Humane Endpoints and Acute Toxicity Testing

Introduction

Toxicology is the science that studies the potential of materials for producing one or more deleterious effects on organisms. Many materials have greatly enhanced our quality of life by improving health and hygiene. These materials have also made significant contributions to agriculture, industry, and most phases of human life. However, some substances are capable of producing adverse effects to exposed humans and animals. Individuals’ and species’ susceptibility, route of exposure, dose, and duration of the exposure may influence the presence or severity of the effect. Acute toxicity testing is conducted to determine the effects of a single exposure to a substance. Acute effects typically become manifest almost immediately after a single exposure, although depending on the causative material and the mechanism of its action, a latent period may precede the manifestation of the effect(s). Subchronic and chronic toxicity testing is conducted to determine the existence of effects that become evident after an exposure of extended duration. It is important that toxic materials, whether they cause acute or longer-term effects, be identified so that procedures and practices can be developed and implemented to prevent injury and disease.

Regulatory bodies associated with governmental agencies and corporations involved in the development of materials for inclusion into products that enter the marketplace must ensure that materials to which the public is or may be exposed are safe. One basic premise in toxicology is that data should be derived from systems indicative of those seen in humans, that is, the output from the source or test system should make possible extrapolation to humans. The historic approximation of the risk to humans has been accomplished by exposing animals to the materials in question. In this article, I discuss methods used for acute toxicity testing, humane endpoint considerations for these tests, and some new approaches to further reduce animal use. Three acute toxicity endpoints are discussed: acute lethal toxicity, dermal irritation/corrosion, and acute ocular irritation.

Acute Lethal Toxicity Testing

According to legend, miners took canaries into the mineshafts to serve as warnings of the presence of poisonous gaseous atmospheres. If the canary succumbed during exposure to the same conditions in which the miners were working, the miners would know to vacate the area. This type of indication provided immediate feedback to the miners that conditions were hazardous. The immediacy of the exposure-effect relation indicated an acute toxicologic effect, one in which “the adverse effects occurring within a short time after (oral) administration of a single dose of a substance or multiple doses given within 24 hours” ( Svendsen 1994 , p. 9).

The goals of acute lethal toxicity testing ( Svendsen 1994 ) include the following: defining the degree of hazard that may result from exposure to a test substance; determining susceptible populations and species; identifying target organs or systems; providing information that can be used in developing risk evaluations; and providing information to clinicians that will enable them to predict, diagnose, and/or provide treatment for acute exposures. Acute toxicity testing requires test materials to be given to animals for a finite but short period of time, usually as a single exposure. A test material can be administered by various routes to determine its ability to induce toxicity, including oral, dermal, and inhalation exposures. For acute lethality testing, rats and mice are usually the species of choice, largely due to the vast amount of background data that has been assembled through history. Death is no longer required as the endpoint indicative of acute toxicity. Endpoints have been developed and validated that permit animals to be humanely killed when in a moribund condition while still fulfilling their role in the development of toxicological information.

LD ₅₀ Lethality Testing

Trevan (1927) introduced the concept of median lethal dose (LD ₅₀¹ ) as a means of biologic standardization of dangerous drugs. For many substances, the test material was administered by gavage and by at least one parenteral route. Precision could be obtained only when large numbers of animals were used. The desirable outcome of an LD ₅₀ assay is obtained when more than half but less than 90% of the animals died after being exposed to the high dose, and less than half but more than 10% of the animals died from the low dose. A third dose would ideally result in the deaths of 50% of the animals. Four doses were used to improve the possibility that some of the doses used would fall within the desired ranges. Although the requirement of large numbers of animals (up to 200) was recognized as a severe drawback, the LD ₅₀ was incorporated into testing protocols for classification of acute toxicity for many classes of chemicals.

