Tumor tissues were obtained from 114 patients with ESCC who underwent surgery at Fujian Medical University Union Hospital from 2015 to 2016. None of these patients received preoperative radiotherapy or chemotherapy before surgery. A retrospective review of the medical records of these patients was performed between January 1, 2015 and June 30, 2020. Relevant clinical data were collected by retrospective review of the files of the patients, and follow-up data were available for all patients. Peripheral blood mononuclear cells (PBMCs) were isolated from six patients with ESCC using Ficoll-Paque PLUS (#171140, GE Healthcare, United States) according to the manufacturer’s instructions. All procedures followed the national and institutional ethical standards, and all samples were obtained in accordance with the institutional policies. All protocols were reviewed and approved by the Research Ethics Committee of Fujian Medical University Union Hospital (No. 2020ky085).

Tumor tissues were fixed in formaldehyde and paraffin-embedded following standard procedures. Tissue sections were incubated with PD-1H antibodies (1:200, #64953, Cell Signaling Technology) and then immunostained using the MaxVision™ HRP-Polymer anti-Rabbit IHC Kit (KIT-5005, Maixin Biol, Fuzhou, China) following the manufacturer’s instructions. Immunostaining of PD-1H was microscopically evaluated (Olympus BX53, Japan). During evaluation of immunostained sections, the investigators were blinded to the clinical status of the patients. After screening the whole section, 10 randomly selected microscopic fields were examined at × 200 magnification (×20 objective lens and ×10 ocular lens; 0.74 mm² per field). Expression levels for PD-1H were scored semiquantitatively based on staining intensity and percentage of stained cells using the immunoreactive score (IRS) as described (Kakavand et al., 2015). The percentage of positively stained cells (PP) was assigned as a numerical score: 0, negative; 1, <25%; 2, 26–50%; 3, 51–75%; and 4, >75% positive cells. The intensity (SI) of the immunostained areas was defined as follows: 0, negative; 1, weak (+); 2, moderate (++); and 3, strong (+++). An immunoreactive score (IRS) ranging from 0 to 12 was calculated using the following formula: IRS = PP × SI. The expression of PD-1H was defined as follows: 0–1, negative; >1–4, mild; >4–8, moderate; >8–12, strong. Based on this formula, the average expression score for each patient was calculated, and the score with the maximum split in survival was chosen as the cutoff using the cutpoint function of Evaluate Cutpoints software (R survMisc package) as described (Ogluszka et al., 2019). Then, the patients were classified into high- or low-expression score groups by a 7.2 score.

Multiplex immunohistochemistry (mIHC) was performed by staining 4-μm-thick formalin-fixed, paraffin-embedded whole-tissue sections with multiple primary antibodies and a TSA 5-color fluorescent IHC kit (Yuanxi Bio, China) as described in a previous study (Stack et al., 2014). The concentration and order of the four antibodies were optimized, and the spectral library was built based on the single-stained slides. Four primary antibody/fluorophore pairs were applied in order: anti-CD4 antibody (#YX32005, Yuanxi Bio)/NEON-TSA 520, anti-CD8 (#YX63005, Yuanxi Bio)/NEON-TSA 620, anti-CD68 (#GM087602, Gene Tech)/NEON-TSA 670, and anti-PD-1H (#64953, Cell Signaling Technology)/NEON-TSA 570. Twelve randomly selected tumor tissue sections were used for mIHC. The slides were deparaffined with a graded series of xylene, and dehydrated in a gradient of alcohols, and then antigen retrieval was performed by microwave. After incubation with 3% H₂O₂ for 10 min, the tissues were blocked in blocking buffer for another 10 min at room temperature. Then, the tissues were incubated with primary antibodies, followed by secondary-HRP and TSA working solutions, respectively, according to the manufacturer’s instructions. Finally, the slides were mounted with ProLong Gold Antifade Reagent with DAPI (#P36931, Invitrogen). All slides were scanned using an Aperio Versa 8 tissue imaging system (Leica, Germany). Images were analyzed using HALO image analysis software (Indica Lab, United States).

RNA sequencing (RNA-Seq) data and clinical information for 182 esophageal carcinoma patients (TCGA ESCC) were collected and downloaded through the TCGA data portal (https://portal.gdc.cancer.gov/). The transcript expression levels were estimated using the fragments per kilobase per million fragments mapped (FPKM) method in HTSeq. Kaplan–Meier survival curve analysis was used to analyze survival between the PD-1H ^high and PD-1H ^low groups as previously described (Rich et al., 2010). Immune cell infiltration was evaluated using Cell-type Identification By Estimating Relative Subsets Of RNA Transcripts (CIBERSORT) algorithm (Newman et al., 2015) based on Tumor Immune Estimation Resource (TIMER) (Li et al., 2020) as described. The gene expression data with standard annotation were uploaded to the CIBERSORT web portal (http://cibersort.stanford.edu/). The algorithm was run by 1,000 permutations and the LM22 signature. Samples with CIBERSORT p < 0.05 were considered eligible for subsequent analysis. The correlation between PD-1H expression and immune infiltration patterns was estimated using linear regression analysis.

