20 Immunohistochemistry quality control
Since its introduction into routine histopathology in the 1980s, immunohistochemistry has become established as an integral part of the diagnostic process. The use of immunohistochemical staining has developed within the diagnostic arena of an increasing range of infectious, neoplastic and reactive disease processes. Immunohistochemistry is now also able to provide prognostic or predictive information such as the likely response to specific treatments. Examples of such ‘pharmaco-diagnostic’ markers include estrogen (ER) and progesterone (PR) receptors and HER2/neu overexpression as well as CD117 (c-Kit) and CD20.
Reflecting the increasing diagnostic and prognostic role played by immunohistochemistry, the markers are known to generate results which may directly impact upon the patient care pathway. Consequently, it is vital that these investigative procedures are properly controlled by monitoring quality, both internally and externally.
External quality monitoring of clinical laboratories is achieved by participation in an external quality assurance scheme, such as UKNEQAS and NordiQC. Whilst in some countries participation in such schemes is a mandatory requirement in order to gain laboratory accreditation, the function of these schemes is primarily to monitor and improve the performance of the participating laboratory over a range of immunohistochemical tests.
It is important to ensure the correct internal quality control measures are in place. Immunohistochemisty is a complicated process and it is essential that the biomedical scientist has a sound understanding of all the requirements and procedures involved. There should be staff experienced in identifying and resolving associated diagnostic procedural problems in order to be able to provide effective and efficient quality control within the laboratory. In addition to technical understanding, the laboratory scientist should also have knowledge of the expected staining patterns for the antibodies in both pathological and non-pathological tissues. Good communication between the biomedical scientist and the pathologist must be maintained, particularly during the introduction and validation of new antibodies and procurement of positive control material.
Detailed documentation and an audit trail throughout the process are necessary for potential back-tracking and troubleshooting. Such audit trail details can include antigen retrieval methods, antibody dilution data, control tissue samples, temperatures and incubation times. The advent of automated platforms for immunohistochemistry has improved this aspect of quality control, but vigilance is still required. These automated platforms generally use standardized protocols for antigen retrieval and staining procedures, which makes overall control of the process easier. The generation and storage of automated run logs by these platforms make full reagent traceability possible. The logs can also be interrogated in the event of abnormal staining to identify errors such as missed steps due to low reagent levels.
Tissue fixation has a significant influence on immunohistochemistry as most antigens are altered during this process (Williams et al. 1997). The purpose of fixation is to preserve tissue and prevent further degradation by the action of tissue enzymes or microorganisms. As discussed in an earlier chapter, good fixation requires tissue to have adequate time in the fixative to allow the solution to penetrate, retaining uniform cellular detail throughout the tissue. However, in the routine laboratory this ideal may be compromised as it is difficult to define a standard tissue size, fixation time, and fixative for each specimen type. Tissues need to be adequately, but not over-fixed, so that antigenicity is preserved without excessive alteration. Prolonged fixation can result in the irretrievable loss of many antigens, particularly membrane-associated antigens such as CD20 and immunoglobulin (Ig) light chains (Miller et al. 1995; Ashton-Key et al. 1996). Lack of adequate fixation, or delay in fixation, may also be equally detrimental to labile antigens (Donhuijsen et al. 1990; von Wasielewski et al. 1998; GEFPICS-FNCLCC 1999).
Any fixative used must be compatible with immunohistochemical staining methods and formalin is still the most universal of fixatives. Formulations differ between laboratories and include 10% neutral buffered formalin (NBF), 10% formalin in tap water, 10% formal saline, and 10% NBF with saline (Angel et al. 1989; Williams 1993; Williams et al. 1997). Even though it may be the pathologist’s choice, it can create a challenge for demonstrating certain antigens. Dabbs (2006) characterizes formalin as:
a satisfactory fixative for both morphology and immunohistochemistry provided that a simple and effective antigen retrieval technique is available to recover those antigens that are diminished or modified.
Williams et al. (1997) investigated the effect of fixation on immunostaining to establish whether a specific preparation schedule would allow for the optimal demonstration of all antigens. Of the fixatives tested, 10% formal saline, 10% NBF (except for CD45RO), and 10% zinc formalin (except for CD3) gave the most consistent results overall and showed excellent antigen preservation. More recently, alcohol-based fixatives have been considered as an alternative to formalin (van Essen et al. 2010) and produced satisfactory immunohistochemistry staining. Other fixatives which may still be used in some laboratories include Bouin’s, B5 (mercury), zinc formalin, 10% formal-acetic and Carson’s, which also have an influence on the reproducibility of staining, each presenting a change in pH, length of required exposure and different artifacts.
