Measurement of Four Uterine NK Cell Subtypes Using Multiplexed Fluorescent Immunohistochemical Staining in Women with Repeated Implantation Failure

Summary

This study presents a pioneering method for quantifying uterine natural killer cell subsets during the window of implantation using advanced multiplexed fluorescent immunohistochemical staining techniques.

Abstract

Immunohistochemistry (IHC) plays a crucial role in biological research and clinical diagnosis, serving as the most commonly used method for identifying and visualizing tissue antigens. However, traditional IHC staining methods have limitations in distinguishing various subtypes of immune cells. This challenge has driven scientists to explore new technologies and methodologies for precise identification and differentiation of immune cell subtypes. In recent years, multiplex IHC has emerged as a solution, enabling the simultaneous detection of multiple antigens and their visualization within the same tissue sample. Uterine natural killer (uNK) cells play a pivotal role in early pregnancy processes, including decidualization, remodeling of uterine spiral arteries, and embryo implantation. Different subtypes of uNK cells exhibit different functions, allowing them to coordinate various biological events for successful embryo development and pregnancy. Therefore, in-depth research on uNK cell subtypes is essential for elucidating immune regulation mechanisms during pregnancy. Such studies provide valuable insights and novel approaches for addressing related conditions such as infertility and recurrent reproductive failure. This paper introduces a detailed multiplex IHC staining protocol for studying the density of four subtypes of uNK cells in endometrial specimens during the window of implantation (WOI). The protocol includes sample preparation, optimization of subtype markers, microscopic imaging, and data analyses. This multiplex IHC staining protocol offers high specificity and sensitivity, enabling simultaneous detection of different uNK cell subtypes, thus providing researchers with a powerful tool to explore the intricacies and mechanisms of immune regulation during pregnancy.

Introduction

The first documented live birth after in vitro fertilization-embryo transfer (IVF-ET) was reported in 1978. Over the past 40 years, there has been a high demand for the assistance of IVF-ET among infertile couples¹. In 2021, 238,126 patients initiated a total of 413,776 IVF cycles in the United States. This marks a 25% increase in cycles from 2 years prior and a 135% increase from 2012². This surge is mainly attributed to the rising prevalence of infertility and delayed planning for pregnancy. Advancements in embryo culture techniques and superovulation protocols have led to an increased live birth rate per ET cycle, reaching 30%-50% in women less than 40 years old and less than 30% in women older than 40 years². However, despite these advancements, over half of transferred embryos still fail to implant. Repeated implantation failure (RIF), typically defined as failure after three or more consecutive attempts of transferring high-quality embryos, affects 15% of women who undergo IVF-ET³. Couples with RIF are extremely vulnerable and more prone to undergo expensive and unnecessary procedures that can expose them to undue risks⁴. Therefore, understanding the causes of RIF and improving embryo implantation is crucial to enhance the success of IVF-ET, particularly for women with RIF. The preparation of endometrium is critical for successful embryo implantation. This process is characterized by a significant accumulation of uterine natural killer (uNK) cells, which transition from constituting 30% of total lymphocytes in the endometrium during mid-secretory phase to 70%-80% in the decidua during early pregnancy⁵. Notably, uNK cells differ from peripheral NK cells, which are cytotoxic lymphocytes critical to the innate immune system for causing the death of the infected cells through lysis or apoptosis. While the exact functions of uNK cells are not yet fully understood, several lines of evidence suggest that they are involved in angiogenesis remodeling, trophoblast invasion, and fetal development⁶. The association between the percentage of uNK cells over stromal cells and RIF has garnered extensive attraction over the past 20 years. A recent meta-analysis, which included 8 studies involving 604 women, demonstrated that the density of CD56+uNK cells during the mid-luteal phase is significantly increased in women with RIF compared to fertile controls⁷. However, it is important to note that the characteristics of uNK cells during the mid-luteal phase differ significantly from those of decidual NK (dNK) cells. Although uNK cells may further differentiate into various subsets of dNK cells post-pregnancy, measuring uNK cells alone does not accurately represent dNK cells⁸. uNK cells undergo dynamic differentiation and play different roles in the menstrual cycle and decidualization processes, making them more complex than can be identified by CD56 alone. Multiple markers are required to achieve a comprehensive understanding of uNK cell behavior during endometrial preparation. Our recent study employed single-cell RNA sequencing to identify the diversity of uNK cells throughout menstrual cycles. The results, validated using flow cytometry, have shown the presence of four distinct subtypes of uNK cells, each exhibiting dynamic changes during the menstrual cycle⁹. Gene enrichment analysis and gene ontology functional enrichment indicate these uNK subsets fulfill different functions at various stages of menstruation. Nevertheless, flow cytometry is not universally accessible in clinical laboratories, and the immediate processing of fresh endometrial tissue for enzyme digestion renders it impossible to repeat experimental steps upon errors.