Human peripheral blood mononuclear cells (PBMCs) were isolated from the whole blood of six patients with ESCC, and then T cells and monocytes were enriched by Rosettesep™ Human T Cell Enrichment Cocktail and Human Monocyte Enrichment Cocktail (#15028_c, Stemcell Technologies, United States), respectively. T cells and monocytes were stained with antibodies against human CD3e (#560352, BD Biosciences), CD4 (#300532, BioLegend), CD8a (#300922, BD Biolegend), CD14 (#301807, BioLegend), PD-1H (#566672, BD Biosciences), and matched isotype controls (#555749, BD Biosciences) at different time points after blood collection (6, 24, and 48 h). PBMCs, purified T cells, and purified monocytes were stimulated with PMA (50 ng/ml, #P1585, Sigma-Aldrich), LPS (200 ng/ml, #L4391, Sigma-Aldrich), and polyI:C (2 µg/ml, #P0913, Sigma-Aldrich), and anti-CD3/anti-CD28 Dynabeads (#11132D, Gibco) for 24 h. T cells were treated with different concentrations of polyhydroxyalkanoate (PHA) (#PZ0135, Sigma-Aldrich) for 5 h. Samples were run on a BD FACSVerse™ (BD Biosciences, United States) and analyzed using FlowJo software version 10 (BD Biosciences, United States).

Human T cells were stimulated with different concentrations of PHA for 5 h. Total RNA of T cells was extracted using TRIzol (#15596-026, Invitrogen) and was reverse-transcribed into cDNA using a One Step TB Green® PrimeScript PLUS RT-PCR Kit (#RR096A, Takara Bio, United States) according to the manufacturer’s instructions. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed with SYBR™ Green PCR Master Mix (#4309155, Invitrogen). PD-1H primer sequences were as follows: forward (5′-ACG​CCG​TAT​TCC​CTG​TAT​GTC-3′) and reverse (5′-TTG​TAG​AAG​GTC​ACA​TCG​TGC-3′). The expression levels of PD-1H were normalized to the expression of GAPDH. The relative expression levels of PD-1H mRNA in treated cells were calculated using their expression in control cells as a reference transcript.

All statistical analyses were performed using GraphPad Prism 8.0 (GraphPad, Canada). All data are shown as the mean ± SD unless otherwise stated. The chi-squared and Fisher’s exact tests were used for interdependence between staining and clinical data. Univariate and multivariate analyses were performed for prognostic factors of overall survival with the Cox regression model, and the Kaplan–Meier method and the log-rank test were used for survival curves with a plot. p < 0.05 was considered statistically significant. The results represent at least two experiments unless otherwise stated.

The clinical pathological features of 114 patients with ESCC are listed in Table 1. A total of 85.1% tumor tissues (97/114 cases of ESCC) were positive for PD-1H expression by IHC staining. According to the IRS-classification scoring systems, predominant expression evaluation was moderate positive staining (46.5%, 53/114) followed by mild positive staining (23.7%, 27/114), strong positive staining (14.9%, 17/114) and negative staining (14.9%, 17/114), and the representative images of four categories of classification are shown in Figures 1A–D. Figures 1E,F indicated that PD-1H protein was predominantly located in the membrane and cytoplasm of positive staining cells. Next, the optimal cutoff for “low” and “high” expression was determined by the inbuilt algorithm in the survMisc R package. A score of 7.2 was used as an optimal cutoff point and all patients were divided into a high-expression group (74.5%, 85/114) and a low-expression group (25.5%, 29/114). Then, the relationships of PD-1H expression level and clinical characteristics of ESCC were investigated. Our results showed that PD-1H expression level was negatively associated with lymph node metastasis (p = 0.020) and Union for International Cancer Control (UICC) stages (p = 0.009) (Table 1). However, the expression of PD-1H exhibited no correlation with sex, age, tumor stage, or tumor grades of differentiation.

