Cross-reactivity of anti-human programmed cell death ligand 1 (PD-L1) monoclonal antibody, clone 28-8 against feline PD-L1

Abstract

Immunotherapy is a breakthrough in human cancer therapy and has become a major concern in veterinary oncology. However, in cats, many unclear points of the tumor microenvironment exist, including immune checkpoint molecules. A reason is that very few monoclonal antibodies have been proven to react with feline molecules. Therefore, this study investigated whether anti-human programmed cell death ligand 1 (PD-L1) monoclonal antibody, clone 28-8, which is currently commercially available, can also recognize feline PD-L1 by flow cytometry, immunoprecipitation, and immunohistochemical (IHC) staining. We confirmed that the antibody’s specificity by flow cytometry and immunoprecipitation using NIH3T3 cells transfected with feline PD-L1. Additionally, we revealed that PD-L1 was expressed on the surface of some feline cell lines by flow cytometry and clone 28-8 antibody unbound to the cells where feline PD-L1 was knocked out. Furthermore, IHC analysis revealed that PD-L1 was expressed in macrophages in the spleen and lymph nodes from healthy cats and mast cell tumor cells. Therefore, we indicated that the clone 28-8 antibody is a valuable tool in detecting feline PD-L1, and further analysis of tumor tissues is expected in the future.

Immunotherapy is a breakthrough in human cancer therapy and is a field of oncology that has made remarkable progress recently [11]. However, the field of veterinary cancer immunotherapy has also become a primary concern in the last decade. It is expected to be a new cancer treatment following surgery, radiation, and chemotherapy [1]. Monoclonal antibodies against immunological checkpoint molecules have been created in veterinary medicine, specifically in dogs. Expression analysis of these molecules and the therapeutic utility of the antibodies are currently being explored [23, 25, 33]. However, although tumors are also a major problem in cats [3], little analysis has been performed on the cancer microenvironment of feline tumors, including immune checkpoint molecules. A reason is that very few monoclonal antibodies that respond to feline molecules have been identified.

One of the primary therapeutic targets for immunotherapy in treating human cancer is programmed cell death-1 (PD-1) and its ligand, programmed cell death-ligand 1 (PD-L1). T-cells that express PD-1 function as an immunosuppressive molecule to prevent overactive T-cell responses. However, PD-L1 expressed on dendritic cells suppresses the excessive reaction of T-cells by binding to PD-1 on T-cells [2]. Cancer cells hijack this PD-1 / PD-L1 immunosuppressive interaction system and express PD-L1 on the cell surface to evade the host’s tumor immunity [17]. Previously, studies have shown that blood lymphocyte in chronically feline immunodeficiency virus (FIV) infected cats express significantly higher PD-1 and PD-L1 proteins than lymphocytes in FIV-negative cats [6]. It has also been reported that upregulation of PD-1 and PD-L1 gene expression in peripheral blood mononuclear cells occurred in cats diagnosed with Feline Infectious Peritonitis [9]. More recently, it has also been reported that serum PD-1 and PD-L1 levels are significantly higher in cats with human epidermal growth factor receptor 2 (HER2) positive and triple-negative (TN) normal-like breast cancer. Furthermore, PD-L1 expression in cancer cells was significantly higher in HER2-positive samples than in TN normal-like tumors [32], using anti-human PD-L1 antibody, clone 28-8. However, the cross-reactivity of the clone 28-8 antibody to feline PD-L1 has not been clearly proved in that study. Furthermore, there are no reports on commercially available monoclonal antibodies against feline PD-L1. Therefore, it is necessary to clarify commercially available antibodies that can specifically recognize PD-1 and PD-L1 molecules in cats for further analysis in the future.

The aim of this study is to demonstrate that commercially available anti-human PD-L1 monoclonal antibody, clone 28-8, cross-react with feline PD-L1 using several different approaches, such as western blotting, flowcytometry, and immunohistochemistry. We demonstrated that the 28-8 antibody is a valuable tool in detecting feline PD-L1.

