This work documents a simple method to create synthetic antigen controls for immunohistochemistry. The technique is adaptable to a variety of antigens in a wide range of concentrations. The samples provide a reference with which to assess intra- and inter-assay performance and reproducibility.
Immunohistochemistry (IHC) assays provide valuable insights into protein expression patterns, the reliable interpretation of which requires well-characterized positive and negative control samples. Because appropriate tissue or cell line controls are not always available, a simple method to create synthetic IHC controls may be beneficial. Such a method is described here. It is adaptable to various antigen types, including proteins, peptides, or oligonucleotides, in a wide range of concentrations. This protocol explains the steps necessary to create synthetic antigen controls, using as an example a peptide from the human erythroblastic oncogene B2 (ERBB2/HER2) intracellular domain (ICD) recognized by a variety of diagnostically relevant antibodies. Serial dilutions of the HER2 ICD peptide in bovine serum albumin (BSA) solution are mixed with formaldehyde and heated for 10 min at 85 °C to solidify and cross-link the peptide/BSA mixture. The resulting gel can be processed, sectioned, and stained like a tissue, yielding a series of samples of known antigen concentrations spanning a wide range of staining intensities.
This simple protocol is consistent with routine histology lab procedures. The method requires only that the user have a sufficient quantity of the desired antigen. Recombinant proteins, protein domains, or linear peptides that encode relevant epitopes may be synthesized locally or commercially. Laboratories generating in-house antibodies can reserve aliquots of the immunizing antigen as the synthetic control target. The opportunity to create well-defined positive controls across a wide range of concentrations allows users to assess intra- and inter-laboratory assay performance, gain insight into the dynamic range and linearity of their assays, and optimize assay conditions for their particular experimental goals.
Immunohistochemistry (IHC) allows the sensitive and specific, spatially precise detection of target antigens in formalin-fixed, paraffin-embedded (FFPE) tissue sections. However, IHC staining results may be affected by multiple variables, including warm and cold ischemia time, tissue fixation, tissue pretreatment, antibody reactivity and concentration, assay detection chemistry, and reaction times1. Accordingly, reproducible performance and interpretation of IHC reactions require rigorous control of these variables and the use of well-characterized positive and negative control samples. Frequently used controls include paraffin-embedded tissues or cultured cell lines known from independent analyses to express the antigen of interest, but such samples are not always available1. Furthermore, the expression levels of the target antigens in tissues and cell line controls are generally understood only qualitatively and may be variable. Controls containing reproducible, precisely known concentrations of target antigen can assist in the optimization of IHC reaction conditions. A general method, adaptable to a variety of antigen types in a physiologically relevant range of concentrations to create synthetic IHC control samples, has been described by the authors2. A detailed protocol is provided here for the creation and use of this type of standard.
Appropriate controls are essential for the valid interpretation of IHC assays1,3,4. Tissues, cultured cells, and peptide-coated substrates have been employed as IHC controls according to the investigators' specific needs. The advantages and limitations inherent in using tissues as IHC controls have been extensively discussed1,4. For many antibodies, appropriate controls can be chosen from selected normal tissues containing cell populations expressing the target antigen over a wide dynamic range. Tissue controls are less suitable when the target antigen is not well-characterized concerning expression site or abundance, or when potentially cross-reacting antigens are co-expressed in the same cells or tissue sites. In these contexts, blocks of cultured cell lines expressing the antigen of interest can be helpful. For providing further evidence of target specificity, cell lines can be engineered to over-or under-express target antigens. For example, such an approach was recently used to evaluate a variety of anti-PD-L1 assays using a tissue microarray of isogenic cell lines expressing a range of PD-L1 antigen5. Practical limitations to the routine use of cell line blocks include the cost and time needed to produce sufficient cell numbers and the fact that the expression of some antigens may not be reliably consistent, even within clonal cell lines6. Synthetic peptides are a third option for IHC controls for antibodies that recognize short linear epitopes. Steven Bogen and colleagues have published extensively on the use of peptides coupled to the surface of glass slides7,8 and glass beads9. One study by this group demonstrated that quantitative analysis of peptide-based IHC controls could detect staining process variation missed by qualitative evaluation of tissue controls analyzed in parallel10. While standards using bead-based antigens could be widely applicable, many details are proprietary to the authors, limiting widespread adoption.
