Differentiation and mitochondrial metabolic characteristics of CD8⁺ T cells in malignant ascites of epithelial ovarian cancer and their correlation with estrogen signaling

Abstract

The tumor microenvironment of epithelial ovarian cancer (EOC) malignant ascites remains incompletely characterized, particularly regarding CD8⁺ T lymphocyte differentiation and metabolic adaptations. This study investigated these characteristics and their relationship with estrogen receptor (ER) expression in 40 treatment-naïve EOC patients. Matched ascitic fluid and peripheral blood samples underwent comprehensive multicolor flow cytometric analysis to evaluate programmed death-1 (PD-1) expression, differentiation subset distribution, and mitochondrial functional parameters. CD8⁺ T lymphocytes demonstrated substantial enrichment among the lymphocyte population in ascitic fluid (27.9 ± 2.0%) compared to peripheral blood (13.1 ± 1.3%, p < 0.0001), accompanied by elevated PD-1 expression (62.4 ± 2.9% versus 40.9 ± 2.1%, p < 0.0001). Effector memory populations (Tem) predominated within ascites (67.9 ± 2.8%), indicating chronic antigenic exposure. Mitochondrial functional assessment revealed a distinctive metabolic phenotype: despite preserved mitochondrial mass, ascitic CD8⁺ T lymphocytes exhibited significantly reduced mitochondrial calcium (Rhod-2 positivity: 44.7 ± 3.8% versus 68.2 ± 4.5%, p < 0.0001) and reactive oxygen species (MitoSOX positivity: 18.7 ± 2.2% versus 28.9 ± 3.0%, p = 0.0287). Notably, ER-positive tumors (≥ 50% immunohistochemical expression) correlated with reduced frequency of CD8⁺ T cells among lymphocytes (22.1 ± 2.2% versus 32.5 ± 2.9%, p = 0.0040) and decreased PD-1 expression (55.7 ± 3.4% versus 74.5 ± 2.5%, p = 0.0003), while maintaining elevated mitochondrial mass (MTG positivity: 77.3 ± 2.8% versus 43.4 ± 6.9%, p < 0.0001). These findings indicate ascitic CD8⁺ T lymphocytes undergo metabolic reprogramming characterized by minimized mitochondrial stress, potentially representing an adaptive survival mechanism within the hostile ascitic microenvironment. The differential immune phenotypes associated with ER expression suggest distinct immune evasion strategies requiring tailored therapeutic approaches. This comprehensive characterization provides critical insights for developing personalized immunotherapeutic strategies incorporating metabolic modulation and hormone receptor status stratification.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00262-025-04174-1.

Introduction

Ovarian cancer represents the most lethal gynecological malignancy globally, with five-year relative survival rates of approximately 32% despite therapeutic advances [1]. Epithelial ovarian cancer (EOC) accounts for 85–95% of ovarian malignancies, with the majority of patients (70–80%) presenting with advanced-stage disease (FIGO III/IV) at diagnosis [2]. Malignant ascites, a hallmark of advanced disease, significantly contributes to morbidity and correlates with poor prognosis through facilitation of tumor dissemination and therapeutic resistance [3]. Beyond serving as an anatomical compartment, malignant ascites constitutes a complex immune microenvironment containing tumor cells, immune cells, and diverse soluble factors that collectively establish a unique ecosystem promoting tumor progression and immune evasion [4].

The critical role of CD8⁺ T lymphocytes in antitumor immunity has been well established. Multiple studies demonstrate that increased tumor-infiltrating CD8⁺ T lymphocyte density correlates with improved survival in ovarian cancer patients [5]. This observation has gained support through comprehensive gene expression profiling in high-grade serous ovarian carcinoma [6] and large-scale clinical cohort analyses [7]. Recent evidence indicates that ascitic CD8⁺ T lymphocytes exhibit distinct phenotypic and functional characteristics compared to solid tumor-infiltrating T cells, particularly enrichment of effector memory T cells (Tem), suggesting sustained antigenic stimulation and unique differentiation trajectories [8]. However, within the tumor microenvironment, T lymphocytes frequently exhibit functional exhaustion characterized by upregulation of inhibitory receptors, particularly PD-1 [9]. A 2023 meta-analysis revealed a disappointing pooled objective response rate of only 6.7% to checkpoint inhibitor monotherapy [10], highlighting critical knowledge gaps in understanding immune evasion mechanisms. This limited therapeutic response underscores the need for deeper mechanistic understanding of T cell dysfunction within the ascitic microenvironment, particularly metabolic alterations.