MATERIALS AND METHODS

Cell lines

The D10 complete medium (Dulbecco’s modified eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/mL penicillin, 100 µg/mL streptomycin, and 55 µM 2-mercaptoethanol) was used to maintain the human kidney cell line (human embryonic kidney (HEK) 293T), packaging cell line (Platinum-E (PLAT-E) [31]), mouse fibroblast cell line (NIH3T3 [18]), feline kidney-derived cell line (Crandell-Rees Feline Kidney Cell (CRFK) [4, 19]), and feline macrophage cell line (fcwf-4 [37]). Additionally, the feline lymphoma cell line, FT-1 [29], feline lymphoblastoid cell line, FL-4 [43], and feline mammary gland tumor cell lines, FONp, FYMp, FONm, FKNp, and FMCm [42], were kept in R10 complete medium (RPMI-1640 supplemented with 10% FBS, 100 units/mL penicillin, 100 µg/mL streptomycin, and 55 µM 2-mercaptoethanol). However, these cell lines were cultured at 37°C in a humid incubator with 5% CO₂. The NIH3T3 cells were provided by the Institute of Development, Aging, and Cancer at Tohoku University provided. Furthermore, Dr. Hajime Tsujimoto generously donated CRFK, fcwf-4, FT-1, and FL-4 (The University of Tokyo). Dr. Takayuki Nakagawa (The University of Tokyo) established all feline mammary gland tumor cell lines. The feline cell lines used in this study are summarised in Supplementary Table 1.

Specimens

The normal tissue specimens except for lymph nodes were obtained from tissue archives in our laboratory, originally obtained from two healthy mixed cats (8 and 12 year old femle cats) for other purposes. Additionally, the normal tissue specimens of lymph nodes and tumor tissue specimens previously used for diagnostic purposes were obtained from the laboratory of Veterinary Pathology of The University of Tokyo and NorthLab, respectively, and mast cell tumor tissue specimens were obtained from the laboratory of Veterinary Pathology of Yamaguchi University as summarised in Supplementary Table 2.

Preparation for expression plasmids

All primers used for the molecular cloning and recombination are described in Table 1. Feline PD-L1 was cloned by PCR using primers, YTM1929 and YTM1930, with cDNA of FL-4, feline lymphoblastoid cell line, as a template. An amplified product was cut with EcoRI and XhoI and ligated into the EcoRI and XhoI sites of pMx-IP (pMx-IP-fPDL1-K#14). To add a FLAG-tag, PCR was carried out with YTM1929 and YTM1935, with pMx-IP-fPDL1-K#14 as a template, followed by another PCR using amplified product with primers YTM1929 and YTM838 and then inserted into the pANT vector (Nippon genetics) to prepare pANT-fPDL1#1. The fragment of pANT-fPDL1#1 cut with BglII and SmaI was inserted into the BamHI-SnaBI site of pMx-IP, resulting in pMx-IP-fPDL1-FL#9.

Table 1. Primers used in this study.

Primer name	For	Direction	Nucleotide sequences (5′ to 3′)
YTM1929	Cloning of full length feline PD-L1	F	TCGAATTCCAGCTCATTAGCGCGAGAAC (an underline indicates EcoRI site)
YTM1930		R	CCCTCGAGTTACGTCTCCTCAAATTGTAGATC (an underline indicates XhoI site)
YTM1935	Addition of FLAG tag to feline PD-L1	R	GTCGATGTCATGATCTTTATAATC- CGTCTCCTCAAATTG
YTM838	Amplification of FLAG tag	R	GTCGATGTCATGATCTTTATAATCGTCGATGTCATG
YTM2106	Knockout oligos of feline PD-L1	F	CACCGACCTGCTGCTGCAGCAGCTC
YTM2107		R	AAACCGAGCTGCTGCAGCAGCAGGTC

For preparing the knockout vector for feline PD-L1, YTM2106 and YTM2107 were annealed and put into lentiCRISPRv2 to make lentiCRISPR-fPDL1#7. lentiCRISPRv2 was a gift from Dr. Feng Zhang (Addgene plasmid #52961; http//n2t.net/addgene:52961; RRID:Addgene_52961).