The aim of this study was, therefore, to investigate the measurement of these four subpopulations of NK cells using a multiplex staining assay, which provides a more practical diagnostic approach. In multiple staining, different specific antibodies against each target are linked to different fluorophore labels that emit different wavelengths of light when excited by a specific wavelength of light. Compared to the traditional IHC staining method, this method can quantitatively compare the relative abundance and distribution of multiple targets in a sample and provide information on the interaction and co-localization of different targets, enabling us to identify different subtypes of uNK cells. This approach will not only deepen our understanding of the relationship between uNK cells and RIF but will also provide insights for investigating other immune cell subpopulations in endometrial-related diseases.

Protocol

The study was approved by the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee (CREC ref no.: 2022.581). Women with RIF were recruited from the Assisted Reproductive Technology Center, Prince of Wales Hospital, Chinese University of Hong Kong. RIF was defined as the failure to achieve a clinical pregnancy after the transfer of at least 4 good-quality embryos in a minimum of 3 fresh or frozen cycles in a woman under the age of 40 years¹⁰. Informed consent was obtained from the participants before collecting the endometrial biopsies.

1. Acquisition and processing of endometrial samples

1. Timing of endometrial specimen collection
   1. Collect endometrial samples following a strict protocol to ensure consistency among the study participants. For patients with natural cycles, conduct urinalysis from the 9^th day of the menstrual cycle and perform endometrial biopsy on the 7^th day after the LH surge (LH+7)¹¹ . For patients undergoing hormone replacement therapy (HRT) cycles, administer 6 mg of estradiol valerate orally daily starting on day 2 of the menstrual cycle.
   2. Assess endometrial thickness by ultrasonography on day 13 of the menstrual cycle. When the endometrial thickness reaches or exceeds 8 mm, administer progesterone transvaginally and conduct an endometrial biopsy 5 days after progesterone administration¹².
2. Endometrial sample collection and fixation: introduce a pipette sampler into the uterine cavity, advancing it to the fundal region; apply suction by pulling down the internal plunger, simultaneously scraping the endometrial surface through rotational and vertical movements within the uterine cavity. Rinse the collected specimens with normal saline to eliminate cervical and uterine secretions. Measure samples and assess the size. For 3 mm x 15-25 mm samples, immediately immerse them in 10 mL of 10% neutral buffered formalin and fix them at room temperature for 24-48 h.
3. Dehydration: Upon completion of fixation, subject samples to a dehydration sequence designed to replace water with ethanol. To do this, immerse tissues in 70%, 80%, 90%, and 100% ethanol baths to ensure complete dehydration. Following the final ethanol bath, clear the samples in xylene to remove any residual alcohol and prepare them for paraffin infiltration.
4. Embedding: After placing the tissue flat on the bottom of the mold, melt paraffin and carefully add it to the samples to ensure that the endometrial tissue is completely encapsulated in the paraffin matrix. After filling the molds with paraffin wax, cool them on a freezer table to promote solidification of the paraffin wax. Once the paraffin blocks harden, store at room temperature until further processing is required.
   NOTE: The molds used were specifically chosen to accommodate the size of the endometrial samples obtained to allow the endometrial tissue to lay flat on the bottom without folding, thus ensuring optimal orientation for subsequent sectioning and without creating unnecessary excess.
5. Sectioning: To prepare for sectioning, refrigerate the paraffin blocks at 4 °C for 2 h. Slice the samples into 4 µm sections using a microtome, followed by flattening the slices on the surface of distilled water warmed to 42 °C and mounting them onto adhesive glass slides. Place the slices on a slide dryer overnight and store at room temperature prior use.

