Application of Immunohistochemistry in Clinical Practices as a Standardized Assay for Breast Cancer

Abstract

Immunohistochemistry (IHC) has become an indispensable tool in the clinical practices for breast cancer; however, to achieve its standardization, numerous issues need to be overcome. In this review, we describe the development of IHC as an important clinical tool, and the challenges in standardizing IHC results for patients. We also present ideas for resolving the remaining issues and unmet needs, along with future directions.

I. Introduction

Immunohistochemistry (IHC) was developed from immunofluorescence techniques introduced by Coons et al. in 1942 [8]. The peroxidase-labeled antibody method, proposed by Nakane [31], broadened the applications of histochemistry to formalin-fixed paraffin-embedded (FFPE) sections, in addition to frozen sections. Further technical advances resulted in the development of the immunoglobulin bridge [24], peroxidase anti-peroxidase (PAP) [49], avidin-biotin complex (ABC) [17], and labeled streptavidin-biotin (LSAB) methods. Development of the heat-induced antigen retrieval (HIER) technique has further facilitated reproducible results with IHC [42]. Although the precise mechanisms of HIER remain unclear, the following hypotheses have been proposed: 1) loosening or breaking of the cross-linkages caused by formalin fixation or renaturation by heating [43], 2) degeneration of protein [6], 3) hydrolysis of Schiff’s bases [46], 4) chelation of calcium ions embedded in calcium complexes produced during formalin fixation [30], and 5) electrostatic effects [44]. Thus, IHC has become an essential tool for research purposes as well as medical diagnosis.

In this review, we describe how IHC has been developed, as an indispensable tool, for clinical practices pertaining to breast cancer over three decades (from around the 1990’s to the 2020’s). Several clinical trials have been undertaken to establish suitable biomarkers (prognostic and predictive factors). Numerous concordant studies have validated the appropriateness of using IHC and in situ hybridization (ISH) for examining these biomarkers; furthermore, several studies have determined that these approaches can facilitate the improved treatment of patients. In addition, there has been considerable discussion and activity regarding the most adequate system for patients with breast cancer in Japan. In this review, we describe the validation of biomarkers for breast cancer management and the challenges involved in developing standardized assays beyond “staining”, as well as the unmet needs and future directions.

II. Clinical Application of Immunohistochemistry for Breast Cancer: ER, PgR, and HER2

The early phase of biomarker development before immunohistochemical examination

1) Hormone receptors (estrogen receptor, ER; progesterone receptor, PgR)

Hormone receptors (ER, PgR) as biomarkers have been developed practically, but not theoretically. The report by Beatson in 1896, indicating that an ovariectomy reduced the size of human breast cancer [4], was followed by the discovery of ER in 1961 [18]; ER and PgR were then established as biomarkers owing to their responsiveness to endocrine therapy [9, 20]. The ER status was initially examined using a ligand binding assay (dextran-coated charcoal method) and an enzyme immunoassay (EIA), followed by immunocytochemical assays (estrogen receptor immunocytochemical assay, ER-ICA; progesterone receptor immunocytochemical assay, PR-ICA) on frozen histopathological sections [10], and finally, using immunohistochemical techniques on routinely processed FFPE [1, 13]. The sensitivity of the immunohistochemical examination has increased beyond that of biochemical assays, a gold standard method until around 2000 [50, 51]. Furthermore, the advantages of immunohistochemical examinations, compared to biochemical methods, including the ability to examine archival materials and/or small tumors and cytologic samples, and observe positive reactions combined with histopathological features, have rapidly promoted IHC as a useful and established tool worldwide.

2) Human growth factor receptor 2 (HER2)

Unlike ER, the history of HER2 as a biomarker started with the discovery of neu, which was identified through DNA transfection studies on nitrosoethylurea-induced rat neuroblastomas [38, 45]. The locus of HER2/c-erbB-2 on chromosome 17q21 [37] and the function of HER2 overexpression have also been revealed. Clinical studies have shown that HER2 overexpression is associated with a poor prognosis [47], treatment resistance or susceptibility to endocrine therapy, cyclophosphamide, methotrexate, 5-fluorouracil (CMF) [11, 22], and doxorubicin [33].

