In this study, we explored the association between TME and favorable outcomes of PD-1/L1 monotherapy in NSCLC. We comprehensively measured soluble immune mediators in plasma samples at ICI initiation and then 6 weeks later. Our analysis identified several potential soluble immune mediators whose pre-treatment levels or post-treatment changes were significantly associated with the prognosis in NSCLC patients treated with ICI monotherapy. Pre-treatment biomarkers were mostly chemokines such as members of the CXCL family and the CCL family, whereas on-treatment biomarkers included cytokines such as those in the IL family. Some of the levels of pre-treatment biomarkers and post-treatment changes in on-treatment biomarkers were correlated, suggesting that they may reflect changes between the pre-treatment and post-treatment TME that were associated with favorable therapeutic efficacy of ICI monotherapy. To understand the complex network formation of cytokines and chemokines in the TME relevant to cancer immunotherapy, comprehensive measurement of soluble immune mediators in the same patient population as in this study was useful [1–3, 7–9]. In addition, analysis of the relationship with other potential biomarkers such as irAEs, PD-L1 expression, CD8 + TIL density and NLR is expected to provide a more detailed understanding of the TME [1–3, 13–17].
Among the pre-treatment biomarkers, CXCL5 was associated with both PFS and OS, and the heatmap showed a survival-related trend. This suggests that CXCL5 may be an important biomarker for the therapeutic efficacy of ICI monotherapy. CXCL5, also known as neutrophil activating peptide 78, is secreted by cancer cells or other host cells in the TME, including macrophages, fibroblasts and dendritic cells [7, 18, 19]. CXCL5 recruits neutrophils into tumor tissue and promotes tumor cell proliferation and metastasis. CXCL5 was also correlated with neutrophils in peripheral blood in this study and may have promoted tumor development. CXCL5 has been reported to promote PD-L1 expression and decrease CD4 + and CD8 + TILs in tumors, but no significant association with either was observed in this study [18]. The association between CXCL5 and cancer immunotherapy is currently under investigation and has not been established [18, 19]. We found a correlation between pre-treatment CXCL5 level and post-treatment changes in IL-34 and CCL25. IL-34 modulates tumor-associated macrophage function, enhances local immune suppression, and promotes survival of cancer cells resistant to ICI treatment [9, 20, 21]. CCL25 attracts mature CD8 + T cells from the thymus into the peripheral blood [7, 12, 22, 23]. The studies on the effects of IL-34 and CCL25 on the TME may provide a better understanding of the clinical significance of CXCL5 in cancer immunotherapy.
Among the on-treatment biomarkers, post-treatment changes in IL-10 and IL-32 are associated with both PFS and OS and they may reflect favorable changes in the TME that are associated with the therapeutic efficacy of ICI monotherapy. The major cellular sources of IL-10 are CD4 + T cells, CD8 + T cells, a subset of Tregs and tumor cells [8]. IL-10 is a potent suppressor of anti-tumor immunity that inhibits tumor antigen presentation [1, 8, 9]. IL-10 acts primarily on dendric cells and macrophages, and it inhibits the differentiation and antigen-presenting properties of dendric cells [1, 8]. Furthermore, in this study, post-treatment changes in IL-10 correlated with pre-treatment TNFSF13B levels, which can promote B cell activation, and post-treatment neutrophil changes in peripheral blood cells. This suggests that pre-treatment B cells and post-treatment changes of neutrophil counts may be important factors in the favorable changes of the TME after ICI administration [10, 24, 25]. IL-32 is derived from NK cells and T cells and has nine different isoforms [9, 26]. IL-32 exhibits both pro- and anti-tumor effects, but the majority of the effects promote tumor growth. IL-32 can modulate the activity of tumor-associated macrophages and induce tumor inflammation [26]. There are few reports exploring the relationship between IL-32 and cancer immunotherapy, and it is necessary to identify the function of each isoform in the cancer immune cycle [9, 26]. Cancer immunotherapy may improve patient survival by altering IL family members in the TME [1, 8, 9, 25, 26].
