Basic Approaches in Anatomic Toxicologic Pathology

1

Introduction

Traditionally trained veterinary and medical pathologists often encounter unanticipated challenges during their transition from diagnostic pathology to the regulatory-driven environment of toxicologic pathology. During their initial years of training and diagnostic effort, pathologists typically serve both clinical and public health functions. Pathologists in diagnostic, hospital, or private laboratory settings serve the clinical community to support therapeutic approaches and disease prognoses. These laboratories are generally governed by internal work practices and procedures that are based on professionally recognized best practices and for the most part experience a limited degree of governmental oversight and influence on individual anatomic pathologist activities. The specific work practices of veterinary pathologists in these environments are seldom dictated by the extensive regulations that mandate proper management and storage of data, quality assurance review, peer review, animal welfare standards, organizational structure and personnel, and study design that are commonly encountered by toxicologic pathologists in the regulatory environment common to the biomedical and agrochemical industries. In addition, various animal species and different spectra of spontaneous background and induced lesions can be encountered in toxicologic pathology venues that are not commonly observed in the traditional training or diagnostic setting.

This chapter will provide a basic introduction to the fundamental considerations, practices, and techniques involved in the proficient generation and interpretation of anatomic pathology data from repeat-dose nonclinical or other animal toxicity studies. Basic aspects of protocol development, necropsy, and tissue fixation will be covered as well as the recording of macroscopic (gross) and microscopic (histopathologic) findings and confounding factors (such as sexual dimorphism and artifacts) that can hinder interpretation. Several of these topics will only be mentioned briefly as they are discussed in greater detail in other chapters (see Nomenclature and Diagnostic Resources in Anatomic Toxicologic Pathology , Vol 1, Chap 25 , Pathology Peer Review , Vol 1, Chap 26 , and Practices to Optimize Generation, Interpretation, and Reporting of Pathology Data from Toxicity Studies , Vol 1, Chap 28 ).

2

General Considerations in Study Protocol Development

Before any toxicity study can get underway, a study protocol must be developed. U.S. Food and Drug Administration (FDA) and U.S. Environmental Protection Agency (EPA) guidance documents state that “each study shall have an approved written protocol that clearly indicates the objectives and all methods for the conduct of the study” ( , ; ). The Organisation for Economic Co-operation and Development (OECD) also recommends that “a written plan should exist prior to the initiation of the study” ( ). Although the guidelines from the three agencies exhibit slight differences in content, all three organizations provide very similar guidance. Briefly, all study protocols should include a descriptive title and statement of the purpose of the study; identification and formulation of the test and control articles (or test items) by name, chemical abstract number, or code number; justification for dose selection; the name of the sponsor and the name and address of the testing facility; and the number, body weight, sex, source, species, strain, substrain, and age of the test animals (see Pathology and GLPs, Quality Control and Quality Assurance , Vol 1, Chap 27 ) . The toxicologic pathologist should be familiar with the protocol requirements of the various regulatory agencies as they will work with study directors, who are typically toxicologists, to support development of the pathology portions of study protocols or study plans.

During the initial stages of developing a study protocol, a number of issues must be addressed so that the ensuing stream of materials and data will meet regulatory requirements, follow established Good Laboratory Practice (GLP) guidelines, and provide reliable data capable of supporting the risk assessment process. These issues include clearly understanding the purpose of a given study; selecting an appropriate test system (i.e., animal species and strain), taking into consideration appropriate animal care and use; incorporating the 3Rs principles (Replacement, Reduction, and Refinement); ensuring appropriate sampling to meet regulatory needs and to satisfy the scientific questions of the study; and addressing key husbandry factors, to name a few. Study protocols should be reviewed by multiple stakeholders and departments involved with the design and conduct of the study to ensure that all parties and separate workgroups have clear direction and that the protocol-directed work is appropriate, necessary, and feasible.

The results of most toxicity studies are included as part of a data package submission to a regulatory agency, such as the U.S. FDA, U.S. EPA, European Medicines Agency, U.K. Medicines and Healthcare Products Regulatory Agency, South Korea's Ministry of Food and Drug Safety, or Japan's Ministry of Health, Labor, and Welfare. These agencies among others are charged with ensuring the safety and efficacy of pharmaceuticals, agrochemicals, medical devices, and food additives as well as numerous other products in their respective host countries. These data packages are part of the review process for these various chemicals and devices.

In the United States, toxicity studies that are included in regulatory submissions must be conducted under the auspices of the GLPs Act of 1975. These practices were adopted by the European Union and Japan as a consequence of the International Council on Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH). GLPs specific to the FDA are detailed in Title 21 CFR (Code of Federal Regulations) Part 58 and are discussed in another chapter of this text, as GLPs are specific to the US EPA Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA, detailed in Title 40 CFR Part 160) and Toxic Substances Control Act (TSCA, detailed in Title 40 CFR Part 792) as well as GLPs specific to the OECD (detailed in ) (see Pathology and GLPs, Quality Control and Quality Assurance , Vol 1, Chap 27 ). GLP guidelines ensure that a study is conducted in a standardized, documented manner, consistent with global regulatory requirements. These practices result in the conduct of a well-designed and properly performed scientific study, with clearly defined expectations and procedures that will consistently yield high-quality, reproducible, and reliable data sets. However, these guidelines were not designed to dictate the most appropriate type of study for a given test article. Given the breadth of potential test articles, combined with the heightened safety standards of global regulatory agencies, the most appropriate study to characterize the toxicity of a compound must be designed with respect to known target pharmacology and test article–related potential liabilities even though, particularly in the early stages of drug development, the full toxicokinetic and toxicologic profile may not be known. In developing an appropriate toxicity study protocol, the study scientists must always ask “What major question(s) is this particular study expected to answer?” Examples of important questions may include but certainly are not limited to the following: What is the expected clinical dosing route and frequency? Is the compound acutely or chronically toxic according to previous investigations? Is there a dose–response to the toxicities, and do they resolve after dosing cessation? Does the test article induce proliferative changes that might lead to neoplasia? What are the expected effects on an individual at a given age range (e.g., juvenile as compared to skeletally or sexually mature animals)? Is the effect in a laboratory animal species of potential relevance to humans? What are the mechanisms of a particular effect in a single organ or across multiple related tissues?