Alternatives to Oral LD ₅₀ Testing

Limitations of LD ₅₀ testing procedures, coupled with increased concern over the use of animals in testing, have resulted in the development and validation of alternative methods of determining acute lethality toxicity ( Table 1 ). The Organization for Economic Cooperation and Development (OECD ¹ ) ( OECD 1981 ) has adopted these methods and advocated them as more humane assays for determining acute toxicity. The principal advantages of using these methods are that fewer numbers of animals are needed to obtain useful information and more humane endpoints (toxic signs and symptoms rather than lethality) are employed. Four methods have been adopted as substitutes for the LD ₅₀ procedure for assessing acute oral toxicity ( Table 1 ; OECD 1987a , 1998 ). These methods have been developed and, in some cases, subjected to validation through controlled interlaboratory trials. Although some of the methods have the capability of providing data that more closely parallel those obtained from the LD ₅₀ test, all are capable of providing the data required under the European Commission hazard classification and labeling system ( CEC 1983 ).

Schlede et al. (1992) conducted a study to validate one of those methods, the acute toxic class method, as an alternative to the LD ₅₀ procedure. The results indicated that the materials could be as effectively classified using the acute toxic class procedure as they had in the LD ₅₀ method. One major advantage to this method was that fewer animals were required and fewer animals were subjected to pain and distress than in the LD ₅₀ test while the same information was provided on toxic signs and symptoms as was obtained in the LD ₅₀ procedure. Yam et al. (1991) compared two of the methods (the up-and-down method and the fixed-dose procedure) against the classical LD ₅₀ to determine their acceptability as replacements for the traditional method. Both methods were found to reduce the numbers of animals used while providing adequate information for ranking the 10 materials tested according to European Economic Commission classifications for acute oral toxicity.

Acute Dermal Corrosion and Irritation Testing

Draize et al. (1944) originally published specifications that are still used to conduct primary irritant testing in rabbits. One rabbit can be used to test up to four materials on both abraded and intact skin. Three rabbits (replicates) are used for each of these trials. After exposure, each animal is observed for 72 hr, and the appearance and disappearance of toxic effects are recorded. In addition to fatalities and the effects seen on the skin at the site of application, any systemic physiologic effects are also noted. If test animals experience severe pain or distress, they are to be humanely killed irrespective of the presence/absence of dermal lesions. This procedure has been reaffirmed by the US Consumer Product Safety Commission (CPSC1) through various issuances of CPSC regulations from 1961 to the latest version in 1984 ( CPSC 1998 ).

Alternatives to Acute Dermal Irritance/Corrosivity Testing

Research has been conducted to refine the dermal testing protocols originally elucidated by Draize et al. (1944) . Calvin (1992) discusses some of the modifications to skin testing procedures directed toward the 3 Rs. For certain materials, either the fuzzy rat or the hairless guinea pig may prove to be acceptable alternatives to the rabbit in dermal testing (replacement with animals deemed lower on the phylogenetic tree). Techniques involving measurement of changes in cutaneous blood flow through treated skin could replace the subjective grading of skin lesions (e.g., edema, necrosis, or erythema). Data from studies using laser Doppler flowmetry, infrared measurements of temperatures in treated areas of skin, or anatomical skin parameters (e.g., skin thickness, magnetic resonance imaging visualization techniques) are being evaluated. Results could lead to the detection and validation of endpoints that precede the development of lesions currently used to assess toxicity. When this occurs, test animals can be humanely removed from the study before the development of pain or discomfort (refinement).

Corrositex®

Corrositex® is a test method for assessing the potential of materials to induce dermal corrosivity on contact with skin. Corrosive materials are those that cause irreversible destruction of tissue at the point of contact. This procedure is based on the ability of a chemical or mixture of chemicals to penetrate a collagen-based biolayer and cause a color reaction in a liquid chemical detection system beneath the barrier. The amount of time required for penetration and reaction with the chemical detector (color producing) medium is of significance in determining the degree of corrosivity of the test material. The Interagency Coordinating Committee on Validation of Alternative Methods (ICCV AM ¹ ) conducted a peer review of the data to evaluate this method as an alternative for in vivo dermal corrosivity testing. It was determined that Corrositex® can serve as a stand-alone assay procedure for evaluating the corrosivity potential of certain chemical classes ( ICCV AM 1999 ). It can accurately assess the corrosivity of acids, bases, and acid derivatives, thereby satisfying standards promulgated by the US Department of Transportation. For other classes of chemicals, Corrositex® can be used as part of a tiered testing strategy in which positive chemicals can be classified as corrosives with no further animal testing. If negative results are obtained, in vivo dermal irritation testing is required to confirm noncorrosivity.