2. Optimization of multiplex immunohistochemistry conditions

1. IHC validation: For the validation of IHC staining parameters, adopt a systematic approach to fine-tune conditions for antigen retrieval, antibody concentration, and incubation times.
   1. Within the manufacturer's recommended dilution ranges, test 2-3 concentrations for each antibody. Optimize antigen retrieval by comparing buffers at pH 6 and pH 9, and evaluate incubation protocols at both room temperature for 1.5 h and overnight at 4 °C.
   2. Assess each condition meticulously, focusing on staining intensity, specificity, and background noise. Select the optimal combination yielding the clearest, most specific staining with minimal non-specific binding. This rigorous validation ensured the highest accuracy and reproducibility in the IHC protocol.
2. Tyramide Signal Amplification (TSA) monoplex staining: After determining the optimal staining parameters for individual antibodies, optimize the pairing of antibodies with different TSA dyes. Pair the marker with higher expression with a dye with a lower fluorescence intensity level, or vice versa. During this phase, make adjustments to antibody concentration and incubation time according to the observed fluorescence signal strength. If significant background or non-specific staining is observed, change the paired dye¹³.
3. TSA multiplex staining: For the multiplex immunohistochemistry (m-IHC) protocol, the sequential order of staining has been carefully defined according to established principles. To ensure the multiplexing process reflects the sensitivity and specificity of single-label staining, prioritized antibodies with lower expression levels at the beginning of the staining sequence. Place antibodies that require base repair for antigen retrieval at the end of the staining sequence, as base repair is robust compared to acid-based methods, which can, in some cases, amplify non-specific staining artifacts.
4. Using the findings from steps 2.1 and 2.2, decide on the antibodies to be used and at what stage, based on their antigen retrieval requirements and expression densities. For example, the CXCR4 antibody, which requires a base repair for antigen retrieval, has been placed in the final stages of staining. Conversely, the CD49a antibody, which is characterized by a lower density of expression, has been given the first position in the staining sequence. Maximize the chances of successful detection and minimize any masking effects that might occur in the later stages of the multiplexing procedure.
5. For the remaining antibodies, test different combinations during the optimization phase to determine the most effective sequence that balances both sensitivity and specificity. The results of these comparative analyses will guide the final composition of the multiplex panel, as detailed in Table 1, ensuring that the selected panel provides optimal performance while minimizing background.

3. m-IHC process

1. Dewaxing, hydration, and fixation: Bake the tissue sections at 60 °C for 2 h, followed by sequential immersion in xylene I (10 min), xylene II (10 min), 100% ethanol (5 min), 95% ethanol (5 min), 80% ethanol (5 min), 75% ethanol (5 min), ddH₂O (2 min for 3x), 10% neutral buffered formalin (at least 20 min), and ddH₂O (2 min for 3x).
2. Antigen retrieval: Add 200 mL of the respective AR6 or AR9 buffer, as listed in Table 1 for each antibody, to a heat-resistant antigen retrieval cassette. Place the sections in the appropriate antigen retrieval buffer and heat in a microwave on high power for about 2 min until boiling. Subsequently, cover the cassette loosely, heat on low power for 15 min, and cool to room temperature. Rinse the sections with ddH₂O and PBST.
3. Drawing hydrophobic barrier: After removing the excess solution with absorbent paper, draw a complete circle around the tissue using a hydrophobic barrier pen.
4. Inactivation of endogenous enzyme activity: Treat sections with 3% hydrogen peroxide and incubate for 8 min at room temperature. Wash with PBST 3x for 2 min each.
5. Blocking of non-specific sites: After drying with absorbent paper, apply a blocking solution (1x Antibody Diluent/Block) and incubate for 10 min at room temperature.
6. Incubation with primary antibody: After removal of the blocking solution, apply 150 µL of the first primary antibody, CD49a, to each slide at a dilution of 1:1000 in 1x Antibody Diluent. Incubate as previously optimized in section 2. After incubation, wash the slides with PBST 3x for 2 min each. For the remaining antibodies, see Table 1 for specific dilution ratios.
7. Incubation with secondary antibody: Before applying the secondary antibody, use absorbent paper to remove excess moisture to prevent dilution of the reagent. Add 4-5 drops (approximately 150 µL) of the 1x Anti-Ms + Rb HRP secondary antibody onto each slide, followed by a 10 min incubation at room temperature. Perform three washes with PBST for 2 min each post-incubation.
   NOTE: Please note that the 1x Anti-Ms + Rb HRP secondary antibody is specific for rabbit and mouse primary antibodies.
8. Signal amplification with TSA dye: After blotting excess moisture, dilute the TSA dye 1:100 in the amplification diluent. Add 150 µL of this diluted TSA dye solution to each slide. Incubate for 10 min at room temperature. Wash with PBST 3x for 2 min each.
9. Repeat steps 3.2 and 3.5-3.8 for the TSA procedure using the corresponding antibody and dye.
10. Staining nuclei: After staining the last antibody, eliminate unbound fluorophores by microwaving sections in 200 mL of AR6 buffer within a heat-resistant cassette, boiling for 2 min, then low-power heating for 15 min. Cool samples to room temperature and rinse with ddH₂O and PBST. Add 150 µL of DAPI working solution diluted in PBST (1:10) dropwise, incubate for 10 min at room temperature, and rinse 2x with ddH₂O.
11. Mounting: Once sections are entirely air-dried in a dark area for approximately 30 min, apply a single drop (around 30 µL) of the anti-quenching fluorescent mountant, and place coverslips.
    NOTE: Protect all operations from light after the first application of TSA dye. Prompt observation and image capture were essential post-staining completion to prevent fluorescence quenching.