Establishment of stably transduced cells

NIH3T3 cells stably expressing fPD-L1, NIH3T3/fPDL1 were established using retrovirus transduction. Briefly, PLAT-E cells (7.5 × 10⁵ cells) were seeded in a 6 well plate, one day before transfection. A mixture of 1.25 µg of pMx-IP-fPDL1-FL#9 was incubated with 5 µL of 1 µg/mL PEI Max (Polysciences, Warrington, PA, USA) in 62.5 µL of OPTI-MEM (Thermo Scientific, Yokohama, Japan) for 15 min at room temperature, then added to the cell culture. Twenty four hours after transfection, medium was replaced with a new D10 medium. After further 24 hr incubation, the supernatant was collected from transfected culture and used for viral transduction into NIH3T3 cells as described by [30]. The transduced cells were cultured for selection in the presence of 10 µg/mL of puromycin (Sigma-Aldrich Japan K.K., Tokyo, Japan).

FYMp cells stably knocked out of fPD-L1, FYMp-kofPDL1, were established using lentivirus transduction. Briefly, HEK293T cells (7.5 × 10⁵ cells) were seeded in a 6 well plate, one day before transfection. A mixture of 0.375 µg of lentiCRISPR-fPDL1#7 was incubated with 0.5 µg of p8.9QV, 0.375 µg of pCVSVG, and 5 µL of 1 µg/mL PEI Max (Polysciences, Warrington, PA, USA) in 62.5 µL of OPTI-MEM for 15 min at room temperature, then added to the cell media. Subsequent steps are the same as above, but 2 µg/mL of puromycin (Sigma-Aldrich Japan K.K.) was used for selection.

Flow cytometry

Cell staining by flow cytometry was performed as explained by [30]. Cells were collected and washed with flow cytometry buffer (Phosphate Buffered Saline (PBS) with 2% FBS and 0.1% NaN₃). Additionally, 2 × 10⁵ cells were stained with anti-human PD-L1 antibody (clone 28-8; Abcam, Tokyo, Japan; dilution 1:100) or isotype control (Normal Rabbit IgG Control; R&D Systems, Minneapolis, MN, USA) for 30 min on ice, followed by Alexa647 labeled anti-rabbit IgG (BioLegend, San Diego, CA, USA) as a secondary antibody. Cells were incubated with propidium iodide immediately before flow cytometric analysis. Results derived from CytoFLEX (Beckman Coulter, Tokyo, Japan) were analyzed using FlowJo software (Treestar, San Carlos, CA, USA). To induce PD-L1 expression, the cell lines were incubated with 10 ng/mL of feline interferon-gamma (IFN-γ; R&D Systems, Minneapolis, MN, USA), a cytokine known to induce PD-L1 expression in other species [21, 28], for 24 hr and collected for staining with anti-human PD-L1 antibody.

Immunoprecipitation and western blotting

Cells were collected and washed once with cold PBS. The cells were lysed for 30 min at 4°C in 250 µL of lysis buffer [1% NP40, 10 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, protease inhibitor cocktails (Nacalai Tesque, Kyoto, Japan), 1 mM Na₃VO₄, and 50 mM NaF], centrifuged at 12,000 × g for 5 min at 4°C, and the supernatant was collected. The cell lysates were precleared with 10 µL of protein A/G agarose (Santa Cruz Biotechnology, Dallas, TX, USA) for 1 hr at 4°C with rotation. One µg of anti-human PD-L1 antibody (28-8) was also incubated with 10 µL protein A/G agarose for 1 hr at 4°C with rotation. Then, the precleared lysates and beads with antibodies were mixed at 4°C overnight with rotation. The immunoprecipitates were washed five times with PBS and dissolved with 1 × loading dye. The samples were subjected to 10% acrylamide gel in non-denaturing condition before proteins were blotted onto the PVDF membrane (Merck, Darmstadt, Germany). After blotting, the membrane was blocked with blocking buffer (Tris-buffered saline with 0.05% Tween 20 and 5% skim milk) for 1 hr at RT. Next, the membrane was incubated with mouse monoclonal anti-FLAG antibody (M2; Sigma-Aldrich Japan KK) in TBST with 0.5% skim milk at 4°C overnight. Horseradish peroxidase (HRP)-conjugated anti-mouse IgG secondary antibody (dilution 1:5,000) was incubated on the membrane for 1 hr after being washed three times with TBS-T for 10 min at a time. The membrane was imaged by Amersham ImageQuant 800 after being washed three times with TBS-T for 10 min (Cytiva, Tokyo, Japan).