Another approach to IHC standards incorporates target antigens into artificially created protein gels. This concept was first demonstrated by Per Brandtzaeg in 1972 in a study in which normal rabbit serum was polymerized using glutaraldehyde11. Small blocks of the resulting gel were then soaked for 1-4 weeks in solutions containing the immunoglobulin antigens of interest at various concentrations. After alcohol fixation and paraffin embedding, sections of the resulting controls were shown to stain with intensities corresponding to the logarithm of the antigen solutions in which they had been soaked. Later, investigators prepared glutaraldehyde conjugates of specific amino acids in dilute BSA or brain homogenate solutions as positive controls in immune-electron microscopy studies12,13. More recent work investigated the use of gels made from formaldehyde-fixed protein solutions as surrogates for FFPE tissue in mass spectrometry analysis14. Another recent work investigated the structure and properties of gels formed by heating human or bovine serum albumin solutions at various concentrations and pH15. These authors found that serum albumin forms three types of gels differing in mechanical elasticity, secondary structure preservation, and fatty acid-binding capability depending on the experimental conditions. Together, these papers demonstrate the general feasibility of the approach employed here. Protein solutions of defined composition create tissue-like gels that can be further processed, sectioned, and stained using routine histological methods.
This protocol describes the formation of a synthetic IHC control made from bovine serum albumin (BSA) polymerized with heat and formaldehyde. The gels can incorporate a wide variety of antigens, including full-length proteins, protein domains, and linear peptides, as well as non-protein antigens including oligonucleotides2. This demonstration uses an example antigen a linear peptide encoding the C-terminal 16 amino acids of the human ERBB2 (HER2/neu) protein TPTAENPEYLGLDVPV-COOH (see Table of Materials). This sequence includes the epitopes recognized by three commercially available, diagnostically relevant antibodies including the Herceptest polyclonal reagent (ENPEYLGLDVP) and the monoclonal antibodies CB11 (AENPEYL) and 4B5 (TAENPEYLGL) (see Table of Materials)16.
The method demonstrated here employs readily available reagents using processes and techniques familiar to any practicing histology laboratory. The most significant limitation is the need to identify and purchase the target antigens, which can be accomplished in many cases at a relatively modest cost. Because these synthetic controls are of wholly defined composition and made with simple methods, they can be implemented in many laboratories with reproducible results. Their use may facilitate the objective, quantifiable evaluation of IHC staining results and allow greater intra- and inter-laboratory reproducibility.
1. Preparation of stock solution and tools
2. Preparation and dilution of peptides
3. Preparing BSA-peptide gels
4. Trimming, processing, and embedding BSA gels
5. Pilot evaluation of the peptide dilution series
6. BSA gel TMA construction
Peptides should dissolve entirely in an appropriate solvent at room temperature to form an optically clear solution. If visible particulate material is still present after 30-60 min, it may be helpful to add additional volumes of the original solvent or an alternative solvent not exceeding the intended volume of the 5x peptide stock solution calculated in Table 1. Likewise, the combined peptide/BSA solution should remain translucent (Figure 1A).
Peptide/BSA gel samples should form an opaque rubbery mass after heating with 37% formaldehyde (Figure 1B–I). Minor fracturing of the gel samples may occur during removal from the microcentrifuge tubes (Figure 3A). These should not interfere with subsequent steps. Paraffin-embedded gel samples should section smoothly without tearing or chatter. Gels that show irregular tear-outs (Figure 3B) may benefit from less aggressive block facing and/or brief soaking in ice water or ammonia water. Areas of variably reduced signal in a watermark pattern may occur when the distribution of one or more staining reagents is uneven (Figure 3C). The focally increased macroscopic signal may be seen if there is reagent trapping under the gel sections at any stage of the staining process (Figure 3C). The use of positively charged glass slides and careful drying of slides after sectioning may reduce this artifact. Microscopic areas of the reduced signal may occur if there is an uneven distribution of antigen in the gel matrix or if variation in the local gel structure limits antibody-antigen interaction (Figure 3D). Incomplete peptide dissolution may result in scattered areas of focally intense signal in a low-intensity background (Figure 3E).
After reacting with an appropriate antibody, negative control BSA-only gel samples should show minimal signal (Figure 4A), optimally less than 2%-3% of the dynamic assay range. Very rarely, there may be significant antibody interaction with BSA gel material lacking any antigen that cannot be eliminated by changing reaction conditions. The signal intensity in an individual gel sample should be relatively uniform, with increasing signal in samples with increasing antigen concentrations (Figure 4A,B). The absolute signal intensity will vary depending on the antigen-antibody combination and conditions used for the IHC stain. Depending on the staining conditions, there may be a threshold antigen concentration below which the signal will be at background and an antigen concentration above which the signal will be saturated. In some assays, the signal above background may be visible in samples with as little as 2.5 E-8 M peptide2. In replicate staining runs, the signal intensity for sections cut from each gel sample should be reproducible from run to run. In this application of anti-HER2 antibody, the widely used MDA-175 and SK-BR-3 cell line controls were stained in the same experiment to compare to the peptide controls. The cell lines show the expected distribution and intensity of signal: >10% of MDA-175 cells (HER2 1+) show faint (Figure 4C,D), incompletely circumferential membranous staining, and >10% of SK-BR-3 cells (HER2 3+) (Figure 4E,F) show intense completely circumferential membranous staining.
