As immunohistochemistry (IHC) becomes more vital to the accurate diagnosis and prognostic implications of pathologic diagnosis, the highest quality of laboratory practices is a necessity for patient care. Standardization of IHC represents a pressing need to close the gap of intralaboratory and interlaboratory variability, and efforts have been focused on both the release of guidelines for assay validation1,2 and through the implementation of IHC proficiency testing.3 Both of these improvements require sensitive and specific controls, whether for generation of a validation set for an assay per the recent guidelines1 or as a challenge to test laboratory protocol and quality. To improve the quality of IHC controls, guidelines for their standardization were recently released.4,5
These guidelines outline that there is regularly an inherent lack of internal controls in cases, where all of the tissue on the slide is malignant and no normal tissue with known epitope expression is present. In other instances, the overabundance of internal control masks the tissue of interest such that the immunostain provides little to no utility. Therefore, there is a constant need for proper external controls.6 Multiple preanalytical variables have been reported to significantly alter the staining performance of the tissue and thus any control: tissue size and thickness, ischemia and fixation time, and constant exposure of the slide to light all significantly alter the effectiveness of IHC stains with different effects based on the analyte.6–10 Further, the procurement of an adequate amount of a tissue as a control can be onerous, especially in cases where large quantities are needed such as on-slide controls and interlaboratory testing. Other important characteristics are that all tissues in the control block be processed identically and the expected staining pattern is defined for each tissue.4
From these findings, the ideal control is one which is produced through a normalized procedure with consistent processing, has a known and reproducible staining pattern, and is in large and easily obtainable quantity. Finding controls from excess diagnostic tissue is the current normal practice, but block to block variability due to differential grossing practices or prosectors, processing protocols between laboratories, and variable staining intensities between tumors is not acceptable for large-scale standardization. Peptide-based IHC where formalin-fixed peptide epitopes are fixed on a bead or glass slide and run as an on-slide control has been proposed as an inexpensive, easily automated, and standardized approach to this issue.11–13 However, this method does not recapitulate tumor morphology and would not be of use in day to day practice for morphologic evaluation or in proficiency testing. So called histoids have also been described in the past that have similar advantages as the peptide-based IHC in that they would be reproducible and with sufficient quantity, but with the additional benefit of preserving the basic morphology of the tumor cells and the surrounding stroma. However, these papers seem to be proof of concepts with little follow-up and did not see widespread use in many immunohistochemical clinical laboratories.14,15 More recently, a few companies have produced tissue microarray controls for general laboratory consumption, including horizon and STATLAB. Herein is presented a tissue microarray control termed a 3D tissue microarray control (3D TMAC) of breast and cervical tumors produced from cultured cell lines to assess for their use in a routine clinical immunohistochemical laboratory as compared with the normal, in house controls.
MATERIALS AND METHODS
3D TMACs were generated by coculturing cancer and stromal cells under defined and controlled cell culture conditions. The cocultured cells were fixed in 10% buffered formalin and embedded in paraffin blocks (referred to as the donor blocks) in accordance with College of American Pathology recommendations. Two donor blocks, representing a breast and cervical cancer cell line and the same stromal cells, were utilized for the construction of 2-core tissue microarray recipient blocks (3D TMAC).
Staining quality was assessed after receipt of 20 precut slides with sections of 4-µm thickness from each 3D TMA block were obtained and immunostained for 11 different antibodies common to breast and cervical tissue (Table 1). Antigen processing was done per laboratory protocol. Staining was performed on either a LEICA BOND III (Leica, Buffalo Grove, IL) or DAKO LINK 48 (DAKO, Carpintaria, CA) with the normal work flow on 3 separate runs over 3 days. A typical in house positive control was processed and run in tandem, in a side-by-side manner, with each run of the 3D TMAC (Table 1). Slides were visually interpreted for staining intensity, diffuseness, and pattern by 2 pathologists separately. Appropriateness of staining of the 3D TMAC was based on manufacturer’s claims of tissue staining characteristics. In stains without manufacturers claims, which were mammaglobin and p63, appropriateness was based on knowledge of the tissue subtype (breast vs. cervical) and the experience of the pathologists with that particular stain. The criteria for HER2 scoring used were according to the 2013 College of American Pathology/American Society of Clinical Oncology guidelines.16 All other stains were graded as: 0—no staining, 1+—light staining, 2+—moderate staining, 3+—intense staining. Discrepancies were resolved by discussion, with both parties agreeing on the final scores.
Reproducibility was assessed after receiving a paraffin-embedded block containing the aforementioned controls. Serial sections were cut and stained every other day for 4 occurrences (excluding weekends) with the normal work flow. Two slides were stained for each occurrence, CK 5 (Leica) and HER2 (DAKO), to assess both tissue controls. The 3D TMAC slides previously received for the staining quality assessment were also used in the reproducibility evaluation. The staining quality slides were denoted as “Run 1,” and the slides procured from the provided blocks were denoted as “Run 2” through “Run 5”. Slides were visually interpreted similarly for staining intensity, diffuseness, and pattern, with discrepancies resolved by consensus.