T lymphocyte functionality depends critically on cellular metabolic status. Mitochondria, as central metabolic organelles, regulate T lymphocyte activation, differentiation, and effector function maintenance [11]. During chronic antigenic stimulation, tumor-infiltrating T lymphocytes demonstrate mitochondrial dysfunction resulting in impaired effector capabilities [12]. Recent investigations have elucidated molecular mechanisms whereby mitochondrial dysfunction drives T cell progression from precursor to terminally exhausted states through HIF-1α-mediated glycolytic reprogramming [12]. Moreover, comprehensive assessment of mitochondrial quality, including mitochondrial mass, membrane potential, calcium homeostasis, and reactive oxygen species production, is crucial for understanding T cell metabolic adaptation [13]. Despite these observations, the metabolic characteristics of T lymphocytes within ovarian cancer ascitic fluid remain poorly characterized. The unique physicochemical environment of ascites, characterized by hypoxia, nutrient deprivation, and elevated lactate, and its impact on CD8⁺ T cell mitochondrial metabolic reprogramming remains poorly understood.

Estrogen signaling represents an underexplored modulator of tumor immunity. Although estrogen receptors (ER) participate in immune cell regulation [14], their specific impact on ascitic T lymphocyte function requires clarification. Accumulating evidence indicates that estrogen signaling modulates tumor immune responses through multiple mechanisms, including regulation of myeloid-derived suppressor cell recruitment and function, influence on tumor-associated macrophage polarization, and direct effects on T cell activation and effector function [15]. In ovarian cancer, ER expression status may represent different immune evasion strategies with important implications for personalized immunotherapy strategy development.

Based on this background, we hypothesized that ascitic CD8⁺ T cells undergo unique metabolic reprogramming regulated by estrogen signaling, which impacts their antitumor function. This investigation systematically evaluated differentiation states and mitochondrial metabolic features of CD8⁺ T lymphocytes in EOC ascitic fluid, examining correlations with ER expression to establish foundations for personalized immunotherapeutic approaches. Our research aims to comprehensively characterize the differentiation status and functional phenotype of ascitic CD8⁺ T cells, systematically evaluate their mitochondrial metabolic characteristics, explore associations between ER expression and T cell immune features, and provide theoretical foundations for personalized immunotherapy based on metabolic modulation and hormone receptor stratification.

Materials and methods

Patient recruitment and clinical characteristics

This prospective study enrolled 40 consecutive treatment-naïve patients with histologically confirmed EOC at the Department of Obstetrics and Gynecology, The First Affiliated Hospital of Soochow University, between November 2023 and November 2024. Inclusion criteria comprised: (1) surgical and histopathological confirmation of EOC, (2) no prior therapeutic interventions including surgery, chemotherapy, radiation, or immunotherapy, (3) extractable ascitic fluid volume exceeding 50 mL, (4) age 18–75 years. Exclusion criteria included: (1) concurrent malignancies, (2) autoimmune conditions, (3) immunosuppressive medication within three months, (4) active infections. The institutional Ethics Committee approved the study protocol (NO.027(2021)), with written informed consent obtained from all participants.

Specimen collection and processing

Sample collection occurred intraoperatively under sterile conditions. Peripheral blood (10 mL) was obtained via antecubital venipuncture before anesthetic induction using EDTA-containing tubes. Ascitic fluid (50–200 mL) was aspirated following laparotomy and transferred to heparinized containers. Specimens were maintained at 4 °C with mononuclear cell isolation completed within four hours.

Mononuclear cells were isolated using density gradient centrifugation with Ficoll-Paque PLUS (density 1.077 g/mL, Dakewe Biotech). For ascitic samples with erythrocyte contamination, cells underwent treatment with erythrocyte lysis buffer (155 mM NH₄Cl, 10 mM KHCO₃, 0.1 mM EDTA, pH 7.4). Following isolation, cells were resuspended in RPMI-1640 medium supplemented with 10% fetal bovine serum, with viability assessed using trypan blue exclusion.