Possible reasons for the absence of signal include: (1) the peptide sequence is not a functional target for the antibody; (2) the staining protocol is not optimized to detect the antigen concentrations present in the gels; (3) the wrong peptide sample was stained; (4) reagent or instrument error. Possible reasons for signal in samples where it is not expected include non-specific reactivity of primary antibody or detection reagents with the BSA gel or peptide samples that were mislabeled during preparation.
Figure 1: Peptide-BSA gel preparation. (A) 25% (w/v) BSA (without added peptide) in PBS, pH 7.2. (B) Gel formed by mixing equal volumes of 25% BSA and 37% formaldehyde solution and heating to 85 °C for 10 min. (C–G) After cooling at room temperature for 5-10 min, the BSA gel can be removed from the microcentrifuge tube using a flexible disposable spatula. (H–I) The intact gel is sliced into discs ~5 mm thick in preparation for processing and embedding. Scale bars are 1 cm. Please click here to view a larger version of this figure.
Figure 2: Preparation of BSA gel samples for processing and embedding. (A–D) Preparation of a paraffin block containing a single disc of BSA gel. (A) The BSA-antigen gel is wrapped in biopsy tissue before processing. (B) After paraffin infiltration, the now-translucent gel disc is placed on the embedding center heated stage and unwrapped. (C) The single gel disc is placed in a 15 mm x 15 mm embedding mold containing liquid paraffin. (D) The completed paraffin block is ready for sectioning. (E–H) Preparation of a paraffin block containing multiple discs of BSA gel. As for Figures A-D above, the remaining gel samples are processed and embedded in one block to provide material for TMA construction. Scale bars are 1 cm. Please click here to view a larger version of this figure.
Figure 3: Gel preparation and staining artifacts. (A) Cores may show minor fracturing introduced during processing and/or embedding steps; focal darkening may reflect reagent trapping under the section. (B) Tearing out of gel material (seen in stained sections as irregular voids) may result from over- or under-soaking the block before sectioning. (C) Irregular watermark staining patterns may be caused by uneven reagent distribution at one or more staining steps. (D) Microscopic patterning of the BSA-peptide structure may be seen, particularly at higher peptide concentrations. (E) Incomplete peptide dissolution may produce a starry sky pattern with a focal signal in a low-intensity background. Scale bars are 200 µm (A–C) or 20 µm (D,E). Please click here to view a larger version of this figure.
Figure 4: Stained ERBB2/HER2 peptide TMA with ERBB2/HER2 1+ and 3+ cell lines. (A) Replicate TMA cores (top and bottom rows) containing no peptide (left column) or ERBB2 peptide at concentrations ranging from 2.5 E-8 M to 2.5 E-4 M were stained with anti-HER2/neu clone 4B5 according to the manufacturer's recommendations. Cores with increasing peptide concentrations show a graded increase in signal. (B) The TMA cores shown in panel A were digitally scanned, and the average pixel intensity [0.01 x (100 – % transmittance) x 255] is plotted vs. peptide concentration. (C–F) Cell line positive controls containing ERBB2/HER2-expressing cell lines MDA-175 (HER2 1+, C,D) and SK-BR-3 (HER2 3+, E,F) were included in the TMA containing the BSA-peptide gels so that the intensity of signal in the cell lines and BSA-peptide gels could be compared on the same slide. Scale bars are 500 µm (A), 250 µm (C,E), and 20 µm (D,F). Please click here to view a larger version of this figure.
A | B | C | D | E | F | G | H | |||
Peptide name | Peptide sequence | MW (Da) | Peptide purity (%) | Peptide mass provided (mg) | Moles of peptide in sample (includes correction for purity) | Volume (microliters) of solvent in which to resuspend peptide to get 1.25 e-2 M stock | Solvent | |||
ERBB2 / HER2 | Ac YGSGTPTAENP EYLGLDVPVGSGC amide |
2424.6 | 95.0 | 20.0 | 7.84E-06 | 626.9 | DMF |
Table 1: Calculations for the preparation of peptide stock.