Slides from both the visual quality and reproducibility study were also scanned into a Leica AT2 scanner (Leica) and manually annotated by a pathologist, with >50% of the target, epithelial cells selected. Similar areas of annotation were made each day. Analysis was done for CK 5 using a nonspecific membrane algorithm provided by Leica (Leica), and HER2 by an in house validated algorithm that used the same membrane algorithm as a base. Output parameters that were considered included number of cells counted, average membrane intensity, and percentage of cells scored as 3+, 2+, 1+, and 0. Membrane intensity represents the average optical density of the membranous stain across all positive cells.
In house controls stain quality was in accordance with typical results. There was appropriate strong, diffuse staining in 7 of 11 tissues, with patchy or weaker than desired, but still positive, staining in GATA3, Ber-EP4, and p16. E-cadherin appropriately lacked staining in the tumor cells but was appropriately positive in background breast tissue. All tissues stained for their appropriate membranous or nuclear pattern (Table 2).
Of the 11 antibodies tested using the synthetic controls, 9 were reported to be positive by STATLAB (McKinney, TX). The staining characteristics of mammaglobin and p16 had no manufacturer’s claims by STATLAB. The breast 3D TMAC had 6 antibodies appropriately staining strongly and diffusely positive with typical pattern. CK 5 and p16 staining were appropriately negative, whereas GCDFP and mammaglobin were inappropriately negative. The E-cadherin was appropriately negative in the tumor cells; however, there was no background “normal” positive control to ensure staining had occurred (Table 3). In the cervical 3D TMAC, 4 of the antibodies were appropriately positive; with the remainder of antibodies appropriately negative (Table 3 and Fig. 1).
Visual reproducibility scores for CK 5 staining of the cervical 3D TMAC were identical across all stains (Table 4), without appreciable difference to either the in house control or the slide previously received for assessment of staining quality. The HER2 staining of the breast 3D TMAC was consistently scored as a 3+ across all runs (Table 4). Overall, these results were not significantly different from those of the slide previously received for assessment of staining quality nor the in house control. Algorithmic investigation of CK 5 staining of the cervical 3D TMAC showed an average membrane intensity of 61.39 with a SD of 2.20 (Table 5) and all interpreted as 3+ by the algorithm. HER2 staining of the breast 3D TMAC was as consistent as the CK 5 staining with an average membrane intensity of 85.54 and a SD of 2.72 (Table 6), with all algorithmic interpretations of 3+. The coefficients of variation for CK5 and HER2 were 3.58% and 3.18%, respectively.
Staining quality performance of the 3D TMAC was similar to the performance of the in house control by both visual and computer-assisted analysis. Both GATA3 and Ber-EP4 of the 3D TMAC outperformed the in house control, scoring strong diffuse 3+ while the in house control were partially or completely 2+. This likely is due to differential preanalytical protocols within the in house laboratory. P16 staining 3+ in the cervical 3D TMAC outperformed the in house control. However, GCDFP-15 which was reportedly positive failed to stain at all in our evaluation, possibly because this is usually a focal stain. Mammaglobin was not reported to be positive in the 3D TMAC and failed to stain in our run. Finally, while the breast tumor cells appropriately failed to stain with E-cadherin, there was no internal control normal breast tissue within the 3D TMAC to pick up E-cadherin stain to show that the stain worked. This is less desirable than the in house control that has peripheral normal breast tissue within the control. A separate normal breast 3D TMAC could solve both this issue and the lack of GCDFP-15 and mammaglobin, and could be developed if needed.
Reproducibility performance by visual analysis was consistently 3+ positive for both CK5 and HER2. This was further borne out by the computer-assisted analysis where each 3D TMAC was identified as 3+ by in house algorithms. As these algorithms may differ between institutions depending on their own algorithm validation studies, the use of the average membrane intensity, a parameter that theoretically should not vary based on laboratory, would be a better indicator. The CK5 and HER2 had average membrane intensities and standard deviation in line with the in house control run with them. 3D TMAC coefficients of variation for CK5 and HER2 were only 3.58% and 3.18%, respectively, below a widely accepted 5% threshold when testing an assay for reproducibility. Although the sample size for this conclusion is small and with only 2 different blocks or “lots” utilized, the initial results show a promising trend.
3D TMAC can meet many of the demands that both excess diagnostic tissue and peptide-based IHC cannot. Its standardized procurement process allows for reproducibility across batches that are not possible consistently from diagnostic tissue sources. 3D TMAC can also be produced as long as the cell line lasts, yielding an almost limitless supply of tissue useful for large-scale quality assurance. Further, the cores are sufficiently sized to make identification easy, but small enough that their utilization as an on-slide control is easily realizable. And while most of the above points can be said of peptide-based IHC as well, 3D TMAC has the benefit of displaying how each stain interacts with the tumor cell morphology. Differential staining intensity, such as PAX5 in hematolymphoid malignancies, or pattern, as in β-catenin evaluation of soft tissue tumors, is important for identification of certain tumors, thus having tissue on slide to demonstrate the proper staining morphology is an important factor for a pathologist interpreting a stain.
