Special Techniques in Toxicologic Pathology

Tissue Cross-Reactivity Studies

TCR studies are nonclinical safety studies conducted to assess the potential for off-target reactivity of therapeutic monoclonal antibody (mAb) and antibody-based therapeutics (see Protein Pharmaceutical Agents, Vol 2, Chap 6). The conduct and applicability of TCR studies is addressed in the scientific literature as well as in various TCR guidelines published by the US Food and Drug Administration (FDA), European Medicines Agency (EMA), and the International Council on Harmonisation (ICH) ( ; ; ). Current industry best practices recommendations indicate that TCR should be performed on human tissues with inclusion of animal tissues if human TCR findings are present ( ). It should be noted that for therapeutic candidates with no mammalian targets such as antiinfectious agents, TCR or other appropriate assessment of off-target binding might be one of the key components of a preclinical safety assessment package.

TCR performed on animal tissue is no longer considered the primary screen for species selection for toxicity and safety testing of therapeutic mAb therapeutics ( ; ). Rather, TCR performed on tissues from toxicology models is useful and often included as part of the larger data package to support selection of a particular species for toxicity testing if the patterns of on- and/or off-target IHC labeling (synonym: IHC staining) are similar to that observed in human tissues and to investigate unexpected or nonpharmacologic toxicity findings in safety studies. In nonclinical toxicity studies in animals, cross-reactivity of the mAb or mAb-based test article with an unexpected tissue in combination with histopathological lesions in that tissue is strong evidence of an adverse result due to unexpected binding of the test article. If the human TCR studies also show a similar distribution of cross-reactivity, additional safety assessments may be needed prior to clinical trials or additional clinical safety monitoring may be warranted. TCR findings should always be interpreted in the context of the entire Investigational New Drug-enabling data package, and a weight-of-evidence approach should be taken in defining whether IHC labeling is linked to the expected expression of the antigen (“on target”), serves as evidence for unexpected antigen localization and potential off-target test article binding, or is spurious ( ).

It is noteworthy that while the technology exists to test other types of therapeutics (such as ligand-binding fusion proteins or even small molecules designed to bind ligands) for binding specificity, human TCR studies are not recommended by the FDA or industry for these nonantibody test articles on a routine basis. In general, TCR studies are required or recommended when the test article includes a complementarity-determining region (CDR) ( ). For test articles that do not contain a CDR, other methods are available that can be used to assess on- or off-target binding and biodistribution (e.g., qualitative whole body autoradiography).

In addition, an ICH S9 Q&A document states that TCR studies are not needed for test articles directed to oncology indications unless there is a specific cause for concern or there is no pharmacologically relevant species for toxicity studies ( ). It should be noted that, as of 2016, there have been no truly tumor-specific antibody or antibody-based tumor therapies available ( ). Given the powerful pharmacologic nature and already precedented adverse effects associated with some of the newer antibody and antibody-based test articles, particularly multispecific molecules and cell-based therapies, some level of risk is indeed inherent in these materials ( ; ). Moreover, for newer multispecific molecules and cell-based therapies, no adequate animal models exist, so TCR testing might still be necessary in these cases.

In a typical TCR study, the mAb or mAb-based therapeutic is applied at two IHC staining concentrations to cryosections of approximately 37 human tissues from at least three unrelated adult donors. Embryonic or fetal human tissues may be included as well for test articles that might be given to infants or women of child-bearing potential. Flash-frozen tissues are preferred for TCR studies as no modification in antigen structure will have been imparted by prior fixation. Ideally, the test article should be assessed in the form it is to be used clinically, but this may not be feasible technically, particularly for chimeric antigen receptor T-cells (CAR-Ts) where a version of the chimeric antigen receptor can be used as a surrogate for the cell-based therapeutic. Positive and negative control slides or cryosections are essential to ensure sensitive and reproducible staining. Such controls might include intact adult tissues, cell lines expressing or not expressing the target of interest, stem cells, or embryonic or fetal tissue that do or do not express the target of interest, or purified target or irrelevant antigens spotted onto and cross-linked to ultraviolet (UV)-activatable resin-coated slides. Other staining controls applied to donor samples include replicate sections stained with species, isotype, and concentration-matched negative control antibody and assay control (omitting the primary antibody) slides ( ; ).

FDA guidance suggests that a positive control antibody against an antigen present on many tissues should be used as a tissue suitability and tissue validation control ( ; ). The tissue suitability controls most commonly used include staining for β ₂ -microglobulin (a ubiquitous antigen expressed on endothelium, many epithelia, and hematopoietic and lymphoid cells) ( Figure 11.1 ); von Willebrand Factor; CD31 (an antigen expressed on endothelium and platelets); or transferrin receptor (a membrane protein on some endothelial, epithelial, endocrine, and Kupffer cells). In some cases, the use of antigen or positive control antibody blocking studies may be useful ( ; ).

TCR slides are interpreted by a pathologist to identify the types and locations of each mAb (i.e., test article)-stained cellular and/or extracellular element. Particular attention is paid to “unexpected” staining (i.e., staining not predicted by known locations of the epitope recognized by the mAb) ( ; ). For example, the T10B9 series of anti-CD3 (T-lymphocyte) monoclonal antibodies recognizes its target CD3 epitope in human tissues, especially T-cell-rich regions of lymphoid organs. It does not react with nonhuman primate (NHP) CD3 but does cross-react with an off-target, cytoplasmic filaments in human and NHP epithelial cells. This off-target cytoplasmic staining is not associated with toxic effects in safety studies ( ).