Episkin™ and EpiDerm™

Episkin™ and EpiDerm™ are tissue culture-based alternatives to in vivo dermal corrosivity testing ( ECV AM 1998a ; Leibsch et al. 1997 ). These methods utilize a three-dimensional model of human skin with a reconstructed epidermis and a functional stratum corneum. Test materials are applied to the layer for prescribed time periods. Their effect on the skin layers are measured by the ability of live cells comprising the model to passively take up the tetrazolium bromide derivative (3-(4,5-dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide) and then act on it through mitochondrial reactions to produce the color response by reduction of the parent dye material. Skin that has been damaged by exposure to the test material will not be able to perform the requisite mitochondrial energy reactions.

Transcutaneous Electrical Resistance Test

The rat skin transcutaneous electrical resistance test is another alternative test for skin corrosivity ( ECVAM 1998b ). In this procedure, test materials are applied for 2 to 24 hr to the epidermal surface of skin discs taken from young rats. Corrosive materials are identified by their ability to produce a loss of integrity to normal stratum corneum, measured by the reduction in the transcutaneous electrical resistance of the layers of the skin to an applied current (5kΩ). Corrosive materials would damage the integrity of the stratum corneum and reduce the electrical resistance of the skin to transmission of current from one side to the opposite. The OECD tiered testing strategy for evaluating dermal corrosion and irritation allows for the use of validated and accepted in vitro methods ( OECD 1998 ) ( Table 2 ).

Ocular Irritation Testing

One of the most contentious areas of toxicity testing involves the assessment of risk when materials enter the eye. Knowledge of toxicities that might affect the eye is important for individuals who might become exposed when accidental entry occurs through spraying, splashing, or entering an atmosphere where the potentially toxic material has been either aerosolized or volatilized. Another major source of eye exposure risk is found in the use of cosmetics. Some cosmetic ingredients are designed for use in the area of the eye; other cosmetic products are to be used on other parts of the face or body but may accidentally enter the eye. Protection of humans from injury from such products requires knowledge of their potential hazards. The rabbit Draize test currently used to assess ocular irritation potential is one of the most controversial procedures because of its potential for severe pain and distress.

Ocular irritation testing is conducted to achieve compliance with national and international labeling requirements for hazardous materials. Countries have promulgated requirements to ensure that materials are shipped safely and properly and that consumers are aware, through labeling, of potential eye hazards from exposure to materials.

The current rabbit ocular irritation test is based on the original protocol developed by Draize et al. (1944) . In this test, the New Zealand White rabbit is the species of choice. Assessment of ocular irritation involves placement of approximately 0.1 mL of the test material (0.1 g of solids) into the conjunctival sac of one eye, after which animals are observed after 2, 24, and 48 hr; the alternate eye is used as the control. If residual injury is present, a subsequent reading is made at 96 hr. Injuries to the eye, including corneal damage (opacity), wounds to the iris (inflammation), and deleterious effects on the conjunctival or palpebral mucosae (redness and chemosis), as well as the length of time required before the conditions are reversed, are scored individually.

Alternatives to Ocular Irritation Testing

Reduction

Gupta et al. (1993) advocated a protocol that was directed toward reduction in the numbers of animals required to obtain data on eye-area toxicity. Structure-activity relationships, assessment of pH extremes, dermal irritation scores, and results from validated in vitro assays are considered before administration of the test article to any animals. When a material has been assessed for its potential to induce ocular irritation, it is administered to two to three animals. The animals are then observed and scored using a modified Draize scoring system at 24, 48, and 72 hr and 7 days. A rabbit is considered positive when it presents with corneal opacity ≥1, iritis ≥1, or conjunctival redness ≥2. Use of this protocol involves fewer animals than the original Draize method (six animals/test material). DeSousa et al. (1984) statistically analyzed the accuracy of predicted results obtained using three, four, or five animals when compared with traditional sixanimal tests. The test viewed studies of 67 different petrochemicals. The results indicated that tests utilizing three animals were 93% accurate in predicting the results obtained from traditional Draize eye testing studies. This reduction in the numbers of animals needed to provide accurate information on the potential of a material to induce eye area damage constitutes a significant step toward the achievement of the 3Rs goals of Russell and Burch (1959) . The success of trials like these influenced OECD member countries to endorse the reduction in the numbers of animals used in ocular testing from six to three ( OECD 1987b ).