4. Image acquisition and analysis

1. Exposure time: Capture images using an imager that inherits excitation and emission wavelengths corresponding to the fluorophores. When setting exposure times, check slides with a wide range of marker expressions to ensure accuracy. For each marker/filter combination, manually focus on different areas of the slide. Use the Auto Exposure feature to set the exposure time.
2. Select filters according to the TSA dyes, and examine multiple slides to determine optimal exposure times per channel. Apply these settings consistently to all subsequent acquisitions to maintain comparability throughout imaging.
3. Selection of capture areas: To ensure unbiased sampling across tissue sections and consistent inclusion of the luminal epithelium in each analyzed field, randomly select the initial area for image acquisition, with the critical criterion being the presence of the luminal epithelial border.
4. Following the capture of this initial field, systemically select subsequent areas by moving one field to the left or right of the starting point (omitting one field between each image captured) while maintaining the luminal epithelial border within the frame. Repeat this process to capture at least five different areas, each containing at least a portion of the luminal epithelium, to maintain a consistent depth of analysis throughout¹⁴. This approach aimed to include a minimum of approximately 10,000 stromal cells across all captured fields to ensure comprehensive and representative data collection.
5. Image analysis
   1. Image preparation: To import and analyze multispectral images, first navigate to File > Open Image in the software interface to load the files and the Autofluorescence (AF) slide. After uploading the images, proceed to unmix the fluorophores by clicking the Select Fluors button and selecting the appropriate spectral library. Once all fluorophores are selected, click OK to confirm the selection. To isolate the AF spectrum, use the AF eyedropper tool to sample a representative area of tissue from the AF image, taking care to exclude any non-tissue pixels. After sampling, click Prepare Image or Prepare All in the lower left corner of the interface¹⁵.
   2. Tissue segmentation: Initiate the segmentation process by manually delineating a baseline of 3-5 regions per tissue type within the image, including epithelial areas containing epithelial cells, stromal areas containing stromal cells, and blank areas without cellular content. If the initial segmentation results are not satisfactory, manually annotate additional regions to improve the training dataset, thereby increasing the accuracy and reliability of subsequent analyses. This dynamic adjustment of regions and parameters ensures that the software can automatically identify similar structures in the current and other imported images with increased accuracy.
      NOTE: The initial annotation serves as a training dataset for the software to learn the characteristics of each tissue component. The software's segmentation capabilities are then iteratively refined based on the complexity and variability of the tissue.
   3. Cell Segmentation: Click Segment Cells to start cell segmentation. In the Cell Segmentation Setting, select Nuclei and Membrane. Choose DAPI to identify cell nuclei and adjust the intensity setting to detect all cell nuclei without background noise. Select CD16, CD49a, and CD56 signals to find the membrane and use this signal to assist in nuclear splitting. Use the nuclear component splitting feature to distinguish closely located nuclei.
      NOTE: DAPI highlights the inside of cells. The nuclear component splitting feature helps to distinguish cell nuclei when they are close together or look merged. This step is important for getting reliable cell counts and maintaining the integrity of the analysis.
   4. Cell phenotyping: Adopt four different cell surface markers including CD56, CD49a, CXCR4 and CD16 to differentiate different subtypes of NK cells. In the Phenotyping Scheme, label these four markers as the phenotypes to proceed. In the Associate View section, assign specific colors to each of the phenotypes for identification. Using the Add button, categorize cells as either positively or negatively expressing the markers; for example, in the CD56 phenotype, cells can be labeled as CD56+ or CD56-. Manually label at least five cells for each phenotype during the training process. Perform additional manual labeling if the initial training does not yield effective results.
   5. After the cell phenotyping configuration is set, each cell in the image will indicate whether the four surface markers are positively or negatively expressed. Export the resulting data to a spreadsheet for further analysis.
   6. Data processing: Import the data obtained from all images into a spreadsheet for processing. Calculate percentages of all uNK cell subtypes relative to the total number of cells in the image. The final cell count represents the average count across multiple images, providing a comprehensive overview of the population of various uNK cell subtypes within the sample.