No single safety study is able to answer all possible questions. Therefore, when developing a study protocol, it is important to focus on specific key questions and consider how the one study will fit into the final regulatory package, given that the package will have been assembled using data from multiple studies that addressed numerous different (though sometimes overlapping) questions. As scientists attempt to gain as much information as possible from a single study in order to reduce animal use and budgets, forethought concerning the information that needs to be conveyed by the final integrated registration package is becoming increasingly important during the design process of individual studies. In addition, GLP studies must comply with the requirements of the regulatory agencies that will evaluate the submission. Depending on the regulatory agency, study requirements may vary according to the species, strain, age, route and frequency of administration, dose range, number of animals per group, study length, and potential need for ancillary data types (e.g., dedicated testing for immunotoxicity, juvenile toxicity, neurotoxicity, or other specialized system-specific endpoints). Thorough knowledge of the requirements of and experience with various regulatory agencies will help preclude costly delays at the time of regulatory package submission, especially in the increasingly frequent desire to seek product registration in multiple parts of the globe.

Husbandry-related factors, such as light exposure (both duration and intensity), temperature, sound, diet, and environmental enrichment, all have the potential of influencing the results of the study or data interpretation, and therefore should be addressed in the study protocol or within facility Standard Operating Procedures (SOPs) or study-specific technical documents in order to assure consistency in experimental manipulations among all study groups (see Issues in Laboratory Animal Science that Impact Toxicologic Pathology, Vol 1, Chap 29 ) . These environmental or husbandry-related factors should be recorded, with records available for the pathologist or other contributing scientists to refer to when interpreting findings. For example, extensive retinal degeneration can occur in rats exposed to excessive ambient illumination. Recommended ambient light levels for rodents (20–25 lux for nonpigmented strains and below 60 lux for pigmented strains) are significantly below those recommended for human working conditions (up to 500 lux in laboratory spaces). Therefore, it may be common for rodents to be inadvertently exposed to high light levels in studies with extensive human interaction. Therefore, if only high-dose animals are routinely housed on the upper shelves of cage racks, where light levels are highest, the degree of retinal degeneration could appear to be dose related ( ; ). Additionally, long photoperiods in rodent rooms can cause atrophy of the ovaries, Harderian glands, and adrenal glands, in addition to anestrus, any of which could erroneously be attributed to a test substance ( ; ). Variations in temperature and sound can cause biologically significant changes such as immunosuppression, hormonal changes, alteration in phenotypic manifestation of musculoskeletal development, cardiovascular parameters, and other physiologic stress responses in animals ( ; ; ; ).

Another husbandry-related factor that can impact anatomic pathology data is diet. Previous studies have documented a direct correlation between average caloric consumption and the incidence of several commonly recognized spontaneous degenerative diseases in rodents, including such conditions as chronic progressive nephropathy, progressive cardiomyopathy, and pancreatic islet fibrosis, as well as some neoplasms (pituitary and mammary tumors) ( ; ; ; ; ; ; ). Furthermore, caloric consumption is inversely correlated with survival in 2-year rodent carcinogenicity studies (see Issues in Laboratory Animal Science that Impact Toxicologic Pathology, Vol 1, Chap 29 , Carcinogenicity Assessment , Vol 2, Chap 5, and Nutrition and Toxicologic Pathology , Vol 2, Chap 20) ( ). More importantly, feeding diets with different compositions has been demonstrated to alter physiological responses to some test articles. Animals fed a calorie-restricted diet have been shown to be more robust, with demonstrable tolerance to known carcinogens as well as reduced sensitivity to shorter-term toxicity when compared with ad libitum-fed, overweight animals ( ; ; , , ; ). Consequently, all toxicity studies of a given test article should be performed using the same diet. Utilizing diets with different compositions or nutrient sources across the spectrum of toxicity studies conducted for the same test article could alter the microbiome, resulting in variability in pathology or toxicokinetic findings. This in turn could lead to conflicting data that could confound data interpretation and potentially lead to repetition of inconclusive studies.

The animal species and strain, as well as the number of animals per sex per dose group, may be determined by customs of the particular laboratory, 3Rs considerations, animal availability (particularly with nonhuman primates [NHPs]), and statistical considerations. Also, the number of treatment groups and the age of the animals should be appropriate to comply with both regulatory and scientific standards. The criteria for selection of an appropriate animal model should weigh a number of factors. It is critical that the animal test system is sensitive to the test article being studied. In this regard, the test animal must exhibit properties relevant to the interaction of the test article (e.g., the target of the test article is expressed in the animal model and the test article is shown to bind to the target, such as in the case of a monoclonal antibody) and the test article must induce a functional pharmacologic response (e.g., as shown by assessment of pharmacodynamic [PD] biomarkers). Species sensitivity may be determined in part by single-dose or short-term repeat-dose exploratory toxicity studies where pharmacokinetic (PK) and PD endpoints are assessed as well as tolerability ( ) (see Volume 1, Part 1 : Principles of Toxicologic Pathology ). Moreover, the PK and metabolic pathways of the test article in the animal model should be well characterized and their comparability to the test article effects in humans should be explored. In addition, the availability of historical control data for parameters measured in animals during the course of a study, such as body weight gain, food consumption, clinical pathology values, and the incidences and severities of background/spontaneous lesions and neoplasia, could be valuable in the recognition of a test article–related effect. Other criteria should include the commercial availability of the animal model, as well as the accessibility of appropriate housing and reagents (e.g., antibodies to quantify protein levels in fluids or detect their localization in tissue sections) that might be needed to probe certain classes of functional and structural effects. Furthermore, investigators should have extensive experience with the husbandry, dosing, and evaluation of the test species.

The number of dose groups should be established during protocol development. Generally, repeat-dose general toxicity studies require a minimum of three test article–treated groups (i.e., low, mid, and high) and at least one control group (usually untreated or vehicle treated), under the same conditions and testing regimen as test article–dosed animals. The lack of control groups, although sometimes advocated by those seeking to reduce the cost of testing and extent of animal use, should be avoided. In the absence of concurrent control animals, data will often be more difficult to interpret, statistical comparisons (for organ weights and clinical chemistry and hematology values) will not be possible, and the study design may not be accepted by regulatory agencies. This is particularly pronounced in nonrodent studies, where individual animal variation (e.g., in genetic background, weight, immune status, responsiveness to compound, social stress–associated changes, etc.) is generally greater than in commonly used rodent stocks and strains. To some degree, this could be ameliorated by including control animals during the in-life phase which could be utilized to obtain clinical pathology values, body weight evaluations, clinical observations, and the like; however, as opposed to euthanizing them for necropsy and histopathology evaluation, they could be recovered and returned to a colony group for use in future studies. This does not provide for concurrent controls in the histopathology or organ weight evaluations of the study, so deliberate consideration should be taken in utilizing this specific approach. The doses for these studies are generally established by prior single-dose or short-term (≤4 weeks) repeat-dose exploratory toxicity studies. Ideally, the high dose should result in evidence of toxicity but not generate more than 10% mortality, with the middose resulting in only slight toxicity and the low dose producing no test article–related effects. This spectrum of dose-dependent effects permits the determination of a No Observed Adverse Effect Level (NOAEL) and/or No Observed Effect Level (NOEL) as measures of the extent to which a test article induces adverse (NOAEL) or any (NOEL) findings in an exposed organism. The meaning and significance of NOAEL and NOEL in risk assessment will be discussed elsewhere in this text (see Interpreting Adverse Effects , Vol 2, Chap 15).