A singular concern in the development and assessment of alternatives to whole animal testing for eye-area toxicity is that little is known about the mechanism by which injuries to the eye develop. For this reason, it is extremely difficult to generate a mechanistically based alternative. Of additional concern is the inability to duplicate or emulate the processes by which repairs to ocular damage occur. Surrogate models, or systems of methods, can only approximate what is seen in intact animals; it is contingent on investigators to develop the most scientifically defensible arguments based on grounded scientific arguments (i.e., data) to support advocacy of an assay method.

Refinement

Bruner et al. (1992) refined the Draize procedure to involve less potential pain and distress by validating low-volume ocular irritation assay. The low-volume assay involves the placement of 1/10th the volume of test material used in the standard Draize (0.01 mL of liquid or 10 mg of solid vs. 0.1 mL of liquid/0.1 g of solid) directly onto the cornea (instead of the conjunctival sac). The traditional Draize eye methodology was found to overestimate human ocular responses. The low-volume exposure more closely duplicated human exposures to the eye and would be more predictive of the human response ( Chu and Toft 1993 ; Freeberg et al. 1984 , 1986a,b ). The data obtained from subsets composed of three animals from 119 different low-volume animal trials utilizing six animals were compared with the results from the sixanimal trials. The three-animal subsets correctly predicted (92%) the results obtained from the six-animal trials. Little accuracy is lost by reducing the numbers of animals used for testing by 50%, and significant refinement results from this 90% reduction in volume of the test substance.

Replacement

In 1993, the Interagency Regulatory Alternatives Group (IRAG ¹ ) sponsored a workshop on eye irritation testing to reflect on the strengths and weaknesses of non-whole animal alternatives for assessment of eye area toxicity. An additional objective of the meeting was to determine the utility of a single assay or a battery composed of multiple assays for obtaining a complete picture of the effects of a material on the eye. IRAG workgroups reviewed 74 data sets from 59 laboratories utilizing 26 different test methods. The conclusion drawn after this review process was that “data are insufficient to support the position that non-whole animal methods, singly or in combination, currently are available to completely replace in vivo ocular irritancy testing” ( Bradlaw and Wilcox 1997 ). Subsequently, OECD developed a tiered strategy for testing and evaluating effects of materials that may cause eye irritation and/or corrosion and included the potential, future use of validated and accepted in vitro methods ( Figure 1 ). Significant efforts are ongoing to find valid in vitro methods, some of which are discussed here.

In vitro ocular assay methods are classified by type. Organotypic models are those that employ eye tissues from animals, including bovine corneal opacity tests, chicken enucleated eye tests, and isolated rabbit eye tests to assess generalized total eye damage or toxicity. The measurable endpoint in these assays is the development, and time to development, of corneal opacity. One major drawback to the use of this type of assay is the test system’s inability to emulate repair mechanisms seen in intact animals. An assay system composed of canine renal cells formed into a cultured barrier has also been developed. When irritant test materials act on this artificial barrier, barrier integrity is disrupted and the indicator material, fluorescein dye, is seen to leak from one side to the other (fluorescein leakage assay). The system was developed to determine whether test materials have an effect on the integrity of the barrier of the eye. One advantage to this system is that cells forming the barrier grow in that configuration; they can conduct some repair activities after removal of the irritant material ( Shaw et al. 1990 ).

Cell function assays such as chorioallantoic membrane (CAM ¹ )-based assays utilize CAMs of embryonated chicken eggs to assess corneal opacity and iritis as a measure of inflammatory responses. After exposure of the CAM to test material, the membrane is observed for signs of hyperemia, hemorrhage, or coagulation induced by the test material. This method has produced results comparable to in vivo tests using both water-soluble and -insoluble test materials ( Chu and Toft 1993 ). Variations in the CAM-based assay are useful for different classes of compounds ( Spielmann et al. 1997 ). EYETEX™ is a product that correlates alterations in physicochemical properties of macromolecules with ocular irritation. A medium containing proteins, glycoproteins, and mucopolysaccharides is exposed to the test material. If the material is an ocular irritant, it will react with components of the mixture and produce turbidity in proportion to the level of ocular irritation and the concentration of the test material ( Gordon and Bergman 1987 ).