Technical difficulties may arise during the conduct and interpretation of TCR studies. Since many therapeutic mAbs are fully human or “humanized” (chimeras of animal-derived variable regions and human-derived Fc domains with or without additional mutations to match human antibody protein sequences) proteins, TCR assays need to minimize background staining that may occur due to binding of the secondary antibody (which is antihuman) to human antibodies present within the human donor tissues being tested. Current best practices include labeling of the antibody with a protein tag (e.g., biotin or digoxigenin) or a fluorophore (e.g., FITC or AlexaFluor, which can be used as epitope tags for enzyme-conjugated detection antibody binding) or indirect labeling with a biotinylated secondary antibody via a precomplexing approach ( ; ; ; ; ) ( Figure 11.2 ). Other technical difficulties arise when the epitope of interest is only expressed at low levels or undergoes a conformational alteration during freezing, storage, thawing, and/or a brief exposure to the tissue fixative during IHC staining ( ; ).

Another challenge related to TCR studies lies in interpreting the pattern of staining observed. If a therapeutic mAb or mAb-based test article binds to unexpected tissues, it does not necessarily indicate whether the antibody will specifically bind to and adversely affect that tissue. Binding to membranes is considered to have greater potential toxicologic significance than binding to intracellular components as such sites in intact cells are considered to be less accessible to therapeutic mAbs in vivo. ICH S6(R1) indicates that binding to areas not typically accessible to the antibody in vivo (e.g., cytoplasm) is generally not relevant ( ; ; ).

Furthermore, when mAb or mAb-based test articles cross-react with many nontarget tissues, the safety risk is not necessarily enhanced. However, in these cases, high levels of off-target binding might require further characterization and interpretation in light of the entire toxicity testing package. In addition, the potential for adverse events often does not relate well to the relative degree of cross-reactivity or abundance of off-target antigen as assessed by IHC. However, in other cases, TCR results have provided valuable information regarding potential sites of toxicity, in particular by identifying tissues that should receive close monitoring in human clinical trials. Greater than 20 case studies have been identified and described where TCR results either confounded, were inessential to, or greatly clarified the risk assessment of mAb drugs ( ).

In some cases, a drug-related adverse event may occur in a tissue with apparently very little antigen, while tissues with abundant antigen may be spared. This has been observed with some toxin-conjugated antibodies (i.e., immunotoxins, a special category of antibody–drug conjugates). In these cases, the addition of the toxin to the therapeutic mAb may alter the biodistribution, uptake, transcellular transport, and clearance of the test article (see below). Other factors that can affect the safety profile include antibody class, which may relate to an antibody's ability to fix complement or elicit antibody-dependent cell cytotoxicity events, manufacturing-related variations in glycosylation and alterations of structure (often by virtue of the effect on pharmacokinetics [PK] or on organ of excretion), and conjugations ( ). Overall, toxin-conjugated antibodies have much poorer safety records than plain antibodies.

There are now several array-based technologies also available for assessment of cross-reactivity of antibodies and antibody-based test articles ( ; ). In general, these technologies involve screening cells transfected with vectors designed to overexpress various membrane-bound proteins. The potential benefits of such technologies include the speed and throughput of the in vitro assay, ability to rapidly screen a variety of constructs, ability to identify the protein that is being bound, and potentially lower cost than conventional TCR studies. However, these technologies are still relatively new and have a few drawbacks including no tissue, cellular, or histologic context; incomplete representation of the membrane proteome; no representation of soluble or secreted proteins; no representation of proteomes from species of interest for toxicity studies; no representation of isoforms; and no context of other intramembrane proteins which might interact with the protein of interest.

Test Article Biodistribution

The most direct application of IHC in the evaluation of an administered test article is in detecting the test article and its metabolites. This detection allows one to determine the spatial distribution of the test article and to specifically localize it to certain cell types or extracellular sites. Distribution to unexpected (off-target) organs also may be detected. The ability to specifically visualize the test article in cellular, subcellular, or extracellular sites is an advantage over other methods to determine the amount but not the precise anatomic distribution, including measurement of test article levels in plasma or digested tissue.

Use of Immunohistochemistry for Biodistribution of Cell or Gene Therapies

IHC can be a powerful tool for spatial localization of cell and gene therapy candidates and provide useful information regarding their safety and activity in vivo. Careful planning during study design is important to ensure that tissues are collected and processed appropriately for these analyses and that the staining method itself is suited to the species and tissues to be stained.

For cell-based therapies such as CAR-T or other modified cells, IHC can be used to localize the therapeutic cells in tissues after administration in toxicity/safety study models to determine their biodistribution and persistence over time (see Stem Cells and Regenerative Medicine , Vol 2, Chap 10). A variety of human-specific markers such as major histocompatibility complex proteins (HLA-ABC) or feature-specific molecules expressed in human mitochondria, stem cells (STEM121), and nuclei (HuNu, Ku80) are available that can facilitate localization of engrafted human cells in animal tissues ( Figure 11.3 ). These analyses can be helpful in determining intratumor versus systemic localization of the cell-based therapeutic candidate over time. This tissue context cannot be demonstrated for cell-based therapies by simple quantification using polymerase chain reaction (PCR)–based biodistribution. In addition, other IHC stains including Ki67 and proliferating cell nuclear antigen (PCNA) can be applied to facilitate evaluation of proliferation and tumorigenicity of cell therapies.