The age of the test species used in repeat-dose toxicity studies is fairly standard for routine studies. Unless other factors dictate another time point, exposure in conventional studies is initiated during young adulthood in rodents or during late adolescence in nonrodents. For example, in rodent studies, initiation of test article administration at 6–8 weeks of age will comply with essentially all regulatory guidelines for products destined for administration to adult animals or humans. Exception to this timing in rodents is required for test articles with potential acute reproductive effects (where initial dosing may begin in older animals to ensure full sexual maturity) or where the products are destined for administration to a pediatric or juvenile population (where initial dosing often is initiated prior to weaning) (see The Role of Pathology in Evaluation of Reproductive, Developmental, and Juvenile Toxicity , Vol 1, Chap 7 ) ( ; ). Nonrodents have longer lives and mature more slowly than rodents. Accordingly, it is acceptable to regulatory agencies to initiate dosing in studies using dogs or minipigs at approximately 4–6 months of age even though animals at this age are often peripubertal rather than fully mature. In like manner, cynomolgus macaques are generally approximately 2–3 years of age at study initiation; at this age, there are a high number of reproductively immature or peripubertal animals, particularly males. Histologically, male cynomolgus macaques may not be reproductively mature until 4–5 years of age; body weight and testicular weights can be additional indicators of maturity ( ; ). For female macaques, reproductive maturity, with regular ovulatory cycling, may not occur until approximately 4 years of age ( ; ). More mature nonrodent animals may be used for specific study purposes, though such studies are more expensive due to the increased husbandry costs incurred by maintaining animals as they age. For all of the larger animal species, full reproductive maturation in a consistent majority of the population may not be observed at these recommended ages for treatment initiation, so studies reliant on evaluating test article effects in reproductively mature specimens may need to source animals months to years older than these general guidelines, depending on the species ( ; ; ) (see Table 9.1 and Vol 1, Part 3: Animal and Alternative Models in Toxicologic Research ).

The proper selection of controls is an important aspect of study design as well. Different study types will require different controls. A standard repeat-dose study in a well-characterized rodent strain will generally require concurrent controls of the same strain, with equal numbers for each sex, all within a 20% body weight range. Animals should be assigned to control and test article–treated groups in a stratified random manner (i.e., ensuring that each treatment group contains a randomized array of animals from separate batches or birthdates or other possible differentiating factors) to minimize bias. For some studies, there may be separate negative control groups for sham-dosed, neutral agent (water or physiologic saline)–dosed, and/or vehicle-dosed groups. In injection studies, an adjuvant-only control group may also be utilized, though this is not required for regulatory purposes if the adjuvant is already well characterized (see Vaccines , Vol 2, Chap 9). Additionally, for some studies, such as transgenic mouse carcinogenicity studies, positive control groups (i.e., exposed to a known carcinogenic agent) may be incorporated to provide internal validation that the model remains responsive (see Carcinogenicity Assessment, Vol 2, Chap 5). Finally, juvenile toxicity studies require careful planning for inclusion of control animals that are to be euthanized at planned and unexpected necropsies in order to allow age-matched comparison of evolving histologic features (see The Role of Pathology in Evaluation of Reproductive, Developmental, and Juvenile Toxicity , Vol 1, Chap 7 ). Careful selection of controls can greatly enhance the ability of the pathologist and other scientists to interpret possible test compound–related findings.

The purpose of a given animal study should be taken into account when formulating the study protocol. Studies may be performed to address a range of questions. Therefore, study procedures, sample collections, and all other study parameters need to be aligned with the primary purpose of the study. Some study types, such as dose range-finding studies or lead optimization studies, do not generally require large numbers of animals, extensive sampling, recovery periods, or numerous anatomic pathology or clinical pathology endpoints as they are focused on providing information for supporting future, more comprehensive (i.e., Investigational New Drug-enabling) studies. Such preliminary studies may be designed to include investigative endpoints that will not be incorporated in later GLP-compliant toxicity studies but that are needed for product registration. Subchronic (usually 3–6 months) and chronic (usually 9–18 months) studies should have sufficient numbers of animals across a range of dose levels in order to characterize toxicity even in the face of early attrition due to unscheduled deaths. Generally, in standard GLP-compliant toxicity studies, a minimum of 10 rodents/sex/dose group and 3–4 nonrodents/sex/dose group are enrolled for shorter-term studies (≤6 months); group numbers may be increased by 50% or more for longer studies, especially in rodents. These numbers also may be increased in instances where a subtle or infrequent finding is expected. Recovery intervals may be included in the study ( ; ; ). Additional animals can be included specifically for serial collection of blood for PK analysis; however, tissues from these animals are typically only examined macroscopically at necropsy and do not undergo histopathologic evaluation. The collective focus of these studies should be evaluation of test article toxicity to facilitate the translatability of these findings to inform human risk. Carcinogenicity studies are conducted in rodents and are meant to determine the potential of a compound under life time exposure to induce neoplastic changes. Numbers of animals for these studies are selected to ensure that there is survival of a sufficient numbers of animals for adequate statistical analysis. Thus, a two-year carcinogenicity study typically is loaded with 60–80 rodents/sex/dose group. For a six-month transgenic mouse carcinogenicity study, where the model has an engineered or spontaneous genetic predisposition to early development of neoplasia, initial animal numbers will be lower (generally 25/sex/group) as age-related mortality should be limited (see Carcinogenicity Assessment , Vol 2, Chap 5). It is obvious that a single study plan cannot cover all potential study types, nor can all questions of interest be addressed in a single study.