Results

To maintain consistency in the timing of endometrial sample collection for women undergoing the natural cycle, a urine test was performed to precisely detect their luteinizing hormone (LH) surge, with endometrial biopsies conducted 7 days after the LH surge. For women undergoing HRT cycles, samples were scheduled precisely 5 days after progesterone supplementation commenced. To quantify different subtypes of NK cells in the endometrium, m-IHC staining was employed. A schematic outlining of the experimental procedure is depicted in Figure 1.

Optimization of the experimental procedure is critical for achieving high-quality m-IHC staining results to ensure accuracy and consistency in experiment outcomes. Initially, IHC was employed to determine the appropriate antibody concentration, with the selection of antigen retrieval solutions AR6 or AR9. For antibodies such as CD56, CD16, and CD49, an antigen retrieval solution with pH 6 is recommended, while for the CXCR4 antibody, superior staining results were achieved using an antigen repair solution with a pH of 9 compared to a pH 6 buffer. Once the parameters, including antibody dilution ratio, incubation time, and antigen retrieval conditions, have been determined, the most suitable TSA dye can be selected based on the antibody's affinity. A dye with weak fluorescence brightness should be selected to complement the high affinity of the antibody. Reference data on the signal strength of TSA dyes can be found in Table 2. Following the completion of the TSA monoplex assay, the order of the TSA multiplex can be determined. Given that the four antibodies are designed to identify different subtypes of uNK cells, co-localization between the antibodies is inevitable. To minimize spectral overlap and prevent interference between fluorescent moieties, the following strategy was employed: selecting fluorescent moieties with wavelengths as far apart as possible, comparing the position and number of positive signals of the TSA multiplex with those of the TSA monoplex to ensure their identity (Figure 2), and ensuring clear background (Figure 3)¹⁶.

Following the optimization and completion of m-IHC staining procedures, the imaging workstation was employed for image acquisition and the image analysis software was applied for in-depth analysis of the images (Figure 4). For the classification of uNK cells, both traditional and innovative methods were utilized. The traditional classification divides NK cells into two categories, CD56+CD16- and CD56+CD16+¹⁷, based on the expression level of CD16. Building on this foundation, we sought to further delineate NK cell heterogeneity using a novel classification strategy. Using single-cell RNA sequencing (scRNA-seq) from our previous research, we isolated individual NK cells by fluorescence-activated cell sorting (FACS) and performed RNA sequencing to capture the transcriptomic profiles of each cell. Our analysis revealed four distinct NK cell subtypes: NK1 (CD56+CD49a+CXCR4-), NK2 (CD56+CD49a+CXCR4+), NK3 (CD56+CD49a-CXCR4-), and NK4 (CD56+CD49a-CXCR4+). These subtypes were characterized by differential expression of CD49a and CXCR4, with CD56 serving as a universal marker for all NK cells. CD49a was specifically detected in NK1 and NK2 subtypes, whereas CXCR4 was present in NK2 and NK4 subtypes. To validate this classification, we performed flow cytometry analysis to ensure the accuracy of our subtyping method (Figure 5)⁹. The proportion of each uNK cell subtype is presented as a percentage relative to the total number of stromal cells.

Figure 1: Diagram of the workflow of endometrial NK cell analysis. The main steps include the acquisition and collection of the endometrium, optimization of the m-IHC panel, m-IHC staining of the endometrium, and acquisition and analysis of the image. Created with BioRender.com Please click here to view a larger version of this figure.