Fixatives dictate many factors for immunohistochemical staining, such as dilution, antibody incubation time, retrieval method (if applicable), type of retrieval solution, and special pretreatments (e.g. pigment removal). Depending upon the type of fixative used, the protocols may require slight modifications. With the advent of heat-induced epitope retrieval (HIER) (Shi et al. 1991) many of the problems associated with fixation have been reduced and, in conjunction with automated techniques, good-quality staining is achievable on most tissue sections.
As with fixation, all tissue must be appropriately processed to produce successful immunohistochemical staining. Tissue that is inadequately processed will potentially produce poor-quality sections, with poor adhesion to the slides – especially fatty tissue such as breast and skin. Modern tissue processors all have the option to include vacuum and temperature variation at each step, allowing for greater optimization of the procedure. However, high temperatures can be detrimental to antigens that are heat labile. It is recommended that paraffin with a low-temperature melting point be used for this reason.
Regarding paraffin processing of tissues for immunohistochemistry, as with fixation, there is no standard protocol for the optimal demonstration of all antigens, as concluded by Williams et al. (1997). In a study of laboratories in the UK, Williams (1993) found nearly as many different schedules as the number of laboratories participating in the survey. Of the nine tissue-processing factors investigated, only two had any significant effect on immunoreactivity. Increasing the temperature of processing from ambient to 45°C, as well as longer processing times for dehydration and wax infiltration, were both found to improve immunostaining. Other factors including type of processor, type and quality of reagents, time in clearing agent, use of vacuum, most of which had been suggested as possible causes of poor processing (Horikawa et al. 1976; Trevisan et al. 1982; Anderson 1988; Slater 1988), were found to have no effect on subsequent immunohistochemistry.
Microwave processing is now being introduced into some laboratories to speed the processing time and reduce turnaround time for diagnostic specimens, and has been used successfully in conjunction with routine antibody staining. Acceptable staining was achieved when compared to tissues processed in a conventional processor (Emerson et al. 2006). As with all processing, if the tissue is not completely fixed then artifacts will be introduced.
The quality and reproducibility of immunohistochemical staining relies upon the reversal of fixation, which results in the targeted epitope being exposed, allowing for the antigen binding site to be available. The revolution of reversing the hydrogen cross-bonds formed by formalin was introduced by Shi et al. (1991). There are now numerous methods for epitope retrieval including protein digestion techniques or, more commonly, heating the slides in a buffered solution. Cattoretti et al. (1993) introduced the solution most commonly used in standardized retrieval methods. They used a citrate buffer at pH 6.0, which is inexpensive, stores easily, and is readily available commercially or easily prepared in the laboratory. Other buffers used include EDTA-based solutions at a higher pH range, which produce more intense staining of some antibodies. These methods have allowed for the successful demonstration of a much greater range of antigens in tumors, including proliferation markers and oncogene expression. The use of automated immunostainers has brought greater standardization of retrieval methods, as these use standard retrieval solutions with defined reproducible protocols. Non-automated laboratories may have a number of variables that require internal standardization in the antigen retrieval technique including the choice of heating method (e.g. pressure cooker, microwave, etc.), retrieval solution, pH, temperature, volume of the fluid, and the temperature and exposure time while heating and cooling slides.
Equipment commonly used to perform epitope retrieval includes the modified pressure cooker, initially reported by Norton et al. (1994), microwaves, waterbath or a pretreatment module. Some automated platforms have on-board retrieval where individual slide bays can be heated with the appropriate solution on the slide.
Other factors required for successful retrieval include the proper drying and complete removal of water from slides. In addition to avoiding wrinkles or tears in the tissue, these factors will all assist the adhesion of tissue to the slide. With respect to enzymatic proteolytic ‘epitope retrieval’, such as trypsin digestion prior to immunostaining, the choice of enzyme usually dictates the temperature and pH of the solution, as different enzymes have different preferential pH and temperatures. For example, the optimal values for a mammalian-derived trypsin are pH 7.8 at 37°C, with 0.1% calcium chloride included as an activator (Huang et al. 1976). The concentration of enzyme required is dependent on the proteolytic qualities of the product being used. A typical concentration used for many commercial trypsins employed in immunohistochemistry protocols is 0.1%. The concentration, pH, and temperature are then usually held constant, while the time of digestion is varied. The time required for optimal digestion will vary, depending on the antigen under investigation, the quality (proteolytic capabilities) of the trypsin and the length of formalin fixation. For antigens that are only present in small amounts, e.g. immunoglobulin light chains on the surface of B cells, the time for optimal digestion may vary from case to case, depending on how long each case has been fixed in formalin.