Another important component of study protocol development is the choice of anatomic pathology endpoints. For instance, the list of essential organs weights to collect and tissues to be saved and processed for histopathologic examination as well as preparation to appropriately preserve samples for other study purposes (e.g., electron microscopy, immunohistochemistry [IHC], etc.) must be compiled in advance of the actual necropsy. Major regulatory agencies and professional organizations have developed guidelines of recommended organ weights, tissues, and other toxicologic pathology endpoints that should serve as a template for study scientists tasked with designing toxicity studies. For example, after reviewing guidelines from regulatory agencies in the United States, European Union, and Japan, working groups of the Society of Toxicologic Pathology (STP) have formulated “best practice” recommendations for organ weight collection and protocol-specified tissue lists ( ; ; ; ). In addition to these general recommendations, the standard tissue list in the protocol should be modified to collect tissues appropriate to the design and objective of the study and reflect potential target-related pharmacology. Reasonable examples of study-specific tissue list modification would be addition of the nasal cavity, larynx, and tracheobronchial lymph nodes when evaluating an inhalation study; inclusion of injection/infusion and/or surgical sites as well as draining lymph nodes for parenteral injection, medical device, or surgical intervention studies; or the inclusion of the site(s) of application for a skin paint (dermal toxicity) study. This modification of the tissue list may also extend to increased sampling (sometimes termed “enhanced histopathology” or “expanded histopathology” by regulatory agencies) of organs and tissues to further characterize and assess findings at a specific site of interest; examples include ocular studies where specific areas are targeted for test article delivery or injection studies where the site of injection may be imperfectly visualized on gross examination. The protocol tissue list also should include tissues that are known or suspected targets of the test article, similar agents in the same chemical class, or that act via the same target molecule. Studies with test articles that are expected to alter the immune system should include collection and examination of lymph nodes, spleen, thymus, and related lymphoid tissues. If a test article or class effect is not known in advance, then it typically is reasonable to conduct a complete necropsy and save a thorough list of tissues even if many of them are archived without histopathological examination. This approach will preclude the loss of valuable data given that the necropsy is a one-time event and that tissues, once discarded, cannot be retrieved. Appropriate collection and preservation of tissues is critical to ensure accurate high-quality sampling and interpretation of the anatomic pathology data set.

Selection of appropriate clinical pathology parameters is another essential step in the development of a study protocol. This choice is particularly critical in shorter studies as hematology, clinical chemistry, and urinalysis results may be sensitive indicators of toxicity (see Clinical Pathology in Nonclinical Toxicity Testing , Vol 1, Chap 10 ). As an example, changes in clinical chemistry values may provide an early indication of hepatic injury (see Liver and Gallbladder , Vol 3, Chap 2). The early identification of hepatic injury can be a challenging but decisive event in pharmaceutical compound development. A retrospective study has demonstrated that approximately 10% of marketed human pharmaceuticals that have either been withdrawn or labeled with a “black box” warning over the past several decades have been as a result of hepatic injury reported as a postmarketing adverse event ( ; ). Current recommendations advise measuring the activities of two serum aminotransferase enzymes, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), as the most sensitive and specific indicators of hepatocellular injury in rats, dogs, and NHPs ( ; ; ). ALT is also considered to be a specific and sensitive test of hepatocellular injury in humans, suggesting that results of nonclinical studies may be predictive of potential hepatic injury in human patients. As with any single parameter, AST values must be used as part of the overall assessment of liver toxicity, including correlation with histopathologic observations in subchronic and chronic toxicity studies. To aid in identifying the clinical pathology parameters that will optimize early recognition of cellular injury, study scientists should be knowledgeable with any potential test article–related toxicities either by perusing data sets from exploratory toxicity studies or by predicting possible effects based on the activities of similar agents in the same class of compounds. Anatomic toxicologic pathologists may analyze, interpret, and communicate routine clinical pathology data from standard toxicity studies or may work with a board-certified veterinary clinical pathologist to discuss findings from their respective analyses and align on how pathology conclusions are to be communicated to the Study Director. More specialized clinical pathology tests typically are evaluated by a clinical pathologist due to their greater familiarity with unusual assays and artifacts that might result from sampling and/or analytical complications.

Biomarkers also may be of use in demonstrating test article–related effects. Certainly, classical clinical pathology parameters such as hematology, blood chemistry, and urinalysis are the most basic of the biomarkers that are routinely included in the toxicologic pathology portions of study designs. However, additional biomarkers such as gene expression (see Toxicogenomics: A Primer for Toxicologic Pathologist s, Vol 1, Chap 15 ) or noninvasive imaging (see In Vivo Small Animal Imaging: A Comparison to Gross and Histopathologic Observations in Animal Models , Vol 1, Chap 13 ) may be incorporated in animal studies to address specific potential findings or to provide more translatable and easily monitored signals for clinical trials. Biomarkers may be used to assess functional responses, direct tissue injury, or to localize toxic insults. Biomarkers are an ever-expanding field of development and may be used to evaluate PD responses, efficacy, or toxicity across a wide range of organ systems, including renal, cardiovascular, musculoskeletal, and nervous systems; hormonal changes; immune function and acute phase reactants; coagulation factors; growth factors; and other metabolic disturbances, to name a few. Biomarker responses vary in timing, as well as among test articles and across species. For example, certain biomarkers of cardiotoxicity may elevate acutely, while others will only elevate over days to weeks, and rodents have different acute phase proteins than humans and other mammals. Some biomarkers may exhibit diurnal variation or fluctuation as a result of feeding or light cycles ( ; ; ). These factors need to be taken into consideration when designing studies utilizing biomarkers or when interpreting biomarker data. Incorporation of biomarkers into the study design should be done in a careful and considered manner, under consultation with a board-certified veterinary clinical pathologist or other subject matter expert to ensure that the correct samples are taken at the optimal time, and the correct panel of biomarkers is used. Avoidance of a random, broad-based (“shotgun”) sampling approach is wise as the potential for producing confounding or nonsensical data—often without sufficient context to aid in its interpretation—is present in the absence of a clearly targeted investigation. As with clinical pathology data, biomarker data should be interpreted in consultation with the study anatomic and clinical pathologists in order to understand changes in the context of the whole animal.