Flow cytometric analysis

Multiparameter flow cytometry was performed using a CytoFLEX instrument (Beckman Coulter). For mitochondrial functional assessment, cells (5 × 10⁵) were incubated with mitochondrial-specific probes: MitoTracker Green (MTG, 50 nM, Invitrogen M7514) for mitochondrial mass; Rhod-2 AM (2 μM, Invitrogen R1245MP) for mitochondrial calcium; MitoSOX Red (5 μM, Invitrogen M36008) for mitochondrial superoxide. Following 30-min incubation at 37 °C under light-protected conditions, cells were stained with fluorochrome-conjugated antibodies: anti-CD8-PE-Cy7 (clone SK1, BioLegend), anti-PD-1-APC (clone EH12.2H7, BioLegend), anti-CD45RA-PerCP/Cy5.5 (clone HI100, BioLegend), and anti-CCR7-APC-Cy7 (clone G043H7, BioLegend).

Data acquisition employed CytExpert software with minimum 10,000 CD8⁺ T lymphocyte events per sample. Analysis utilized FlowJo software (version 10.8.1) with sequential gating for lymphocyte identification, doublet exclusion, and subset characterization. Differentiation subsets were defined according to established criteria [16] as: naïve (Tnaive, CCR7⁺CD45RA⁺), central memory (Tcm, CCR7⁺CD45RA⁻), effector memory (Tem, CCR7⁻CD45RA⁻), and effector memory RA⁺ (TEMRA, CCR7⁻CD45RA⁺).

Statistical analysis

Statistical analyses employed GraphPad Prism 9.0 and SPSS 26.0. Data are presented as mean ± standard error of the mean (SEM). Normality was assessed using Shapiro–Wilk testing. Matched samples underwent paired t-test comparison; unmatched samples were analyzed using independent t-tests. Multiple comparisons employed one-way ANOVA with Tukey's post-hoc testing and Bonferroni correction. Pearson correlation coefficients assessed variable relationships. Two-sided p < 0.05 indicated statistical significance.

Results

Patient clinical characteristics

A total of 40 treatment-naïve patients with epithelial ovarian cancer were enrolled in this study, with their clinicopathological characteristics summarized in Table 1. The median age of the cohort was 54.0 years (interquartile range: 51.0–58.3 years), with 77.5% (31/40) of patients aged ≥ 50 years and 70.0% (28/40) being postmenopausal. Histopathological classification revealed high-grade serous ovarian carcinoma (HGSOC) as the most prevalent histotype, accounting for 72.5% (29/40) of cases, followed by clear cell carcinoma and low-grade serous carcinoma at 10.0% (4/40) each, and endometrioid carcinoma at 7.5% (3/40). According to FIGO staging criteria, 82.5% of patients presented with advanced-stage disease at diagnosis (stage III: 40.0%, stage IV: 42.5%). Tumor estrogen receptor expression assessment demonstrated high expression (≥ 50% positive cells by immunohistochemistry) in 57.5% (23/40) of patients, while 42.5% (17/40) exhibited low or negative expression.

Table 1.

Baseline clinicopathological characteristics of patients (n = 40)

Clinicopathological variables	Categories	n (%)
Age, median (IQR)	54.0 (51.0–58.25)
Age	≥ 50	31(77.50)
	< 50	9 (22.50)
Menopausal status	Postmenopausal	28 (70.00)
	Premenopausal	12 (30.00)
Histological type	Clear Cell Carcinoma (CCC)	4 (10.00)
	Endometrioid Carcinoma (EC)	3 (7.50)
	High-Grade Serous Ovarian Carcinoma (HGSOC)	29 (72.50)
	Low-grade serous carcinoma (LGSC)	4 (10.00)
FIGO stage	I	2 (5.00)
	II	5 (12.50)
	III	16 (40.00)
	IV	17 (42.50)
Ascites volume	≥ 500	21 (52.50)
	< 500	19 (47.50)
Tumor size	≥ 10	18 (45.00)
	< 10	22 (55.00)
ER status	High (≥ 50%)	23 (57.50)
	Low/Negative (< 50%)	17 (42.50)
Total(n)		40 (100)

Ascites volume was categorized using a 500 mL cutoff

Tumor size refers to the maximum diameter of the primary lesion, with a cutoff at 10 cm

ER status was classified as high (≥ 50%) or low/negative (< 50%) based on the percentage of positive tumor cells

Frequency and PD-1 expression of CD8⁺ T Cells in ascites versus peripheral blood

To accurately identify CD8⁺ T cell populations, we established a stringent flow cytometry sequential gating strategy (Fig. 1a). Analysis of paired samples from 40 EOC patients revealed that CD8⁺ T cell infiltration among lymphocytes in malignant ascites reached 27.9 ± 2.0%, significantly higher than the 13.1 ± 1.3% observed in paired peripheral blood (p < 0.0001) (Fig. 1b). Further analysis of PD-1 expression demonstrated that 62.4 ± 2.9% of CD8⁺ T cells in ascites expressed PD-1, markedly elevated compared to 40.9 ± 2.1% in paired peripheral blood (p < 0.0001) (Fig. 1c), indicating that CD8⁺ T cells within the ascites microenvironment exist in a distinct functional state.

