T-cell exhaustion and senescence are two dominant dysfunctional states in the TME. Exhausted T-cells and senescent T-cells both accumulate in cancer patients and are correlated with clinical features and prognosis[15]. We describe here the predictive value of PD-1+CD8+ TILs and CD57+CD8+ TALs for patients with advanced HGSOC. The combination of higher PD-1+CD8+ TILs and lower CD57+CD8+ TALs seemed to predict treatment efficacy and chemosensitivity, not only prognosis.
A large amount of malignant ascites, which serve as a transporter to facilitate the spread of tumor cells and contain multiple immune cells, is an important feature of advanced HGSOC[22]. Recent studies have found that T-cells in tumors and ascites share a variety of T cell receptor (TCR) clones, and these TCR clones are basically undetectable in T-cells from peripheral blood[23]. Ascites-derived memory T cells could migrate into tumor sites and remodeling the ecosystems [23]. Therefore, we focused on the expression of PD-1 and CD57 on CD8+ TILs and CD8+ TALs from patients with advanced HGSOC. Compared with that in ascites, the proportion of CD8+ T-cells was significantly elevated in tumors, reflecting an active antitumor immune response[5, 10]. The expression of PD-1 was significantly upregulated in CD8+ TILs and CD8+ TALs. Consistent with previous research, CD8+ TILs had more PD-1+ cells than CD8+ TALs, which further showed a higher immunoreactive status in tumors than in ascites[10, 23]. The proportion of CD57+ in CD8+ TILs and CD8+ TALs was approximately 20% and showed a significant positive correlation, indicating similar signals inducing T-cell senescence in tumors and ascites. In this study, the median CD57 expression on PD-1+CD8+ T cells was approximately 20%, suggesting a relatively limited overlap between exhaustion and senescence.
In HGSOC, the number of CD8+ TILs were positively correlated with patient survival[11, 24]. In addition to content, changes in the function of CD8+ T-cells also impact the prognosis of cancer patients. In breast cancer, renal cancer, and non-small cell lung cancer (NSCLC), PD-1+ TILs are associated with poor patient survival[4]. Nevertheless, PD-1+ TILs are associated with a good prognosis in HPV+ head and neck cancer[25] and follicular lymphoma[26]. In previous studies, PD-1+ TILs were not associated with the prognosis of HGSOC patients[9, 10] or correlated with good prognosis[11, 13]. In our study, high levels of PD-1+CD8+ TILs were found to be a positive prognostic predictor for patients with advanced HGSOC. Patients with higher PD-1+CD8+ TILs had more TEM CD8+ T cells in tumor, which may reflect an active anti-tumor state. We further explored the function of PD-1+CD8+ TILs and found that these cells exhibit many features expected of tumor-reactive T-cells, including (i) mainly in the EM state, (ii) upregulation of costimulatory CD28, (iii) upregulation of activation- and exhaustion-related molecules, (iv) enhanced proliferation ability, and (v) favorable ability to secrete IL-2 and inflammatory cytokines (e.g., IFN-γ, TNF-α, and granzyme B). These results demonstrate that PD-1+CD8+ TILs of patients with advanced HGSOC are still in a state of precursor exhaustion of T-cells with stem cell properties and high proliferative ability[27, 28]. However, a systematic description of the phenotype and function of PD-1+CD8+ TILs in different types of cancer is still lacking. Though there is no significant correlation between the levels of PD-1+CD8+ TILs and PD-L1 expression in our research, the PD-L1 expression level in ovarian cancer was significantly lower than that in established immunosensitive tumors[29], which partly contributes to maintaining PD-1+CD8+ TIL function.
Our preliminary work found increased senescent CD8+ PBLs of advanced HGSOC patients compared with healthy volunteers, suggesting that T-cell senescence may be associated with cancer diagnosis. However, senescent CD8+ PBLs showed no correlation with progression free survival in advanced HGSOC[20]. Therefore, we explored the correlation between CD57+CD8+ TILs and CD57+CD8+ TALs and prognosis. High level of CD57+CD8+ TALs was a biomarker of poor prognosis. Patients with higher CD57+CD8+ TALs had less TEM CD8+ T cells and more TEMRA CD8+ T cells in ascites, which may reflect impaired anti-tumor immunity. CD57+CD8+ TALs showed obviously senescent features. The loss of CD28 expression results in the failure of CD57+CD8+ T-cells to activate and proliferate after receiving cancer antigen stimulation[30]. Low level of IL-2 may preferentially promote Treg according to the high IL-2Rα expression[31]. Meanwhile, lowering IL-2 levels impairs TEM CD8+ T cells development[32]. Excessive IFN-γ and TNF-α promote tumor cell stemness and the expression of immune suppressive factors[33–35]. These results suggest that CD57+CD8+ T-cells not only lose the antitumor immune response ability but also aggravate the tumor immunosuppressive microenvironment, promote the progression of ovarian cancer, and lead to poor prognosis. On the basis of first-line treatment, preventing or reversing T-cell senescence may be able to improve treatment efficacy, inhibit chemotherapy resistance and ameliorate outcomes for patients with advanced HGSOC.
Since PD-1+CD8+ TILs and CD57+CD8+ TALs both showed predictive value of prognosis in patients with advanced HGSOC, we further grouped the patients according to these two biomarkers and found they could be complementary tools. The composite score can help distinguish advanced HGSOC patients with different prognoses and adjust the treatment procedure in time.
Immune checkpoint inhibitors (ICIs), a key method of immunotherapy, are effective in patients with multiple cancers. However, in ovarian cancer, ICIs monotherapy efficacy is only approximately 10%, and the efficacy of multidrug combination therapy is only approximately 20%[29]. PD-1high TILs have been associated with better prognosis under ICIs[4]. Nevertheless, ICIs rely on CD28 signaling to restore the function of CD8+ T-cells[36, 37]. The expression of CD28 and PD-1 on CD57+CD8+ T-cells was significantly downregulated, suggesting that their function could not be restored by ICIs. In vitro, ICIs could not restore the proliferation ability of CD57+CD8+ T-cells[38]. In NSCLC and melanoma, patients with high pretreatment senescent T-cells in peripheral blood have poor ICIs efficacy[39–41]. The above studies also indicate that the composite score may be a more reliable biomarker compared to assess efficacy from ICI. Prevention or reversal of T-cell senescence may improve ICIs therapeutic responsiveness and prolong patient survival. The time window for ICIs therapy needs to be further clarified. Since none of the patients in our study received ICIs, the influence of PD-1+CD8+ TILs and CD57+CD8+ TALs on the efficacy of ICIs still needs to be further verified.
To our knowledge, this study is the first to explore the T cell exhaustion and senescence in newly diagnosed HGSOC simultaneously, and demonstrate that they are significantly associated with low CR rate, chemotherapy resistance, and poor survival upon first-line treatment with or without PARPi maintenance. Our study also has some shortcomings. First, the samples in this study were all from a single institution. A multicenter study is needed to obtain a more definitive conclusion. Second, T cells with the same phenotype in tumor and ascites do not have the same prognostic prediction effect, which may be related to the differences in the cellular composite or other factors.
In conclusion, we propose that PD-1+CD8+ TILs and CD57+CD8+ TALs reflects immunological status, which is important for prognosis in patients with advanced HGSOC. These cells were determined before treatment using medical waste with few technical requirements. Further studies are needed to better explain the mechanisms we observed.