Figure 2: Representative images for quantification of biomarkers via DAB, TSA monoplex, and TSA multiplex staining. Representative images for each of the antibodies stained by DAB (left panels), TSA monoplex IHC (middle panels), and TSA multiplex IHC (right panels) demonstrated minimal deviation through the development of m-IHC. The TSA simulated IHC images underneath highly imitates the corresponding bright field images. Please click here to view a larger version of this figure.

Figure 3: Experimental results after m-IHC staining of the endometrium. (A) An unmixed image was acquired using the imager at 200x magnification. (B) Further magnification of this unmixed image. (C-F) Images were observed in the TSA dye 570 (CD16), TSA dye 650 (CD56), TSA dye 620 (CXCR4), and TSA dye 520 (CD49a) channels after unmixing using the imaging software. Please click here to view a larger version of this figure.

Figure 4: Analyze m-IHC stained images. The images were analyzed using imaging software's (A) tissue segmentation, (B) cell segmentation, and (C) cell phenotype functions. Please click here to view a larger version of this figure.

Figure 5: Subtype analysis of endometrial NK cells. (A) Conventional endometrial NK cells can be classified into CD56+CD16-NK cells and CD56+CD16+NK cells based on CD16 expression. (B) The innovative subtype analysis method classified endometrial NK cells into four categories: NK1 (CD56+CD49a+CXCR4-), NK2 (CD56+CD49a+CXCR4+), NK3 (CD56+CD49a-CXCR4-), and NK4 (CD56+CD49a-CXCR4+). Please click here to view a larger version of this figure.

Staining cycle	Marker	Clone	Company	Product	Antibody dilution	AR buffer	Fluorophore
1	CD49a	CL7207	Novus Biologicals	NBP2-76478	1:1000	AR6	TSA 520
2	CD56	CD564	Leica	NCL-L-CD56-504	1:100	AR6	TSA 650
3	CD16	SP175	abcam	ab183354	1:100	AR6	TSA 570
4	CXCR4	44716	R&D	MAB172	1:100	AR9	TSA 620

Table 1: List of the concentrations of the antibodies and TSA Dye used in the m-IHC.

Fluorophore	Mantra
TSA dye 520	High
TSA dye 540	Medium
TSA dye 570	Medium
TSA dye 620	Medium
TSA dye 650	High
TSA dye 690	Low

Table 2: TSA dyes and corresponding fluorescence intensity.

Discussion

Embryo implantation involves a complex interaction between the embryo and the endometrium. The immunological status of endometrial homeostasis plays a pivotal role in determining endometrial receptivity. During WOI, the predominant leukocyte population in the endometrium is NK cells. Approximately 90% of uNK cells exhibit high CD56 expression but lack CD16. However, a minor subset of uNK cells resembles peripheral blood NK cells, displaying low CD56 expression but positive CD16 expression¹⁸. These two uNK cell subtypes demonstrate distinct functions. The CD56+CD16- subtype primarily contributes to cytokine production, vascular remodeling, and interactions with trophoblast cells, which are crucial for establishing and maintaining pregnancy. In contrast, the CD56+CD16+ subtype exhibits higher cytotoxicity. Using single-cell RNA sequencing, our recent study further categorized uNK cells into four subsets, further subgrouping CD56+CD16- regulatory NK cells into three subpopulations based on their specific phenotypes and functions⁹. The NK1 subset, characterized by CD56+CD49a+CXCR4-, is involved in tissue remodeling. The NK2 subset, identified by CD56+CD49a+CXCR4+, regulates immune response. The NK3 subset, marked by CD56+CD49a-CXCR4-, interacts with trophoblasts. The NK4 subset, exhibiting CD56+CD49a-CXCR4+, shares similarities with CD56+CD16+ NK cells, displaying cytotoxic functions⁹. In the present study, for the first time, we successfully established a protocol to identify these four distinct NK cell subsets utilizing m-IHC staining.