Fig. 1.

Comparison of CD8⁺ T cell frequency and PD-1 expression in malignant ascites versus peripheral blood. a Sequential gating strategy for flow cytometric identification of CD8⁺ T cells. Lymphocytes were identified based on forward scatter (FSC) and side scatter (SSC) properties, followed by doublet exclusion (FSC-A vs FSC-H), and CD8⁺ T cell identification based on CD8 expression. b Frequency of CD8⁺ T cells among total lymphocytes (LYM) in malignant ascites (MA) compared to paired peripheral blood (PB) samples (n = 40, paired t-test). c PD-1 expression on CD8⁺ T cells in malignant ascites versus paired peripheral blood from epithelial ovarian cancer patients (n = 40, paired t-test). Data represent mean ± SEM. ****p < 0.0001

Association between CD8⁺ T cell mitochondrial metabolic characteristics and functional status

To comprehensively understand the impact of the ascites microenvironment on CD8⁺ T cell metabolism, we systematically analyzed mitochondrial functional parameters of CD8⁺ T cells and explored their relationship with T cell functional status. Flow cytometric analysis revealed that all mitochondrial function probe staining displayed clear bimodal distributions, enabling unambiguous discrimination between positive and negative populations. The results demonstrated that MTG positivity was significantly higher in PD-1⁺ CD8⁺ T cells compared to PD-1⁻ cells in both ascites (63.9 ± 4.4% vs 60.7 ± 4.2%, p = 0.0097) and peripheral blood (67.4 ± 3.0% vs 57.2 ± 3.6%, p < 0.0001) (Fig. 2a, b), indicating that a higher proportion of CD8⁺ T cells with elevated PD-1 expression maintain detectable levels of mitochondrial content.

Fig. 2.

Mitochondrial metabolic profiles of CD8⁺ T cells in ascites and peripheral blood from epithelial ovarian cancer patients and their association with PD-1 expression. a, b MitoTracker Green (MTG) positivity in PD-1⁺ versus PD-1⁻ CD8⁺ T cells from malignant ascites a and paired peripheral blood b of epithelial ovarian cancer patients (paired t-test). c Mitochondrial calcium (Rhod-2) and mitochondrial reactive oxygen species (MitoSOX) positivity in total CD8⁺ T cells from ascites versus paired peripheral blood (paired t-test). d Rhod-2 and MitoSOX positivity in PD-1⁺ CD8⁺ T cells from ascites versus peripheral blood (paired t-test). e Rhod-2 and MitoSOX positivity in PD-1⁻ CD8⁺ T cells from ascites versus peripheral blood (paired t-test). Data represent mean ± SEM. *p < 0.05, **p < 0.01, ****p < 0.0001; ns, not significant

Comparison of CD8⁺ T cell mitochondrial function between ascites and peripheral blood environments revealed that ascites-derived CD8⁺ T cells exhibited significantly reduced Rhod-2 positivity (44.7 ± 3.8% vs 68.2 ± 4.5%, p < 0.0001) and MitoSOX positivity (17.0 ± 2.2% vs 23.4 ± 3.2%, p = 0.0287) compared to peripheral blood (Fig. 2c). This suggests that a smaller proportion of CD8⁺ T cells in ascites experience mitochondrial calcium overload and oxidative stress. Further analysis revealed that these differences were primarily observed in the PD-1⁺ CD8⁺ T cell subset: ascites-derived PD-1⁺ CD8⁺ T cells showed significantly lower Rhod-2 positivity (44.3 ± 3.6% vs 68.6 ± 4.7%, p < 0.0001) and MitoSOX positivity (18.0 ± 2.4% vs 27.0 ± 3.6%, p = 0.0053) compared to peripheral blood (Fig. 2d), while no significant differences were observed between ascites and peripheral blood in the PD-1⁻ CD8⁺ T cell subset (Fig. 2e).