For gene therapy, IHC can be used for assessing the biodistribution of the test article (see Gene Therapy and Genome Editing, Vol 2, Chap 8). It is important that a suitable marker for the therapeutic protein is available and/or that there is an antibody available that is specific for the human-derived test article that does not cross-react with any endogenous animal protein. Protein tags (i.e., artificial nonendogenous antigens) such as human influenza HA or FLAG-tag that can easily be detected with IHC methods can be used as a marker for the test article. These data provide tissue, cellular, and subcellular context for the expression of the gene therapy candidate and can complement PCR-based quantitative biodistribution data. In addition, other IHC stains can be used to characterize downstream changes in protein expression or tissue effects. For example, in a mouse model of mucopolysaccharidosis Type IIID, IHC for glial fibrillary acidic protein (GFAP) has been used to demonstrate reduction in astrocytic activation after delivery of adeno-associated virus 9 (AAV9) vectors expressing N-acetylglucosamine 6-sulfatase ( ). Similarly, in a mouse model of Alzheimer's disease, immunohistochemistry for GFAP has been used to demonstrate neuroinflammation in animals that received AAV vectors designed to express antibodies directed against β-amyloid. The observed neuroinflammation was attributed to presence of the expressed antibodies in the tissues ( ). Having a sensitive and specific staining method for such analyses is critical to their utility for assessing safety and functionality of cell and gene therapies.

Use of Immunohistochemistry for Biodistribution of Small or Large Molecule Therapies

IHC can be used to assess the biodistribution of large or small molecules as well as cell and gene therapies in dosed animals. Unlike TCR studies, which are required for CDR-containing biotherapeutics to characterize potential off-target binding, biodistribution studies are not a standard component of a safety program but rather are performed on a case-by-case basis. These studies can provide valuable information about the distribution, persistence, association with observed lesions, or association with other tissue-based markers of effect of a therapeutic candidate in dosed tissues.

For large and small molecule therapies, the ideal IHC test article biodistribution study has the following characteristics ( ):

• a.

  Provides sensitive, specific, and robust demonstration of the test article in paraffin sections (usually replicate sections from tissues obtained during a toxicity or efficacy study);

• b.

  Demonstrates if the test article binds to expected target sites after in vivo dosing to test species;

• c.

  Demonstrates if the test article binds to unexpected or undesired off-target sites after in vivo dosing to test species;

• d.

  Can be used to compare test article distribution produced by different dosing routes;

• e.

  May be employed to evaluate temporal shifts in test article distribution;

• f.

  Demonstrates if the test article is associated with histopathologic findings; and

• g.

  Demonstrates if the test article is associated with immune complex formation and deposition in vivo.

Sensitive and specific test article demonstration is easiest when the test article is antigenically distinct from any endogenous molecule found in the test species tissue and when a commercially available antibody is available with specificity for the test article, and this detection antibody and the staining method have been qualified. For example, IHC can localize microcystin-LR in mouse liver after intraperitoneal injection and associate distribution of this toxin with the temporal onset of hemorrhage and apoptosis affecting hepatocytes ( ). Similarly, a commercially available antiantibiotic antibody has been utilized to localize accumulation of the test article in activated lung macrophages ( ).

Biodistribution studies are more difficult when the biomolecule test article is closely related to an endogenous protein. For example, when the test articles is a human enzyme administered to a heterologous species, that species may have an endogenous enzyme that is closely related to the administered human protein. Thus, even when an antienzyme antibody is available, it might detect both the administered human-derived test article and the endogenous animal enzyme in tissue.

Another example is the IHC evaluation of paraffin sections from animals administered human monoclonal IgG antibodies (mAbs), particularly when the test species of choice is an NHP. Sensitive and specific detection of the human mAb against the cross-reactive background of monkey IgG can be accomplished but only following extensive cross-adsorption of the antihuman IgG antibody ( ). Additionally, the exogenous human mAb becomes incorporated into the endogenous monkey IgG pool and is transported, taken up, and cleared according to the physiologic processes that govern IgG biodistribution (e.g., FcR-mediated clearance, etc.). Thus, in order to localize the human mAb (human IgG), one must also use extensively cross-adsorbed, antimonkey IgG antibodies to identify the patterns of monkey IgG staining ( Figure 11.4 ). These patterns are then “subtracted” from the patterns of human mAb staining to identify potential sites of target or off-target binding of the test article ( ; ).

IHC may also allow evaluation of the temporal differences in distribution of the test article, including the determination of shifts between organ compartments and the total time of duration of the agent in certain tissues (see Principles of Pharmacokinetics and Toxicokinetics , Vol 1, Chap 5 ). Antisense oligonucleotide (ASO) test articles have been monitored in this manner using IHC, sometimes in combination with ISH ( ; ). When ASO test articles are administered to an animal, basophilic granular material can be seen in the kidneys, liver, and lymph nodes on H&E-stained sections, particularly when higher doses are given ( ) (see Nucleic Acid Pharmaceutical Agents , Vol 2, Chap 6). Using an antibody directed toward the antisense test article demonstrated that these granules represent either the oligonucleotide itself or its metabolite. When new antisense test articles are developed, IHC can be used to evaluate their distribution. This method is particularly useful when tissue concentrations are not high enough to visualize the granule accumulation using standard H&E staining. Small interfering RNA test articles can be localized in tissues using fluorescent dye tags. Localization of these test articles can be further characterized using IHC and colocalization for other cell-specific markers ( ; ).

A potential caveat to using IHC to detect administered test articles in tissues is that in some cases the method may not be sensitive enough to detect low levels of the test article. For example, IHC may not detect the small amounts of mAb that reach other tissues via the systemic circulation following intravitreal exposure. Combining IHC with other, more sensitive methods such as ELISA to measure administered mAb in serum or using amplification techniques in the IHC protocol (e.g., ABC, CARD, and TSA) could resolve such problems.

Lesion Identification and Characterization

When a morphological alteration is identified by routine H&E evaluation in a nonclinical safety study, it may be important to further characterize that alteration. Characterization should always be designed on a case-by-case basis using a focused approach. IHC staining should only be done on slides from a safety study when the question to be answered is well defined and the appropriate methods development, qualification, and/or validation have been performed to ensure interpretable and reproducible data.

