The field of cancer immunotherapy has undergone a renaissance due to a greater understanding of the complex pathways that regulate tumor-induced immunosuppression. Immune checkpoint inhibitors (ICIs) targeting programmed cell death protein 1 (PD-1), its ligand PD-L1, and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) pathways have become the most potent and durable cancer immunotherapy, which has been shown to increase tumor control and extend life in many cancer types (1, 2). As ICIs have become an essential treatment modality in non-small cell lung cancer (NSCLC) treatment for those patients without EGFR or ALK mutations (3, 4), it becomes necessary to develop reliable methods of predicting and monitoring immune responses to these drugs (5, 6). Currently, the only validated biomarkers for ICI treatment are derived from tumor tissue: PD-L1 protein expression or tumor mutational burden (7). However, obtaining tumor biopsies may not be feasible in patients who are critically ill or whose tumors are in inaccessible locations.

Alternatively, peripheral blood offers a minimally invasive approach for examining tumor biology and tracking the evolution of molecular and immune biomarkers throughout cancer therapy (8, 9). We and others have shown that early changes in blood immune cells can correlate with clinical response to ICIs (10–12). However, the assessment for the phenotype changes of immune responses in lung cancer patients following ICI therapy has been hampered by limitations in the multiplexing capability of traditional flow cytometry, the high skill requirement for multiplexing a higher number of fluorochromes, and cost of equipment and filters for advanced multiplexing. Other options, such as mass cytometry (13) (e.g., CyToF), allow for significantly greater multiplexing of antigens, but are accompanied by higher costs and decreased sensitivity compared to conventional flow cytometry (14).

Traditional flow cytometry relies on matched filter and detector sets to measure the fluorescence emission peak for each fluorochrome marker. However, this method has limited resolution and multiplexing capabilities, which has driven the development of more sophisticated and sensitive instruments, often at increased costs when advanced multiplexing is required. In contrast, spectrum flow cytometry employs prisms to capture the complete emission spectrum across the entire visible light wavelength range for each fluorophore using an array of detectors.

Instead of using compensation to correct for fluorescence spillover like conventional flow cytometry, spectrum flow cytometry employs a technique known as spectrum unmixing. This process involves a mathematical algorithm that leverages the distinct spectrum signature of each fluorophore to differentiate multiple fluorophore signatures within a sample. Consequently, fluorophores with nearly identical peak emissions but differing off-peak emissions can be distinguished and utilized simultaneously in a panel (15, 16). As a result, spectrum flow cytometry can differentiate fluorophore combinations that conventional systems are unable to discern. This increased differentiation capability allows for enhanced flexibility when designing highly complex multicolor panels for comprehensive immunophenotypic analysis. Additionally, since more markers can be evaluated concurrently, spectrum flow cytometry effectively addresses the issue of limited sample availability by expanding the range of immune cell subtypes that can be assessed from a single sample. The aim of this study was to showcase the application of a 24-color spectrum flow cytometry panel in monitoring immune cell alterations in peripheral blood mononuclear cells (PBMCs) obtained from freshly collected or cryopreserved samples from lung cancer patients undergoing treatment with ICIs.

Although the tumor-specific cytotoxic T cells are the prime targets of ICIs, they can modulate the expression of additional immune cells and molecules in the tumor microenvironment (TME) and lymphoid organ frequencies (24, 25). These immune cell subtypes include macrophages, T cells, NK cells, monocytes, B cells, dendritic cells, myeloid-derived suppressor cells (MDSCs), and neutrophils. These components can collectively contribute to the regulation of tumor cell immune evasion. Moreover, imbalances in the levels of specific immunoregulatory cells and cytokines within the TME may influence tumor responsiveness (26). Cancer cells may employ several immune escape mechanisms, including inadequate presentation of tumor neoantigens, secretion of inhibitory chemokines or cytokines (e.g., TGF-β, IL-10), mutations that make molecules unrecognizable by the immune system, and recruitment of suppressive cells (e.g., Treg, MDSC) (27). A growing body of evidence suggests circulating immune cells in peripheral blood of patients may reflect the dynamic immune cell changes in the tumor tissue (28).

