorter the in vitro STANDARD MITOGEN ASSAY  incubation time, the less the chance of amplified effects on viability.
In Vitro Manipulation-Adoptive Transfer In a recent report concerning natural killer cells, Kasai et al. (4) demonstrate how the concept of short in vitro culture time might be applied. Natu-I NYLON ral killer (NK) cells have become the focus of many recent studies on antitumor activities of the immune system because they appear to be one of the firstline defenses against virus-or chemical-transformed cells and have a functional counterpart in humans. Kasai et al. attempted to demonstrate that cell subpopulations which expressed NK activity in vitro were actually responsible for NK activity in vivo. The antitumor role of NK cells in vivo had not yet been conclusively shown. The experimental approach taken by these investigators demonstrates how several new technologies could be applied to immunotoxicologic investigations. Figure 2 depicts the separation scheme used by Kasai et al. Their procedure, which could be applied to chemicalexposed animals as well, was as follows.
Spleen cells were removed from mice. Some cells were placed in plastic dishes coated with antimouse Fab antibody; in this way, immunoglobulin (Ig)bearing cells (B-cells) were enriched and non-Igbearing cells were removed. These populations were retained for testing. Other spleen cells were passed through a nylon wool column known to retain Igbearing B-cells. Thus, T-cells bearing Thy 1 antigen were enriched as were other non-Ig-bearing cells. Some of these cells were treated with anti-Thy 1.2 antibody + C', to specifically lyse the mature, Thy

1-bearing T cells. Monoclonal antibody technology
currently assures the purity of such antibody preparations. The Thy 1-cells were found to have Ly 5 surface antigen by fluorescent antibody methods. The Thy 1-Ly 5+ cells were coated with anti-Ly 5 antibody, but without complement because no lysis was desired. The coated, viable cells were placed in dishes with protein A-coated sheep erythrocytes. Protein A specifically binds to the Fc portion of antibody molecules; the Ly 5+ cells were bound to erythrocytes by the protein A-antibody bridge and could easily be separated from nonbound cells. The erythrocytes were lysed by distilled H20 shock and the Ly 5 + cells were freed and retained for testing. These cells were 82-87% pure as determined by fluorescence testing with specific antisera. The total in vitro time for these cell separations was relatively short, probably less than 4 hr. Figure 3 shows how the isolated cell populations from each step of separation were mixed with RL1 or YAC tumor cells and adoptively transferred to appropriate mice. The in vitro manipulated cells were thus returned to an in vivo environment. Table 3 depicts the results of the adoptive transfer conditions. Spleen cells alone, although they contained the NK cells, would not demonstrate protection, probably because of insufficient numbers of NK cells. Not until the Ly 5+ Thy 1-cells were highly enriched could protection be seen, but then it was quite remarkable. Further, lysis of this population by Ly 5 antisera showed that the population carrying this surface antigen was definitely involved in protection from the tumor. Such manipulations are impossible in a totally in vivo situation.
By such separations, NK cells or other desired populations could be isolated in vitro, exposed to chemical treatment to assess direct effects, particularly if in vitro metabolic activation could be carried out, and returned to the natural in vivo environment in a syngeneic host for testing. Likewise, such tests could be done on cells from chemical-pretreated animals. 112 Combining in vitro manipulation with adoptive transfer and in vivo functional analysis is one way in which immunotoxicologic studies might be carried out using currently available methodology and averting possible erroneous conclusions caused by problems associated with in vitro survivability.

New Approaches to Screening for Immunotoxicants
Recently, Kutz et al. (5) have applied the Mishell-Dutton in vitro antibody-forming cell system to screening a large number of environmentally relevant compounds. Many of these chemicals were shown to be immunosuppressive but directly cytotoxic to lymphoid cells in vitro at suppressive levels ( Table 4). These chemicals included the organotins, several heavy metals and polychlorinated biphenyls. Others, shown in Table 5, were suppressive but not cytotoxic. Most notably, azathioprene was suppressive in vitro at levels easily attained in vivo  and with similar suppression kinetics. Thus, some degree of predictability was achieved with this system. Cyclophosphamide also had no effect on this system, owing to the lack of the required microsomal activation of this chemical. The need for an in vitro activation step in this potentially useful procedure is discussed by Tucker (6). Vos (7) and Boonnan (8) also report on new methods for assessing the effects of chemicals on the adult immune system and the developing immune system, respectively.

Conclusion
Immunology as related to toxicology is an exciting field of research. We are currently learning which methods are applicable to toxicity testing and improving and adapting others to make them suitable. While no one magic test will suffice, it may be possible in the very near future to report the development of more rapid, reliable, economical screening methods for a wide range of immune functions which have predictive value to the whole animal.