Immunohistochemistry

NIH3T3 and NIH3T3/fPDL1 cells were fixed with 10% neutral buffered formalin overnight, followed by addition of agar (FUJIFILM Wako Pure Chemical Corp., Osaka, Lapan) at a final concentration of 1.7%, and embedded in paraffin. Sections were cut from paraffin-embedded tissue or cell block with 4 µm thickness and placed on a glass slide (MAS; Matsunami, Osaka, Japan). After dewaxing in three xylene baths and rehydration in alcohol, antigen retrieval was performed in 10 µg/mL of proteinase K for 10 min at 37°C. After washing three times in PBS for 5 min, the sections were incubated with 3% H₂O₂ in methanol for 30 min at RT. After washing in distilled water, the sections were rinsed with PBS. First, the sections were incubated with 5% BSA in PBS for 30 min at RT to block non-specific binding. Second, the sections were washed in PBS for 5 min and incubated with anti-human PD-L1 antibody (28-8; dilution 1:200 for cell block specimens, 1:100 for tissue specimens) in 2% BSA in PBS at 4°C overnight. Next, the sections were washed three times in PBS for 5 min and incubated with HRP-conjugated anti-rabbit IgG secondary antibody (Histofine Simple Stain MAX PO: Nichirei Corp., Tokyo, Japan) for 30 min at RT. After washing in PBS, brawn color was developed with diaminobenzidine (Nacalai Tesque, Kyoto, Japan) and counterstaining with Mayer’s hematoxylin. The melanin granules were decolorized for melanoma specimens with 10% H₂O₂ at 65°C for 10 min before the staining procedure, and the color was developed by AEC (Nacalai Tesque) instead of diaminobenzidine.

RESULTS

To investigate if clone 28-8 cross-react with feline PD-L1, NIH3T3/fPDL1 cells, NIH3T3 cells transfected with feline PD-L1 were used by flow cytometry. The 28-8 antibody is bound to feline PD-L1 (Fig. 1A). Since 28-8 antibody was expected to recognize the three-dimensional structure of feline PD-L1, NIH3T3/fPDL1 cells were immunoprecipitated with 28-8 antibody, and PD-L1 was detected by wetstern blotting with anti-FLAG antibody, which resulted in two different bands at around 50 kDa and 75 kDa (Fig. 1B).

Fig. 1.

The cross-reactivity of anti-human programmed cell death ligand 1 (PD-L1) monoclonal antibody, 28-8 to feline PD-L1. (A) Flow cytometric analysis of feline PD-L1 expression by 28-8 antibody. NIH3T3 (/mock) and NIH3T3/fPDL1 (/fPDL1) cells were collected and stained with 28-8 antibodies, followed by secondary antibodies. The red and blue histograms indicate isotype control and 28-8 antibody staining, respectively. (B) Immunoprecipitation and western blotting analysis of feline PD-L1 expression. Cell lysates from NIH3T3 and NIH3T3/fPDL1 were extracted and used for immunoprecipitation with 28-8 antibody, followed by western blotting using an anti-Flag antibody. Anti-actin antibody was used as a loading control. (C) Flow cytometric analysis of feline PD-L1 expression in feline cell lines. Cells were treated with or without interferon-gamma (IFN-γ) for 24 hr. After incubation, each cell line was collected and stained with 28-8 antibodies, followed by secondary antibodies. (D) Flow cytometric analysis of feline PD-L1 expression in feline PD-L1-knocked out cell line, FYMp/ko-fPDL1 cells. FYMp cells and FYMp/ko-fPDL1 cells were incubated with feline IFN-γ for 24 hr. After incubation, cells were collected and stained with 28-8 antibodies, followed by secondary antibodies. (E) NIH3T3 cells (A–B) and NIH3T3/fPDL1 cells (C–D) prepared from paraffin-embedded cell block were immunohistochemically stained using rabbit IgG isotype control antibody (A and C) and 28-8 antibody (B and D). Scale bars are 50 µm.