Much has been written about animal welfare issues and their potential impact in toxicologic pathology (see Issues in Laboratory Animal Science that Impact Toxicologic Pathology, Vol 1, Chap 29 ) . Increased awareness of animal welfare and the regulatory oversight supplied by the U.S. Department of Agriculture (USDA) and equivalent ex-U.S. agencies as well as through organizations such as the American Association for Accreditation of Laboratory Animal Care have generally resulted in improved treatment of laboratory animals. No study can be legally initiated without the prior approval of the detailed study design by an Institutional Animal Care and Use Committee (IACUC). Failure to adhere to animal welfare regulations or lack of appropriate IACUC oversight can result in severe consequences for a research facility and the acceptability of its studies for regulatory submission. In addition, animal welfare concerns have intensified research into ways to utilize fewer animals, such as using transgenic mice in short-term bioassays as well as alternative methods of testing such as in silico (computer based), in vitro (isolated cells or tissue portions), or in vivo testing in nonstandard species ( ). These advances are expected to lessen our dependence on the use of large numbers of laboratory animals in the future; indeed, the EPA has expressed an ambition to reduce in vivo testing to essentially zero in the next two decades ( ). Toxicologic pathologists may be involved in the application and interpretation of alternative models of toxicity including cultured cells, microphysiologic systems (“organs on a chip”), organoids, and tissue slices (see In Silico, In Vitro, Ex Vivo, and Non-Traditional In Vivo Approaches in Toxicologic Research, Vol 1, Chap 24 ).

3

In-Life Evaluations

Toxicity studies in laboratory animals may vary from single-dose acute studies to both short- and long-term repeat-dose studies. Results from acute toxicity studies are generally used to estimate doses for short-term repeat-dose studies that may vary from one to 4 weeks in duration. Long-term repeat-dose studies that extend to 3 months are considered to be subchronic, and those that proceed from 6–12 months or more are regarded as chronic studies. Repeat-dose studies of 24-month duration in rodents are considered to inform on carcinogenicity risk as they largely span the lifetime of mice and rats ( ; , ; ).

Prior to the initiation of any toxicity study, it is important to assess the health of every animal. Following any quarantine period necessitated by shipment from another facility, all animals should be subjected to a thorough health screening prior to study initiation. This assessment minimally must include a physical examination by trained veterinary staff and baseline total body weight measurement. Ophthalmologic examination, clinical pathology assessment, and pathogen screening are highly recommended. Any animal deemed to be unhealthy should be removed from the study; consequently, it is recommended to acquire at least 10% more animals than will be needed to fill the study groups. In addition, in rodent studies, a sentinel group of 5–10 animals of each sex should be established in the room along with the study animals to monitor for pathogens. Sentinel animal serum antibody titers for species-specific pathogens of concern should be measured at the initiation of the study and at the end of the in-life phase, and sometimes at interim time points for longer-term studies or experiments with immunocompromised animals. In the event of a suspected pathogen, sentinels may be used to measure antibody titers and/or euthanized for gross and histologic examination to assess for evidence of infection without disturbing the animals on study (see Issues in Laboratory Animal Science that Impact Toxicologic Pathology, Vol 1, Chap 29 ).

Upon arrival at a testing facility, each animal should receive a unique individual animal identification number that will be linked to the animal throughout its tenure at that facility. Additional alternate identification numbers may be assigned for use within a specific study, based on group number, sex, or similar parameters, with individual animal study-specific identification numbers assigned after the establishment of groups based on uniform body weight distribution. This unique animal number is used to identify all hematology and clinical chemistry samples, tissues, and all other data (e.g., body weight, organ weights, food consumption, clinical observations, etc.) collected and recorded from the animal during and following the in-life portion of the study. This number is permanently coupled with the animal to preclude misidentification of samples as well as to support the restoration of the study for peer review purposes or in the event of a quality assurance audit. There are a number of methods used to apply these unique animal identification numbers. The most up-to-date method, and one that is becoming virtually ubiquitous in modern laboratory animal facilities, is subcutaneous microchip implantation. This technique facilitates electronic recording of the animal identification numbers into a database for permanent storage, thus minimizing the potential for errors when hand-recording data. This method is well established with few notable side effects. Migration of the transponder microchip within the subcutis has been noted to occur, and there is some evidence to suggest that transponder-associated soft tissue sarcomas can develop in some species (e.g., rats and dogs), potentially complicating carcinogenicity study results unless the potential for this background neoplasm is taken into account during study design ( ; ; ). Older identification methods used in rodents include toe clipping (a small portion of the toe is removed in a specified coded manner), ear punching (small holes in the ear are punched in a specified coded manner), ear tags, and tattooing (at the base of the tail or digits). Similar methods have been used in larger animals, such as tattooing (of the inner pinnae of the ear or the armpit), ear punching/clipping, or ear tags. These methods are currently much less accepted and may require IACUC approval prior to implementation.

Following the allocation of unique individual identification numbers, all animals must be randomized into study groups. Proper randomization is essential to minimize bias and improve the recognition of statistical differences between groups in the study. Popular approaches to accomplishing this are through the use of random number tables or generators.

Several parameters are commonly measured and recorded throughout the in-life phase of a study. Each animal should be observed at least twice daily by laboratory animal staff for overt signs of toxicity, moribundity, and mortality. Ideally, observations should be made once in the morning with a subsequent surveillance in the evening. During these inspections, animals should be observed for changes in hair coat, posture, behavior, mobility, and activity. In the event of suspected toxicity, the animal should be removed from the cage for a more detailed physical examination. These examinations should record the date, approximate time, severity, and duration of observed changes in the skin, eyes, mucous membranes, and respiration. Marked changes in motor function, excretory function (e.g., increased or decreased urine output; changes in fecal color, consistency or amount), respiration, and moribundity should prompt an examination by a laboratory animal veterinarian. Severe signs of toxicity, such as profound ataxia, tremors, or seizures; respiratory distress; reduced body temperature; hemorrhage; severe weight loss (>10% compared to controls in 1 week for rodents or dogs or >6% for NHPs) ( ); or other signs of ongoing pain or distress, could lead to outcomes such as temporary cessation of treatment (i.e., a dosing holiday), removal of the animal from the study, or humane euthanasia. All clinical signs of toxicity should be uniformly recorded using a glossary of terms that are defined by an approved SOP. The terms should be simple and descriptive, using minimal medical or diagnostic terminology.

The body weight should be measured at least once a week for short-term studies. In chronic studies, body weight is generally measured weekly for the first 13 weeks and then monthly thereafter. A weekly measurement is recommended because rapid body weight loss is one of the most sensitive clinical indicators of a test article effect during the in-life phase of the study. It may reflect declining health and may be prodromal to death. Rapid body weight loss may be due to decreased feed and/or water consumption, disease, dental maladies, or specific toxic effects. Severe, rapid body weight loss can also prompt the removal of an animal from a study. Importantly, changes in body weight not related to test article treatment can impact the interpretation of organ weight data at the conclusion of a study.

