Multiparameter fluorescence immunohistochemistry can be used to assess the number, relative distribution, and localization of immune cell populations in the tumor microenvironment. This manuscript describes the use of this technique to analyze T-cell subpopulations in oropharyngeal cancer.
The four-color fluorescence immunohistochemistry (IHC) technique is a method to quantify cell populations of interest while taking into account their relative distribution and their localization in the tissue. This technique has been extensively applied to study the immune infiltrate in various tumor types. The tumor microenvironment is infiltrated by immune cells that are attracted to the tumor site. Different immune cell populations have been found to play different roles in the tumor microenvironment and to have a different impact on the outcome of disease. This manuscript describes the use of multiparameter fluorescence IHC on oropharyngeal squamous cell carcinoma (OPSCC) as an example. This technique can be extended to other tissue samples and cell types of interest. In the presented study, we analyzed the intraepithelial and stromal compartment of a large OPSCC cohort (n = 162). We focused on total T lymphocytes (CD3+), immunosuppressive regulatory T cells (Tregs; i.e., FoxP3+), and T helper 17 (Th17) cells (i.e., IL-17+CD3+) using a nuclear counterstain to distinguish tumor epithelium from stroma. A high number of T cells was found to be correlated with improved disease-free survival in patients with a low number of intratumoral IL-17+ non-T cells. This suggests that IL-17+ non-T cells may be correlated with a poor immune response in OPSCC, which is in agreement with the correlation described between IL-17 and poor survival in cancer patients. Currently, novel multiparameter fluorescence IHC techniques are being developed using up to 7 different fluorochromes and will enable the more precise characterization and localization of immune cells in the tumor microenvironment.
Oropharyngeal squamous cell carcinomas (OPSCCs) are a heterogeneous group of squamous cell cancers originating in the oropharynx. Risk factors for OPSCC include human papillomavirus (HPV) infection and alcohol and tobacco use1,2. The role of the immune response and how to use this in a clinical setting is just starting to be explored. The tumor microenvironment is infiltrated by immune cells that are attracted to the cancer site. Although a high CD8+ cytotoxic T-cell frequency has been correlated with improved survival in OPSCC patients3, the role of other T-cell subsets, including Tregs and Th17 cells, is still unclear4. Whereas Th1 and Th17 cells are supposed to aid in the immune response targeting tumor cells, Tregs are well known for their abilities to suppress the activity of other T cells5. However, the presence of Tregs has been found to correlate with both favorable and unfavorable responses in different tumor types6. Since not all immune cells present in the blood infiltrate the tumor to the same extent, studying the local tumor microenvironment provides the most reliable measure of the immune response directed against the tumor. The aim of this study is to determine the correlation between the numbers and types of immune cells and the clinical outcome. We used four-color fluorescence IHC imaging to analyze the number and localization of various T-cell subpopulations in human OPSCC.
We focused on total T lymphocytes (CD3+), Th17 cells, and immunosuppressive FoxP3+ Tregs, whose differentiation pathway is closely related to Th17 cells. Th17 cells are characterized by the combination of CD3 and IL-17. The cytokine IL-17 can also be produced by non-T cells7. We determined the distribution of intra-epithelial and stromal T cells, Tregs, Th17, and IL-17+ non-T cells in a large series of OPSCC cases and analyzed the correlations with patient survival. Multicolor fluorescence IHC was used to identify the expression of CD3, Foxp3, and IL-17, in combination with a DAPI counterstain. This assay allowed for the easy and clear identification of both tumor cells (using DAPI nuclear staining) and the infiltrating T-cell populations (using a combination of different markers). Following sample preparation and staining, a fluorescent microscope and imaging software were used to separate the different fluorescent colors used and to determine the number and type of cells present in both the tumor epithelium and the tumor-associated stroma.
An alternative assay to quantify and phenotype immune cell populations is flow cytometry analysis, or cytometry by time of flight (CyTOF) analysis, of tumor or peripheral samples (i.e., blood or ascites). Using this technique, all information about localization and relative distribution of the different cell types is lost. The use and analysis of peripheral samples also does not provide information about which cells are able to infiltrate the tumor microenvironment. Blood and ascite immune-cell analyses have been shown not to reflect the phenotype and frequency of immune cell infiltration of the tumor tissue8,9.
Another alternative is the use of bright-field microscopy. An advantage of this technique over fluorescence imaging is the absence of tissue autofluorescence. Although some samples contain more autofluorescence—particularly erythrocytes, but also other cell types, including neutrophilic granulocytes—these areas can easily be removed in the analysis of almost all samples. Immunofluorescence offers the advantage of analyzing multiple markers in one sample by using a panel of targeted fluorescent wavelengths. This is currently impossible to the same extent for bright-field microscopy due to the lack of a sufficient number of labels and commercially available antibody isotypes for a particular antigen.
The multicolor fluorescence IHC technique described here has been used in many different cancer types and antibody combinations to study different immune cell populations, as well as tumor cell-expressed molecules, such as human leukocyte antigens (HLA) and PD-L110,11,12. The protocol has been established and validated using many different types of samples and antibodies.