Depending upon the histopathologic alteration, IHC staining can provide valuable information regarding many aspects of the lesion pathogenesis. For example, IHC markers may be employed to identify the cell type, determine the lineages and/or functional states of responding cells (especially resident or infiltrating immune elements), assess the relative abundance of intracellular proteins such as hormones, provide evidence of cell death or proliferation, localize test article, and detect the presence of immune complexes. These data can add context and aid in risk assessment of observed tissue changes.

Finally, IHC staining using panels of antibodies can be useful in determining the histogenesis of a tumor. Use of marker panels expected to be expressed or not expressed in a given tissue can be useful in determining the origin of an observed mass and can offset the pathologist's internal bias ( ). Epithelial cell markers (e.g., cytokeratins) as well as markers for mesenchymal cells (e.g., vimentin) and hematopoietic cells (e.g., CD45) should be included in such a staining panel ( ). For example, in a U.S. National Toxicology Program (NTP) study, mouse uterine cervical granular cell tumors examined using IHC with a panel of antibodies were found to be positive for smooth muscle actin and desmin and negative for S-100 and other neural markers, suggesting myogenic rather than neural origin ( ).

Evaluation of Responses to Test Article Administration

IHC is also valuable for evaluating alterations in host response to target organs following administration of a test article. In this respect, it can generate much information about the mechanisms of action and toxicity of an agent.

Biomarkers of Exposure and Effect

Evaluation of Leukocyte CD Markers or Other Surface Markers (e.g., Chemokine Receptors) in Response to Immune-Modulating Agents

Current international guidelines include IHC as a potentially useful method to evaluate the immunotoxicity of a test article in nonclinical and clinical studies ( ). In current toxicologic practice, IHC is used to evaluate changes in leukocyte cluster of differentiation (CD) antigens or other cell type–specific surface markers in response to the administration of agents that modulate the immune system. These types of changes might be difficult to observe with evaluation of H&E-stained sections alone as changes in numbers of some cell types after treatment might be subtle, especially for some immune-modulating test articles.

Flow cytometry provides a sensitive and specific monitor of CD, cytokine, and chemokine receptors on peripheral blood cells or ex vivo tissue suspensions but provides limited information regarding distribution of these markers with respect to tissue architecture. IHC staining for different CD antigens in lymphoid or nonlymphoid tissues can illustrate normal architecture ( Figure 11.5 ) or can provide vital information regarding toxicologic or pharmacologic alterations. For example, IHC staining was used as part of the weight-of-evidence approach to demonstrate potential efficacy of combination therapy with the B-cell immunomodulators atacicept and rituximab in monkeys ( ).

When IHC labeling is used as a predictor of efficacy in nonclinical studies, a similar IHC assay also may be useful as a predictor of responsiveness in clinical trials, especially when the marker can be used as a selection criterion for clinical trial entry. The number of CD antigens and the availability of antibodies directed against them is constantly growing, and investigators should thoroughly review whether the CD antigen in which they are interested is appropriate for the question being asked as well as for the test species (and ultimately for translation to support the human clinical trials indication) under consideration ( ; ).

If the test article targets a certain CD or chemokine receptor molecule, changes in the presence of that marker on cells may be evaluated by IHC ( ). On numerous occasions in our experience, no apparent changes were present in lymphoid tissues based on evaluation of H&E-stained sections, whereas changes in the numbers and/or distribution of various leukocytes were obvious using IHC evaluation of CD markers. However, because subtle differences in cell numbers may not be detectable by conventional semiquantitative evaluation of IHC-stained sections, quantification of cell numbers may be better approached by flow cytometry on peripheral blood cells or cells isolated from lymphoid organs or by digital image analysis of sections containing chromogenic or immunofluorescent signals for specific cell types.

Evaluation of CD markers not only has value in adult animals but also can give information about the effects of a test article on the developing immune system. In our experience, IHC evaluation of CD marker expression has illustrated on numerous occasions that there may be obvious effects on the numbers and distribution of various cell types without the presence of clinically detectable immunosuppression (including the lack of detectable opportunistic infections, a hallmark of clinical immunosuppression). This is often useful information for the developers of immunomodulatory compounds since it demonstrates that their therapeutic candidate has a desired, but not overwhelming, effect on regulating the function of the immune system.

One limitation of using IHC to evaluate changes in CD, cytokine, and chemokine receptor molecule expression is that it detects changes in numbers of cells and their distribution but does not generally give any information about cellular function. The exception to this is the use of antibodies against CD markers found on activated cells of various types where IHC effectively detects activated cells expressing activation state-specific markers ( ; ). A limitation of this approach is that many activation markers are expressed on multiple immune cell types; therefore, multiplex IHC labeling with fluorescently labeled antibodies is often used to detect subsets of activated cells. Commercial antibody suppliers are often very helpful in choosing appropriate activation markers via their catalogs and technical service representatives. Other methods including flow cytometry and functional assays that more specifically evaluate immune cell function can be combined with IHC to provide an overall functional assessment so that effects on the immune system are not missed if assays are used in isolation.

Cell Proliferation Studies and Evaluation of Cell Cycle–Related Proteins

IHC evaluation of cell proliferation is commonly performed, particularly for agents known or suspected to cause changes in cell proliferation rates or neoplasia. Numerous methods are available to perform these studies, each of which has advantages and disadvantages ( ). They all have an advantage over simple evaluation of the numbers of mitotic figures in routinely stained H&E sections because the IHC-stained cells are easier to see and count, and for some markers cells are detectable in various stages of replication in addition to mitosis. Originally, tritiated thymidine was used to evaluate cell proliferation, but this method has mostly been replaced by nonradioactive methods. For example, BrdU, a thymidine analog, is incorporated into cellular DNA in the S (synthesis) phase of the cell cycle and can be detected by IHC using antibodies to BrdU resulting in an accurate and reliable technique for assessing cell proliferation. However, BrdU must be administered to test animals either by injection 1 to 2 hours prior to necropsy, by means of subcutaneously implanted minipumps, or by administration in the drinking water. As such, BrdU is not the ideal method, particularly in the context of toxicity studies, where there is often a preference to avoid techniques that require administration of an agent to the animals during the course of the study.