Food consumption is another important parameter that should be measured regularly during the in-life phase of a study. In rodents, it is generally measured once per week during the entirety of subchronic studies and the first 3 months of chronic studies; food consumption may be measured less frequently after the first 3 months. The accurate measurement of food consumption is critical for studies in which the test article is administered in the diet given that amount of food consumed and the dietary concentration of the test material are used to calculate the dose of test article consumed by the animals on study. As mentioned above, the amount of food consumed may be an excellent indicator of animal health as reduced levels could indicate a test article effect. Reduced food consumption may also be the consequence of an unpalatable diet (possibly as a result of the addition of the test article) or poor dental health. Food consumption data should be correlated with body and organ weight data during study interpretation.

All animals should receive an ophthalmologic examination prior to study initiation, as well as at the completion of the dosing period (see Eye , Vol 3, Chap 10). These examinations should be performed by a trained veterinary ophthalmologist experienced in examination of the test species. Slit lamp examination and the use of an indirect ophthalmoscope are two commonly used methods. Animals are excluded from the study prior to initiation if retinal atrophy; corneal, retinal, or lens dysplasia; or other confounding ophthalmic findings are observed. This avoids possible confusion when having to determine whether ocular findings in the dosing phase are preexisting or potentially test article related in both the ophthalmologic examination and the pathology examination. The outcome of such ophthalmologic assessments typically should be made available to the anatomic pathologist before eye sections are evaluated.

Hematology, clinical chemistry, and urinalysis are important clinical pathology data sets that inform on the general health of the test animal and are capable markers of test article effect (see Clinical Pathology in Nonclinical Toxicity Testing , Vol 1, Chap 10 and Interpretation and Reporting of Clinical Pathology Results in Nonclinical Toxicity Testing , Vol 2, Chap 14, for a detailed discussion of clinical pathology endpoints in toxicity testing) ( ). In addition to the routine clinical pathology endpoints used for safety monitoring, novel biomarkers of organ function and injury may be included (e.g., urinary biomarkers of renal tubular injury) as well as markers of drug PD and efficacy (see Biomarkers: Discovery, Qualification and Application , Vol 1, Chap 14 ). Study scientists should identify the clinical pathology parameters of interest prior to the beginning of the study and select the collection site that will provide the most consistent results, keeping in mind that sampling site selection may (1) affect future ability to evaluate those tissues clinically or microscopically or (2) limit collection of further clinical pathology samples as trauma associated with collection may obscure possible findings ( ). In addition, the effect of serial blood sampling as well as the sample size that can be obtained for a given time interval should be considered. The recommended blood volume for nonterminal blood sampling is 55–70 mL/kg for all species, which limits the sample size for smaller species ( ). An important point to consider when planning blood collection (and the choice of plasma or serum) for clinical pathology analysis is that these samples should not be obtained after animals have been dosed in intravenous administration studies, in order to avoid sample dilution; after animals have been handled for behavioral or other physical manipulations that might stress the animals; or after they have been bled for PK measurements, as the prior bleedings will affect both hematology and clinical chemistry parameters. Additionally, collection of clinical pathology samples in animals given dosing holidays (skipped doses), that have had midstudy dose reductions, or that were added to treatment groups during the course of a study will generally limit the ability to interpret clinical pathology data generated from these animals, as it may well result in a lack of clear controls for the affected groups. Certainly, these manipulations are occasionally required to allow assessment of reversibility, toxicity, tolerability, or other parameters, but clinical pathology data assessment may be compromised as a result.

In rodent GLP toxicity studies, clinical pathology assessments are usually performed for 10 animals of each sex per group and should be performed for all animals in nonrodent toxicity studies. Clinical pathology measurements are recommended to be conducted at least once during the course of 2- to 6-week studies for nonrodent species (typically within 7 days of dosing initiation). For rodent and nonrodent subchronic and chronic studies, ¹

1 The authors differentiate chronic studies (6–12 months) from carcinogenicity studies (generally 24 months). The utility of measuring clinical pathology values at the termination of carcinogenicity studies is debatable and is not recommended as a standard procedure in current best practices ( ).

clinical pathology parameters should be measured at the termination of the dosing phase. These endpoints may also be measured prior to study initiation or at interim stages (typically at 4 weeks in subchronic studies and 13 weeks in chronic studies) to evaluate the progression of test article effects. Clinical pathology measurements may also be conducted during and at the end of the recovery period, if included in the study.

4

Necropsy

Necropsy is a pivotal process in any toxicity study as it is one of the few events that cannot be repeated, in contrast with blood collection, which can be redone. Tissues and samples that are not collected, are inadvertently damaged as a result of mishandling, or are improperly fixed can result in a permanent loss of data that variably impacts the interpretation of the study. Therefore, careful planning, detailed SOPs, and familiarity with the study protocol (including any amendments) in addition to a coordinated team approach utilizing well-trained necropsy technicians and pathologists are all necessary components for completion of a successful necropsy. While the study protocol takes precedence, institutional SOPs that standardize practices (e.g., performing euthanasia, collecting and fixing tissues, obtaining organ weights, describing gross lesions, and trimming tissues) are necessary to ensure consistency and to eliminate variables that could subsequently complicate the interpretation of macroscopic observations, organ weight changes, and/or microscopic findings.

Necropsies can be conducted at scheduled intervals (i.e., predetermined interim time points or at the end of the dosing or recovery phases) or at unscheduled time points throughout the course of the study (i.e., when animals are found dead or humanely euthanized due to clinical signs or severe moribundity). Postmortem autolysis, the partial or complete destruction of cells or tissues following death due to self-produced enzymes, can result in a loss of cellular detail, thereby impairing histopathologic examination and interpretation. A study in neonatal Wistar rats determined that delays as short as 30 min can result in histological evidence of autolysis ( ). Therefore, necropsies should be performed as quickly as possible following the death or euthanasia of an animal, preferably beginning within 5 min. Study animals found dead during nonoperational hours should be immediately refrigerated at 4°C (never frozen) until a necropsy can be performed as ice crystals that form during freezing destroy microscopic features of cells and tissues.

The initial steps to a productive necropsy will vary depending on the species, size of the study, and availability of well-trained necropsy technical support. In large studies that require multiple days to complete, the order of necropsy should be randomized among dose groups, as it minimizes the potential for systematic technical errors to mask or exacerbate test article–related effects by distributing procedure-related consequences across the groups (including controls). For example, if animals are euthanized sequentially (i.e., all control males on Day 1 of necropsy, followed by low-, mid-, and high-dose males on subsequent days), a technical error on any single day, such as improper calibration of a balance or misformulation of a fixative, could result in the loss of organ weight data or reduced tissue quality, respectively, for an entire group or inaccurate attribution of a test article effect to a group. In addition, randomization minimizes the effects of fasting on clinical pathology parameters, organ weights, and tissue architecture. For instance, the amount and distribution of hepatocellular glycogen can vary depending upon the length of the fasting, thereby affecting liver weights and histopathologic features of hepatocytes ( ; ; ); an artifactual dose–group relationship could result if the animals were not euthanized in random order. Though still preferred, randomization of necropsy order is less crucial in smaller studies as they can be completed in a short period of time.