Patient samples were handled according to the medical ethical guidelines described in the Code of Conduct for Proper Secondary Use of Human Tissue of the Dutch Federation of Biomedical Scientific Societies (www.federa.org).
1. Prepare Slides
2. Perform Antigen Retrieval
3. Stain Tissue Slides
4. Mount and Counterstain Tissue Slides
5. Analyze Stained Tissue Slides
6. Statistical Analysis
NOTE: All statistical analyses should be discussed with a statistician to assure quality of data. For general statistics, use any medical statistics guide14.
A series of FFPE pretreatment tumor samples obtained from primary oropharyngeal tumors diagnosed in the Leiden University Medical Center, Leiden, the Netherlands between 1970 and 2011, selected as described before (n = 162), was stained using the protocol described13. One to four random images of each slide were analyzed (Figure 1). Some autofluorescent cells are indicated, which can clearly be distinguished by their complete yellow appearance. Cells truly double- or triple-positive only stain positive for a particular antibody at the expected localization, which is at the cell membrane or the nucleus, in this case. Negative control slides did not show any specific staining above background tissue autofluorescence (Figure 2 and Figure 4), confirming that all signal is specific to the targets recognized by the primary antibodies. All tumor samples were found to be infiltrated by CD3+ T cells to varying extents. FoxP3+ Tregs always expressed CD3 and were one of the major infiltrating T-cell populations. CD3 cells expressing IL-17 (Th17 cells) were a minor population of the infiltrating T cells. IL-17 expressed by CD3– cells was another abundant infiltrating cell population. IL-17+ FoxP3+ cells comprised no more than 0.01% of all FoxP3+ cells.
All statistical tests were two-sided, and p-values below 0.05 were considered significant14. The correlations between intra-epithelial or total cell numbers and survival are presented, since the correlations between stromal or total cell numbers and patient survival were similar. A high number of infiltrating total CD3+ T cells showed a trend toward a correlation with improved disease-free survival (0.086, Figure 5A) as compared to a low number of T cells (i.e., lowest quartile). When specifically studying patients with a low number of IL-17+ cells, a high number of total infiltrating T cells was correlated with improved disease-free survival (p = 0.012, Figure 5B). The prognostic effect of tumor-infiltrating T cells was lost in the group of patients with a high number of tumor-infiltrating IL-17+ cells (data not shown). Thus, the effect of tumor-infiltrating T cells in OPSCC may be related to the low number of IL-17+ cells present.
Figure 1: FFPE Oropharyngeal Cancer Tissue Stained by Four-color Fluorescence IHC. Representative image of an immunofluorescent stain for CD3 (A, red), IL-17 (B, green), and FoxP3 (C, blue), as well as the merged image combined with DAPI (gray) (D). Two autofluorescent cells are indicated by the arrows. Please click here to view a larger version of this figure.
Figure 2: Negative Control for CD3. Representative image of FFPE oropharyngeal cancer tissue, stained as described in the protocol, substituting the anti-CD3 antibody with a rabbit Ig isotype control antibody with unknown specificity. No staining for rabbit Ig (A, red) combined with staining for IL-17 (B, green), FoxP3 (C, blue), and DAPI (D, gray) is shown. Please click here to view a larger version of this figure.
Figure 3: Negative Control for IL-17. Representative image of FFPE oropharyngeal cancer tissue, stained as described in the protocol, substituting the anti-IL-17 antibody with a goat Ig isotype control antibody with unknown specificity. No staining for goat Ig (B, green) combined with staining for CD3 (A, red), FoxP3 (C, blue), and DAPI (D, gray) is shown. Please click here to view a larger version of this figure.
Figure 4: Negative Control for FoxP3. Representative image of FFPE oropharyngeal cancer tissue, stained as described in the protocol, substituting the anti-FoxP3 antibody with a mouse IgG1 isotype control antibody with unknown specificity. No staining for mouse IgG1 (C, blue) combined with staining for CD3 (A, red), IL-17 (B, green), and DAPI (D, gray) is shown. Please click here to view a larger version of this figure.
Figure 5: Kaplan-Meier Survival Curves. For patients with a below-median number of IL-17+ cells/mm2, disease-free survival curves are shown for a very low (i.e., lowest quartile) versus high number of total T cells (A) and a low (i.e., below median) versus high number of total T cells (B) in the total tumor area. Reproduced with permission from the Cancer Immunology Immunotherapy journal13. Please click here to view a larger version of this figure.