Our study presents several important clinical implications through the immunophenotypic analysis of immune subtypes in patient PBMCs before and after ICI treatment by the multicolor spectrum flow cytometry. Firstly, we found that at baseline monocytes, CD8+ naive T cells, double-negative CD3+CD4-CD8- T cells, and B cells were significantly lower in the patient’s samples, whereas the CD14+CD16- monocytes, CD8+ CM cells, and lineage-negative cells were higher in patient’s samples compared to healthy donors. Of these markers, NSCLC patients with a lower baseline proportion of CD4+ CM cells (<31%) or a higher baseline proportion of DN (CD4-CD8-CD3+) T cells (≥3%) experienced longer median PFS to ICI treatment. The function of CD4+ helper T cells has been shown to be crucial for effective CD8+ T cell responses to ICI therapy. CD4+ T cells can promote tumor regression through various mechanisms, including cytokine secretion (IL-2), enhancing tumor-specific CD8+ T-cell function, or directly eliminating cancer cells (29). Several preclinical studies have demonstrated that tumor-specific CD4+ T cells might recognize immunogenic mutations (30, 31). Consistent with our result, Tada et al. showed that a decrease in baseline CD4+ CM T cells in 7 out of 10 patients with head and neck squamous cell carcinoma exhibiting PR or SD to PD-1 inhibitor nivolumab (32). However, due to the small sample size, the result did not achieve statistical significance. Furthermore, the expansion of CD4+ naïve T cells to CD4+ CM T cells after ICI therapy predicted long-term survival benefits in patients with malignant melanoma (33). CD4+ CM T cells, characterized by CCR7 or CD62L expression and lacking CD45RA, can circulate within secondary lymphoid organs. While the functional interplay of CD4+ memory cells directly shapes the effects of PD-L1/PD-1 inhibitors on CD8+ anti-tumor responses, the exact mechanisms mediating this response remain elusive (29). Additionally, various studies have shown other T cell subtypes involved in response to ICI therapy. For instance, a significant decrease of CD4+ FOXP3- PD-1 high T cells during the initial stages of therapy correlated with an improvement in OS (30). There was a trend suggesting that a lower proportion of circulating CD4+ T cells, including CD38+CD4+ T cells in peripheral blood, could indicate a favorable prognosis (31). Further studies are needed to characterize the role of different T cell subtypes in mediating response to ICIs.

Secondly, we observed that individual NSCLC patients had unique patterns of blood immune cell expression and changes to ICI treatment ( Figure 3 ). Although all 3 poor responders had decreased frequencies of CD8+PD1+ T cells, 6 out of 11 (54.5%) good responders also had decreased frequencies of CD8+PD1+ T cells ( Figure 4A ). Overall, there was a correlation between increased frequencies of cytotoxic CD8+PD1+ T cells in PBMCs and good clinical responses to ICIs. This finding aligns with previous reports that PD-1 expression in CD8+ tumor infiltrating lymphocytes (TILs) within tumor samples had clonally expanded tumor-reactive lymphocytes (34, 35). A majority of patients experiencing clinical benefits demonstrated CD8+ PD-1+ T-cell responses within four weeks of therapy (36). We further showed that ICI treatment modulated the expression of subtypes of NK cells and monocytes. Although there was no statistically significant difference in the expression of NK cell subtypes in NSCLC patients before ICI treatment compared to healthy donors ( Table S2 ), we observed ICI treatment led to significant increase in the frequencies of CD16+CD56dim NK cell (P=0.006) ( Figures 4C, D ) and CD16+CD56- NK cells (P=0.03) ( Figures 4I, J ) compared to poor responders. Furthermore, an increase in the CD16+CD56dim NK cell frequency of ≥3% after ICI treatment compared to before treatment was associated with a longer PFS (NR vs. 5 months, HR=5.4, 95% CI 1.1-25.7, P=0.03) ( Figure 5B ). Our findings are consistent with several reports showing that NK cells, either alone or in combination with cytotoxic T cells, play an essential role in mediating tumor response to ICI treatment in multiple cancer types (32, 33, 37). In mouse models, NK cells promote the function of cytotoxic T cells in response to anti-PD-L1 treatment (38, 39). NK cells also acted as the primary cytotoxic cells in tumors with low MHC expression, even in PD-L1 negative tumors (40). A single dose of tumor vaccine targeting resistant tumors by dual T and NK cells was able to increase CD8+ and CD4+ T cell frequencies by 17.9 and 29.3 fold, and NK cell counts by about 40 fold, respectively, compared to the control in murine models (41). Recently, the expression of NK cell frequency in the peripheral blood was independently associated with longer survival of gastric cancer patients and colorectal patients (33, 37). The frequency of NK cells was positively correlated to the frequencies of T and B lymphocytes (33). However, these reports did not analyze the CD16 and CD56 subtypes of NK cells. In this study, we found that CD16+CD56dim NK cells played a more significant role in mediating response to ICI compared to CD16+CD56- NK cells. CD16+CD56dim NK cells comprise the majority of circulating human NK cells, which are the most cytotoxic NK cells. Upon target recognition, CD16+CD56dim NK cells release perforin and granzyme granules and mediate antibody-dependent cellular cytotoxicity through CD16 (FcɣRIII) to clear cancer cells. In contrast, increased CD16+CD56- NK cells were associated with immune escape from innate immunity during AML progression (42).