In addition to the clinical use of HER2 as a prognostic factor, the development of a novel drug, trastuzumab, a humanized anti-HER2 antibody [5], has necessitated HER2 examination as a treatment decision factor [55]. HER2 amplification or overexpression has been historically examined using Southern blotting, ISH, followed by fluorescence or non-fluorescence labeled probe, and IHC.

Standardization of immunohistochemistry

1) Primary antibodies and procedures

Various factors in the IHC process may affect the results, that is, the condition of fixation, type of primary antibody, condition of antigen retrieval, and type of detection method. For example, the primary antibodies against ER, which are commercially available, are mouse monoclonal antibodies from clones 6F11 and 1D5, and rabbit monoclonal antibodies from clones SP1 and EP1; the sensitivity and specificity of these antibodies are not the same. To standardize these variables, in vitro assay kits approved by the Pharmaceuticals and Medical Devices Agency (PMDA) are commercially produced. It is further recommended to follow the provided protocol, which is the most adequately optimized, and to use automated machines for immunostaining [54].

2) Evaluation systems

Regarding evaluation, the cell counting system, by which positive cells are counted among a certain number of cancer cells, is simple and highly reproducible. Another line of scoring systems involves the determination of staining intensity and positive-cell populations, that is, German immunoreactive score (IRS) [21, 34, 35] and Allred score [1]. A task force from the Japanese Society of Breast Cancer proposed a J-score system for ER based on clinicopathological evidence [52, 54]. The perspectives considered were: 1) tissue handling, 2) concordance rate between EIA and IHC using different staining methods [2], 3) inter-observer diversity [54], 4) threshold for the predictive value of endocrine therapy in primary breast cancer [15, 16], and 5) threshold for the predictive value of endocrine therapy in recurrent/metastatic breast cancer [58].

Guidelines from academic associations (ASCO/CAP) [12, 57], non-profit organizations (UK-NEQAS) [21, 36], and commercial companies are also useful for standardizing staining procedures and evaluation.

3) Quality assessment systems

External quality assessment systems (EQA) have been developed in each country and regional network, as well as globally, including at the College of American Pathologists (CAP), UK National External Quality Assessment Service (UK-NEQAS), Nordic Immunohistochemical Quality Control (NordQC), and The Royal College of Pathologists of Australasia (RCPAQAP). In Japan, the Japan Pathology Quality Assurance System (JPQAS) (http://www.jpqas.jp/) was established in 2014 [53]. It was initiated as a collaborative study by the Quality control committee from the Japan Society of Pathology, Japan Breast Cancer Society, and research groups of the Ministry of Health, Labor and Welfare. Several trials were undertaken to investigate the quality validation of immunohistochemical staining, with evaluation among hundreds of institutions, and to determine how an organization should work, along with the expenses for establishing and maintaining an EQA system in Japan. Finally, the Japan Society of Pathology and Japanese Association of Medical Technologists established the JPQAS Currently, approximately 400 institutions participate in this EQA system; JPQAS has also become a member of the International Quality Network for Pathology (IQN for Pathology) (http://www.iqnpath.org/).

III. Understanding Tumor Heterogeneity Using Histochemical Techniques

Histochemical analysis of tumor heterogeneity

Willis [56] defined a neoplasm as “an abnormal mass of tissue, the growth of which exceeds and is uncoordinated with that of the normal tissues and persists in the same excessive manner after cessation of the stimuli which evoked the change.” Currently, it is known that the tumor consists of more heterogeneous cells than previously thought. Understanding the heterogeneity of solid tumors is significant based on both biological and clinical aspects. As heterogeneity of the primary tumor indicates diverse sensitivity and resistance levels to adjuvant therapy, cancer cells with an inherent biological nature to survive may recur or metastasize [41]. However, understanding phenotype-genotype correlations with retained topological information in the tumor is technically challenging [25]. Before the development of next generation sequencing (NGS), the following attempts were made to detect genetic alterations with correlated phenotypes: heterogeneous glucose‑6‑phosphate dehydrogenase on the X chromosome based on the Lyonization theory, androgen receptor polymorphism, loss of heterozygosity (LOH), comparative genomic hybridization, and somatic mutation of the mitochondrial DNA D-loop region [19, 26]. NGS-based analyses clearly revealed that the tumor comprises heterogeneous cancer cells as well as their lineages [7, 19].