For the protocol described, one of the most critical steps is to determine the correct dilution of the primary antibodies used. The dilution of the labeled secondary antibodies was 1:200, as recommended by the manufacturer of these specific Alexa-labeled antibodies. The dilution of the primary antibodies should then be determined by a serial dilution, preferably using two different samples of the intended tissue type (in this case, OPSCC). The optimal working dilution is the dilution at which a clear signal is obtained, without background noise and in the absence of a signal for the negative control at the same concentration. Modifications can be made with regard to the type of secondary antibodies and mounting medium used, depending on the type of microscope available and the preference of the researcher. If the intended results are not achieved, we recommend looking into the type of glass slides used, as some are known to cause problems with background signals for particular kinds of staining solutions (e.g., Perma Blue). Second, some antigens may be disrupted by the use of heat-induced epitope retrieval. In that case, the researcher can replace the heating step of the buffer with other techniques, such as enzyme-induced epitope retrieval. Third, if using the hydrophobic pen, the marking should not be made too close to the tissue, as this will prevent the staining solution from working properly. Regarding the number of sections per sample and the required number of images per section, it is up to the researcher and the specific research question to determine the number required to answer the research question with sufficient statistical power. For our analysis, we used one section per sample and four images per section. Finally, when using immunofluorescent staining, care should be taken when using nail polish to seal the cover slips, as alcohol present in the nail polish solution can affect the fluorescent signal.
The protocol described here does not involve high-throughput analysis. We have tried to automate this step, but the subjective interpretation of cellular size and morphology complicates the automation of the analysis with the software packages available at the time. Although tissue preparation and staining can be automated to a great extent, the counting of the cells was performed manually.
We were able to study the correlations between the infiltrating T-cell subsets studied and clinical outcome using the described protocol. A high number of T cells was found to be correlated with improved disease-free survival in patients with a low number of intratumoral IL-17+ non-T cells. This suggests that IL-17+ non-T cells may be related to a poor immune response in OPSCC, which is in agreement with the correlation described between IL-17 and poor survival in cancer patients16.
Improved software programs have now become commercially available to automate this step and are currently replacing the manual cell observation and counting, which will lead to more reliable, objective, reproducible, and fast results.
In addition, the availability of commercial kits allowing for the multiplexing of fluorescent immunohistochemical markers is currently on the rise17. This will allow for the multiplexing of up to 7 different immune markers/biomarkers in combination with a nuclear counterstain. However, we believe that the protocol reported here will still be more efficiently and easily applied in laboratories lacking complex and expensive infrastructure, microscopes, and software packages for multispectral staining analysis.
The authors have nothing to disclose.
Simone Punt was supported by grant UL2010-4801 from the Dutch Cancer Society. We would like to thank all the co-authors of the original paper that this JoVE protocol is based on: Emilie A. Dronkers, Marij J. P. Welters, Renske Goedemans, Senada Koljenović, Elisabeth Bloemena, Peter J. F. Snijders, Arko Gorter, and Sjoerd van der Burg.
Pathos Delta Ultra Rapid Tissue Processor | Milestone and Histostar | Automated tissue processor | |
Histostar | Thermo Scientific | Tissue block embedding machine | |
Formaldehyde | Baker | ||
Xylol | Merck | ||
Ethanol | Merck | ||
milliQ water | Elaga Purelab Chorus | ||
Paraffin wax/Paraclean | Klinipath | 5079A | |
Microtome tissue holder | Leica | RM225 | |
Flex IHC side | Dako | Tissue slide | |
Tris | Merk-Milipore | 1,083,821,000 | |
EDTA | Baker | 1073 | |
PBS | Bio-Rad | BUF036A | |
BSA | Sigma | A9647 | |
Rabbit anti-CD3 | Abcam | ab828 | Titrate required antibody dilution |
Mouse IgG1 anti-FoxP3 | Abcam | ab20034 | Titrate required antibody dilution |
Goat IgG anti-IL-17 | R&D Systems | AF-317-NA | Titrate required antibody dilution |
Rabbit Ig isotype control antibody | Abcam | ab27472 | Use at same final concentration as anti-CD3 |
Mouse IgG1 isotype control antibody | Abcam | ab91353 | Use at same final concentration as anti-FoxP3 |
Goat IgG isotype control antibody | ThermoFisher Scientific | 02-6202 | Use at same final concentration as anti-IL-17 |
Donkey anti-rabbit IgG A546 | ThermoFisher Scientific | A10040 | Dilute 1:200 in 1% BSA/PBS |
Donkey anti-mouse-A647 | ThermoFisher Scientific | A31571 | Dilute 1:200 in 1% BSA/PBS |
Donkey anti-goat IgG A488 | ThermoFisher Scientific | A11055 | Dilute 1:200 in 1% BSA/PBS |
VectaShield containing DAPI | Vector Laboratories | H-1200 | |
LSM700 confocal laser scanning microscope | Zeiss | ||
LCI Plan-Neofluar 25x/0.8 Imm Korr DIC M27 objective | Zeiss | 420852-9972-720 | |
LSM Zen Software | Zeiss | version 2009 | |
LSM Image Browser | Zeiss | version 4.2.0.121 | Available to download at www.zeiss.com/microscopy/int/website/downloads/lsm-image-browser.html. |
SPSS | IBM Corp. | version 20.0 | |
ImageJ | version 1.50i | Available to download at http://rsb.info.nih.gov/ij. |