Another class of methods for assessing ocular irritation potential are those that measure the effect of test materials on cells in culture. Of all the classes of methods tested, the cell cytotoxicity methods appear to possess the greatest ability to parallel results obtained in vivo. These methods utilize color reactions, including neutral red uptake, alamar blue reactions, and MTT assays to detect parameters indicative of cell viability, alterations in cellular functions, and cell damage after exposure to test materials. Continued work in developing a system that combines multiple assay procedures that assess a variety of cell parameters may result in identifying the materials that induce eye toxicity.

Worth and Fentem (1999) reviewed the use of a tiered testing strategy for the evaluation of ocular irritation potential of materials. Such a tiered approach may assist in the refinement and/or reduction of animal testing procedures by eliminating test materials from further testing based on assessments made in the tier. After each step in the experimental process, the data obtained to that point must be assessed to determine whether a decision on hazard classification for the test material can be made. No scientific methods can adequately replace animals for acute toxicity testing; however, when tiered approaches are used, the number of animals necessary to determine a hazard is reduced significantly. The initial step in this tiered procedure is an examination of structure-activity relations and physicochemical properties between the test material and related compounds. If related compounds have been shown to elicit toxic effects, appropriately selected in vitro assays can document effects similar to those seen in the related materials. When it is seen that the test material behaves like those compounds previously tested, no further testing is required.

Acceptance of Valid Alternatives

The development of alternative methods for refinement, reduction, or replacement of animals in acute toxicity testing is the goal toward which investigators should strive. When a method has demonstrated promise as an alternative, it must be validated and accepted before its use becomes conventional. Validation is the process by which reliability and relevance of the method are demonstrated. Reliability is attained when the method can produce reproducible results within and between laboratories. Relevance occurs when a test is demonstrated as being meaningful for its intended purpose. Criteria for determining both validation and acceptance of any method have been enumerated by the ICCVAM ( NIEHS 1997 ).

Other Humane Endpoint Considerations

The foregoing discussion has focused on the use of animals for specific acute toxicity testing procedures, citing both classical methods and current practices for refinement and reduction that have gained varying levels of awareness and/or acceptance. The search for humane endpoints as a means of refinement in toxicologic testing procedures represents a modification both in perceiving and in applying the ethics of research.

Guidance on the practice of humane techniques, including the selection of humane endpoints, is based on best current knowledge available through personal contacts with investigators, documentation provided in peer-reviewed literature, and presentations at meetings and symposia. Any guidance is intended to be flexible so that it can change with improved or more complete knowledge. It is expected that with increasing knowledge, investigators will be able to identify more specific and earlier humane endpoints. This facilitated identification would permit harmonization of earlier humane endpoints that would be acceptable to investigators, regulators, and member states.

In 1998, OECD assembled an international working group to discuss the concept of application of humane endpoints to toxicology testing. This group, after deliberations, promulgated a draft entitled Guidance Document on Humane Endpoints for Experimental Animals Used In Safety Evaluation Studies ( OECD 1999 ). In the Preamble, it is acknowledged that “there is strong scientific evidence that pain, distress, and suffering can exist in animals as in man” (Section III, Guiding Principles, p. 3) and that certain principles must be applied when developing, designing, and conducting toxicology testing and research involving animals. These principles focus on the parameters that must be considered when designing a study, with an eye toward successful completion of the study while addressing the conditions under which the animals are being held during the study. The design of the study must be such that the earliest possible endpoint is utilized, that pain and distress are minimized, and that a thorough review of the ethics of the study has been conducted and all members of the research team adhere to these protocol-contained elements. In designing a toxicity study to meet the objectives intended while minimizing pain, distress, and suffering, it is essential to gather as much information as possible about the substance to be tested. This information would assist in establishing the appropriate objectives of the test. A comprehensive review of the literature will indicate whether the test material has been evaluated in the past and the outcomes of prior studies and will aid in selecting the most appropriate species for the study. More important, knowledge of prior and related studies might enable the investigator to predict likely clinical signs and the timing of their occurrence. Consideration of expected clinical signs and time to appearance before a study commences enables investigators to develop and implement procedures to minimize or eliminate any pain or discomfort in the animals before their appearance. Documentation of these expected signs will also prepare all members of the research team and provide the ethical review body with expected pain and distress parameters against which the conduct of the study could be measured.