Production of high-quality staining is dependent upon the correct storage, handling and application of the reagents used. Once a protocol has been developed, it is important to ensure the reproducibility of the stain. To achieve this, the storage conditions and expiration dates of in-house and commercial reagents must be monitored as the preparation and use of each reagent must be consistent. Details of the storage and preparation of all reagents used in each staining run must be documented as part of the audit trail to allow back-tracking and troubleshooting.
Reagent monitoring is one area in which the use of an automated staining system with bar-code reagent labeling can be of assistance both in alerting the operator to reagents which have reached their expiry date and in the automated creation of audit trails and quality control documentation.
The buffer and diluent used for wash steps and antibody dilution will affect the results of immunostaining. The pH of these reagents needs to be monitored and must be checked and documented prior to use. If it falls outside of the range proscribed by the established protocol, corrections must be made or the reagent discarded.
Many antibody diluents contain additives such as sodium azide to stabilize and maintain the protein. Although this extends the shelf life of the antibody, the additives may interfere with, or inhibit, staining if present at excessive levels.
Concentrated antibodies tend to have a longer shelf life than pre-diluted ‘ready to use’ antibody preparations. Concentrated antibodies can be mixed with glycerine to prevent ice crystal formation, aliquoted into cryovials, then ‘snap’ frozen and stored in a −80°C freezer. This greatly extends their shelf life; compliance with the expiry date printed on the antibody packaging is important. The storage temperature for antibodies and reagents should also be monitored closely, as any fluctuation in temperature may cause increased deterioration of reagents. Frost-free −20°C freezers should hence be avoided for the storage of antibodies due to damage caused by the freeze-thaw cycles that these types of freezer perform.
The automation of immunostaining is perhaps the easiest way of improving the reproducibility and consistency of the staining. Although the use of automation is increasing, many laboratories still perform at least some immunohistochemical staining manually. Other laboratories perform semi-automated staining (for example manual epitope retrieval). The production of good manual procedures is important for these stains and also for use in case of failure of automated staining machinery. These procedures must be clear, easy to follow and sufficiently detailed to ensure a minimum of inter-operative variation. Adherence to these protocols and the reduction of human error is the goal in order to ensure consistent stain quality.
Processed blocks should be kept in a cool dry place. Resealing paraffin blocks following cutting can help protect the tissue from everyday elements such as air drying, excessive moisture or physical damage. The use of fresh cut controls for each run would be ideal but this is not always feasible in a busy laboratory. Repeated sectioning of a control tissue block can also result in loss of usable tissue, although some laboratories now have a control on the same slide as the test section.
Pre-cutting control slides is more time efficient and serial sections result in minimal tissue loss. The correct storage of pre-cut slides is important and frequently overlooked as a potential source of error in staining. Some studies have found deterioration of antigens in stored sections (Raymond & Leong 1990; Bromley et al. 1994; Prioleau & Schnitt 1995). Others have found no deterioration of some markers investigated, including estrogen receptor (ER), CD3, CD20, CD45RO, vimentin, and Ig light chains, in sections stored for up to four months at room temperature (Williams et al. 1997; Eisen & Goldstein 1999).
The viability of the antigen and speed of antigen deterioration in cut tissue sections is highly dependent upon the antigen under consideration and the temperature used for section adhesion. For example, whilst antigens such as CD30 and PSA seem relatively robust, CD117 (C-kit) deteriorates quickly and should be cut fresh.
The temperature at which the slides are dried can also affect the immunoreactivity of the antigens. It is advisable to dry all slides for immunohistochemical staining at 37°C. Urgent cases can generally be dried at 60°C for up to 4 hours. The exception to this is HER2, where it is recommended that sections should not be dried at 60°C for more than 1 hour.