More popular methods for IHC evaluation of cell proliferation, especially for toxicity studies, include staining for PCNA, Ki67, and phosphohistone H3 (PhH3). PCNA and Ki67 staining offer advantages compared to BrdU staining since these techniques do not require prior injections and can often be performed retrospectively on formalin-fixed, paraffin-embedded tissue. Furthermore, BrdU stains only detect cells in S phase, while PCNA and Ki67 detect cells in all stages of the replicative portion of the cell cycle (G1 + S + G2 + M). Some investigators believe that PCNA-positive cells can be separated according to each of these stages based on characteristic differences in the pattern of nuclear staining. Studies using characteristic differences in nuclear staining of PCNA to compare S phase labeling of PCNA to that of BrdU have found that the labeling indices are similar for the two methods and that the methods are equally useful for evaluating cell proliferation. Histone H3 (PhH3) is a nuclear protein that is a component of chromatin. Its phosphorylation is a key event in chromatin condensation during mitosis and meiosis, and thus, it can serve as a useful marker of mitotic cells ( ). PhH3 has been used as an early marker of carcinogens inducing cell proliferation in a number of tissues in rats treated with carcinogens targeting different organs ( ). In another study, PhH3 was used to demonstrate the role of p53 in microcystin-LR tumor promotion in mice ( ). One method may be preferred over the other depending on the specific situation ( ; ; ; ).

Examples of the use of cell proliferation markers in toxicologic pathology abound in the literature and it should not be difficult to find information appropriate for a desired situation. Experts in this field caution that with regard to evaluating cell proliferation in an attempt to assess for a potential carcinogenic effect, an increase in cell proliferation does not necessarily indicate that such proliferation will lead to neoplasia. For example, a specific study in which both proliferation and apoptosis were evaluated for two different agents determined that the relative balance of the two different processes might be important in determining why carcinogenesis occurs with some nongenotoxic agents and not others ( ). It should be noted that a single assessment for cell proliferation might not be sufficient for prediction of tumor outcome; rather, understanding the extent of cell proliferation over the entire course of the disease via assessment at multiple time points might provide more valuable information in this regard.

IHC also is used to evaluate changes in other molecules involved in cell cycle control, particularly caspases, cyclins, and cyclin-dependent kinases. Numerous antibodies are available commercially for these molecules. Various investigators have used IHC to evaluate changes in these cell cycle–related proteins in response to administration of various toxicants. The application of digital pathology techniques (see Digital Pathology and Tissue Image Analysis , Vol 1, Chap 12 ) should provide the ability to rapidly gather more quantitative cell proliferation data ( ; ).

Evaluation of Apoptosis and Necrosis

Apoptosis, or programmed cell death (PCD), is an active cell process that plays a critical role in a number of normal physiological events, including embryonic modeling, organization of the central nervous system, metamorphosis, etc ( ). PCD has received widespread attention as a crucial component of a number of pathologic processes, including neoplasia and other disturbances in growth, inflammation, and immune responses ( ; ). The best-defined biochemical alteration in cells undergoing PCD is generation of an endonuclease that cleaves nuclear DNA at the regions between the nucleosomes. This cleavage yields characteristic DNA fragments with sizes that are multiples of approximately 180–200 base pairs. Gel electrophoresis of DNA extracted from apoptotic cells shows a characteristic “DNA ladder” that is diagnostic for PCD. Morphologic criteria for PCD are also very specific at the ultrastructural level. Apoptotic cells are characterized by the shrinkage of the cytoplasm and nuclear chromatin condensation. The cellular shrinkage results in apoptotic bodies, which are typically phagocytosed by macrophages or other surrounding cells ( ; ).

Although the morphologic criteria of PCD are well defined, apoptotic cells can be very difficult to recognize or confused with other cellular changes in standard histologic sections. This difficulty stems from both their relative rarity in most tissues and their transience—commonly a few hours only. Two methods have been described that utilize the presence of DNA breaks to label single cells undergoing PCD. One method uses DNA polymerase I (Pol 1), and is called either in situ nick translation (ISNT) or in situ end labeling (ISEL) in the literature, where “in situ” reflects the location where the method is employed (i.e., in a tissue section), “nick translation” indicates that the method leads to extension (by translation) from a site of DNA strand breakage, and “end labeling” denotes that the extension occurs at the broken end ( ; ). The second method uses terminal deoxynucleotidyl transferase (TdT) and has been called either the tailing reaction or TdT-mediated bio-dUTP nick end-labeling (TUNEL) ( ). While these methods involve a combination of molecular biology and IHC techniques, they will be covered here in order to include them in a discussion of IHC techniques used to evaluate various stages of the apoptotic pathway.

ISEL or ISNT relies on the ability of DNA Pol 1 to add labeled nucleotides to the 3′-hydroxyl end of a DNA strand in the presence of a template, thereby extending the strand in a 5′ to 3′ direction. It is hypothesized that ISEL probably detects the 3′ recessed DNA fragments formed in apoptosis. Because DNA Pol 1 is primer- and template dependent, it cannot label the blunt-ended or 5′ recessed DNA fragments that may also occur in PCD. DNA Pol 1 also has an exonuclease activity that, starting at the site of a single-strand break, may remove unlabeled nucleotides ahead of the enzyme, allowing their replacement by labeled nucleotides ( ; ). Many vendors now offer the large subunit of DNA Pol 1, known as the Klenow fragment, for use in molecular biology and related applications. The Klenow fragment lacks any 3′ to 5′exonuclease activity while retaining the primer extension and proofreading capabilities of DNA Pol 1, and has also been used with success in the ISEL procedure ( ; ).