A | B | C | D | E | F | G | H | I |
Tube | 5x Peptide stock | 1x Peptide stock | Dilution 1 | Dilution 2 | Dilution 3 | Dilution 4 | Dilution 5 | Negative control |
[Peptide] stock (M) | 1.25*10-2 | 2.5*10-3 | 5*10-4 | 5*10-5 | 5*10-6 | 5*10-7 | 5*10-8 | None |
Peptide solution (in solvent) |
>30 uL | 30 uL from 5x (1.25E-2 M) peptide stock |
140 uL from 1x (2.5E-3 M) peptide stock |
70 uL from Dilution 1 |
70 uL from Dilution 2 |
70 uL from Dilution 3 |
70 uL from Dilution 4 |
None |
Solvent volume | 120 uL | |||||||
Final peptide stock volume |
150 uL | |||||||
31.3% BSA / PBS volume |
560 uL 31.3% BSA |
|||||||
25% BSA / PBS volume |
630 uL 25% BSA |
630 uL 25% BSA |
630 uL 25% BSA |
630 uL 25% BSA |
700 uL 25% BSA |
|||
Volume of diluted peptide in BSA | 700 uL | 700 uL | 700 uL | 700 uL | 700 uL | 700 uL |
Table 2: Calculations for the preparation of BSA-peptide dilutions.
A | B | C | D | E | F | G |
Tube | Dilution 1 | Dilution 2 | Dilution 3 | Dilution 4 | Dilution 5 | Negative control |
Volume remaining of peptide in diluted BSA | 630 uL | 630 uL | 630 uL | 630 uL | 700 uL | 700 uL |
[Peptide] (M) in diluted BSA | 5*10-4 | 5*10-5 | 5*10-6 | 5*10-7 | 5*10-8 | None |
Volume of 37% Formaldehyde to add | 630 uL | 630 uL | 630 uL | 630 uL | 700 uL | 700 uL |
Final [Peptide] (M) in gel |
2.5*10-4 | 2.5*10-5 | 2.5*10-6 | 2.5*10-7 | 2.5*10-8 | None |
Table 3: Calculations for the preparation of BSA-formaldehyde gels.
This method allows the user to create uniform samples of known composition and antigen concentration as standards in IHC reactions, using materials and techniques familiar to most histology laboratories. The most crucial step is to identify the epitope to which the antibody of interest binds. This protocol describes using a linear peptide antigen from the ERBB2/HER2 ICD. The same protocol can be used to form BSA gels containing oligonucleotides, fluorescent labels, protein domains, or full-length proteins. This latter approach can be helpful for antibodies binding to conformational epitopes that are not recreated by a single linear peptide sequence. For example, BSA gel standards containing 0.1 mg/mL naive IgG from mice, rats, and rabbits may be used as process controls to confirm that secondary detection and staining steps have worked as expected.
The antigen-BSA gel method described here complements other techniques to control and standardize IHC reactions. Cell lines and tissue samples with well-characterized expression levels of target antigens have been essential to well-controlled IHC protocols1,3,4. Peptides coupled to the surface of glass slides and glass beads have been proposed as quantitative controls7,8,9. Each of the potential methods has overlapping advantages and limitations. The BSA-antigen gels have several advantages. They are simple to make and adaptable to various antigen compositions and concentrations. They reproduce some of the three-dimensional architecture of tissue samples while controlling for the inherent heterogeneity found in biological samples. Because the synthetic antigen gels can be made with user-selectable antigen concentrations extending from below the limit of detection to the highest concentrations found in biological samples (~10-4 M)2, they offer opportunities to calibrate and standardize assays performed with modern techniques, e.g., AQUA18 and imaging mass spectrometry19, whose dynamic range far exceeds traditional chromogenic IHC. In addition, the method allows controls incorporating more than one antigen, which can be used in multiplex assay development and standardization.
While the method requires the user to know the epitopes for the antibodies of interest, many antibodies bind to well-documented linear epitopes. Epitope mapping techniques20,21,22 can often identify undefined linear epitopes, and synthetic peptides including known linear epitopes can be purchased at a modest cost. Other antibodies bind to conformational epitopes that are not reproduced by linear peptides but are preserved in whole proteins or protein domains. For some of these antibodies, recombinant forms of the target antigens are commercially available.
For creating control samples with optimal uniformity and reproducibility, the peptide or other target antigen must be completely dissolved and carefully diluted into BSA solutions. It is also essential that efficient, rapid mixing of the BSA/peptide solution with 37% formaldehyde occurs before the sample starts to gel, and that the mixed sample is then placed promptly at 85 °C for 10 min. For avoiding over-fixation and processing artifacts, gels should be taken through processing and paraffin embedding according to the recommended schedule.