Previous studies suggested that CD14+HLADRhigh, low or neg monocytes play distinct roles in the regulation of inflammatory and immune-suppressive conditions (43, 44). Patients who responded to ipilimumab had significantly lower levels of pre-treatment CD14+HLA-DRlo/neg monocytes compared to their non-responsive counterparts (45). Another independent study linked lower pre-treatment frequencies of CD14+HLA-DRlo/neg monocytes to improved OS of patients (46). In this study, we did not observe any significant difference in the pre-treatment frequency of CD14+HLADRhigh and low/neg monocytes in NSCLC patients compared to healthy donors ( Table S2 ). Instead, we found that there were significant increases in the frequency of CD14+HLADRhigh monocyte (P=0.04) ( Figures 4E, F ), and CD11c+HLADR+ monocytes (P=0.02) ( Figures 4K, L ) in good responders compared to poor responders after ICI treatment. Of these two monocyte subtypes, only an increase in CD14+HLADRhigh monocyte frequency by ≥0.1% following ICI treatment was associated with improved PFS (7.8 vs. 3.8 months, HR=5.7, 95% CI 1.0-31.7, P=0.04) ( Figure 5C ). Consistent with our results, single cell analysis revealed that high levels of CD14+CD16- HLADRhigh monocytes before ICI correlated with significantly increased PFS in melanoma patients (47) and NSCLC patients (48). Further study is warranted to validate these findings in NSCLC patients and delineate underlying mechanisms.

Individual NSCLC patients had unique patterns of immune cell expression and changes to ICI treatment and other cancer therapy ( Figure 6 ). The 24-color spectrum flow cytometry assay may be used to monitor treatment response during the disease course with the optimal goal of improving the prediction of patient responses to cancer therapy and identify those who may benefit most from a specific treatment. However, this study has several limitations, including a small sample size, a retrospective design, and no adjustment for multiplicity due to its exploratory nature. Additionally, we did not analyze the changes in T cell receptor repertoire and various immunoregulatory cytokines in the blood. Potential selection bias and imbalances in patients’ baseline characteristics and treatment history could have influenced the outcomes. Compared to the liquid biopsy for tumor genomic profiling of plasma circulating tumor DNA, flow cytometry of blood immune cells requires special skills and the cost of appropriately collecting and processing blood immune cells in a timely manner. Before this method can be used in routine practice, the sensitivity, turnaround time, and cost need to be evaluated. Further study is warranted to use the established multicolor flow cytometry and RNA sequencing tools to monitor the dynamic changes of blood immune cells during cancer immunotherapy treatment.

Multicolor spectrum flow cytometry can simultaneously evaluate 1 million blood immune cells for changes in major immune cell subpopulations in NSCLC patients receiving ICIs. Our data support the critical role of subsets of T cells, NK cells and monocytes in mediating response to ICI. Further studies are needed to validate the predictive biomarkers and assays to select the appropriate patients for ICI therapy.

The RNA sequencing data presented in the study are deposited in the Gene Expression Omnibus (GEO) repository, accession number GSE235048. Other original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by University of California, Davis Institutional review board (IRB)-approved protocol (Protocol No. 226210). The patients/participants provided their written informed consent to participate in this study.

TL contributed to the conception and design of the study. WM, SW, and TL contributed to the patient sample and data collection. WM, SW, SL, ET, BM, MJ, DM, VV, JC, QZ, JVD, SC, and TL contributed to the acquisition, analysis, or interpretation of data. WM, SW, SL, DM, SC, and TL drafted and revised the manuscript. All authors contributed to the article and approved the submitted version.