The TUNEL method is based on the specific binding of terminal TdT to 3′-OH ends of DNA, with subsequent incorporation of labeled deoxyuridine at the sites of DNA breaks. Biotin and digoxigenin are the common choices for labels. Because TdT is template independent, it can label blunt-ended, 3′-recessed or 5′-recessed DNA fragments at the hydroxylated 3′ ends ( ; ; ).

These methods have become very popular, and commercial kits are available to specifically label apoptotic cells. However, controversy has arisen regarding the specificity of these enzymatic reactions. Both methods rely on the enzymatic detection of DNA fragments in tissue sections. Although DNA fragmentation is a key feature of PCD, DNA fragmentation also occurs in necrosis, albeit at a later stage in the death of the cell. Some groups suggest that these methods do not reliably detect apoptotic cells in tissue sections, and in particular are unable to differentiate apoptotic from necrotic cells, or even from early postmortem autolysis. Other authors emphasize the value of these methods but caution that they must be used in conjunction with the morphologic assessment of the labeled cell in order to distinguish necrosis from PCD ( ; ; ; ).

Many antibodies and methods are now available for detecting earlier changes in the apoptotic pathway by IHC ( ). Some IHC markers for apoptosis include CC3, cleaved cytokeratin 18, phosphorylated histone H2AX, and cleaved lamin-A ( ). Some of these are more specific than the TUNEL assay, although all have their limitations. CC3 is frequently used as a marker for apoptotic cells. The epitope recognized by antibodies directed against CC3 is not present in healthy (i.e., nondegenerating) cells and is similar across species, making this an attractive marker for apoptotic cells ( ). However, using multiple methods in combination may provide more valuable information than using any single method alone. In our experience, use of CC9 and CC3 provides useful information about early and later events in the apoptotic cascade, respectively ( ; ). Additionally, the application of digital pathology methods to apoptosis studies may facilitate the acquisition of more quantitative data and improve intergroup comparability in toxicity studies.

Necrosis is cell death usually resulting from acute cell injury and is characterized by loss of cell and organelle membrane integrity which leads to cell swelling, release of proteolytic enzymes, and ultimately cell death. The subsequent leakage of cellular components generally results in inflammation ( ). Necrosis is most often diagnosed by evaluation of H&E-stained slides and is generally characterized by loss of cellular detail, often associated with cellular debris and inflammation. Other changes associated with necrosis include cell swelling, nuclear swelling, karyolysis, karyorrhexis, nuclear pyknosis, pale eosinophilic cytoplasm, and cytoplasmic vacuoles ( ).

In some cases, apoptosis and necrosis can be difficult to distinguish. In these cases, transmission electron microscopy (TEM) is considered the “gold standard” for differentiation of these processes ( ). Other methods used for confirmation of necrosis include measurement of enzyme release into serum (e.g., beta-glucuronidase for hepatic necrosis and creatine kinase for myocardial necrosis) ( ; ); exclusion dyes (e.g., propidium iodide, ethidium homodimer III) ( ), or other methods such as flow cytometry and western blotting ( ). There are commercially available kits that use dyes to distinguish apoptotic cells from necrotic cells in suspension or adherent cells, but these kits cannot be used on whole tissue sections.

Use of IHC or exclusion dyes to detect necrosis in whole tissue can be difficult or impossible. IHC on tissue sections with necrotic areas can be difficult to interpret since antibodies often bind nonspecifically to necrotic tissue. In addition, there is no one marker that is truly a specific marker of necrosis. Use of exclusion dyes on tissue sections is also not possible since the nuclear membrane is cut during histological sectioning, thereby exposing nuclear DNA. However, exclusion dyes have been reported to detect necrosis in renal tissue after perfusion of the whole organ with an exclusion dye ( ). In addition, there are reports where IHC has been used successfully for elucidation of necrosis in tissue sections in conjunction with other markers for apoptosis and etosis (i.e., leukocyte death leading to release of an extracellular trap lattice composed of DNA, histones, and cytoplasmic proteins) ( ). However, if IHC is being considered for detection of necrosis in tissue sections, careful method development is necessary to ensure that any observed binding is not due to nonspecific “sticking” of the antibody reagents to necrotic material.

Evaluation of Infectious Disease Agents

Immunohistochemistry using antibodies against various infectious agents is commonly performed in diagnostic pathology and can be useful in toxicologic pathology (see Basic Approaches in Anatomic Toxicologic Pathology , Vol 1, Chap 9 ). If immunomodulatory compounds are administered, it is possible that increased susceptibility to infectious agents may occur in the test species due to immunosuppression, and such agents may be detected using IHC. Evaluation of background infectious diseases in a research animal colony also may be undertaken in some instances. Animal models of infectious disease are used frequently to evaluate the efficacy of potential antiinfective test articles, and IHC for the particular infectious agent may help evaluate responses to therapy. Finally, viral vectors often are used to deliver therapeutic genes, and IHC may be used to detect the presence of the viral capsid proteins in the vector to ensure delivery to the target organ (though ISH for the transgene sequence is often used for this purpose as well) ( ).