Careful consideration must be given when deciding upon the appropriate method of euthanasia for a toxicity study. Several acceptable methods are available; however, the preferred method will vary depending upon the capability of the facility, the purpose of the study, and the species and age of the animal used (see Issues in Laboratory Animal Science that Impact Toxicologic Pathology, Vol 1, Chap 29 ). The method of euthanasia should adhere to the guidelines established in the American Veterinary Medical Association Guidelines on Euthanasia ( ) and the Guide for Care and Use of Laboratory Animals ( ), with review and approval of the euthanasia method in the study protocol by the local IACUC. Regardless of the method chosen, it should seek to induce a loss of consciousness and rapid death with little to no pain, distress, or anxiety to the animal; be reliable; be easy to perform consistently with minimal risk of injury and/or emotional distress to the technician; and reduce the production of confounding tissue artifacts. It is essential that personnel are appropriately trained and able to perform the procedure in a professional, compassionate manner.

In general, inhalant (e.g., carbon dioxide) or noninhalant (e.g., sodium pentobarbital) chemicals are preferred to physical methods (e.g., cervical dislocation, decapitation). Carbon dioxide inhalation is the most commonly utilized technique for the euthanasia of small laboratory animals, while the preferred method for dogs, NHPs, and other small nonrodents (e.g., rabbits, Göttingen minipigs) is intravenous injection of a barbituric acid derivative, such as sodium pentobarbital.

Since gross and histopathologic parameters can be altered depending upon the method of euthanasia chosen, the study objectives must be taken into consideration when determining the appropriate method. For example, the use of barbiturates in dogs or Göttingen minipigs can result in variable degrees of splenic congestion which could affect interpretation of gross observations and organ weight data. Therefore, if the spleen is an organ of interest and/or test article–related effects are expected in blood-forming, blood-storing, and/or lymphoid organs, alternative methods of euthanasia may need to be considered during generation of the study protocol. In rodents, focal alveolar hemorrhage can occur in the lungs of animals euthanized via carbon dioxide asphyxiation; such regions need to be differentiated from genuine test article effects.

Following euthanasia, animals are commonly exsanguinated prior to necropsy. This not only serves as a secondary (insurance) method of euthanasia, but improves the microscopic quality of certain tissues, such as the liver and spleen, by removing most of the blood. Failure to exsanguinate can result in increased organ weights compared to historical ranges, while inconsistent exsanguination can result in variability in organ weights among animals, particularly in liver and kidney ( ; ).

In order to minimize tissue artifacts associated with postmortem autolysis, dissection should begin within 5 min of euthanasia and ideally be completed within 20 min; larger nonrodent species may require the implementation of technical teams in order to meet this time frame. In addition to autolysis, prolonged postmortem intervals have been associated with increased relative and absolute liver organ weights and hepatocellular centrilobular vacuolization with delays of only 25 min ( ).

The goal of the necropsy is to create a detailed individual animal record that documents all essential findings of the event in a simple, concise, and unambiguous manner. The necropsy record must include the animal's unique individual animal number, time and date of death, method of euthanasia, body weight, organ weight data, macroscopic (gross) tissue findings, a checklist of the saved tissues, and signature lines for all personnel of the prosection team, which generally consists of a prosector, an assistant who weighs tissues, and a supervising pathologist; the “weigher” and pathologist may interact with multiple prosectors serially during a single necropsy session. It is important to thoroughly plan and communicate the duties of each member of the team well before the day of necropsy. Careful planning of the necropsy details can save time (and ultimately cost by eliminating the need to potentially rerun a study) and prevent the aggravation of inadequate supplies, missing specimens, or incorrectly preserved samples. A simple but often overlooked method of facilitating a smooth necropsy is to ensure that each member of the team reads and thoroughly understands the necropsy portion of the study protocol, prior to initiating the necropsy. Some institutions formally document this understanding by requiring delegated personnel to sign statements to this effect.

Each member of the prosection team will have a specific set of assigned duties that should be defined in the study laboratory's SOPs and best practices. A skilled prosector is responsible for the examination, removal, and trimming of the organs identified on the protocol-specified tissue list. In some instances, especially studies that involve larger nonrodent species, one or more “trimmers” may assist the prosector, either by dissecting a subset of tissues or by trimming organs prior to weighing. The “weigher” is responsible for ensuring that the weight of all protocol-specified organs is uniformly measured on a validated scale and in accordance with SOP guidelines; one weigher is often capable of recording organ weights from several prosection teams at the same time. A “recorder” may be required if the necropsy data are manually written. The recorder's task is to document all weights, as well as any other findings, such as macroscopic observations. Depending on the study design and the resources of the study laboratory, this role may be fulfilled by the trimmer or another individual on the necropsy team. While it is not a regulatory requirement that a pathologist be present at the necropsy, one should be available to provide expertise on the identification and nomenclature of gross findings, in addition to assisting with any other important aspects of the necropsy process. Although trained necropsy prosectors are capable of recognizing an abnormal tissue, they are generally not trained in assessing its significance or selecting the appropriate descriptive terminology for a particular lesion or set of lesions. When there are several teams of prosectors working during a necropsy, it is the pathologist's responsibility to record the macroscopic findings from all animals. Ultimately, the pathologist must sign the necropsy record for each animal and is legally bound to the accuracy of its content.