3D TMAC has both intralaboratory and interlaboratory applications. Customization of epitopes would allow individual laboratories to generate positive controls for the validation and titration of new assays, especially in cases of rarer tumor subtypes where sufficient tissue procurement can be an issue. Interlaboratory applications revolve around standardization among different sites. The increase of hospital mergers creating large hospital systems should bring with it a need to normalize laboratory testing amongst each site, something 3D TMAC can provide. Proficiency testing would also benefit from the standardization of fixation, abundance of supply, and preservation of morphology inherent in 3D TMAC. There are some potential limitations of the test, including the current lack of normal epithelial tissue to act as an internal, negative, or positive control where the absence of stain is a positive result. Cost is another possible prohibitive factor, especially if large quantities are needed for an on-slide control protocol. For assays that require quantification for therapeutic determination such as HER2, a spectrum of staining intensities is needed to assure the validity of results within different result ranges. 3D TMAC has the potential to solve this issue, but the present version can only accommodate the 3+ range common among current laboratories. The current control reservoir of choice, excess diagnostic tissue, has failings in many of the strong suits of tissue microarray controls: inconsistent fixation and processing lends itself to variance in immunohistochemical performance, differences in tissue selection for controls can lead to further variance, and once the block of the control runs out, a new block with different antigen characteristics must be selected, causing a lot to lot variation. However, due to the ubiquity of tissue available, barring the more uncommon antigens, and the relative lack of cost since the tissue is procured from the practices specimen, excess diagnostic tissue remains the top choice for many clinical laboratories, especially those in nonacademic settings. In summary, the 3D TMAC has the potential to markedly improve the quality of immunohistochemical analysis through standardization.
1. Fitzgibbons PL, Bradley LA, Fatheree LA, et al. Principles of analytic validation of immunohistochemical assays: Guidelines from the College of American Pathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med. 2014;138:1432–1443.
2. Fitzgibbons PL, Murphy DA, Hammond MEH, et al. Recommendations for validating estrogen and progesterone receptor immunohistochemistry
assays. Arch Path Lab Med. 2010;134:930–935.
3. Vyberg M, Nielsen S. Proficiency testing in immunohistochemistry
—Experiences from Nordic Immunohistochemical Quality Control (NordiQC). Virchows Arch. 2016;468:19–29.
4. Torlakovic EE, Nielsen S, Francis G, et al. Standardization
of positive controls
in diagnostic immunohistochemistry
: recommendations From the International Ad Hoc Expert Panel. Appl Immunohistochem Mol Morphol. 2015;23:1–18.
5. Torlakovic EE, Francis G, Garratt JRT, et al. Standardization
of negative controls
in diagnostic immunohistochemistry
: recommendations from the International Ad Hoc Expert Panel. Appl Immunohistochem Mol Morphol. 2014;22:241–252.
6. Torlakovic EE, Nielson S, Vyberg M, et al. Getting controls
under control: the time is now for immunohistochemistry
. J Clin Pathol. 2015;68:879–882.
7. Miller RT, Swanson PE, Wick MR. Fixation and epitope retrieval in diagnostic immunohistochemistry
: a concise review with practical considerations. Appl Immunohistochem Mol Morphol. 2000;8:228–235.
8. van der Brook LJ, van de Vijver MJ. Assessment of problems in diagnostic and research immunohistochemistry
associated with epitope instability in stored paraffin sections. Appl Immunohistochem Mol Morphol. 2000;8:316–321.
9. Eisen RN. Quality management in immunohistochemistry
. Diagnostic Histopathol (Oxf). 2008;14:299–307.
10. Taylor CR. Predictive biomarkers and companion diagnostics. The future of immunohistochemistry
: “In Situ Proteomics,” or Just a “Stain”? Appl Immunohistochem Mol Morphol. 2014;22:555–561.
11. Sompuram SR, Kodela V, Zhang K, et al. A novel quality control slide for quantitative immunohistochemistry
testing. J Histochem Cytochem. 2002;50:1425–1434.
12. Bogen SA, Vani K, McGraw B, et al. Experimental validation of peptide immunohistochemistry controls
. Appl Immunohistochem Mol Morphol. 2009;17:239–246.
13. Sompuram SR, Vani K, Tracey B, et al. Standardizing immunohistochemistry
: a new reference control for detecting staining problems. J Histochem Cytochem. 2015;63:681–690.
14. Ishikawa K, Miyamoto M, Yoshioka T, et al. Method for the validation of immunohistochemical staining using SCID mouse xenografts: expression of CD40 and CD154 in human non-small cell lung cancer. Oncol Rep. 2013;29:1315–1321.
15. Kaur P, Ward B, Saha B, et al. Human breast cancer histoid: an in vitro 3D co-culture model that mimics breast tumor tissue. J Histochem Cytochem. 2011;59:1087–1100.
16. Wolff AC, Hammond MEH, Hicks DG, et al. Recommendations for human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice Guideline Update. J Clin Oncol. 2013;31:3997–4013.
Keywords:Copyright 2018 Wolters Kluwer Health, Inc. All rights reserved.
immunohistochemistry; controls; standardization