We also performed flowcytometric analysis to examine the binding of the 28-8 antibody to feline PD-L1 endogenously expressed in feline cell lines. Feline lymphoid cell lines (FT-1 and FL-4), kidney-derived cell line (CRFK), macrophage-derived cell line (fcwf-4), and mammary gland tumor cell lines (FONp, FYMp, FONm, FKNp, and FMCm) were available for this purpose. As shown in Fig. 1C, feline PD-L1 was detected in FL-4, CRFK, fcwf-4, and 3 out of 5 mammary gland tumor cell lines (FONp, FYNp, and FKNp). Furthermore, twenty-four hours after treatment with IFN-γ, cell-surface expression of PD-L1 was increased in all five cell lines of mammary gland tumor (Fig. 1C), though the expression levels were not altered in the four cell lines (FT-1, FL-4, CRFK, and fcwf-4). Additionally, fPD-L1 was knocked out using FYMp, the cell line with the highest expression of PD-L1 among feline cell lines (FYMp/ko-fPDL1), and it was confirmed that binding of clone 28-8 to fPD-L1 was lost in FYMp/ko-fPDL1 cells by flow cytometry (Fig. 1D). To determine whether clone 28-8 can be used to detect PD-L1 in immunohistochemistry, paraffin-embedded cell block sections of NIH3T3 cells and NIH3T3/fPDL1 cells were prepared, and stained with clone 28-8. Clone 28-8 clearly stained for NIH3T3/fPDL1 cells, suggesting that clone 28-8 can detect PD-L1 in formalin-fixed paraffin-embedded tissue (Fig. 1E).

Next, the immunohistochemical analysis determined the expression of PD-L1 in feline normal tissue specimens. Feline PD-L1 expression was detected in none of the tissues, including the cerebrum, cerebellum, heart, lungs, liver, pancreas, stomach, small intestine, large intestine, and uterus, except the lymph nodes and spleen where cytoplasms of macrophages were intensively stained (Fig. 2 and Supplementary Fig. 1).

Fig. 2.

Immunohistochemical staining of programmed cell death ligand 1 (PD-L1) in feline normal tissue specimens. Lymph node (A–B) and spleen (C–D) were immunohistochemically stained using isotype control rabbit IgG antibodies (A and C) and 28-8 antibodies (B and D). Scale bars are 50 µm.

Furthermore, immunohistochemistry analysis of feline tumor samples was used to examine PD-L1 expression. We used six samples of each of seven different feline tumor types, including skin mast cell tumor, melanoma, mammary carcinoma, renal cell carcinoma, squamous cell carcinoma (three developed on the skin and the others in the oral cavity), lymphoma, and fibrosarcoma. PD-L1 was expressed in 4 out of 10 samples of mast cell tumors (Fig. 3). However, none of the other tumor tissues showed the positivity of feline PD-L1 (Supplementary Fig. 2).

Fig. 3.

Immunohistochemical staining of programmed cell death ligand 1 (PD-L1) in feline mast cell tumor specimens. Mast cell tumor (MCT) tissues from four different feline MCT cases (A and B, C and D, E and F, and G and H) were Immunohistochemically stained using isotype control rabbit IgG (A, C, E, and G) and 28-8 antibody (B, D, F, and H). These four cases were shown as representative staining patterns. Feline PD-L1 was intensively stained (B and D), less stained (F), and none of cells were stained (H). Scale bars are 50 µm.