Variations on the method described here may be helpful in specific contexts. For instance, alternative N-and C-terminal peptide sequences may be used to optimize the detection of peptides bound to BSA2. It should be anticipated that antibodies recognizing epitopes from the extreme C-termini of proteins may be variably sensitive to C-terminal modifications of the native sequence. The peptide-BSA gel samples may also be formed by heating at 85 °C for 10 min without fixative, and then prepared as frozen blocks or post-fixed with a variety of chemistries2. In addition, peptide conjugation can be accomplished with maleimide chemistry or other methods than described here.
Biological samples such as cell lines and tissues have the advantage as controls of representing the complexity of the many variables found only in life. On the other hand, because tissue and cell standards are both internally heterogeneous and variable from sample to sample, the interpretation of variable staining from run to run is confounded. In addition, the antigen concentrations in tissues and cells are often known only qualitatively, though, with the increasing use of quantitative mass spectrometry, absolute protein concentrations are reported more frequently23. The BSA-antigen controls described here intentionally eliminate the spatial and biological context of proteins in cells and tissues. For this reason, the correlation between IHC signal intensity and antigen concentration found in these controls may imperfectly reproduce that seen in tissues. The synthetic controls described here may be helpful to optimize and standardize some aspects of IHC assay performance. Still, they do not prevent the need for the appropriate use of other control samples. For instance, potential non-specific staining in a target tissue can be evaluated only by the parallel use of a negative tissue control sample. In addition, BSA-antigen gel controls do not reproduce tissue-specific pre-analytic variables, including warm and cold ischemia time, proteolysis, or fixation and processing conditions that may affect IHC performance. Accordingly, and as has been discussed elsewhere in more detail1,3,4, investigators should make thoughtful use of cell and tissue standards to precisely characterize the performance of the IHC system.
BSA-antigen controls may be used to develop or optimize staining protocols and serve as intra- and inter-laboratory reference samples when assessing the reproducibility of established protocols. Access to well-defined standard samples should allow users to more rigorously characterize IHC reaction behavior, identify variation in staining performance, and optimize reaction conditions to achieve specific objectives.
The authors have nothing to disclose.
The authors gratefully acknowledge their colleagues Jeffrey Tom and Aimin Song for peptide synthesis, Nianfeng Ge for TMA construction, Shari Lau for IHC staining, Melissa Edick for digital microscopic scanning, and Hai Ngu for digital image quantification.
Anti-HER2/neu clone 4B5 | Ventana | 5278368001 | |
Biopsy Wraps | Leica | 3801090 | |
Bovine Serum Albumin, ultra pure | Cell Signaling Technology | BSA #9998 | |
50 mL Conical Tube | Corning | 352070 | |
Disposable base mold (15 mm x 15 mm) | Fisher | 22-363-553 | |
Disposable base mold (24 mm x 24 mm) |
Fisher | 22-363-554 | |
Disposable spatula | VWR | 80081-188 | |
Eppendorf Thermomixer | Eppendorf | 22331 | |
37% Formaldehyde | Electron Microscopy Sciences | 15686 | |
ERBB2 / HER2 peptide | UniProt P04626-1; a.a. 1240-55 | ||
Leica Autostainer XL | Leica | ST5010 | |
Magnetic Stir Bar | |||
NanoZoomer 2.0 HT whole slide imager | Hamamatsu | ||
10% Neutral Buffered Formalin | VWR | 16004-128 | |
Nuclease-free microfuge tubes 1.5 mL | |||
Paraplast paraffin | Leica | 39601006 | |
Peptide parameter calculator | Pep-Calc17 | https://www.pep-calc.com/ | |
Peptide suppliers | ABclonal Science | Users should contact peptide vendors for details of mass, purity and cost. | |
Anaspec Peptide | Users should contact peptide vendors for details of mass, purity and cost. | ||
CPC Scientific | Users should contact peptide vendors for details of mass, purity and cost. | ||
New England Peptide | Users should contact peptide vendors for details of mass, purity and cost. | ||
Phosphate Buffered Saline pH 7.2 | |||
Reagent Alcohol | Thermo Scientific | 9111 | |
Single Edge Razor | VWR | 55411-050 | |
Superfrost Plus positively charged microscope slides | Thermo Scientific | 6776214 | |
TMA Tissue Grand Master | 3DHISTECH | ||
Xylenes | VWR | 89370-088 |