Distribution of CD8⁺ T cell differentiation subsets in EOC patient ascites versus peripheral blood

Based on CCR7 and CD45RA expression, CD8⁺ T cells were classified into four differentiation subsets (Fig. 3a). Significant differences in differentiation patterns were observed between ascites and peripheral blood CD8⁺ T cells: the proportion of Tem cells was significantly elevated in ascites (67.9 ± 2.8% vs 44.5 ± 2.8%, p < 0.0001), while the proportions of naive T cells (7.0 ± 1.2% vs 23.4 ± 2.4%, p < 0.0001) and TemRA cells (17.1 ± 2.1% vs 24.4 ± 2.3%, p = 0.0028) were significantly reduced; no significant difference in Tcm cell proportions was observed between the two groups (Fig. 3b, c). This Tem-dominated differentiation pattern reflects persistent antigen stimulation experienced by T cells in the ascites environment.

Fig. 3.

Distribution of CD8⁺ T cell differentiation subsets in malignant ascites and paired peripheral blood from epithelial ovarian cancer patients. a Flow cytometric gating strategy for defining CD8⁺ T cell differentiation subsets based on CCR7 and CD45RA expression: naive T cells (Tnaive, CCR7⁺CD45RA⁺), central memory T cells (Tcm, CCR7⁺CD45RA⁻), effector memory T cells (Tem, CCR7⁻CD45RA⁻), and effector memory RA⁺ T cells (TEMRA, CCR7⁻CD45RA⁺). b Representative flow cytometry plots showing the distribution of CD8⁺ T cell differentiation subsets in malignant ascites and paired peripheral blood. c Frequencies of each CD8⁺ T cell differentiation subset in ascites versus paired peripheral blood from epithelial ovarian cancer patients (n = 40, paired t-test). Data represent mean ± SEM. **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant

Functional status and mitochondrial metabolic characteristics of CD8⁺ T cell differentiation subsets

Analysis of functional phenotypes across differentiation subsets (Fig. 4a) revealed that PD-1 expression rates in Tem and TemRA subsets from ascites were significantly higher than those from peripheral blood (Tem: 51.9 ± 2.2% vs 42.3 ± 2.7%, p = 0.0076; TemRA: 70.3 ± 3.5% vs 59.6 ± 3.3%, p = 0.0091), while no significant differences in PD-1 expression were observed in naive T cells and Tcm subsets between the two environments (Fig. 4b).

Fig. 4.

Mitochondrial metabolic heterogeneity among CD8⁺ T cell differentiation subsets in ascites versus peripheral blood from epithelial ovarian cancer patients. a Flow cytometric gating strategy for analyzing PD-1 expression and mitochondrial markers (MTG, Rhod-2, MitoSOX) within each CD8⁺ T cell differentiation subset. b PD-1 expression on different CD8⁺ T cell differentiation subsets in ascites versus paired peripheral blood (n = 40, paired t-test). c MTG positivity across differentiation subsets. d Rhod-2 positivity across differentiation subsets. e MitoSOX positivity across differentiation subsets. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant

Mitochondrial functional analysis revealed distinct metabolic characteristics among subsets (Fig. 4c–e). Ascites-derived Tem cells exhibited higher MTG positivity (70.5 ± 4.0% vs 60.9 ± 3.4%, p = 0.0258) and lower MitoSOX positivity (14.0 ± 2.4% vs 22.3 ± 3.1%, p = 0.0128), suggesting that a greater proportion of cells in this subset maintain higher mitochondrial mass with reduced oxidative stress. Conversely, ascites-derived TemRA cells showed lower MTG positivity (45.9 ± 3.8% vs 60.1 ± 3.9%, p = 0.0017) and higher MitoSOX positivity (37.5 ± 4.3% vs 20.2 ± 3.2%, p = 0.0016), reflecting severe functional impairment. Rhod-2 positivity was significantly reduced in ascites-derived naive T cells (40.4 ± 3.7% vs 63.6 ± 5.4%, p < 0.0001), Tcm (42.6 ± 3.3% vs 57.5 ± 4.8%, p = 0.0008), and Tem (33.9 ± 3.0% vs 63.7 ± 4.3%, p < 0.0001) subsets.