Availability of Antibodies

A significant problem facing investigators wishing to use IHC to detect target molecules is the availability of antibodies specific for the molecule. Proprietary mAb or mAb-based test articles generally will not have commercially available antibody reagents to enable IHC, so the antibodies needed for IHC will have to be specially generated for this purpose by the institution. In addition, even if antibodies to the target molecule are available commercially, they may not work in IHC but only in homogenized tissues (e.g., western blots). Commercial antibody vendors will generally have information regarding the applicability of their reagents to IHC and the ability of their products to detect related antigens from various species (animal and human); for specially generated antibodies designed to detect a novel test article, this information will have to be worked out experimentally. Commercial antibody reagents for IHC may have cross-reactivity with off-target antigens that may complicate interpretation.

Due to the widespread use of rodents in research, large numbers of antibody reagents specific for rodents have become available commercially. This is good news for those investigators using rodents in their studies and has been important in understanding many biologic responses (e.g., rat pituitary cell type determination). Some antihuman antibodies may cross-react with one or more animal species, but in the absence of such information, trial and error testing must be performed to determine the level of cross-reactivity across species. Many antihuman antibodies will cross-react with NHP tissue because of the close relationship between these species. If long-term studies with a critical test article are planned, a company may decide to synthesize (or have synthesized) a custom antibody reagent for the specific purpose of developing an IHC technique.

For some molecules, myriad antibodies are available, and the investigator must choose between a mAb (directed against a single epitope of the antigen) made in one species and a polyclonal antibody (directed against multiple epitopes of an antigen) made in a different species. There is no general rule of thumb regarding which type of antibody performs the best in IHC, and the choice of antibody must often be made empirically by testing several. In addition, some commercially available antibodies will only work in frozen tissues. Therefore, it is important to plan ahead with respect to tissue collection and fixation conditions if it is at all possible that IHC will be performed.

Tissue Collection and Fixation for Immunohistochemistry

Sampling of desired tissues should be considered before a toxicity study is even designed in order to have adequate tissue samples for both standard histological examination and IHC. For larger species, a particular tissue usually can be split into two parts, or in the case of lymph nodes, two or more of a specific type of node may be collected, if the fixation for routine processing and IHC requires different conditions. However, due to the small size of rodents and fetuses, it is not always possible to split tissues. Therefore, additional animals may need to be added to the study, with some dedicated to IHC and the others to routine histology.

It may be desirable in larger institutions that perform IHC on an extensive scale to establish banks of appropriately collected and fixed tissues (both frozen and paraffin embedded) from normal animals. In this way, suitable control tissue is always available in the event that the cross-reactivity of a commercially available antibody reagent needs to be tested or for small pilot experiments in which concurrent control animals may not have been included in the study.

Target molecules in tissue sections need to be immobilized prior to the IHC because of the possibility of postmortem antigen diffusion. Target immobilization is usually accomplished through fixation (as in the case of formalin-fixed, paraffin-embedded tissue). The choice of fixative may impact the ability of the antibody reagent to bind to its target as not all epitopes or antigens survive all fixation procedures.

Some antibody reagents will only bind to their intended target in fresh frozen tissues, while others might be used successfully in both frozen or fixed tissues ( ). If fixation may be used, overfixation should be avoided, which typically is accomplished by limiting the time spent in fixative and/or by using coagulating fixatives (also termed denaturing or precipitating fixatives, e.g., acetone, ethanol) rather than cross-linking (e.g., aldehydes) fixatives ( ). As mentioned above, it is important to plan ahead during the study design phase to ensure that tissues will have been collected and processed appropriately for any potential IHC staining that might be needed to interpret any histopathologic findings.

For tissues that have been fixed in neutral buffered 10% formalin (NBF, commercial formulations of which contain about 4% formaldehyde mixed with approximately 1% methanol as a stabilizing agent) or methanol-free 4% formaldehyde (known colloquially as 4% paraformaldehyde [PFA] since it is made from PFA pellets or powder), IHC may be problematic, depending on the antibody and depending on the length of time the tissues have been in fixative. It is most desirable that tissues only be fixed in aldehydes for no more than 24–48 h and then transferred to 70% ethanol if IHC is to be performed. Lengthy fixation in formaldehyde-based fixatives may destroy any hopes of performing IHC using many antibodies due to overcross-linking of protein. For many antibodies, retrieval of antigen in formalin-fixed, paraffin-embedded tissue may be necessary, using any of a variety of methods including enzymatic digestion and heat ( ; ; ; ). Delicate antigens may be detected only if fixation occurs after acquisition of frozen sections. In such cases, sections are immersed in fixative (typically acetone, NBF, or other suitable fixative as determined during method development) for a few seconds up to several minutes.

Controls

Anyone who has performed IHC is aware that there are many potential pitfalls to accurate performance of the assays and interpretation of the results. Proper controls are always essential to the correct interpretation of the data. At a minimum, a serial section of tissue should be stained with an isotype-matched control antibody (for monoclonal antibodies) or pre- or nonimmune serum (for polyclonal antibodies) to demonstrate specificity of the primary antibody. In some cases, the IHC reagents may be the cause of nonspecific background staining. In this case, staining a series of sections where a different component of the IHC reaction is withheld would be useful in identifying the problematic reagent ( ).

Knowledge of the expected staining results based on prior studies and a thorough understanding of the target protein is helpful ( ). Expression patterns in various tissues gained by western blotting may be helpful in defining the list of expected binding sites for novel antibodies. Otherwise, trial and error and thorough investigation of the reasons for unexpected staining are often necessary to ensure accuracy.