DISCUSSION

In this study, we verified anti-human PD-L1 antibody, the clone 28-8, crossreacted to feline PD-L1 from several different approaches. However, in immunoprecipitation with the clone 28-8 antibody shown in Fig. 1B, two bands were larger than the molecular weight of 33.4 kDa, which was determined from the expected amino acid sequence of feline PD-L1. The exact reason is unclear, but one possibility is that glycocylation may affect the molecular weight sizes. Previous reports have demonstrated that human PD-L1 is extensively N-glycosylated. Western blotting of PD-L1 in human cancer cells revealed two bands with different molecular weights [22], as shown in this study. Among five predicted N-glycosylation sites in feline PD-L1 (Fig. 4), four were conserved between species, indicating that these two molecular bands were possibly different glycosylated forms. Or, as shown in the previous study that the crystal structure of full-length PD-L1 created a homodimer [45], the larger molecular weight band may be the dimer form of PD-L1 because the samples were electrophoresed without denaturing in this experiment. Concerning the cross-reactivity of the clone 28-8 antibody, this antibody was also used to detect murine PD-L1 in immunohistochemistry (IHC) and western blotting [8, 15] and human PD-L1 in western blotting [38], but in this study, it could not detect endogenously expressed feline PD-L1 (data not shown). Since the clone, 28-8 antibody, defines the three-dimensional structure of extracellular epitopes [20], differences in amino acids in the secondary structure of PD-L1 may affect detection sensitivity in western blotting (Fig. 4).

Fig. 4.

Multiple human, cat, mouse, and dog programmed cell death ligand 1 (PD-L1) amino acid sequences are aligned. Multiple alignments of human, cat, mouse, and dog PD-L1 amino acid sequences were made by CLUSTAL O (1.2.4) multiple sequence alignment. * Indicates the same amino acid as humans, and the discontinuous epitopes of human PD-L1 recognized by 28-8 antibodies are shown in bold. The predicted N-glycosylation sites are boxed.

Previous research has reported that IFN-γ treatment for 24 hr induced cell-surface expression of PD-L1 in canine tumor cell lines [23, 39]. In this experiment, IFN-γ induced PD-L1 expression in the mammary adenocarcinoma cell lines. A mammary carcinoma is a solid tumor exposed to several cytokines released from immune cells that have infiltrated the tumor tissue, and it has been reported that IFN-γ induced PD-L1 expression on the cell surface in human mammary tumor cell lines [40]. In addition, we investigated whether PD-L1 is expressed in other kinds of available cell lines, and found that it was expressed in lymphoblastoid, kidney, and macrophage cell lines. However, in this experiment, PD-L1 was not induced by IFN-γ in these cell lines which may be less sensitive to IFN-γ than the mammary tumor cell lines. However, this result was contradicted to previous reports that IFN-γ induced PD-L1 expression in lymphocytes and antigen-presenting cells [7, 44], suggesting that PD-L1 expression was not induced by IFN-γ in lymphoblastoid and macrophage cell lines in this study maybe the phenomenon specific to the cell lines used in this study.

Previous studies have reported that PD-L1 is expressed in the normal placenta, heart, lungs, liver, spleen, lymph nodes, thymus, pancreas, and small intestine of mice [5, 7]. In normal canine tissues, PD-L1 is detected in the cerebrum, cerebellum, stomach, duodenum, bladder, uterus, thymus, spleen, and lymph nodes [39]. In this study, expression of PD-L1 in normal feline tissues was limited to macrophages in the lymph nodes and spleen. The reason for this discrepancy is unknown; however, it may suggest that the localization of feline PD-L1 in normal tissues and its function may differ from other species.