To further investigate cell types with PD-1H expression in the tumor immune microenvironment, we randomly selected 12 tumor sections from 114 ESCC patients for multiplexed immunohistochemistry. Colocalization staining of PD-1H and multiple immunohistochemical markers was measured in tumor tissues. Our results documented that the majority of the PD-1H protein were colocalized with the CD68 macrophage marker. In contrast, CD4 and CD8 colocalized with PD-1H to significantly lower levels, although some areas of colocalization could be found (Figure 2A). The Halo image analysis platform was used to quantify the density and colocalization of positively stained cells. The results showed that the amount of PD-1H colocalized with CD68 was significantly higher than that colocalized with the amount of PD-1H colocalized CD4 (p = 0.0069) or CD8 cells (p = 0.0115) (Figure 2B). Likewise, among PD-1H-positive cells, the percentage of CD68⁺ cells were significantly higher than the percentage of CD4⁺ cells (p = 0.0142) (Figure 2C). Consistent with this result, the absolute number of CD4⁺ T cells and total immune cells was higher in the high PD-1H expression group than in the low PD-1H expression group (Figures 2D,E).

As our results indicated that PD-1H was highly expressed in CD68⁺ myeloid cells of human ESCC tumors, we next assessed possible relationships of PD-1H mRNA expression level and the infiltration of nine types of immune cells by searching ESCC datasets from TCGA. The data showed that PD-1H mRNA (C10orf54) expression levels were significantly positively correlated with cell infiltration of CD4⁺ T cells (Rho = 0.186, p = 0.012), CD8⁺ T cells (Rho = 0.639, p < 0.001), CD68⁺ macrophages (Rho = 0.155, p = 0.038), CD1c⁺ myeloid dendritic cells (Rho = 0.201, p = 0.007), CD14⁺ monocytes (Rho = 0.147, p = 0.048), CD56⁺ NK cells (Rho = 0.196, p = 0.008), CD19⁺ B cells (Rho = 0.236, p = 0.001), and CD31⁺ endothelial cells (Rho = 0.272, p < 0.001) (Figure 3A). In contrast, the PD-1H mRNA expression level was negatively correlated with CD11b⁺CD33⁺HLA-DR⁻ myeloid-derived suppressor cells (MDSCs) (Rho = −0.353 p < 0.001) (Figure 3B).

Recent studies demonstrate that high PD-1H expression correlates with a favorable prognosis in patients with colorectal cancer, high-grade serous ovarian cancer, triple-negative breast cancer, hepatocellular carcinoma, and esophageal adenocarcinoma (Zhang et al., 2018; Loeser et al., 2019; Cao et al., 2020; Zong et al., 2020; Zong et al., 2021). Here, 114 patients who were followed for 65 months were subjected to survival analysis, and the correlation of PD-1H expression with clinicopathologic parameters was analyzed by Kaplan–Meier survival and Cox regression models (Tables 2, 3). In the validation cohort, univariate Kaplan–Meier analyses showed that tumor stage (p = 0.001), lymph node metastasis (p = 0.0 02), UICC stage (p = 0.001), and PD-1H expression (p = 0.029) were associated with the overall survival of patients with human ESCC. However, only tumor stage remained significant on multivariate Cox regression analysis. Although no statistically significant association was found between PD-1H mRNA expression and overall survival in ESCC from TCGA (Figure 4A), Kaplan–Meier analyses showed that ESCC patients with high PD-1H protein expression had significantly improved overall survival (HR = 0.60; 95% CI, 0.27–0.93; p = 0.029; Figure 4B) in the exploratory cohort, especially among the G2/3 subgroup (Figures 4C,D). These results indicated that the PD-1H expression level is positively associated with a favorable prognosis in patients with ESCC.

Given that high PD-1H expression was found in the tumor microenvironment of ESCC, we further examined PD-1H expression in peripheral T cells and monocytes derived from six ESCC patients by FCM and qRT-PCR. No positive staining was detected on the surface of CD4⁺ and CD8⁺ T cells PD-1H expression at 6, 24, and 48 h of blood collection. In contrast, PD-1H expression was found in CD14⁺ monocytes, and the mean fluorescence intensity (MFI) of PD-1H positive staining on monocytes at 24 and 48 h of blood collection was lower than MFI of PD-1H positive staining on monocytes at 6 h (Figure 5A). Although the expression levels of PD-1H mRNA in CD3⁺ T cells were upregulated by the stimulation with PHA for 24 h (Figure 5B), PD-1H protein on the surface of T cells was not detectable (Figure 5C). Moreover, no significant PD-1H expression was detected on the T cell surface after T cells were stimulated with different stimulators, including PMA, LPS, poly (I:C), and TCR signaling (anti-CD3/anti-CD28 Dynabeads). PD-1H expression on CD14⁺ cells was transiently downregulated 24 h after isolation, and membrane PD-1H on CD14⁺ monocytes was decreased by the treatment with PMA, LPS, or poly (I:C) (Figure 5D). Intriguingly, PD-1H expression was maintained on CD14-positive cell surface after monocytes were mixed with T cells and treated with the PMA for 24 h. However, no detectable PD-1H membrane expression was found in activated T cells (Figure 5E).