Conventional IHC staining and flow cytometry are the most commonly used methods for identifying immune cells, each with its own strengths and limitations¹⁹. While conventional IHC is widely used in clinical settings due to its easy accessibility, it faces challenges in simultaneously detecting multiple antigens and quantifying positive signals. This represents a significant hurdle in accurately characterizing immune cell phenotypes, which often necessitate the use of multiple markers for precise measurement. On the other hand, flow cytometry is capable of simultaneously detecting multiple cell markers, but it requires a large volume of fresh samples and does not provide spatial information about cell location within the tissue. Notably, endometrial immune cells are not uniformly dispersed throughout the endometrial tissue; for instance, uNK cells are more densely aggregated near the basal layer of the endometrium compared to areas near the luminal cavity²⁰. These limitations of conventional IHC and flow cytometry can be addressed through m-IHC and the image analysis platform conjugated with machine learning algorithms to achieve tissue segmentation, cell segmentation, and the identification of positive cells. This approach enhances efficiency by analyzing tens of thousands of cells within minutes, minimizes observer bias by automatically identifying positive cells, and enables more accurate analysis using multiple markers. Furthermore, it eliminates the need for fresh samples and allows for batch experiments or repeated analyses when necessary.

Conducting a comprehensive assessment of uNK cell phenotypes using the m-IHC technique can provide a deeper understanding of the correlation between uNK cells and embryo implantation, as well as their changes in women with RIF. Apart from uNK cells, other endometrial immune cells also contribute to local immune homeostasis. For instance, regulatory T lymphocytes (Treg cells) maintain immune tolerance by releasing inhibitory factors, such as IL-10 and TGF-β, which suppress the activity of Th1 and Th2 cells²¹. The differentiation of M2 macrophages, regulated by steroids and Th2 cytokines, contributes to the immunotolerant status by upregulating arginase activity and inhibiting nitric oxide production²². Moreover, dendritic cells (DCs) differentiate into mature tolerogenic cells within the uterus, releasing factors such as sFLT1 and TGF-β1 that regulate vascular neogenesis and inhibit the function of CD8+ T cells, thus maintaining immune homeostasis and supporting embryo development¹⁸. Similar to uNK cells, these immune cells are also heterogeneous, and a single marker cannot accurately identify their phenotypes. The application of m-IHC for assessing uNK cell subsets illustrates the potential for identifying subsets of these endometrial immune cells.

In addition to endometrium, this method can be applied to various tissues for identifying immune cell phenotypes. When applying it to other tissues, only minor adjustments are necessary to ensure optimal staining. Furthermore, if researchers intend to replace certain markers, the original panel can be modified to meet the specific experimental requirements. The versatility and flexibility of this panel render it an adaptable tool for studying different tissues or different cell subsets.

This technology not only facilitates understanding the detailed changes in endometrial immune cells to dissect the underlying mechanism of endometrial non-receptivity but also provides potential for the development of diagnostic tests. The relative ease of obtaining and preserving paraffin blocks and sections has paved the way for the widespread use of this technique. With standardized staining panels, hospitals can leverage automated staining technology in conjunction with sophisticated software algorithms to significantly enhance the efficiency of their analysis procedures. The advantages of m-IHC technology make it a viable option for clinical practices.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
CD49a	Novus Biologicals	NBP2-76478	Primary antibodies
CD56	Leica	NCL-L-CD56-504	Primary antibodies
CD16	abcam	ab183354	Primary antibodies
CXCR4	R&D	MAB172	Primary antibodies
Amplification diluent	Akoya Biosciences	FP1498 series	Fluorophore dilution buffer
Antibody diluents	Akoya Biosciences	ARD1001EA	Dilute the antibody
Citrate buffered solution, pH 6.0 /9.0(10x)	Akoya Biosciences	A6001/A9001	Antigen retrieval solution
inForm advanced image analysis software	Akoya Biosciences	inForm Tissue Finder Software 2.2.6	Data analysis software
Mantra Workstations	Akoya Biosciences	CLS140089	Spectral imaging
microwave	Akoya Biosciences	inverter	Microwave stripping
TSA 520	Akoya Biosciences	FP1487001K	Suitable tyramide-based fluorescent reagents
TSA 620	Akoya Biosciences	FP1495001K	Suitable tyramide-based fluorescent reagents
TSA 650	Akoya Biosciences	FP1496001K	Suitable tyramide-based fluorescent reagents
TSA 570	Akoya Biosciences	FP1488001K	Suitable tyramide-based fluorescent reagents
Poly-L-lysine coated slides	Fisher Technologies	120-550-15	Slides for routine histological use
PolyHRP Broad Spectrum	Perkin Elmer	ARH1001EA	Secondary antibodies
Invitrogen™ Fluoromount-G™ Mounting Medium	ThemoFisher Science	495802	Installation
Spectral DAPI	Akoya Biosciences	FP1490A	Nucleic acid staining