The expression of PD-L1 was widely expressed in human tumor tissues, including melanoma, renal cell carcinoma, non-small cell lung cancer, blast cancer, colorectal cancer, thymic cancer, ovarian cancer, multiple myeloma, and associated with clinical responses to immune checkpoint inhibition with anti-PD-1 and PD-L1 agents [36]. Additionally, PD-L1 expression has been observed in canine melanoma, mammary gland tumor, mast cell tumor, renal cell carcinoma, and lymphoma, making it a promising candidate for PD-1/PD-L1 inhibitor treatment [24, 39]. The exact reason why PD-L1 is expressed in various tumors in humans and dogs, whereas in cats, its expression is limited to only some mast cell tumors, are unknown. The limitted expression in normal tissues compared to that in other species raises the possibility that IHC’s sensitivity may not be sufficient. However, since there was clear staining of the tumor cells from some of mast cell tumor cases and the macrophages in lymph nodes and spleen, the function of PD-L1 in cats may be very different from that in other species. Therefore, in the future, it is expected that PD-L1 expression will be intensively analyzed in various feline tumors. Previous studies using 28-8 antibodies have reported that HER2-positive tumors tended to express a higher amount of PD-L1 than TN normal-like mammary carcinoma in cats [32]. Although all mammary carcinoma tissues used in this study was grade III (Supplementary Table 2, [27]), but the subtypes of mammary carcinoma such as the positivity for HER2 were undetermined, all of them may be TN normal-like mammary tumors, because none of the six mammary gland tumors used in this study demonstrated the expression of PD-L1. In this study, the expression of PD-L1 in tumor cells was detected in 4 out of 10 samples of mast cell tumors. All 10 tumor cases used in this study were primary skin tumors with a well-differentiated pathologic grade, and there were not any differences of clinical information between PD-L1 positive cases and negative ones. In the previous study in dogs, mast cell tumors also expressed PD-L1 [24, 39]. Mast cell tumor is the second most frequent malignant cutaneous tumor in cats, and high-grade mast cell tumors have a high risk of metastasis and a demanding prognosis [34]. Further analysis of PD-L1 expression in feline mast cell tumors in more tumor tissue samples as well as cellular expression should be performed whether the tumor is a potential candidate for antibody therapy against PD-1 and PD-L1.

Among several antibodies raised against human PD-L1, clone 28-8 was approved by FDA as a companion diagnostic to identify the patients with metastatic non-small cell lung cancer to make them eligible for the treatment with nivolumab in combination with ipilimumab [13], because the clone 28-8 exhibits high to moderate staining sensitivity and cell membrane staining in PD-L1 expressed on tumor cells [14, 16, 26, 41]. By examining the optical condition for immunohistocchemstry analaysis for feline PD-L1 in this study (data not shown), proteinase K showed better staining results than Universal HIER antigen retrieval reagent (Abcam, Tokyo, Japan), as shown in immunohistochemical analysis of human PD-L1 in the previous report [12]. However, immunohistochemical analysis with clone 28-8 stained not only the cell membrane, but also the cytoplasm diffusely. In a previous report, PD-L1 expressed on macrophages infiltrating lung cancer specimens in immunohistochemistry with clone 28-8 showed diffuse cytoplasmic staining similar to the present results [35], which may be a feature of PD-L1 expressed on macrophages stained by clone 28-8. In addition, previous studies have reported diffuse PD-L1 immunohistochemical expression in the cytoplasm of human cutaneous mastocytomatosis [10], and the diffuse PD-L1 staining seen in feline mast cell tumor tissue specimens may be part of the PD-L1 staining pattern.

Summarily, this study revealed the cross-reactivity of the monoclonal antibody against human PD-L1 to the feline PD-L1 molecule using cell lines, and PD-L1 is expressed on the surface of some feline cell lines. Additionally, immunohistochemical research showed that mast cell tumors, normal cat lymph nodes, and spleens expressed PD-L1 in macrophages. We will further increase the number of samples, analyze the expression of PD-L1 in feline tumor tissue specimens, and explore the potential efficacy of PD-1 / PD-L1 inhibition therapy in feline tumors.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.