Validation of findings in HGSOC cohort

Given that HGSOC represented the predominant histological subtype in our cohort (72.5%, 29/40), we performed subgroup analysis to validate our primary findings within this molecularly distinct entity. Analysis of the HGSOC subset demonstrated consistent patterns with the overall cohort. The frequency of CD8⁺ T cells among lymphocytes in HGSOC ascites remained significantly elevated compared to paired peripheral blood (p < 0.0001), with similarly elevated PD-1 expression patterns (p < 0.0001) (Supplementary Fig. 1a, b). The mitochondrial metabolic adaptations observed in the total cohort, characterized by preserved mitochondrial mass with reduced mitochondrial calcium and reactive oxygen species, were consistently maintained in the HGSOC subset (Supplementary Fig. 1c–j). The Tem-dominated differentiation pattern was also preserved in HGSOC patients (Supplementary Fig. 1 k–m). Furthermore, the differentiation subset-specific metabolic profiles, including differential PD-1 expression patterns and mitochondrial functional parameters across T cell subsets, remained consistent with the overall cohort findings (Supplementary Fig. 2a–d). Other histological subtypes, including clear cell carcinoma, endometrioid carcinoma, and low-grade serous carcinoma, had insufficient sample sizes for meaningful statistical analysis.

Association between estrogen receptor status and CD8⁺ T cell differentiation and mitochondrial activity in ascites

Correlation analysis exploring the relationship between clinicopathological parameters and T cell function (Fig. 5a) revealed significant associations between tumor ER expression and multiple mitochondrial metabolic indicators. While ER expression demonstrated the strongest associations with CD8⁺ T cell characteristics, progesterone receptor (PR) expression showed moderate positive correlation with the frequency of CD8⁺ T cells among lymphocytes (r = 0.433, p = 0.011). Other immunohistochemical markers including WT1, P53, and CK7 exhibited no significant correlations with T cell metabolic parameters. Based on these findings, patients were stratified into ER-positive (immunohistochemistry ≥ 50% positive cells, n = 23) and ER-negative/low expression (immunohistochemistry < 50%, n = 17) groups.

Fig. 5.

Association between estrogen receptor expression and mitochondrial metabolism of CD8⁺ T cells in ascites from epithelial ovarian cancer patients. a Heatmap displaying Pearson correlation coefficients between CD8⁺ T cell markers (CD8, PD-1, MTG, Rhod-2, MitoSOX) and clinicopathological parameters. Flow cytometric markers are expressed as percentages. Clinical parameters include: age (years), menopausal status (0 = premenopausal, 1 = postmenopausal), ascites volume (mL), tumor size (cm), FIGO stage (I = 1, II = 2, III = 3, IV = 4), CA125 (U/mL), HE4 (pmol/L), and immunohistochemical markers (Ki-67, P53, WT1, ER, PR, CK7, P16) quantified as continuous variables. Red indicates positive correlations; blue indicates negative correlations. Asterisks denote significant correlations (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). b Frequency of CD8⁺ T cells among total lymphocytes in ascites from ER-positive (ER ≥ 50%, n = 23) versus ER-negative/low (ER < 50%, n = 17) patients (unpaired t-test). c PD-1 expression on CD8⁺ T cells in ascites from ER-positive versus ER-negative/low patients. d MTG positivity of CD8⁺ T cells in ascites from ER-positive versus ER-negative/low patients. e Rhod-2 positivity of CD8⁺ T cells in ascites from ER-positive versus ER-negative/low patients. f MitoSOX positivity of CD8⁺ T cells in ascites from ER-positive versus ER-negative/low patients. g PD-1 expression on CD8⁺ T cell differentiation subsets in ascites from ER-positive versus ER-negative/low patients. Data represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant

The frequency of CD8⁺ T cells among lymphocytes in ascites from ER-positive patients were significantly lower than those from ER-negative patients (22.1 ± 2.2% vs 32.5 ± 2.9%, p = 0.0040) (Fig. 5b), with PD-1 expression rates also significantly reduced (55.7 ± 3.4% vs 74.5 ± 2.5%, p = 0.0003) (Fig. 5c). Analysis of differentiation subsets showed that PD-1 expression in both Tem (46.3 ± 2.8% vs 57.9 ± 3.3%, p = 0.0124) and TemRA (60.9 ± 3.7% vs 76.0 ± 4.5%, p = 0.0043) subsets from ER-positive patients was significantly lower than that from ER-negative patients (Fig. 5g).