Multiple staining is one of the solutions available for pathological sections. Simultaneous detection of multiple proteins in the FFPE sections is now possible using commercially available systems. For example, MultiOmyx^TM (NeoGenomics Laboratories) is a direct immunostaining method that, in a single run, uses two primary antibodies labeled with Cy3 and Cy5, captures images, converts to DAB color images, and decolorizes using alkaline solutions containing H₂O₂; more than 60 molecules can be detected in a single section in this system [14]. The Toponome imaging system (TIS) (ToposNomos GmbH) is an automated robotic system that, in a single run, uses fluorescence labeled monoclonal antibodies, captures images using a CCD camera, and decolorizes via low energy photobleaching; this system can detect up to 100 molecules in a single section [39, 40]. The Mantra 2^TM Quantitative Pathology Workstation (Akoya Biosciences) is a spectrum analyzing system for tissue sections stained using the Opal^TM multiplex immunostaining kit. The antigen-antibody reaction is labeled using a fluorescent dye involving a tyramide signal amplification system; the secondary antibody is then removed through microwave treatment, whereas the fluorescent dye remains at the reaction site [32, 48]; up to nine primary antibodies are available for observation in a single section. These techniques enable the observation of multiple proteins in single FFPE sections.

Proposal for a simultaneous gene protein assay

We introduced a gene protein assay (GPA) for simultaneously detecting proteins (HER2 and ER) and genes (HER2 and CEP17) in a single section (Fig. 1) [27]. This system enables the observation of heterogeneous cancer cells in breast cancer, that is, ER+/HER2−, ER+/HER2+, ER−/HER2+, and ER−/HER2− cells (Fig. 2), especially in ER+/HER2+ tumors. Simultaneous observation in a single section allows the detection of cancer cells showing discrepancy between HER2 amplification and protein overexpression independent of ER expression, and demonstrates that ER+/HER2+ tumors have the most heterogeneous cancer cells compared with other tumor types. Furthermore, we clarified that the cancer cells constituting ER+/HER2+ tumors are more predominantly HER2-positive cells with or without ER expression, rather than ER-positive cells with or without HER2 overexpression (Fig. 3). However, the mechanisms by which ER+/HER2+ tumors gain both ER expression and HER2 amplification remains unclear. Such a precise observation of individual cancer cells with heterogeneous phenotypes in a tumor mass can contribute to the advancement of relevant investigations.

Fig. 1.

Gene protein assay (GPA) for HER2 plus the estrogen receptor (ER). In a single FFPE section, HER2 on the cell membrane (brown), ER protein in the nuclei (pink), HER2 signal (black), and CEP17 signal (green) are detected.

Fig. 2.

Heterogeneous cancer cells constituting the tumor mass. Cancer cells with different phenotypes including ER+/HER2− (a), ER+/HER2+ (b), ER−/HER2+ (c), and ER−/HER2− (d) were detected.

Fig. 3.

Cancer cell population in a HER2+/ER+ tumor. Each column of the circle represents a single tumor sample. The phenotypes of cancer cells are shown in different colors, that is, HER2−/ER+ (pink), HER2+/ER+ (green), HER2+/ER− (blue), ER+ with HER2 microheterogeneity (yellow), and ER− with HER2 microheterogeneity (light pink). The length of the curved column shows the proportion of cancer cells within a particular subpopulation in the respective color among the total number of cancer cells counted in each tumor. The percentage of ER-positive cells (HER2−ER+ and HER2+/ER+) and HER2-positive cells (HER2+/ER+ and HER2+/ER−) are shown.

IV. Immunohistochemistry as a Standardized Assay: an Issue of Tissue Thickness

Tissue thickness

Considering the history of IHC, the inclusion of antigen retrieval in the protocol made IHC a stable and versatile method to observe protein histopathological sections and simultaneously, facilitate protein quantitation. As the use of IHC increased in routine practice, the strictness for the methods used for staining and evaluation tended to decrease. To develop IHC as a reliable assay for detecting the biological characteristics of obtained tissues accurately, efforts need to be made to standardize most of the variables in the IHC process. Despite being the most basic factor in the staining process, the thickness of the FFPE tissue sections has been given little attention. Only a few studies have addressed the issue of tissue thickness. Leong and Leong described the significance of tissue thickness as a factor that needs to be standardized [23]. Barker et al. reported that, to manage EQA, tissue thickness of control slides is a critical factor that should be standardized [3]. McCampbell addressed the correlation between tissue thickness and staining intensity using whole slide imaging of Ki67, BCL6, CD7, and CK [29].