Since irAEs reflect immune activation by ICI administration and are associated with therapeutic efficacy, we analyzed their association with the pre- and on-treatment biomarkers identified in the present study [13, 14]. Our analysis revealed that pre-treatment biomarkers were not associated with development of irAE, whereas post-treatment changes in CCL23 were associated with the development of irAE. CCL23 is also known as CKbeta8, macrophage inflammatory protein 3 and myeloid progenitor inhibitory factor-1 [7, 11]. CCL23 is produced by eosinophils, monocytes, and monocyte-derived cells, and it acts as a chemoattractant for monocytes and dendritic cells [7, 11]. Our results support the possibility that post-treatment changes in monocytes or dendritic cells induced by CCL23 may be involved in the development of irAE associated with the therapeutic outcome of ICI monotherapy. Considering that post-treatment changes in CCL23 may play an important role in irAE development, it is expected that more detailed investigation will improve the management of irAE.
It is well established that PD-L1 expression is associated with ICI treatment outcome [1–6]. In this study, pre-treatment levels of TNFSF13B and post-treatment changes in CCL25 were associated with PD-L1 expression. TNFSF13B, also known as B-cell activating factor, is produced by myeloid cells, activated T cells, and bone marrow stromal cells to promote B-cell development and survival [10, 24, 25]. In solid tumors, TNFSF13B expression varies among different cancer types, and its prognostic and functional roles are not well understood [24]. An association between the presence of intra-tumoral B cells and the therapeutic efficacy of anti-PD-L1 antibodies in NSCLC has been reported, but there are no reports yet on TNFSF13B and PD-L1 expression in NSCLC specimens [25]. Since the functions of B cells in cancer immunity are not as well understood as those of T cells, further investigation is needed to determine whether TNFSF13B regulates PD-L1 expression in NSCLC. PD-L1 expression was also associated with post-treatment changes in CCL25. CCL25, also known as thymus-expressed chemokine, is expressed in the thymus, intestinal tract and tumor cells [7, 12, 22, 23]. CCL25 binds to its receptor on mature CD8 + T cells in the thymus and enhances their migration to secondary lymphoid organs such as lymph nodes. This chemoattraction of CD8 + T cells to secondary lymphoid organs may promote the therapeutic efficacy of ICI treatment [12, 22, 23]. Further studies are needed as CCL25 may be a more direct therapeutic target as a surrogate for pathological PD-L1 expression.
CD8 + TIL density is also a pathologic predictor of ICI treatment [1–3]. In this study, pre-treatment CXCL10 was associated with CD8 + TILs, but on-treatment biomarkers were not. CXCL10 showed a trend toward higher levels in the low CD8 + TIL group. CXCL10 is a chemokine that is mainly produced by intra-tumoral myeloid immune cells and it correlated with monocytes in this study. In the cancer immune cycle, CXCL10 promotes T-cell trafficking to tumors, whereas it does not promote T-cell infiltration associated with CD8 + TIL density [1, 27]. This suggests that CXCL10 in this study may have been secondarily upregulated to recruit CD8 + T cells by negative feedback, reflecting the low density of CD8 TIL + cells. Considering that CD8 + TIL density is an important pathological predictor for cancer immunotherapy, its association with biomarkers may identify novel therapeutic targets [3].
The NLR is a predictor of ICI monotherapy that can be routinely measured in daily practice [15–17]. In this study, pre-treatment CCL17 and CCL13 were correlated with the NLR. Although not significantly correlated with the NLR, pre-treatment CXCL5 and CCL19 were correlated with neutrophil and lymphocyte counts. These chemokines may play a role as pre-treatment biomarkers by regulating the chemoattraction of lymphocytes or neutrophils in the peripheral blood [7, 11, 12, 18, 19]. Elucidation of the role of these biomarkers in cancer immunotherapy may allow prediction of the immune status of the TME from the numbers of peripheral blood cells.
In summary, we conducted a comprehensive measurement of soluble immune mediators and identified several potential pre-treatment and on-treatment biomarkers in patients with NSCLC who received ICI monotherapy. It is expected that pre-treatment biomarkers might help to identify those patients who could benefit from ICI monotherapy. Moreover, on-treatment biomarkers might help our understanding of the favorable changes in TME associated with therapeutic efficacy. In addition, we found several associations and correlations between these pre- and on-treatment biomarkers and irAE, PD-L1 expression, CD8 + TIL density, and the NLR. Combining these biomarkers identified in the exploratory analysis with other predictors could provide a more detailed understanding of cancer immunotherapy and potentially identify new therapeutic targets. Nevertheless, our study had several limitations. First, this study was a retrospective single-center study. Second, the analysis in this study was exploratory and needs to be verified prospectively. Further large-scale studies are needed to provide a larger number of patients to better characterize the benefits of ICI monotherapy.