Regarding mitochondrial metabolism, CD8⁺ T cells from the ER-positive group demonstrated higher MTG positivity (77.3 ± 2.8% vs 43.4 ± 6.9%, p < 0.0001) (Fig. 5d) along with lower Rhod-2 positivity (36.3 ± 4.4% vs 47.9 ± 4.0%, p = 0.0384) and MitoSOX positivity (13.6 ± 2.2% vs 22.6 ± 3.2%, p = 0.0255) (Fig. 5e, f). These differences primarily originated from the PD-1⁺ subset: PD-1⁺ CD8⁺ T cells from ER-positive patients showed significantly higher MTG positivity compared to the ER-negative group (78.9 ± 2.8% vs 43.8 ± 6.9%, p < 0.0001), while exhibiting lower Rhod-2 positivity (38.9 ± 3.9% vs 52.3 ± 4.3%, p = 0.0471) and MitoSOX positivity (12.8 ± 2.6% vs 22.4 ± 3.1%, p = 0.0110).

Discussion

This investigation provides comprehensive characterization of CD8⁺ T lymphocyte differentiation and metabolic adaptations within EOC malignant ascites, revealing novel associations with estrogen receptor expression. Our findings demonstrate that ascitic CD8⁺ T lymphocytes exhibit a distinctive phenotype characterized by substantial infiltration, elevated PD-1 expression, effector memory predominance, and unique metabolic adaptations that minimize mitochondrial stress while preserving mitochondrial mass.

The observed enrichment of CD8⁺ T lymphocytes within ascites accompanied by elevated PD-1 expression aligns with previous observations by Landskron and colleagues [17], confirming functional compromise in the ascitic microenvironment. However, our data reveal metabolic adaptations distinct from those reported in solid tumor infiltrates, potentially reflecting fundamental differences between ascitic and solid tumor microenvironments. While Scharping et al. [18] reported severe mitochondrial dysfunction in solid tumor-infiltrating T lymphocytes, we observed preserved MTG positivity in ascitic Tem cells, suggesting retention of metabolic plasticity and potential for functional recovery.

The distinctive metabolic phenotype—characterized by reduced Rhod-2 and MitoSOX positivity despite maintained mitochondrial mass—represents a novel adaptation pattern diverging from conventional exhaustion paradigms. This "metabolically quiescent" state likely reflects survival adaptations to the unique ascitic microenvironment, which presents multiple metabolic challenges including hypoxia, nutrient deprivation, and elevated lactate concentrations [19]. Recent work by Wu and associates [12] demonstrated that mitochondrial dysfunction is associated with terminal exhaustion through HIF-1α-mediated glycolytic reprogramming. Our findings suggest ascitic T lymphocytes employ alternative adaptive strategies, minimizing oxidative stress while maintaining mitochondrial capacity. This adaptation may be mediated through selective downregulation of electron transport chain activity, enhanced antioxidant defense systems, and optimized mitochondrial quality control mechanisms [20].

The heterogeneity among differentiation subsets carries important functional implications. High MTG/low MitoSOX characteristics of Tem cells suggest enrichment for "stem-like" precursors with self-renewal capacity [21, 22], while the opposite pattern in TEMRA cells aligns with terminal differentiation [23]. Recent single-cell sequencing studies support the functional significance of this heterogeneity, demonstrating that T cell subsets with stem-like characteristics play crucial roles in checkpoint blockade responses [24]. This metabolic heterogeneity may contribute to the limited efficacy of checkpoint blockade monotherapy and suggests subset-specific targeting strategies. Our findings extend this concept, suggesting that metabolic characterization may provide more accurate functional predictions than conventional phenotypic classification.

Our discovery of ER-associated immune phenotypes provides critical insights into hormone-mediated immune regulation. The comprehensive review by Gjorgoska and Rižner [14] highlighted estrogen's complex involvement in ovarian cancer microenvironments. Our empirical data extend these concepts, showing an association between ER-positive tumors and reduced frequency of CD8⁺ T cells among lymphocytes and their activation. This finding aligns with recent studies demonstrating estrogen-mediated suppression of antitumor immune responses through multiple mechanisms [25, 26]. Multiple mechanisms may contribute, including estrogen-mediated suppression of pro-inflammatory mediators [27] and altered chemokine expression profiles [28]. The reduced PD-1 expression and elevated MTG positivity in ER-positive patients suggest relatively quiescent T lymphocyte states, potentially reflecting alternative immune evasion strategies distinct from exhaustion-based mechanisms.