National laws in most countries demand that a body duly constituted either by the government or the institution performing the research conduct an ethical review of the study in accordance with legal requirements. Careful scrutiny of the ethical aspects of a study, as well as the science, will help ensure the investigator, the scientific world, and the general public that a study is being conducted in the most humane manner possible. The course of the study, as described by the protocol, can then be monitored for adherence to the conditions agreed on in advance.

Pilot Studies

It may be useful to implement pilot studies to identify the earliest decision point for successful completion of an experiment or to determine criteria for the humane killing of animals in the study. Pilot studies can help determine whether the doses proposed in the study are likely to cause unexpected or unacceptable problems. Pilot studies can also indicate whether other parameters (e.g., hematologic biochemical information obtained from telemetric devices) can provide information useful to the identification of earlier endpoints either for the completion of the study or for the humane killing of experimental animals before they experience pain or suffering. These signs then become potential endpoints to the test procedure. If it can be determined that the clinical sign(s) are precursors to the end result of the toxic effect, it may be possible to use any or all of these as endpoints to the assay and future studies can be terminated at the appearance of these observable parameters. It is not necessary for the animals to proceed to points where pain, stress, or distress result.

Humane Killing

Humane killing of animals as a means of alleviating pain or distress is an alternative that may be considered during the conduct of an acute toxicity study. Recognition of signs before death or the time when animals become moribund allows investigators to end the study before the level of pain or distress increases. Humane killing may be the best way to preserve the welfare of the animals if pain, stress, or distress, unpredicted before the initiation of the study and unmitigated through the use of anesthetics or analgesics, occurs. Clinical signs and conditions in which humane killing may be appropriate have been documented in many reports, including documents published by Morton and Griffiths (1985) . All members of the study team should be aware of the presentation of any of these signs or symptoms. The observations and impressions of animal technicians are crucial to providing information on the status of the animals. The principal investigator and attending veterinarian can determine the significance of the condition presented by the animal and whether predetermined means for amelioration should be implemented. The use of analgesics or anesthetics for elimination of pain or distress that cannot be removed by other means should be described in the approved protocol.

Training

The principal investigator has the obligation for ensuring that all individuals involved in the study have the training necessary to fulfill their roles under the study design. This is further emphasized by the protocol, which serves as a contract between the principal investigator, his/her coworkers and staff, and the IACUC or ethical review committee. When the roles described in the protocol involve the determination of the endpoint beyond which animals will experience pain and/or suffering, the individuals accountable must be sufficiently experienced in observing animals, assessing their physiology, behavior, and appearance, and initiating prescribed actions.

Summary

The quest for humane endpoints to studies involving the use of animals may sound altruistic: The goal is admirable but aside from philosophical implications, the achievement of the goal is essential. Mankind has benefited from research using animals. The safety of the food and drug supplies of the world, as well as other chemicals used in today’s societies, has been ensured at significantly higher levels because materials have been evaluated through the use of toxicity testing in animals. There is no doubt that mankind has benefited substantially from toxicity testing. Convention or traditions, as applied to the use of animals in testing, are used to generate acceptance of older methods. These methods have been validated and documented as being relatively accurate predictors of hazards to which humans may be exposed. The drawback to the use of traditional methods is that the public, people concerned about animal welfare, and the scientific community realize that change is possible. This concept was not original with Russell and Burch, although the initiation of the 3 Rs and the search for alternatives are most often ascribed to them. The goals of toxicology have not changed; however, they have been suffused with the need for more humane ways to accomplish the same results—proof of safety or hazard from materials to which humans may be exposed. The public is alert to any actions taken by scientists. All appropriate measures must be taken to assure the public that (1) their safety is being considered, and (2) no animals are suffering from needless pain, stress, or distress in efforts to assure this safety. The search for more humane means of obtaining applicable data, including refinement of experimental procedures, reduction in the numbers of animals used, or replacement of animals with validated alternatives, must be practiced by investigators and enforced by ethical review bodies charged with the approval of animal use protocols.

References