Whenever a new batch of reagent is used, its details and the efficacy must be recorded, checked and compared with the previous batch. Reagents from different batches should never be mixed. Any enzymes used must be validated prior to use, due to a high degree of inter-lot variation. It should be noted that poor storage or shipping conditions can result in a reduction in enzyme activity. Antibody validation documentation should be available and is best kept in combination with the antibody specification sheet. This may be in either paper or electronic form. Detection system reagent validation is best performed by using a panel of several different antibodies. This should include different antigen retrieval methods and cover a variety of different staining patterns (nuclear, cytoplasmic, and membrane staining). A clean (i.e. non-reacting) negative control is just as important as the intensity of positive staining patterns, and such controls must be carefully checked.
It is important that all antibodies used in diagnostic testing are fully validated in the laboratory and detailed records are kept of the validation process. This should include demonstration of the reactivity of the antibody, validation of the procedure and quality checking of positive and negative controls used with that antibody. Before introducing a new antibody into the laboratory repertoire it is important to research details of the clone required. Most commercial antibodies have data sheets available on-line which should include a number of facts for consideration. These should indicate the host in which the antibody was raised (e.g. rabbit, mouse, or goat), location of the target antigen, concentration of the antibody, recommended application (e.g. frozen tissue, formalin-fixed paraffin-embedded tissue), recommended positive and negative control tissue sources, classification (e.g. analyte-specific reagent, research use only, or in vitro diagnostic) and reference materials for the application of the antibody. The specification sheet also typically includes suggested staining protocols.
Most antibodies used in diagnostic pathology laboratories are classified as ‘research use only’ or ‘analyte-specific reagent’. This means that it is the responsibility of the diagnostic laboratory to validate and document the sensitivity and specificity of the antibody. This process can be simplified if an appropriate antibody classified for ‘in vitro diagnostic’ (IVD) use can be sourced. The vendor has assumed responsibility for the validation and application of these antibodies.
The suggested protocol should serve as a baseline whilst working up the antibody and each laboratory should optimize the stain. If no protocol is suggested by the manufacturer, journal articles can be a good source of baseline protocols.
Selection of the most appropriate epitope retrieval method is important to ensure the maximum sensitivity of the stain. Antibodies are pH sensitive and it has been shown that staining intensity is better when the epitope retrieval step is performed at a pH specific to each antibody. It has also been shown that an antibody may not be specific or work as well if a proper pH is not maintained during the retrieval step. Selection of the detection complex used may also be influenced by the primary antibody. Some antibodies have been found to work better with an alkaline phosphatase detection system than with a horseradish peroxidase system and vice versa. Antibodies commonly used in pigmented tissues such as melanoma markers may benefit from a red (or other) chromogen end-point, as the endogenous pigment can disguise the more commonly used brown DAB chromogen.
Staining protocol development generally consists of trial and error, making sequential alterations in order to achieve optimal signal-to-noise ratio. In real terms this means strong, crisp target antigen staining with little or no background staining. If the staining is too strong, then further dilution of the primary antibody may improve staining specificity. If target antigen staining fades and the background remains, then the addition of a blocking step, change of epitope retrieval or changing the detection system used (e.g. to a polymer-based system) may improve the result. Changing the antibody diluent may help to increase the antibody’s reactivity, while at the same time lowering background staining. Weak staining may be improved with an increase of the antibody concentration (e.g. from 1 : 50 to 1 : 25).
Negative staining can be more difficult to resolve and may be the result of multiple factors. It is advisable to rule out human or mechanical error by repeating the staining before making further modifications. If the stain remains negative, change one variable at a time and document each reagent, step and reaction time. Negative staining may be resolved by increasing the antibody concentration, changing the epitope retrieval method, changing the solution pH, changing the antibody diluent to one of a different pH, changing the base composition or using an amplification step as part of the detection system. It is also important to ensure that the tissue stained should exhibit positive staining and the slides are freshly cut.
The duration of formalin fixation to which the tissue has been exposed may affect the reversal of protein cross-linking. Over-fixation of tissue may require more aggressive epitope retrieval methods to achieve satisfactory retrieval. This must be considered if the control or test tissue available for protocol development and validation has been stored in formalin for an extended period of time.
The final stain protocol must be confirmed on both positive and negative tissue controls prior to implementation. False-positive staining of negative control tissue suggests that the concentration of the antibody is too high.
Once an antibody dilution and staining method has been established and validated, each step of the procedure must be clearly documented and maintained in written or electronic form. This should include the antibody lot, expiration date, dilution, details of blocking steps performed (serum, avidin-biotin, and hydrogen peroxide), secondary and label (detection), chromogen tested, and the duration of each step.