These findings carry substantial therapeutic implications. Ascitic Tem cells, with preserved mitochondrial mass and minimal oxidative stress, represent promising targets for functional restoration. Their metabolic characteristics suggest potential responsiveness to metabolic interventions combined with checkpoint blockade. Evidence supporting this approach includes studies demonstrating enhanced T lymphocyte function through PGC-1α overexpression [29] or IL-10 treatment [30], both targeting mitochondrial metabolism. Recent studies have demonstrated that mitochondrial transfer via nanotubes can enhance T cell metabolic fitness and antitumor function, providing novel therapeutic avenues [31].

Patient stratification based on ER expression may optimize therapeutic selection. ER-negative patients, exhibiting enhanced CD8⁺ T lymphocyte infiltration and activation, may demonstrate superior checkpoint blockade responses. Conversely, ER-positive patients likely require initial strategies to enhance T lymphocyte infiltration, potentially through CDK4/6 inhibitors or anti-estrogen therapy to "inflame" the tumor microenvironment [32, 33]. This hormone receptor-based immunotherapy stratification strategy, showing promise in breast cancer, warrants extension to ovarian cancer treatment.

Our flow cytometric methodology enables real-time metabolic monitoring during treatment, providing dynamic assessment capabilities [34]. Furthermore, ascitic fluid represents an accessible "liquid biopsy" for longitudinal monitoring [35, 36], potentially informing treatment decisions and predicting responses.

Several limitations warrant consideration. First, while our CCR7/CD45RA-based classification follows established conventions, future studies could benefit from incorporating additional functional markers beyond our current PD-1 analysis, such as co-stimulatory molecules (CD27, CD28) and other exhaustion-related markers (TIM-3, LAG-3), to achieve more comprehensive functional subset characterization. Second, the absence of functional validation experiments, including cytotoxicity assays and cytokine profiling, limits our ability to directly link metabolic alterations with functional consequences. Third, the predominance of HGSOC (72.5%) in our cohort, while allowing for subgroup validation, limits the generalizability to other histological subtypes. Fourth, the cross-sectional design precludes causal inference between metabolic changes and functional states. Finally, limited sample availability prevented comprehensive functional studies that would strengthen our metabolic findings.

Future research priorities include functional validation of ascitic CD8⁺ T cell metabolic phenotypes through cytotoxicity assays and metabolic functional analyses in larger cohorts; preclinical evaluation of combined metabolic-checkpoint blockade strategies; development of metabolic signature-based stratification systems [37]; mechanistic dissection of ER-mediated T lymphocyte regulation[38]; and establishment of dynamic monitoring platforms for real-time therapeutic optimization.

Conclusions

This investigation delineates unique differentiation and mitochondrial metabolic characteristics of CD8⁺ T lymphocytes within EOC malignant ascites, revealing critical associations with estrogen receptor expression. Ascitic T lymphocytes exhibit metabolic characteristics consistent with minimized mitochondrial stress while preserving functional capacity, potentially representing adaptive responses to microenvironmental challenges. Differential ER expression correlates with distinct immune phenotypes, suggesting the potential value of tailored therapeutic approaches. Our findings not only advance mechanistic understanding of T cell regulation within tumor microenvironments but also provide important theoretical foundations for developing personalized immunotherapy strategies based on metabolic modulation and hormone receptor stratification. These discoveries highlight the potential of combined therapeutic strategies integrating metabolic intervention, hormonal modulation, and immune checkpoint blockade, potentially opening new avenues for improving outcomes in advanced ovarian cancer patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Z.J. and Q.Q. designed the experiments. Z.J. and Y.Z. collected the samples. Z.J., Y.S., and Y.Z. performed the flow cytometry experiments. Z.J., D.T., and Y.H.W. analyzed the data. Z.J. wrote the original draft. J.W. and Q.Q. supervised the study. J.W. acquired funding. All authors reviewed the manuscript.

Funding

This work was supported by the Fundamental Research Funds for Soochow University (Grant Nos. H230337 and H240578) and the Open Research Fund of Suzhou Key Lab of Multi-modal Data Fusion and Intelligent Healthcare (Grant No. 25SZZD05).

Data and code availability

Flow cytometry data files (.fcs format) and analysis templates are available from the corresponding authors upon reasonable request. Data will be shared following institutional data sharing agreements and de-identification procedures.

Declarations

Conflict of interest

The authors declare that they have no competing interests.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (NO.027(2021)).

Consent to participate

Written informed consent was obtained from all individual participants included in the study.