Another reason for addressing tissue thickness is the development of artificial intelligence for automated image analysis. Under a microscope, histopathological tissue sections can be determined as thin or thick. In contrast to the tissue sections, which are three-dimensional, the images captured from these tissue sections are two-dimensional (Fig. 4). Therefore, estimation using captured images carries the risk of underestimation or overestimation of the IHC results. An example of this is shown in Figure 5. As the tissue thickness increased, the Ki67-positive cells estimated using the image analysis system, especially the strongly positive cells, also increased in a statistically significant manner. The issue of tissue thickness should be of particular concern, because the algorithms for artificial intelligence systems used for image analyses are based on images captured in 2D photos. If the depth of 3D tissue sections is considered adequately, the results provided by artificial intelligence systems may not be reliable.

Fig. 4.

Schema explaining the influence of tissue thickness on microscopic observation and captured images. Microscopic observation is three dimensional (3D), whereas the captured image is two dimensional (2D). The differences between 3D and 2D results may lead to different captured images.

Fig. 5.

Correlation between tissue thickness and the results of Ki67-labeling index. HE sections, immunohistochemical images for Ki67, and image analyses for different tissue thicknesses (2 μm, 4 μm, and 6 μm) are shown. In the image analyses, positive cells are labeled in red, orange, and yellow to indicate strong, moderate, and weak staining, respectively, whereas negative cells are labeled in blue.

Proposed management system to control tissue thickness

We paid special attention to managing the tissue thickness for the HER2 protein in breast cancer, because estimation of this biomarker directly affects the clinical decision regarding whether targeted therapeutics are applicable (Fig. 6). To investigate the correlation between tissue thickness and staining intensity, we prepared an experimental system, where variable factors for estimating IHC results, except for tissue thickness, that is, the light of a microscope, condition of the captured images, and methods for image analysis, were standardized as strictly as possible [28]. Tissue thickness and staining intensity were found to be well correlated (SK-BR3; R² = 0.9282, MDA-MB-453; R² = 0.9268) [28]. Thus, we proposed a potential management system to control tissue thickness in routine practice. The key concept of this idea was to use a standardized control for tissue thickness in addition to controls for adequate staining. As shown in Figure 6, only sections cut to an adequate thickness could be subjected to the staining and evaluation processes. As a standardized control for tissue thickness, we evaluated two systems: the use of a cell line block stained with β-actin, which is a well-known internal control in biochemical assays, and use of a non-cellular control. For the first approach, a polyclonal antibody was raised against a synthetic β-actin peptide (2–11) (Ac-DDDIAALVVDC) with a specific epitope site different from that of α-actin (cardiac, smooth, and skeletal muscles) and γ-actin (Immuno-Biological Laboratories, Co., Ltd., Gunma, Japan). However, the staining intensity tended to decrease with increasing tissue thickness [28]. This was attributed to the flattened cytoplasm of cultured cells in mounted sections, and the unparalleled staining intensity of cytoplasmic β-actin with increasing tissue thickness. In contrast, the non-cellular control, a blue colored urethane foam, showed a superior parallel correlation between tissue thickness and color intensity.

Fig. 6.

Current situation, facts, and future directions considering the variable factor of tissue thickness in immunohistochemical evaluation. The upper row shows the current situation, the middle row shows the known facts, and the lower row shows the expected future. The images show cell blocks comprising SK-BR3 cells stained with anti-HER2 polyclonal antibody. Tissue thickness measured using interferometry is shown.

V. Concluding Remarks

Overall, in this review, we described how IHC has become one of the indispensable tools used in clinical practices for breast cancer, and how clinicians and pathologists have struggled to standardize IHC results for patients. IHC has the potential to be a more automated, precise, and useful method, which is applicable in the clinical practice for many diseases. To develop IHC as a standardized assay more usefully in the future, we should be aware of and maintain scientific reliabilities on every factor contributing the process of IHC.

VI. Conflicts of Interest

The authors declare that there are no conflicts of interest.