It is crucial to assure that all protocol-required tissues, in addition to any gross lesions, are collected at necropsy. There are several ways to ensure that this occurs. The necropsy record should contain a list of all of the protocol-specified tissues. These tissues should be harvested in the same order for each animal. The use of a tissue collection template (typically a card with an area marked for all required tissues) under a sheet of plastic, a clear plastic tray with dividers, or an ice cube tray labeled for various tissues is recommended as it helps to ensure that all required tissues are collected from each animal. It is also recommended that two technicians be involved in placement of the tissues into their appropriate fixative. The first technician (a prosector or trimmer) verifies the tissues with the recorder (the second technician) as they are placed in their appropriate fixative one tissue at a time. The recorder subsequently checks off the tissue on the necropsy record as it is being placed in the proper fixative by the first technician. To prevent loss or misidentification during processing for histopathologic examination, small tissues may need to be placed in labeled cassettes. Flaccid organs and tissues, such as open segments of the gastrointestinal tract, skin, skeletal muscle, and nerve, should be placed on a dry piece of stiff white blotter paper or card stock to prevent curling. It is important to remember that collection of protocol-specified tissues typically includes the identification and retrieval of all macroscopic findings irrespective of the location of the finding ( ). During the necropsy process, it is crucial that the collected tissues do not dry out prior to their immersion in fixative. Drying can adversely affect the quality of tissue preservation and result in artifacts, such as shrinkage and fractures. To prevent this, tissues can be sprayed with isotonic saline (phosphate buffered or “physiologic” [0.9% NaCl]) or gently covered with saline-soaked gauze while on the necropsy table or in the collection tray. It is critical that tap water not be used as its hypotonicity can result in the production of tissue artifacts, such as cell swelling ( ).

Prosectors must be appropriately trained in and experienced with proper tissue handling techniques to ensure that they understand how to correctly manipulate unfixed tissues during the collection process. Excessive manipulation of tissue, such as increased digital pressure; trauma caused by crushing, puncture, or squeezing of tissues with dissection instruments; or distortion due to pulling or stretching, can create artifacts that confound histopathologic examination. For example, markings caused by forceps can be seen in the liver, lung, and other parenchymatous organs, and nerve fiber deformation can be seen in nervous tissue such as brain, spinal cord, and sciatic nerve that have been aggressively pulled/stretched during removal ( ).

Necropsy data can either be manually written or electronically recorded. While manual recording of data on a necropsy record is acceptable, it is fraught with potential challenges such as illegible characters, general uncleanliness (due to smudged ink or fixative spots), and an increased risk of missing or misplaced data. It is currently common practice to employ commercially available electronic pathology data management programs such as Provantis (Instem) or Pristima (Xybion). These programs possess specific modules that allow critical information, such as animal identification number, body and organ weights, and gross findings, to be recorded into an online database. When properly validated, these programs have features that help to prevent the loss of data by prompting the user to verify that all listed tissues and organ weights of interest have been recorded prior to proceeding to the necropsy of the next scheduled animal.

Documentation of gross findings is a critical stage in study evaluation. Generally, all gross abnormalities should be recorded. In some cases, normal physiological changes should be documented as well; an example of this is dilatation of the uterine horns in mice and rats during proestrus and estrus. Although this is a normal change, the degree may be affected by compounds, such as endocrine modulators, and therefore should be recorded in order to determine if there is a test article–related effect. However, the recording of various gross findings is left to the discretion of the laboratory based upon their internal best practices. It is important that gross findings are described rather than merely diagnosed, as an official diagnosis requires microscopic examination. Gross descriptions should include the following: location/distribution, number, shape, size (as maximal or three dimensions), color, consistency, and any additional special features (e.g., presence of a capsule) that characterize the finding. While gross findings can be recorded manually, this may result in variations in descriptive terminology of comparable gross findings. Therefore, the use of commercial pathology data management programs allows for the advanced creation of a consistent and uniform glossary of descriptive terminology. For more information on pathology nomenclature, please consult Nomenclature and Diagnostic Resources in Anatomic Toxicologic Pathology , Vol 1, Chap 25 ).

Total body weight and select organs weights are routinely collected for scheduled necropsies from GLP-compliant toxicity studies, in addition to some other animal studies. Organ weights from found dead or dying (moribund) animals are typically not collected as they are considered of limited value given the absence of matched concurrent control data and differences in nutritional status, exsanguination, and tissue congestion and edema. The STP recommends that the brain, liver, kidneys, heart, adrenal glands, and testes be weighed in all repeat dose GLP toxicity studies of 7 days to 1-year duration; organ weights from acute, single-dose studies are considered of limited value and they are not recommended for carcinogenicity studies, including alternative mouse bioassays ( ). Additional organs, such as thyroid/parathyroid glands, ovaries, spleen, uterus, thymus, and lung, also may be included depending on the species, mechanism of action of the test article, and institutional preference. However, it is important to note that interpretation of data derived from these ancillary organs can be complicated due to residual blood (spleen, lung); age-related involution (thymus); hormone-related cyclical changes (ovaries, uterus); and difficult prosection (thyroid /parathyroid glands in mice). In general, organ weights of reproductive tissues are most valuable in sexually mature animals due to the fact that variability in age, sexual maturity, and cycle stage can complicate or limit interpretation, especially in nonrodent species. While the STP recommends that the testes from all species be weighed in repeat dose general toxicity studies, weights for the epididymides and prostate are only recommended in rat studies; in nonrodent and mouse studies, weights can be considered on a case-by-case basis ( ). Organs should be weighed as soon as possible following removal from the test animal but should be trimmed free of fat and connective tissue first. Paired organs are often weighed together. For some organs, such as brain and thyroid/parathyroid glands, there is an advantage to waiting to weigh them until after they have been fixed as this can aid in minimizing the creation of handling artifacts that might confound the histopathologic evaluation. When weighing a given organ, the fixation status of that organ (fresh or fixed) should be consistent for all animals within the study.

Organ weight data must be normalized to facilitate proper interpretation. This is typically done by expressing absolute organ weight relative to either total body weight or brain weight. Normalized (“relative”) organ weights eliminate the influence of normal variations in animal growth on organ weight data; however, these relative data have limitations that must be taken into account during analysis and interpretation. For example, body weight loss can impact body-to-organ weight ratios as it can result in an apparent false-positive test article effect on organ weights. Therefore, the body-to-organ weight approach is generally considered less reliable than comparing organ weight data to brain weight; brain weight remains relatively constant in mature, nonsenescent adult animals and is generally not affected by changes in body weight ( ). The brain must be collected similarly for all animals when normalizing organ weights to brain weight in rodent studies, by either always including or excluding intact olfactory bulbs, as the paired bulbs account for 6%–7% of the entire brain weight in common stocks and strains of these species ( ). When interpreting organ weight data, it is recommended to consider both absolute organ weights and organ weight ratios (organ weight: body weight and/or organ weight: brain weight). Whenever possible, an attempt should be made to correlate organ weight changes with histopathologic findings. However, dramatic changes in organ weight are often necessary before histomorphologic evidence is observed; subtle changes in organ weight are generally not associated with histopathologic changes. For example, in the liver, microscopic evidence of hepatocellular hypertrophy and/or increased liver enzymes often require liver weight increases of up to 20% relative to the controls before it can be appreciated ( ). Similarly, dose-related changes in kidney weight commonly occur without histopathological changes.