Bright-Field Versus Immunofluorescence Microscopy for Immunohistochemistry

The choice of bright-field versus immunofluorescent microscopy for IHC can be made based on user preference or may be dependent on the staining method used. Multiplex IHC often is conducted more readily using immunofluorescence, for several reasons. First, distinct primary antibodies, each labeled with a different fluorophore having a unique emission spectrum from the other dyes, can be applied simultaneously. In constrast, for multiplex IHC using chromogenic methods, more care must be taken since the immunobridges to detect different antigens need to be built sequentially. Second, immunofluorescence might be the method of choice if there is interference from endogenous pigment (e.g., lipofuscin) or molecules used to label IHC reagents (e.g., biotin) in a chromogenic method. Third, in some instances, immunofluorescent staining is more sensitive than chromogenic staining and thus may better detect proteins expressed at very low levels. However, potential disadvantages of immunofluorescence methods include tissue autofluorescence (e.g., collagen and erythrocytes) and lack of fine tissue/cellular detail since the immunofluorescence method is dark field and other cellular/tissue structures are not readily visible with counterstain, thus limiting tissue or histologic context. Further, fluorescent staining is not permanent, so optimally whole-slide images (WSIs) would need to be prepared for long-term retention of fluorescent staining data. Consideration for long-term storage of large data files for high-resolution images would be necessary in these cases. In these instances, chromogenic IHC, where the colored product is more stable over time, might be the preferred method for long-term retention of the study materials needed to reproduce the data set.

Multiplex Labeling and Combination Techniques in Immunohistochemistry

Multiplex IHC, the staining of a single sample using multiple primary antibody reagents detecting distinct antigens, is an effective method to characterize the spatial distribution and coexpression of more than one biomarker in cells and tissues. The technique is becoming increasingly important as it enables the simultaneous collection of data for expanded lists of protein targets that have emerged to improve disease profiling or guide treatment while preserving limited sample material.

Multiplex labeling can be performed either using fluorescent or bright-field chromogenic techniques. Chromogenic immunohistochemistry labeling typically makes use of different enzymes (e.g., HRP, AP) and different soluble noncolored dyes that can be converted to colored precipitating products ( Figure 11.6 ). The chromogenic double labeling technique is extremely useful in toxicologic pathology to determine the spatial location of an administered test article and a substance changed as a result of exposure to the test article, or to look for activated leukocytes infiltrating into tissue as a result of treatment with a certain test article or in response to a particular lesion (e.g., a tumor). However, study of biomarker colocalization is limited as commonly available chromogenic dyes used for IHC have poorly separated spectra features, which makes it challenging to unambiguously differentiate these dyes when they overlap.

In contrast, immunofluorescence labeling circumvents this limitation as fluorophores with disparate emission peaks can be separately imaged using fluorescence filters with restricted bandwidths so that binding of each labeled antibody can be interpreted in isolation, or paired antibodies can be assessed in any combination. Traditionally, immunofluorescence multiplex staining is performed with primary antibodies directly conjugated with fluorophores. However, it is cost inefficient to conjugate multiple primary antibodies, and sensitivity is often poor with low expression antigens. There are now commercially available and easy-to-use tyramide-based reagents that permit the staining of seven or more targets on a single tissue section (e.g., Opal multiplex assay, TSA kits) regardless of the species of the primary antibody host. Such techniques have also been applied successfully for the multiplex of ISH and immunofluorescence labeling ( ; ).

Quantitative Analysis in Immunohisotchemistry

A major limitation of IHC is that it has historically been more qualitative and not quantitative, demonstrating the presence but not the amount of the antigen. An exception to this is in studies where numbers of positively stained nuclei or cells (i.e., discrete objects) may be counted. Specific applications of quantitative IHC analysis in toxicologic pathology include cell proliferation assays and distinguishing the presence of human cells (cell-based therapies) in animal tissues. In most instances, however, a qualitative evaluation of changes in staining patterns and intensity has been sufficient to provide valuable information to the researcher. Digital pathology and computerized image analysis are rapidly evolving methodologies with the ability to objectively quantify IHC data through the use of algorithms (see Digital Pathology and Tissue Image Analysis , Vol 1, Chap 12 ). An important concept in IHC quantification of protein expression by digital image analysis is that true quantification cannot be done by counting positive pixels and their intensities. Instead, the algorithm must be calibrated to a known standard (e.g., cell pellets or tissues with known levels of protein expression) processed using the same IHC conditions and at the same time as the test samples.

When performing any type of quantitation, whether through manual counting or through the use of image analysis, it is important to design the study appropriately and understand when bias may be introduced into the quantitative methods. For example, manual counting of positively labeled cells in a certain number of “randomly chosen” fields of view contains significant inherent user bias, as the eye will always be drawn to fields with higher positivity. Care must be taken when choosing fields of view to count since bias related to sampling and processing of the tissue as well as geometrical features of the objects being counted may affect the accuracy of data acquired by image analysis. It is important to capture samples that are truly representative to minimize sampling bias, and unbiased sampling is the proper way to accomplish this task. It is also important to take tissue shrinkage, particularly when paraffin is used, into account when determining the density of positive cells within the tissue as packing density increases with increased tissue shrinkage. Finally, geometrical bias of the objects of interest can be important, particularly for larger objects that are irregularly shaped. Additional explanations of bias are given below in Section 9 (Stereology).

Regardless of whether the endpoints are qualitative or quantitative in nature, it is important to have an adequate sample size to ensure statistical power and suitably cover the range of normal staining in control animals. At least 5 animals/sex/groups are needed to come to any reasonable conclusion about the results with certainty, and 10 sex/groups are preferable to ensure adequate statistical power. Because quantitative analysis is more sensitive than qualitative evaluation, results may be more reliable if fewer samples per group are present for evaluation. In addition, the interpretation of TCR studies is primarily descriptive with no statistical analysis performed, so fewer donors (usually three) per tissue are generally included in those studies. It is also important to consider staining variation within the tissue, particularly if the tissue is heterogeneous (i.e., staining for islets of Langerhans in the pancreas or staining for a particular neuron type in the brain). In these instances, it is imperative to capture several sections of the region of interest for evaluation, such as samples through the head, body, and tail of the pancreas.